https://cpw.cvlcollections.org/files/original/4d39ef53e50c14bab61789a8af09d358.pdf a73978f7d9a8fa37f23d1fbf5c972867 PDF Text Text Colorado Division of Wildlife Wildlife Research Report •July 2001 and July 2002 JOB PROGRESS REPORT State of Colorado Division of Wildlife-Mammals Research Work Package No. -~0_6~6_2_ _ _ _ _ __ Preble's Meadow Jumping Mouse Conservation Task No. Effects of Resource Addition on Preble's Meadow Jumping Mouse (Zapus hudsonius preblei) Movement Patterns 2 Period Covered: July I, 2000- June 30, 2002 Author: Anne Trainor Personnel: T. M. Shenk, G. C. White, K. Wilson Interim Report - Preliminary Results This work continues, and precise analysis of data has yet to be accomplished. Manipulation or interpretation of these data beyond that contained in this report should be labeled as such and is discouraged. ABSTRACT Preble's meadow jumping mouse (Zapus hudsonius preblei; PMJM) is federally listed as threatened under the Endangered Species Act (ESA). Habitat conservation plans (HCPs) as defined in Section IO of the ESA, allow for 'take' of species and their habitat on private property. HCPs attempt to minimize take and provide for mitigation. Collection of reliable information and an increased understanding of PMJM habitat requirements are essential for the development of effective mitigation strategies for this species. Thus, our objectives are to (I) determine how the presence ofresource additions influences the distribution of individual PMJM within a population, and (2) to quantify and compare microhabitat characteristics among areas PMJM used heavily to areas of no use. A manipulation experiment will be conducted in sections of riparian habitat and adjacent grasslands in Douglas County, Colorado in 2002 and 2003. ��3 Research Prospectus Effects of Resource Addition on Preble's Meadow Jumping Mouse (Zapus hudsonius preble1) Movement Patterns Anne Trainor, Tanya Shenk, and Kenneth Wilson Problem: The U.S. Fish and Wildlife Service (USFWS) listed the Preble's meadow jumping mouse (Zapus hudsonius preb/ei; PMJM) as a threatened species in 1998 under the Endangered Species Act (USFWS 1999). Upon listing, little was known about the biology and habitat requirements of this subspecies within its range along the Front Range of Colorado and southeastern Wyoming. Since listing, a number of projects (e.g., long-term monitoring, surveying, and movement studies) have collected valuable information throughout Colorado (Schorr 2001, Meaney 2000, Shenk and Sivert 1999). However, information on specific habitat requirements and their relationship to the distribution, density, survival and reproduction of PMJM is still lacking. The threatened status of PMJM requires management decisions be made despite our limited knowledge. In particular, the species and its habitat are subject to habitat conservation plans (HCPs). HCPs are written for endangered and threatened species to compensate for authorized "take" with mitigation practices (Bingham and Noon 1998). HCPs require the use of the "best available" science to determine the biological needs of target species (Harding et al. 2001). Collection of reliable information for the species will improve the mitigation practices developed for HCPs. Well-designed habitat manipulation experiments provide the strongest inference to determine cause and effect relationships. Understanding of the species habitat requirements will enable the development of effective mitigation strategies. A manipulation experiment will be conducted in Douglas County, Colorado (Columbine Open Space) during 2002 and 2003 to advance our understanding of PMJM habitat requirements. We will manipulate sections of the riparian habitat and adjacent grassland within the 100-year flood plain. The site will be manipulated by adding patches (3 m x 2.43 m) of artificial resources (food and cover). Time limitations (2 field seasons are inadequate for vegetation to establish) and funding (cost of planting and sustaining vegetation) will restrict this manipulation experiment to simulating habitat with temporary structures and food supplementation. The treatments will be placed in areas of low use based on past monitoring studies conducted by the Colorado Division of Wildlife (CDOW) during 1998-2000 within 60 m of East Plum Creek. PMJM will be radio tracked before and after the manipulation to determine if PMJM locations can be altered through the addition of resources. Research Objectives: We propose two primary objectives: 1) determine how the presence ofresource additions influences the distribution of individual PMJM within a population, and 2) quantify habitat characteristics of PMJM on a micro habitat scale. Desired outcome: We want to examine if the distribution of individual PMJM can be altered in response to the addition ofresources (food and cover) and to quantify relevant microhabitat characteristics where PMJM have been detected. Approach: A field experiment will be conducted during 2002-2003 (June-August) to test if PMJM can be attracted to areas where they haye not previously been detected within the 100-year flood plain. Study Site- Riparian habitat within the Columbine Open Space, owned by Douglas County Open Space managed by the CDOW and the adjacent grassland. Columbine Open Space was selected �4 because PMJM were monitored for 3 years by the CDOW ( 1998-2000), providing site-specific information on PMJM locations before this manipulation experiment. Methods- PMJM will be trapped using non-folding Sherman live traps (7.6 cm x 8.9 cm x 22.9 cm) placed 5 m apart along approximately 0.5 km transects adjacent to both sides of East Plum Creek for a minimum of 5 consecutive nights. Trapping procedures will be in accordance with the guidelines published by the USFWS ( 1999). Species other than PMJM will be recorded with trap location and immediately released. The following information will be recorded for captured PMJM: unique identification, trap location, weight, sex, age, and reproductive condition. PMJM will be scanned for a passive integrated transponder (PIT) tag. Newly captured individuals will have a unique PIT-tag injected and individuals ~18 grams will be anesthetized with isoflurane to fit a 1-g radio transmitter (Holohil Systems Ltd Ontario, Canada). All methods were approved by the Animal Care and Use Committee of Colorado State University (Authorization Number A3572-0I). Radio telemetry will be used to monitor locations of individuals for a 21-day period, the battery life of the radio transmitters. Observers will attempt to stay approximately 3 m from the radio-tagged individual to avoid influencing PMJM movement. Observations taken 3 m or greater from PMJM did not influence movement (T. Shenk, CDOW personal comm.). The following information will be recorded at each relocation: individual identification, time, weather, and surrounding vegetation. All data will be combined into a geographical information system (GIS) database using ArcView®3.2 (Environmental Systems Research Institute, Redlands, California, U.S.A.). The manipulation experiment will consist of 5 phases: I) select areas of little or no previous use by PMJM based on CDOW location data (1998-2000) collected at Columbine Open Space, 2) record pre-treatment location data ofradio-tagged individuals for 6 nights, 3) select placement of treatment plots based on pre-treatment and CDOW location data, 4) add resources to treatment plots, and 5) record post-treatment location data ofradio-tagged individuals. Two sessions (June and July) of the manipulation experiment will be conducted each year. A digital map with a grid cell size of 9 m x 9 m has been constructed for the entire study site with ArcView®3.2 (Environmental Systems Research Institute, Redlands, California, U.S.A.) software. CDOW location data was pooled into a single coverage over the grid to establish areas~ 1,000 m2 containing only low use cells (<2 locations/cell based on CDOW location data) within the 100-year flood plain. Location of treatments will be selected with a stratified random design from a set of candidate cells meeting a criteria developed based on PMJM biology (sparse vegetation and little food source) within 60 m of East Plum Creek, and low historical use. The artificial cover, simulating vertical complexity, will be constructed with wheat straw and tree branches distributed in a patch (3 m x 2.43 m). Burlap cloth will be suspended 30 cm over the tree branches and straw. Food supplements composed of an equal mixture of whole wheat, dehydrated alfalfa pellets and sweet feed will be placed on cardboard trays (0.16 m x 0.3 m) within the straw and branches as an attractant and a source of high protein. The dimensions of the treatments were selected to balance the manageability of construction and decrease the chance of inter and intraspecies domination within a treatment. Quantification of microhabitat variables in areas of high use will be examined by comparing a random sample of cells (9 m x 9 m) containing~ 99 % of PMJM locations for each session and a random sample of cells with no locations detected. Two line transects will be randomly placed in each selected cell with 6 quadrat frames (50 cm x 20 cm) evenly distributed per line transect (Daubenmire 1959). The variables measured in each cell will include percent bare ground, shrub, grass, and forb cover and vegetation composition . .J �5 Analysis- The location data will be analyzed with linear regression. The response variable will be the number of locations detected in a cell. A suite of candidate models will be developed as predictors of the response variable. Akaike's information criterion (AIC) will be applied to select the best "approximating" model (Burnham and Anderson 2002). The independent habitat variables of interest for the models include distance from the center of the cell to the nearest water, area and juxtaposition of nearest shrub, and presence of wetland grasses in the cell. Additional variables to be included in the models are period (pre- or post-treatment), sex, session, and year. The microhabitat data collected from the Daubenmire plots will be analyzed with Proc GLM (SAS 2002) to test for differences in means among areas of high use and no use by PMJM. Schedule: Fall 2001.. ............... Formation of committee and write study plan development Spring 2002 .............. Completion of study plan and preparation for field season Summer 2002 ........... Begin data collection Fall 2002 ................. Begin data analysis Spring 2003 .............. Continue data analysis, begin thesis and complete comprehensive oral examination Summer 2003 ........... Complete data collection Fall 2003 ................. Complete data analysis Winter 2004 .............. Complete thesis Budget: Fiscal Year 2001-02 Refurbished Holohil radio collars Technicians Housing Vehicles Supplies Computer Tuition GRA Faculty Support FY 2001-02 Total $2,000 1@$1,250/month for 3 months $833/month for 3 months 1@$200/month for 3 months (including mileage) $700 $2,000 $2,880 $1,300/month for 12 months $3,830.00 $33,779 Fiscal Year 2002-03 Refurbished Holohil radio collars PIT tags Technicians Technicians Supplies Housing Vehicles Stipend Tuition Faculty Support FY 2002-03 Total $500 $1,000 1@$1,250/month for 3 months 1@$1,250/month for 2 months $500 $833/month for 3 months 1@$200/month for 3 months (including mileage) $1,300/month for 12 months $679.00 . $3830.00 $31,458 �6 Fiscal Year 2003-04 Stipend FY 2003-04 Total Project Total $1,300/month for 8 months $10,400 $75,637 Potential cooperators: Funding is provided by the CDOW; Douglas County has given permission to use the Open Space; Colorado State University has provided office space, equipment, computers, and adviser. Alternative and obstacles: Alternatives considered include modeling habitat utilization at a microhabitat and site specific scales. Potential obstacles include 1) low number of radio-tagged PMJM resulting in low power for the manipulation experiment, 2) other species deterring PMJM from using the additional resources, and 3) radio-tagged mice do not detect the additional resources. PMJM have demonstrated general site fidelity to daytime nesting sites and nighttime feeding sites (Shenk and Sivert 1999). It is possible PMJM have established use areas and do not easily alter their use patterns. Literature Cited: Bingham, B. B., and B. R. Noon. 1998. The use of core areas in comprehensive mitigation strategies. Conservation Biology 12:241-243. Burnham K. P., and D.R. Anderson. 2002. Model selection and multimodel inference. Second edition. Springer, New York, New York, USA. Daubenmire, R. 1959. A canopy-coverage method ofvegetational analysis. Northwest Science 33:4364. Harding, E., E. Crone, B. D. Elderd, J.M. Hoekstra, A. J. McKerrow, J. D. Perrine, J. Regetz, L. J. Rissler, A.G. Stanley, E. L. Walters and NCEAS Habitat Conservation Plan Working Group. 2001. The scientific foundations of habitat conservation plans: a quantitative assessment. Conservation Biology 15:488-500. Meaney, C. A. 2000. Monitoring for Preble's meadow jumping mice along South Boulder Creek and Four Ditches. Boulder, Colorado, USA. Report prepared for the Colorado Division of Wildlife. SAS Institute. 2002. SAS Version 8.2. SAS Institute, Cary, North Carolina, USA. Schorr, R. 2001. Meadow jumping mice (Zapus hudsonius preblei) on the U.S. Air Force Academy, El Paso County, Colorado, USA. Shenk, T. M., and M. Sivert. 1999. Movement patterns of Preble's meadow jumping mouse (Zapus hudsonius preblei) as they very across time and space. Annual Report to the Colorado Division of Wildlife. Fort Collins, Colorado, USA. U. S. Fish and Wildlife Service. 1999. Interim Survey Guidelines for Preble's meadow jumping mouse. U.S. Fish and Wildlife Service. Denver, Colorado, USA. �13 ketamine (3 mg/kg) administered intramuscularly (IM) with either an extendible pole-syringe or a pressurized syringe-dart fired from a Dan-Inject air pistol. Immobilized lynx were monitored continuously for decreased respiration or hypothermia. If lynx exhibited decreased respiration 2mg/kg ofDopram was administered under the tongue. lfrespiration was severely decreased, the animal was ventilated with a resuscitation bag. If medetomidine/ketamine were the immobilization drug, the antagonist Antisedan was administered. Hypothermic (body temperature < 95° F) animals were warmed with hand warmers and blankets. While immobilized, the lynx were fitted with a replacement VHF/satellite collar and blood and hair samples were collected. Once the animal was processed recovery was expedited by injecting the antagonist Antisedan IM if medetomodine/ketemine was used for immobilization. The lynx was monitored until it was sufficiently recovered to move safely on its own. No antagonist is available for Te)ezol so lynx anesthetized with this drug were monitored until the animal recovered on its own. If captured and in poor body condition the lynx was anesthetized with Telezol (2 mg/kg) and returned to the Colorado holding facility for rehabilitation. Reproduction Reproductive status of all female lynx was determined prior to release through radiographs. Pregnancy was confirmed through radiographs if the bones of the fetuses had begun to ossify. All females known to be pregnant or thought to possibly be pregnant on release were monitored closely from their release through the following August to determine reproductive success. Females remaining within a limited area immediately after release through August were located and observed to look for accompanying kittens or a den site. Females that had been released in 1999 and were alive in spring 2000 were monitored for proximity to males during breeding season and for site fidelity to a given area during the denning period of May and June 2000. Each female lynx from the 1999 releases was directly observed in summer 2000 over 3-5 different visits to look for accompanying kittens or evidence of denning. Each female alive in May 2001 that exhibited stationary movement patterns in June 2001 was observed in summer or fall 2001 to look for accompanying kittens. Females were also snow-tracked in winter months to look for accompanying kitten tracks. Hunting Behavior Snow-tracking of released lynx provided preliminary information on hunting behavior by documenting location of kills, food caches, chases, and diet composition estimated through scat analysis. Snow-tracking was conducted during February-May 1999 (Year 1), November 1999-May 2000 (Year 2), and November 2000 -April 2001 (Year 3). Prey from failed and successful hunting attempts were identified by either tracks or remains. Scat analysis also provided information on foods consumed. Scat samples were collected wherever found and labeled with location and individual lynx identification. Only part of the scat was collected, the remainder was left where found so as not to interfere with the possibility that the scat was being used by the animal as a territory mark. Habitat Use Gross habitat use was documented by recording canopy vegetation at aerial locations. More refined descriptions of habitat use by reintroduced lynx were obtained through snow-tracking and sitescale habitat data collection. Specific objectives for the site-scale habitat data collection included: 1. Describe and quantify site-scale habitat use by lynx reintroduced to Colorado. 2. Compare site-scale habitat use among types of sites ( e.g., kills vs. long-duration beds). 3. Compare site-scale habitat use between sexes. 4. Compare habitat use over years. 5. Develop methodology that will result in data that will be comparable to data collected in studies investigating the ecology of snowshoe hare in Colorado. �14 Snow-tracking Locations from aerial- and satellite-tracking were used to help ground-trackers locate lynx tracks in snow. Snowmobiles, where permitted, were used to gain the closest possible access to the lynx tracks without disturbing the animal. From that point, the tracking team used snowshoes to access tracks. Once tracks were found, the ground crew back- or forward-tracked the animal if it was far enough away not to be disturbed. Back-tracking generally avoided the possibility of disturbing the lynx by moving away from the animal rather than towards the animal. However, monitoring of the lynx through radiotelemetry was used to assure that the ground crew was staying a sufficient distance away from the lynx in the event the lynx might double back on its tracks. Radio-telemetry was also used in forward-tracking to make sure the team did not disturb the animal. If it appeared the lynx began to move in response to the observers, the observers stopped following the tracks. If the lynx began to move and the movement did not appear to be a response to the observers, the ground crew continued following the track. An attempt was made in Year I and Year 2 to track each lynx. In Year 3 we attempted to track all lynx within the Core Release Area. Ground crews were instructed to track lynx only where it was safe to travel. Restrictions to safe travel included avalanche danger and extremely rugged terrain. Ground crews worked in pairs and were fully equipped for winter back-country survival. Data Collection For each day of tracking the date, lynx being tracked, slope, aspect, UTM coordinates, elevation, general habitat description, and summary of the days tracking were recorded. Aspect was defined as the direction of 'downhill' or 'fall line' on a slope. This is the direction along the ground in a dihedral angle between the horizontal and the plane of the ground surface. Units were compass direction that most closely defined the cardinal points (e.g., N, NW, etc.). Slope was defined as the dihedral angle between the horizontal and the plane of the ground surface (e.g., 45° ). There were 4 levels of intensity of human activity recorded. They included: 1. None: track was not found off an existing snowmobile, ski, or snow shoe track. Distance to nearest human track is greater than 1.0 km 2. Low: track was found near low human activity (e.g., existing snowmobile or ski track) 3. Medium: track found near medium human activity (detected the presence of other people in the area during tracking effort). 4. High: track found near high human activity (e.g., detected presence of many people nearby, near major road, near housing). There were 2 categories for recording detection of tracks of other species. They included "M" for tracks from multiple animals of the same species and "T" for detection of tracks of only a single animal of the species. Once a track was located there were 2 types of'sites' that were encountered. Site I areas needed documentation but either did not reflect areas lynx selected for specific habitat features, or sites that occurred too frequently to measure each in detail. Site II areas were places where lynx may have selected habitat features. At each of the 2 types of sites the date, lynx tracked, slope, aspect, forest structure class, UTM coordinates, and elevation was recorded. Forest structure classes included grass/forb, shrub/seedling, sapling/pole, mature, and old growth as defined in Table 3. For Site I areas, the only additional data that was collected was identification of what the site was used for (e.g., short-duration bed), and a brief description of the site. These sites included the start and end of the track being followed, the location of scat, and short-duration beds defined as being small in size (approximating an area a lynx would crouch), and with little ice formed in the bed indicating little time spent there. The Site II areas included areas that might reflect specific habitat features lynx selected for. These sites required habitat sampling (see below) and included locations where the following were found: kills, start of chases, territory marks (e.g., spray sites, buried scat, scat placed on prominent locations), long-duration beds (encompasses an area where a lynx would have lain for an extended period, iced bottom), travel (if no other sites sampled in last hour), and road crossing (both sides of road). �15 Table 3. Definitions of forest structure classes used to describe habitat sites (Thomas 1979). Forest Structure Class Definition Grass/forb The grass/forb stage is created naturally by a catastrophic event, such as wildfire, and is typified by the near complete absence of snags, litter or down material in the aspen and ponderosa pine types, or vice versa in the lodgepole or subalpine forest types. Shrub/seedling The shrub/seedling stage occurs when tree seedlings or shrubs grow up to 2.5 cm at diameter breast height (DBH), either naturally or artificially through planting. Sapling/pole The sapling/pole stage is a young stage where tree DBH's range from 2.5-17 .5 cm with tree heights ranging 1.8-13 .5 m. These trees are 5-100 years of age, depending on species and site condition. Mature The mature stage occurs when tree diameters reach a relatively large size (25-50 cm) and the trees are usually 90 or more years old. Forest stands begin to experience accelerated mortality from disease and insects. Old-growth The old-growth stage occurs when a mature stand is at advanced age (100 years for aspen or 200 years for spruce), is very slow growing, and has advanced degrees of disease, insects, snags, and down, dead material. An exception to this occurs in ponderosa pine and aspen types where these old-growth stands typically experience low densities of down dead material or snags. Description of the Habitat Plot A habitat sampling plot was completed wherever a Site II was encountered. The habitat plot consisted of a 12 m x 12 m square defined by a series of 25 points placed in 5 rows of 5 with the center point being on the object that defined the site (e.g., a kill) (Figure 1). Each point was 3 m apart. The 12 m x 12 m sampling square exceeded the minimum requirement of0.01 ha. Recommended by Curtis (1959) for sampling trees. Measurements taken at each of the 25 points included: 1. Snow depth - measured vertically by an avalanche probe marked in cm. , 6 11 16 2i.... 2. Understory - measured from top of snow to 150 cm above snow in a column of 3-cm radius i + t + t !2 • • • • • around the avalanche probe. Because understory i + t t + t measurements were influenced by vegetation outside !3 • • • • • the perimeter of the 25 sampling points (12 m x 12 i + t + t m) the area used for esfimating undersofy cover was 4. •-•-• • 15 m by 15 m. At each point, crews recorded all + t shrubs, trees and coarse woody debris (CWD) that -~-~-~•-•-•-•_2F fell within this column and was visible above the 12 m snow. Crews also recorded number of branches of 15 m each species that fell within the column at 3 different height categories (0-0.5 m, 0.51-1.0 m, 1.01-1.5 m). 3. Overstory: measured at 150 cm above snow with a sighting tube. The tube was made of PVC pipe, with a curved viewing end and a crosshair made of wire on the opposite end. The sighting tube Figure 1. Design of site-scale habitat sampwas attached to the avalanche probe used to measure ling plot. Each point was 3 m apart. The object snow depth. Species that hit the crosshair were that triggered the habitat sampling (e.g., a kill) recorded at each of the 25 points in the vegetation was located at the center point. plot. Ganey and Block (1994) found this method of h•-•-• •-• �16 measuring canopy cover (with ~ 20 sample points per plot; Laymon 1988) provided greater precision among observers. 4. Species composition: all the different species of tree or shrub that hit. the crosshair of the sighting tube at each of the 25 points were recorded. • Tree composition of the vegetation plot was recorded by species and diameter at breast height (DBH). Snow depth was used in conjunction with this recorded DBH to estimate true DBH. Within the 12 m x 12 m square all conifers and deciduous trees were recorded by DBH size class (A= 0-15 cm, B = 15.1-30 cm, C = 30.1-45 cm, D = 45.1-60 cm, E = »60 cm). Area for the tree composition analysis was 12 m x 12 m. Understory was estimated as: (1) percent occurrence within the vegetation plot (number of points with understory/total number of points surveyed) and (2) mean percent occurrence and variance by species and height category over the total points sampled within the vegetation plot. Overstory was estimated as percent occurrence over the vegetation plot (number of points with overstory/total number of points surveyed). Results Assessment of Release Protocols A total of 41 lynx were released in Colorado in 1999 under 5 different release protocols (Table 2). Release protocols were modified as new information became available from monitoring the released lynx through radio-telemetry and snow-tracking. Each modification of the release protocols decreased the percent of animals dying from starvation (Table 4). Three of the 4 animals released under Protocol 1 died of starvation within 6 weeks of their release and the fourth was recaptured and returned to the holding facility where she recovered and was later re-released. Reevaluation of the condition of animals released under the Protocol I suggested that these animals might not have been in optimal physical condition when released. Therefore, Protocol 2 was initiated. Most lynx gained considerable body weight while in captivity (Wild 1999). Nine lynx were released under this second protocol. Of these, 1 juvenile female died of starvation 7 weeks after release. After the starvation death under Protocol 2, Protocol 3 was developed (3-week minimum holding time, high quality diet, no release prior to May 1). Twenty lynx were released under Protocol 3 with no starvation deaths of these animals occurring within 6 months post-release. Six females were released under Protocol 3P (known to be pregnant) and 2 under Protocol 3P? (possibly pregnant). Two of the 6 pregnant lynx released died of starvation within 6 months post-release. An assessment of the fates of each lynx under all 5 release protocols used in 1999 led to release protocols for lynx released in 2000. Release Protocols 2 and 3 resulted in the fewest starvation mortalities up to 8 months after release date. The common element in both protocols 2 and 3 was increased captivity time in the Colorado holding facility. The single starvation mortality for lynx released under Protocol 2 in 1999 was also the only juvenile released under that protocol and the only animal released in February (the other 8 Protocol 2 lynx were released in March 1999). Thus, all lynx released in 2000 were released under either Protocol 2 or 3 but not before April 1. Because of the high percentage of starvation mortalities in females pregnant on release, we also attempted to avoid reintroducing lynx that were known to be pregnant. This was best accomplished by trying to have animals captured for the reintroduction effort in Canada prior to their breeding season. A series of 11 models (Table 5) were developed using various combinations of the hypothesized factors that may have affected survival up to 8 months post-release: (1) whether the release was in winter or spring (Rel), (2) whether the released lynx was an adult or kitten (age), (3) sex of lynx released (sex), ( 4) whether or not females were released while pregnant (preg) and the interaction of pregnancy and age of the female (adult vs. kitten), and (5) the duration of holding time in the Colorado facility (DCF). Survival time and DCF were modeled with and without a log transformation (Ln) because of possible threshold effects overtime. We used AICc as the model selection criterion to select the model that best explains the data (Table 5). The model that best fit the data was {S(age+preg+Rel+LnT+LnDCF}, �17 which suggested pregnancy had a deleterious effect on survival of females, with the effect being stronger on kittens than adults (Figure 2). This model also indicated that winter releases led to higher mortality than spring releases for both non-pregnant kittens (Figure 3) and non-pregnant adults (Figure 4), with no sex effects on either age class. Lastly, long stays in the Colorado holding facility increased survival if the duration was at least 21 days with no significant decrease or increase in survival for stays longer than 21 days (Figures 2, 3, 4). Table 4. Starvation mortalities and recaptures of poor body condition lynx reintroduced to Colorado under the 5 release Erotocols over 2 }:'.ears. Release Protocol Year Total Number Released I 1999 4 2 2 3 3 3P? 3P? 3P 3P 1999 2000 1999 2000 1999 2000 1999 2000 9 41 20 10 2 3 6 I Number of. Starvation Mortalities a 3b 1c 1c % Mortality Number of Recaptures in Poor Body Condition a 75 11 0 0 0 0 2d 0 100 11 2 0 0 0 0 33 0 I 0 0 0 0 0 0 2 0 0 0 0 33 0 % Failure of Release Protocol 0 0 within 8 months of release. b I juvenile, 2 adults. c juvenile. d adults. 8 Table 5. Model selection results of the a priori models concerning the effects of age, sex, pregnancy, season ofrelease, and amount of time spent in the Colorado holding facility on survival oflynx 8 months Eost-release. Ranking based on AICc values. AICc #Model AICc ~Cc Weight Deviance Pars. {S(age+preg+Rel+LnT+LnDCF} {S(age*preg+Rel+LnT+LnDCF} {S(age+preg+Rel+T+LnDCF} {S(age+preg+Rel+T+T2+LnDCF} {S(age+Rel+preg+LnDCF} {S(age+preg+Rel+T'+T2'+LnDCF} {S(age+preg+Rel+T"+T2"+LnDCF} {S(age*preg +Rel+LnDCF} {S(age*Rel+preg+LnDCF} {S(age+Rel+preg+DCF} {S(age+ReI+ereg+DCF+DCF2} 200.120 201.027 202.702 203.225 203.266 204.069 204.936 205.265 205.289 205.609 205.760 0 1.91 2.58 3.10 3.15 3.95 4.82 5.14 5.17 5.49 5.64 0.28305 0.10908 0.07784 0.05993 0.05871 0.03930 0.02547 0.02161 0.02135 0.01819 0.01687 6 7 6 7 5 7 7 6 6 5 6 188.036 187.914 190.618 189.113 193.206 189.957 190.824 193.181 193.205 195.549 193.676 �18 Kittens -; > .E Adults 1 = ~ 0.8 > i 0.6 =E= u 50 Two-week Interval Figure 2. Effects of pregnancy and tiine spent in the Colorado holding facility on survival of pregnant kittens and adult females. Wmter Release Spring Release Two-week lnlemll Figure 3. Effects ofrelease season and time spent in the Colorado holding facility on survival of non-pregnant kittens. Spring Rd~ T~weeklnterval Figure 4. Effects of release season and time spent in the Colorado holding facility on survival of non-pregnant adults. �19 Movement Patterns A total of 2,158 aerial VHF locations for all 96 reintroduced lynx have been collected to date (Figure 5, Figure 6). An additional 4,020 satellite locations (1,375 satellite locations if multiple locations for a single night were averaged and counted as only I location) for 49 of the 51 lynx fitted with dual collars have been collected. Two satellite collars never worked after the lynx were released. The majority of movements in 1999 away from the an area encompassed by alO0-km radius area centered on the release sites (Core Release Area) were to the north (Figure 5), although some movements occurred to the south into New Mexico and west into Utah as well. A single male from the 1999 releases traveled to Nebraska where he was shot in violation of Nebraska regulations. Initial dispersal habitats used by lynx released in 1999 were highly variable, from high elevation Engelmann spruce/subalpine fir to Nebraska agricultural lands. Dispersal movement directions for lynx released in 2000 were similar to those of lynx released in 1999 (Figure 6). Most movements away from the Core Release Area were to the north. However, more animals remained within the Core Release Area. Numerous travel corridors have been used repeatedly by more than one lynx, possibly suggesting route selection based on olfactory cues. These travel corridors include the Cochetopa Hills area for northerly movements, the Rio Grande Reservoir-SilvertonLizardhead Pass for movements to the west, and southerly movements down the east side of Wolf Creek Pass to the southeast through the Conejos River Valley. Lynx appear to remain faithful to an area during winter months, and exhibit more extensive movements away from these areas in the summer. Such movement patterns have also been documented by native lynx in Wyoming and Montana (Squires and Laurion 1999). Most lynx currently being tracked are within the Core Release Area (Figure 7). Mortalities occurred throughout the areas through which lynx moved. However, mortalities occurred in New Mexico in higher proportion to all lynx locations in that area than elsewhere (Figure 8). Survival and Mortality Factors Of the 96 lynx released, 39 mortalities have been recorded to date. From the 1999 releases (41 animals) we have had 24 known mortalities (Table 6). From the 2000 releases (55 animals) we have 15 known mortalities (Table 6). Of the total 9 confirmed starvation deaths, 3 were associated with animals released in less than ideal body condition (released under Protocol I) and 2 were lynx less than I-year old (Table 4). Fourteen of the mortalities died of unknown causes. In 4 of these cases starvation could be ruled out as cause of death by evidence of good body condition through examination of bone marrow. Pneumonic plague could be ruled out in all 14 cases. Delayed retrieval of carcasses resulted in advanced deterioration of the body, making determination of cause of death impossible. Necropsy results for 3 female lynx released in 2000, indicate they died from pneumonic plague. Two of these lynx were in good condition, with abdominal fat, no muscle wasting, and fat in the bone marrow. The only gross lesions were an acute fibrinous pneumonia (i.e., lung infection of short duration). These lynx had probably only been sick a few days before they died. A third female was in poore~ body condition when found. Plague was diagnosed by flourescent antibody tests and isolation of Yersinia pestis from lung and spleen samples. A fourth lynx was also diagnosed with plague after she was hit by a car. A male lynx, recaptured near Laramie, Wyoming, tested positive for plague titers but did not have active plague. Thus, he had been exposed to plague but either did not contract the disease or recovered from the disease. Recaptures Seven lynx have been recaptured and 6 subsequently re-released since their initial release. Lynx BC99F6 was released in 1999 under Protocol 1. Her behavior and incidental sightings by the public suggested the lynx was in poor condition. We trapped her using a Tomahawk™ live trap baited with rabbit. She was recaptured the first night (March 25, 1999) we set the trap. On capture, we found she was severely emaciated. We anesthetized her with Telezol (2 mg/kg) and returned her to the Colorado holding facility. She was rehabilitated through diet. The lynx gained weight steadily and was re-released on May �20 . . --------i l J __) I ----M• ~-- •• • • Locations of lynx Released in 1999 ('JHigh-ys • •• 'c:::J Colorado Counties [-=-j New Mexico Counties ···'·'··'··· Wyoming Cowities NNe!nslm Counties s 300 - 0 300 Figure 5. Locations oflynx released in 1999, obtained through aerial telemetry. 600 Kilometers �21 • VHF·locations of lynxrele~ed in 200:0 • • PTJ locations of lynx released. In 2000 Highways. ·Colontdo Counties N_ew Mexico C ouities r---; ~oming Counties • A/ t:J c:J 75/Nebras1ca Counti~ N ./\./ utah C~unties W*E 300 - s 0 600 .Kilometers Figure 6. Locations of lynx released in 2000. Gray circles indicate locations obtained from satellite collars. Black circles are locations obtained through aerial telemetry. �22 Last Known Locations • alive ■ collar off .~ missing /'/Highways Colorado Counties .-.-, New Mexico Counties t:::=J s -- 300 0 300 600 Kilometers Figure 7. Last known locations oflynx. Circles depict locations oflynx currently being tracked. Triangles are last known locations of missing lynx. �23 • .4· ·Mortality looation·s of lynx releasecl·in 2000 • Mort.lity looations of lynx released iii 1999 / \ / Highways t::J Colorado Counties CJ NeW. Mexico· Counties N . W * E s -- 300 0 300 600 Kilometers Figure 8. Locations of lynx mortalities. Circles depict mortalities of lynx released in 1999, triangles depict mortalities from lynx released in 2000. I �24 Table 6. Causes of death for lynx released into southwestern Colorado in 1999 and 2000. 1999 1999 2000 2000 2000 Male Female Male Female Unknown Cause Starvation 1 6 1 1 Road-kill 2 2 2 Shot 1 Human-caused a Trauma - unknown cause Possible predation Plague 3 Unknown 2 3 2 2 Unknown - not starvation b 1 2 Total Mortalities 7 17 4 10 1 Total 9 5 5 2 1 3 9 4 39 a Cut collar found, no carcass. b Starvation ruled out by condition of bone marrow. 28, 1999. She was hit by a car on Interstate 70 on July 19, 1999. Necropsy results indicated she was in excellent body condition at her time of death. Lynx AK99M9 was released on May 12, 1999 and recaptured on March 24, 2000. Field observations by the lynx monitoring crew suggested that the lynx was severely emaciated. Live-trapping the lynx failed, so the lynx was darted with Telazol (3 mg/kg) using a Dan-Inject CO 2 pistol. Physical examination revealed severe emaciation (6 kg). The lynx was returned to the Colorado holding facility and rehabilitated through diet. The lynx gained weight steadily and was re-released on May 3, 2000 but has not been located since and is listed as missing. Lynx AK99F2 was released on May 7, 1999 and recaptured on April 18, 2000. Field observations by the lynx monitoring crew suggested that the lynx was emaciated. She was live-trapped with a Tomahawk™ live trap with one night's effort. On capture, we found she was emaciated. We anesthetized her with Telezol (2 mg/kg) and returned her to the Colorado holding facility. She was rehabilitated through diet. The lynx gained weight steadily and was re-released on May 22, 2000. This lynx is currently in the Core Release Area. Lynx BC00F7 was released on April 2, 2000 and recaptured on February 11, 2001. She was severely emaciated and was captured by anesthetizing her with Telazol delivered IM by a jab-pole. She was returned to the Frisco Creek Wildlife Rehabilitation Center but died that night from emaciation and hypothermia. Lynx BC00M13 was released on April 2, 2000 and recaptured on March 21, 2001 near Laramie, Wyoming. He had been observed by a homeowner on his porch. We recaptured the lynx because this type of behavior was not considered normal. On examination he was in good body condition. After a period of observation this lynx was re-released at the Rio Grande Reservoir on April 24, 2001. This lynx had previously been listed as one of our 15 missing lynx as he had not been located since Sept 2000. This lynx is currently in the Core Release Area Lynx YK99F5 was recaptured on Aprill 9, 2001 to have her radio collar changed. She was captured in a live trap baited with one of her own kills. She was in very good body condition. We anesthetized her with Telazol (3mg/k.g), processed and released her on the same site where she was captured. Only her cut collar was found on October 17, 2001, cause of death is assumed human-caused. Lynx AK99F5 was treed by hounds and anesthetized with Telazol (3 mg/kg) on September 2, 2001. Her collar was exchanged, hair and blood samples were collected. She was in very good body condition and showed no evidence of lactation. She was re-released on the site she was recaptured once she recovered from the Telazol. This lynx remains in the same area as her recapture, within the Core Release Area. �25 Reproduction Six lynx released in 1999 were known to be pregnant (Table 2, Release Protocol 3P), and 2 other females released may have been pregnant (Table 2, Release Protocol 3P?). Three of the 6 lynx known to have been pregnant on release in 1999 died within 2 months after release: 2 starved and 1 was killed on the road. Long-distance movements and lack of stationary movement patterns of the other 3 lynx known to have been pregnant on release in 1999 suggests these females did not have young with them by July 1999. Of the 2 females that might have been pregnant, movement patterns were not suggestive of a female rearing young. It is not known if any other females bred and/or had young once released, however no females snow-tracked in Year 2 had young with them. Beginning in March 2000 both male and female lynx began to exhibit extensive movements (> 100 km) away from areas they had used throughout the winter. For example, 1 male moved from the area near Frisco he used in the winter to the area west of Lizardhead Pass, a straight line distance of approximately 270 km (Figure 9). Such movements by both females and males put them in close (< 5 km) proximity to a lynx of the opposite sex. These extreme movements may have been related to breeding behavior. All 7 females alive in spring 2000 were documented in close(< 5 km) proximity to a male during the breeding season and could have bred. Two isolated males did not move during March or April and thus were not in close proximity to a known _female during the breeding season. This was a male that had used the area in and adjacent to the northwest comer of Rocky Mountain National Park and a male that used the area around Cuchara, Colorado throughout the winter. The 7 females in the wild during breeding season 2000 were monitored for site fidelity to a given area during the denning period of May and June. Each of these 7 females was directly observed in summer 2000 over 3-5 different visits to look for accompanying kittens. No kittens were found. The question of whether they successfully bred or had kittens at some point in 2000 is unknown. However, no kittens were found during the following winter through snow-tracking. From radiographs taken of the 35 females released in 2000, after breeding season,1 female was known to be pregnant and 3 were possibly pregnant. Movement patterns suggested none of these 4 females had kittens with them by July 2000. Of the 49 lynx being tracked on a regular basis during the March 2001 breeding season, there were 29 females and 20 males. We documented movements that may have been related to breeding. The largest movement observed was a male that moved to Laramie, Wyoming and was subsequently recaptured, rehabilitated and re-released in the Core Release Area in Colorado after the breeding season. Other movements were of a much smaller scale, 10-30 km. These movements were primarily movements of males towards a female. We documented 10 potential 'pairs' where a pair was defined as a male and female within 5 km of each other and in the same drainage. More pairs could have occurred which we did not document from aerial- or ground-tracking because of the time delays between lynx locations. To date, no reproduction has been documented in 2001 from direct observations of females. Snow-tracking efforts this winter will focus initially on females in an attempt to document possible kittens through tracks. Current Status Of the total 96 lynx released we have 39 known mortalities (Table 7). We currently are listing 16 lynx as missing - 11 males, 5 females. We have not heard signals on 13 (11 males, 2 females) of these lynx since at least December 2000. The remaining 3 missing lynx are females that have been lost for less than 1 year. Possible reasons for not locating these missing lynx include (1) long distance dispersal, beyond the areas currently being searched, (2) radio failure, or (3) destruction of the radio (e.g., run over by car). We continue to search for all missing lynx during both aerial and ground searches. There have been 4 incidents whe;:re lynx missing for over a year have returned to the Core Release Area and are now once again being monitored on a regular basis. Thus, it is premature to consider missing lynx as lost to the Colorado lynx program. However, of the 16 missing lynx, 3 have collars whose battery life expired spring 2001 and will probably never be located through telemetry. At least 1 of the missing lynx is a mortality where we know a collar was found on a road kill bot the collar was not returned to the �26 NHighways -□ Colorado Counties Lynx YK99M3 Movements 19990513 • 19990521 • 19990521 • 19990715 • 19990715 • 19990923 • 19990923 • 19991027 • 19991027 • 19991229 • 19991229 • 20000229 • 20000229 • 20000329 • 20000329 • 20000512 N W*E s -- 300 0 300 600 Kiloinete Figure 9. Movements of a male lynx in breeding season 2000. Straight-line distance from winter use area to the area used during breeding season in approximately 270 km. The larger the circle the more recent the date, uptoMay2000 �27 CDOW for identification. One female is known to have slipped her collar. Thus, we are currently tracking 41 lynx. Table 7. Current status of lynx reintroduced to Colorado. Females Males Unknown Released 57 39 Known Dead 27 11 1 Missing 5 11 Slipped Collar 1 Tracking 24 17 TOTAL 96 39 16a 1 41 a l is unknown mortality. Hunting Behavior Snow-tracking of released lynx provided preliminary information on hunting behavior by documenting location of kills, food caches, chases, and through scat analysis. Prey from failed and successful hunting attempts were identified by either tracks or remains. Scat analysis also provided infonnation on foods consumed. During Year 1 a total of 10 kills were located. All the snow-tracking effort was conducted on 9 lynx released under Protocols 1 and 2. Any lynx released under Protocol 3 were released too late to track. In Year 2, ground crews tracked 13 of the lynx released in 1999. Two other lynx were being located during this time but were not in areas covered by snow. We found 64 kills and collected 109 scat samples that will be analyzed for content. Lynx released in 2000 were released too late to snow track in Year 2. In Year 3, field crews snow-tracked 48 lynx, documented 86 kills and collected 189 scat samples. Data collected on kills (Figure 10) suggests the reintroduced lynx are feeding on their preferred prey species, snowshoe hare (Lepus americanus) and pine (red) squirrel (Tamiasciurus hudsonicus) in similar proportions as those reported for northern lynx during lows in the snowshoe hare cycle (Aubry et al., 1999). Caution must be used in interpreting the proportion of identified kills. Such a proportion ignores other food items that ~e consumed in their entirety. For example, through snow-tracking we have some evidence that lynx are mousing and several of the fresh 70 carcasses have yielded small mammals in 60 the gut on necropsy. 1131999 ■ 1999-00 ■ 2000-01 However, the extent of small 50 ~ mammals in the diet are not accurately 'o 40 portrayed by information collected based a.. _8 30 on prey remains in snow. Nearly all the e::S 20 scat samples collected have been found z through snow-tracking efforts and thus will 10 be representative of winter diet only. The summer diet of lynx elsewhere has been 0 documented to include less snowshoe hare Snowshoe Cottontail Other Red and more alternative prey than in winter hare squirrel (Mowat et al. 1999). Species = Habitat Use Gross habitat use was documented from 2441 aerial locations oflynx I Figure 10. Winter diet of reintroduced lynx estimated from snow-tracking data. �28 collected from February 1999 through December 2001. Throughout the year 80 Engelmann spruce (Picea engelmannii) 70 / subalpine fir (Abies lasiocarpa) (S/F) "'= 60 was the dominant cover used by lynx -~ (Figure 11 ). A mix of Engelmann cc 50 CJ spruce, subalpine fir and aspen 0 40 ~ (Populus tremuloides) (S/F/A) was the = ll> 30 second most common cover type used CJ i.. ll> throughout the year. Various riparian p.., 20 and riparian mix areas was the third 10 most common cover type where lynx were found during the daytime flights. 0 Use of S/F and S/F/A was similar J J A S O N D J F M A M throughout the year. There was a trend Months in increased use of riparian areas beginning in July, peaking in Noveml ■ s/F DS/F/A ■ Riparian ber, and dropping off December through June. Figure 11. Percent aerial locations in Engelmann spruce A total of 4 73 site-scale habitat subalpine fir forests (S/F), Engelmann spruce- subalpine firplots were completed in Year 3. The aspen forests (S/F/A), and riparian areas by month. majority understory species at all 3 heights was Engelmann spruce, followed by subalpine fir, willow (Salix spp.) and aspen (Figure 12). Various other species such as Ponderosa pine (Pinus ponderosa), lodgepole pine (Pinus contorta), cottonwood (Populus sargentii), birch 100 (Betula spp.) and others were also found in less than 5% of the habitat 9 0 ,□ LOW aMEDIUM •HIGH plots. Coarse woody debris was also 8 0 present in 10-35% of plots. If present, 70 willow provided the greatest percent cover within a plot (Figure 13) 6 0 = ll> followed by Engelmann spruce, CJ 50 ~ subalpine fir, aspen and coarse woody i:i... 40 debris. 30 Engelmann spruce provided a mean of35.87 % overstory within 20 86.68% of the plots (Figure 14). 10 Subalpine fir and aspen provided 0 overstory for< 50% of the plots, but SF CWD WI AS LO ES when present provided approximately Species the same mean percent cover as Engelmann spruce (Figure 14). Willow and lodgepole pine provided fewer than l 0% of the plots with Figure 12. Mean percent cover of habitat plot by understory cover, but when present provided tree/shrub species Engelmann spruce (ES), subalpine fir (SF), nearly the same percent cover as the willow (WI), aspen (AS), lodgepole pine (LO), and coarse other tree species (Figure 14). woody debris (CWD) if species is present. Mean percent The most common tree species cover is estimated for 3 height levels above the snow (low = in the habitat plots was Engelmann 0-0.5 m, medium= 0.51-1.0 m, high= 1.1-1.5 m). - I - �29 50 45 ,□ LOW ■ MEDIUM ■ HIGH I 40 35 30 = . 25 .,u ., I>. 20 15 10 5 0 ES SF CWD WI Species AS LO Figure 13. Mean percent cover of habitat plots by understory tree or shrub species Engelmann spruce (ES), subalpine fir (SF), willow (WI), aspen (AS), lodgepole pine (LO), and coarse woody debris (CWD) if species is present. 100 90 80 70 EIES 60 c...u ...u "- ■ ES Snag ■ SF 50 □ SF Snag ■ AS 40 ■ AS Snag ■ WI 30 EILO 20 10 0 Percent Plots With These Species Mean Percent Cover if Species Present Figure 14. Percent plots with overstory tree species Engelmann • spruce (ES), subalpine fir (SF), willow (WI), aspen (AS), lodgepole pine (LO), and coarse woody debris (CWD). Mean percent overstory cover if tree species present. spruce (Figure 15). Subalpine fir and aspen were also present in> 35% of the plots. Most habitat plots were vegetated with trees of DBH < 6" (Figure 15). As DBH increased, percent occurrence decreased within the plot. The larger DBH trees (> 18") within the plots were generally Engelmann spruce with fewer subalpine firs of that DBH class present in the habitat plots. No willow or aspen of DBH > 18" were present in any of the plots. Of the 5 most common tree species in the habitat plots, mean number of trees for each DBH size class ranged from 0.18 to 10.82 except for willow which averaged 74.83 plants per plot (Figure 16). Areas of willow used by lynx are typically dense willow thickets. Discussion Of the 96 lynx released in Colorado in 1999 and 2000 we are currently monitoring 41 lynx on a regular basis and an additional 16 lynx may still be alive, although not being monitored. We have 39 confirmed mortalities. Survival oflynx released in the second year has been higher than lynx released in the first year. Human-caused mortalities due to vehicle collision, gunshot, and the mortalities where only a cut collar was found comprise the greatest known cause of mortality for the reintroduced lynx (31 % ). Mortalities due to starvation (23%) were minimized with improved release protocols. Only 2 of the 55 lynx released in 2000 died of starvation and 1 of those �30 died 8.5 months post-release. Three lynx died of plague, 1 road kill tested positive for plague, and 1 lynx had plague positive titers while healthy. Carnivores are most often exposed to plague by eating infected rodents or by being bitten by rodent fleas (Biggens and Kosoy 200] ). Although it is known that felids are highly susceptible to plague (Aiello 1998), the 5 cases of plague in lynx reintroduced to Colorado are the first documented for this species. Dispersal movement patterns for lynx released in 2000 were similar to those of lynx released in 1999. However, more animals remained within the Core Release Area. This increased site fidelity may be due to the presence of con-specifics in the area on release. Numerous travel corridors have been used repeatedly by more than ] lynx, possibly suggesting route selection based on olfactory cues. These travel corridors include the Cochetopa Hills area for northerly movements, the Rio Grande Reservoir-SilvertonLizardhead Pass for movements to the west, and southerly movements down the east side of Wolf Creek Pass to the southeast to the Conejos River Valley. Lynx appear to remain faithful to an . area during winter months, and exhibit more extensive movements away from these areas in the summer. Most lynx currently being tracked are within the Core Release Area. During the summer of 2000 and 2001, several lynx that had been faithful to a 100 D0-6"DBH 90 .6.1-12" DBH 80 •12.1-18" DBH D 18.1-24" DBH 70 • > 24" DBH 60 = "' :::'. 50 "' 0.. 40 30 20 I0 0 ES SF CWD WI AS LO Species Figure 15. Percent of habitat plots with tree species Engelmann spruce (ES), subalpine fire (SF), willow (WI), aspen (AS), lodgepole pine (LO), and coarse woody debris (CWD) by diameter at breast height (DBH) size class. 100 90 □ 0-6"DBH .6.1-12" DBH 80 .12.1-18" DBH 70 = ." "' D18.1-24" DBH . > 24" DBH 60 50 "' 0.. 40 30 20 10 0 ES SF CWD WI AS LO Species Figure 16. Mean number of trees or shrubs in habitat plots with tree species Engelmann spruce (ES), subalpine fire (SF), willow (WI), aspen (AS), lodgepole pine (LO), and coarse woody debris (CWD) by diameter at breast height (DBH) size class. �31 given area during the winter months made large movements away from their winter-use areas. Extensive summer movements away from areas used throughout the rest of the year have been documented in native lynx in Wyoming and Montana (Squires and Laurion 1999). In winter, lynx reintroduced to Colorado appear to be feeding on their preferred prey species, snowshoe hare and red squirrel in similar proportions as those reported for northern lynx during lows in the snowshoe hare cycle (Aubry et al., 1999). Caution must be used in interpreting the proportion of identified kills. Such a proportion ignores other food items that are consumed in their entirety and thus are biased towards larger prey and may not accurately represent the proportion of smaller prey items, such as microtines, in lynx winter diet. Through snow-tracking we have evidence that lynx are mousing and several of the fresh carcasses have yielded small mammals in the gut on necropsy. Nearly all the scat samples collected have been found through snow-tracking efforts and thus are representative of winter diet only. However, the summer diet of lynx has been documented to include less snowshoe hare and more alternative prey than in winter (Mowat et al., 1999). Reproduction is critical to achieving a self-sustaining viable population of lynx in Colorado. Although females have been monitored and observed during each denning season, no kittens have been found to date. Snow-tracking has also not provided evidence that any of the females tracked had kittens with them. However, the question of whether they successfully bred or had kittens at some point is unknown. With only 7 females from the 1999 releases in the wild in spring 2000 it was expected that there _might not be successful reproduction in 2000. However, the extreme movements observed by both females and males in March and April 2000 may have been related to breeding behavior. March and April are the natural breeding periods for northern lynx (Tumlison 1987). From observations of the 29 females alive in summer 2001, we have not yet documented kittens. We may still find evidence of kittens through snow-tracking efforts in winter 2001-02. Mowat et al. (1999) suggest lynx and snowshoe hare select similar habitats except that hares select more dense stands than lynx. Very dense understory limits hunting success of the lynx and provides refugia for hares. Given the high proportion of snowshoe hare in the lynx diet in Colorado, we might then assume the habitats used by reintroduced lynx also depict areas where snowshoes hare are abundant and available for capture by lynx in Colorado. From both aerial locations taken throughout the year and from the site-scale habitat data collected in winter, the most common areas used by lynx are in stands of Engelmann spruce and subalpine fir. This is in contrast to adjacent areas of Ponderosa pine, pinyonjuniper, aspen and oakbrush. The lack oflodgepole pine in the areas used by the lynx may be more reflective of the limited amount oflodgepole pine in southwestern Colorado, the Core Research Area, rather than avoidance of this tree species. Hodges (1999) summarized habitats used by snowshoe hare from 15 studies as areas of dense understory cover from shrubs, stands that are densely stocked, and stands at ages where branches have more lateral cover. Species composition and stand age appears to be less correlated with hare habitat use than is understory structure (Hodges 1999). The stands need to be old enough to provide dense cover and browse for the hares and cover for the lynx. In winter, the cover/browse needs to be tall enough to still provide browse and cover in average snow depths. Hares also use riparian areas and mature forests with understory. Site-scale habitat use documented for lynx in Colorado indicate lynx are most commonly using areas with Engelmann spruce understory present from the snow line to at least 1.5 m above the snow. The mean percent understory cover within the habitat plots is typically less than 15% regardless ofunderstory species. However, if the understory species is willow, percent understory cover is typically double that,.with mean number of shrubs per plot approximately 80, far greater than for any other understory species. • In winter, hares browse on small diameter woody stems (<0.25"), bark and needles. In summer hares shifts their diet to include forbs, grasses, and other succulents as well as continuing to browse on woody stems. This shift in diet may express itself in seasonal shifts in habitat use, using more or denser coniferous cover in winter than in summer. The increased use ofriparian areas by lynx in Colorado from July to November may reflect a seasonal shift in hare habitat in Colorado. Major (1989) suggested lynx hunted the edge of dense riparian willow stands. The use of these edge habitats may allow lynx to hunt �32 hares that live in habitats nonnally too dense to hunt effectively. The use of riparian areas and riparianEngelmann spruce-subalpine fir and riparian- aspen mixes documented in Colorado may stem from a similar hunting strategy. However, too little is known about habitat use by hares in Colorado to test this hypothesis at this time. Lynx also require sufficient denning habitat. Denning habitat has been described by Koehler ( 1990) and Mowat et al. ( 1999) as areas having dense downed trees, roots, or dense live vegetation. No den sites have been located as yet in Colorado for comparison. Through extensive monitoring of released animals we were able to continuously evaluate and modify release protocols to improve survival of released lynx. The primary element in later, more successful release protocols was increased time in captivity at the Colorado holding facility. Increasing the amount of time lynx were held in the Colorado holding facility provided each lynx with an opportunity to increase body weight and acclimate to the climate, elevation, and local conditions of the environment they would be released into. Although most lynx were housed in individual pens, with a few sharing a pen with one other lynx, the holding facility also allowed the lynx to hear and smell each other throughout this acclimation period. Such contact may have provided time for social interactions to occur. Such social interactions may improve the likelihood these animals could fonn a breeding population. If additional lynx are released in Colorado the following guidelines are recommended in establishing release protocols. Translocated animals should be adults and females should not be pregnant on release. Once lynx are moved from their place of origin they should be held a minimum of 3 weeks in a local holding facility to provide a high quality diet for gaining optimal body condition prior to release in the new area, acclimation time to adjust to local conditions, and possible social interactions. Animals should be released in the spring to ensure the highest prey abundance in the release area. These release protocol guidelines may also prove useful if other states attempt lynx reintroductions or augmentations. Future Research Future research will include the continued monitoring of lynx released in Colorado that have remained in the Core Release Area. Such monitoring will include continued data collection and analysis on survival and mortality factors, reproduction, habitat use, winter and summer diet, and movement patterns. If additional funding becomes available, reintroduced lynx that have moved beyond the Core Release Area should also be monitored, particularly those lynx using areas near the Interstate Highway 70 corridor. We will continue to attempt to recapture lynx to replace radio collars that are either malfunctioning or scheduled to stop functioning. Any Colorado born lynx will be radio collared once they reach a minimum of 10 months of age. Studies have been initiated to refine mark-recapture techniques to estimate abundance oflynx from hair-snag data. Such an approach would provide a non-invasive technique for estimating abundance. A snowshoe hare ecology study was initiated in 2001 to describe density of hares in various forest stands and which habitats and topographic features are most important to hare density and survival. From this research, management prescriptions may be designed to better manage forests for optimal hare populations. Maintaining abundant and widespread snowshoe hare populations is essential to establishing lynx in Colorado. Through funding provided by Colorado Department of Transportation (COOT) a detailed analysis of lynx movement patterns as they relate to highways has been initiated. The feasibility of augmenting this reintroduction effort by releasing additional animals from Canada and Alaska is being considered by CDOW to improve the likelihood of establishing a viable population oflynx in Colorado. Funding is being sought to develop protocols for collecting data on lynx summer diet by using dogs trained to locate lynx scat. �33 If viable, self-sustaining populations oflynx are established in Colorado, habitat manipulation studies will be needed to more fully understand how lynx respond to their habitat and how best to alter habitats to maintain and enhance lynx populations. Acknowledgments The lynx reintroduction program involved the efforts of literally hundreds of people across North America, in Canada and the U.S. Any attempt to properly acknowledge all the people who played a role in this effort is at risk of missing many people. The following list should be considered to be incomplete. CDOWCLAWS Team (1998-2001): Bill Andree, Tom Beck, Gene Byrne, Bruce Gill, Mike Grode, Rick Kahn (Program Leader), Dave Kenvin, Todd Malmsbury, Jim Olterman, Dale Reed, John Seidel, Scott Wait, Margaret Wild CDOW: John Mumma (Director 1996-2000), Conrad Albert, Jerry Apker, Cary Carron, Don Crane, Larry DeClaire, Phil Ehrlich, Lee Flores, Delana Friedrich, Dave Gallegos, Juanita Garcia, Drayton Harrison, Jon Kindler, Ann Mangusso, Jerrie McKee, Melody Miller, Mike Miller, Kirk Navo, Robin Olterman, Jerry Pacheo, Mike Reid, Ellen Salem, Eric Schaller, Mike Sherman, Jennie Slater, Steve Steinert, Kip Stransky, Suzanne Tracey, Anne Trainor, Brad Weinmeister, Nancy Wild, Perry Will, Brent Woodward, Kelly Woods, Kevin Wright. Lynx Advisory Team (19982001): Steve Buskirk, Jeff Copeland, Dave Kenny, John Krebs, Brian Miller (Co-leader), Mike Phillips, Kim Poole, Rich Reading (Co-leader), Rob Ramey, John Weaver. U.S. Forest Sen,ice: Kit Buell, Joan Friedlander, Jerry Mastel, John Squires, Fred Wahl. U.S. Fish And Wildlife Sen,ice: Lee Carlson, Gary Patton (1998-2000), Kurt Broderdorp. State Agencies: Gary Koehler (Washington). National Park Sen,ice: Steve King. Colorado State University: Alan B. Franklin, Gary C. White. Colorado Natural Heritage Program: Rob Schorr, Mike Wunder. Alaska: ADF&G: Cathie Harms, Mark Mcnay, Dan Reed (Regional Manager), Wayne Reglin (Director), Ken Taylor (Assist. Director), Ken Whitten, Randy Zarnke, Other:Ron Perkins (trapper), Dr. Cort Zachel (veterinarian). British Columbia: Dr. Gary Armstrong (veterinarian), Mike Badry (government), Paul Blackwell (trapper coordinator), Trappers: Dennis Brown, Ken Graham, Tom Sbo, Terry Stocks, Ron Teppema, Matt Ounpuu. Yukon: Government: Arthur Hoole (Director), Harvey Jessup, Brian Pelchat, Helen Slama, Trappers:Roger Alfred, Ron Chamber, Raymond Craft, Lance Goodwin, Jerry Kruse, Elizabeth Hofer, Jurg Hofer, Guenther Mueller (YK Trapper's Association), Ken Reeder, Rene Rivard (Trapper coordinator), Russ Rose, Gilbert Tulk, Dave Young. Alberta: Al Cook. Northwest Territories: Albert Bourque, Robert Mulders (Forbearer Biologist), Doug Steward (Director NWT Renewable Res.), Fort Providence Native People. Colorado Holding Facility: Herman and Susan Dieterich. Pilots: Dell Dhabolt, Larry Gepfert, Al Keith, Jim Olterman, Matt Secor, Whitey Wannamaker, Dave Younkin. Field Crews: Bryce Bateman, Bob Dickman, Denny Morris, Gene Orth, Chris Parmater, Jake Powell, Jeremy Rockweit, Jennifer Zahratka. Photographs: Tom Beck, Bruce Gill, Mary Lloyd, Rich Reading, Rick Thompson. Funding: CDOW, GOCO, Turner Foundation, U.S. Forest Service, Vail Associates. Literature Cited Aiello, S. E., editor 1998. The Merck Veterinary Manual. Eighth Edition. Merck & Co., Inc. Whitehorse Station, New Jersey. Aubry, K. B., G. M. Koehler, J. R. Squires. 1999. Ecology of Canada lynx in southern boreal forests. Pages 373-396 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U.S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. Biggins, D. E. and M. Y. Kosoy. 2001. Influences of introduced plague in North American mammals: implications from ecology of plague in Asia. Journal of Mammalogy 82: 906-916. Burnham, K. P. and D.R. Anderson. 1998. Model Selection and Inference: A Practical InformationTheoretic Approach. Springer-Verlag, New York, New York. �34 Byrne, G. 1998. Core area release site selection and considerations for a Canada lynx reintroduction in Colorado. Report for the Colorado Division of Wildlife. Curtis, J. T. 1959. The vegetation of Wisconsin. University of Wisconsin Pres, Madison. Ganey, J. L. and W. M. Block. A comparison of two techniques for measuring canopy closure. Western Journal of Applied Forestry 9:1: 21-23. Hodges, K. E. 1999. Ecology of snowshoe hares in southern boreal and montane forests. Pages 163-206 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. Koehler, G. M. 1990. Population and habitat characteristics oflynx and snowshoe hares in north central Washington. Canadian Journal of Zoology 68:845-851. Laymon, S. A. 1988. The ecology of the spotted owl in the central Sierra Nevada, California. PhD Dissertation University of California, Berkeley, California. Major, A. R. 1989. Lynx, Lynx canadensis canadensis (Kerr) predation patterns and habitat use in the Yukon Territory, Canada. M. S. Thesis, State University of New York, Syracuse. Mowat, G., K. G. Poole, and M. O'Donoghue. 1999. Ecology of lynx in northern Canada and Alaska. Pages 265-306 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U.S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. Pollock, K. H., S. R. Winterstein, C. M. Bunck, and P. D. Curtis. 1989. Survival analysis in telemetry studies: the staggered entry design. Journal of Wildlife Management 53: 7-15. Poole, K. G., G. Mowat, and B. G. Slough. 1993. Chemical immobilization oflynx. Wildlife Society Bulletin 21:136-140. Seidel, J., B. Andree, S. Berlinger, K. Buell, G. Byrne, B. Gill, D. Kenvin, and D. Reed. 1998. Draft strategy for the conservation and reestablishment of lynx and wolverine in the southern Rocky Mountains. Report for the Colorado Division of Wildlife. Shenk, T. M. 1999. Program narrative: Post-release monitoring of reintroduced lynx (Lynx canadensis) to Colorado. Report for the Colorado Division of Wildlife. Squires, J. R. and T. Laurion. 1999. Lynx home range and movements in Montana and Wyoming: preliminary results. Pages 337-349 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University Press of Colorado, Boulder, Colorado. Thomas, J. W. Ed. 1979. Wildlife habitats in managed forests - the Blue Mountains of Oregon and Washington. USDA Agricultural Handbook No. 553. U.S. Government Printing Office. Washington, D. C. Tumlison, R. 1987. Mammalian Species: Fe/is lynx. American Society ofMamrnalogists. U. S. Fish and Wildlife Service. 2000. Endangered and threatened wildlife and plants: final rule to list the contiguous United States distinct population segment of the Canada lynx as a threatened species. Federal Register 65, Number 58. White, G. C. and K. P. Burnham. 1999. Program MARK: Survival estimation from populations of marked animals. Bird Study 46 (suppl):120-138. Wild, M.A. 1999. Lynx veterinary services and diagnostics. Job Progress Report for the Colorado Division of Wildlife. Fort Collins, Colorado. �1 JOB PROGRESS REPORT Stateof_ _ _ _ _ _~C=o=l=or=a=d=o_ _ _ __ Division of Wildlife - Mammals Research Work Package No. --~0~6~62~------- Preble's Meadow Jumping Mouse Conservation Task No. - - - - - ~ 2 ~ - - - - - - - Effects of Resource Addition on Preble's Meadow Jumping Mouse (Zapus hudsonius preblei) Movement Patterns Period Covered: July 1, 2002 - June 30, 2003 Author: Anne M. Trainor. Personnel: T. M. Shenk, K. Wilson, G. C. White Interim Report - Preliminary Results This work continues, and precise analysis ofdata has yet to be accomplished. Manipulation or interpretation of these data beyond that contained in this report should be labeled as such and is discouraged. ABSTRACT The Preble's meadow jumping mouse (Zapus hudsonius preblei; PMJM) is a federally threatened species. Improving our understanding of PMJM habitat is essential for the development of effective management strategies for conservation of the species. Thus, the objectives of our research were to compare microhabitat characteristics among low and high use areas within PMJM habitat and to determine how the addition of artificial resources influence the movement patterns of PMJM. A comparison of microhabitat characteristics from a random sample of "high-use" and "no-use" areas indicated a greater (P < 0.0001) shrub canopy cover in "high-use" areas verses "no-use" areas (47.7% ± 29.8%, 12.6% ± 14.11 %, respectively). Further, "high-use" areas had greater basal cover (P = 0.013) and bare ground (P = 0.0459) and "no-use" areas contained a greater (P = 0.0331) abundance of forb canopy cover. We conducted a manipulation experiment where we constructed patches of artificial resources (food and cover) in areas without previous PMJM activity. PMJM were radio collared and located hourly before and after the addition of food and cover. The majority of PMJM movements were not influenced by the addition of resources in 2002. These results may be due to site fidelity or lack of exploratory movement to locate the additional resources �2 Effects of Resource Addition on Preble's Meadow Jumping Mouse (Zapus hudsonius preb/ei) Movement Patterns Anne M. Trainor Colorado State University INTRODUCTION The U.S. Fish and Wildlife Service (USFWS) listed the Preble's meadow jumping mouse (Zapus hudsonius preblei; PMJM) as a threatened species in 1998 under the Endangered Species Act (USFWS 1999). Upon listing, little was known about the biology and habitat requirements of this subspecies within its range along the Front Range of Colorado and southeastern Wyoming. Since listing, a number of projects (e.g., long-term monitoring, surveying, and movement studies) have collected valuable information throughout Colorado (Schorr 2001, Meaney 2000, Shenk and Sivert 1999). However, information on specific habitat requirements and their relationship to the distribution, density, survival and reproduction of PMJM is still lacking. The threatened status of PMJM requires management decisions be made despite our limited knowledge. In particular, the species and its habitat are subject to habitat conservation plans (HCPs). HCPs are written for endangered and threatened species to compensate for authorized "take" through mitigation practices (Bingham and Noon 1998). HCPs require the use of the "best available" science to determine the biological needs of target species (Harding et al. 2001). Collection ofreliable information for the species will improve the mitigation practices developed for HCPs. Well-designed habitat manipulation experiments provide the strongest inference to determine cause and effect relationships. Understanding of the species habitat requirements will enable the development of effective mitigation strategies. A manipulation experiment was conducted in Douglas County, Colorado (Columbine Open Space) during 2002 and 2003 to advance our understanding of PMJM habitat requirements. We manipulated sections of the riparian habitat and adjacent grassland within the I 00-year flood plain. The site was manipulated by adding patches (3 m x 2.43 m) of artificial resources (food and cover). Time limitations of only a 2-year study were inadequate for vegetation to establish and limited funding (cost of planting and sustaining vegetation) restricted this manipulation experiment to simulating habitat with temporary structures and food supplementation. The treatments were placed in areas of low use based on past monitoring studies conducted by the Colorado Division of Wildlife (CDOW) during 1998-2000. PMJM were radio tracked before and after the manipulation to determine if PMJM movements were altered through the addition of resources. We propose two primary objectives: 1) determine how the presence ofresource additions influences the distribution of individual PMJM within a population, and 2) to quantify habitat characteristics of PMJM on a microhabitat scale. We want to examine if the distribution of individual PMJM can be altered in response to the addition of resources (food and cover) and to quantify relevant microhabitat characteristics where PMJM have been detected. �3 STUDY AREA The study was conducted within the riparian habitat within Columbine Open Space, owned by Douglas County Open Space managed by the CDOW and the adjacent grassland. Columbine Open Space was selected because PMJM were monitored for 3 years by the CDOW ( 1998-2000), providing sitespecific information on PMJM locations before this manipulation experiment.· METHODS PMJM were trapped using non-folding Sherman live traps (7.6 cm x 8.9 cm x 22.9 cm) placed 5m apart along approximately 0.5 km transects adjacent to both sides of East Plum Creek for a minimum of 5 consecutive nights. Trapping procedures were in accordance with the guidelines published by the USFWS (1999). Species other than PMJM were recorded by trap location and immediately released. The following information was recorded for captured PMJM: unique identification, trap location, weight, sex, age, and reproductive condition. PMJM were scanned for a passive integrated transponder (PIT) tag. Newly captured individuals were marked by inserting a unique PIT-tag. Individuals ;::18 grams were anesthetized with isoflurane and fitted with a 1-g radio transmitter (Holohil Systems Ltd Ontario, Canada). All methods were approved by the Animal Care and Use Committee of Colorado State University (Authorization Number A3572-0l). Radio telemetry was used to monitor locations of individuals for a 21-day period, the battery life of the radio transmitters. Observers attempted to stay approximately 3 m from the radio-tagged individual to avoid influencing PMJM movement. Observations taken 3 m or greater from PMJM did not influence movement (T. Shenk, CDOW personal. comm.). The following information was recorded at each relocation: individual identification, time, weather, and surrounding vegetation. All data were combined into a geographical information system (GIS) database using ArcView®3 .2 (Environmental Systems Research Institute, Redlands, California, U.S.A.). The manipulation experiment consisted of 5 phases: 1) selection of areas oflittle or no previous use by PMJM based on CDOW location data (1998-2000) collected at Columbine Open Space, 2) recording of pre-treatment location data of radio-tagged individuals for 6 nights, 3) selection of treatment plot location based on pre-treatment and CDOW location data, 4) addition ofresources to treatment plots, and 5) recording of post-treatment location ofradio-tagged individuals. Two sessions (June and July) of the manipulation experiment were conducted each year. A digital map with a grid cell size of 9 m x 9 m was constructed for the entire study site with ArcView®3.2 (Environmental Systems Research Institute, Redlands, California, U.S.A.) software. CDOW location data was pooled into a single coverage over the grid to establish areas::: 1,000 m2 containing only low use cells (<2 locations/cell based on CDOW location data) within the 100-year flood plain. Location of treatment plots was selected with a stratified random design from a set of candidate cells meeting criteria developed to describe poor PMJM habitat (sparse vegetation and little food) within 60 m of East Plum Creek, and low historical use. The artificial cover, simulating vertical complexity, was constructed with wheat straw and tree branches distributed in a patch (3 m x 2.43 m). Burlap cloth was suspended 30 cm over the tree branches and straw. Food supplements composed of an equal mixture of whole wheat, dehydrated alfalfa pellets and sweet feed were placed on cardboard trays (0.16 m x 0.3 m) within the straw and branches as an attractant and a source of high protein. The dimensions of the treatments were selected to balance the manageability of construction and decrease the chance of inter and intra-species domination within a treatment. Quantification of microhabitat variables in areas of high use were examined by comparing a random sample of cells (9 m x 9 m) containing ::: 99 % of PMJM locations for each session to a random �4 sample of cells where no PMJM locations detected. Two line transects were randomly placed in each selected cell with 6 quadrat frames (SO cm x 20 cm) evenly distributed per line transect (Daubenmire 1959). The variables measured in each cell included percent bare ground, shrub, grass, and forb cover and vegetation composition. The location data were analyzed using linear regression. The response variable was the number of locations detected in a cell. A suite of candidate models was developed as predictors of the response variable. Akaike's information criterion (AIC) was applied to select the best "approximating" model (Burnham and Anderson 2002). The independent habitat variables of interest for the models included distance from the center of the cell to the nearest water, area and juxtaposition of nearest shrub, and presence of wetland grasses in the cell. Additional variables -included in the models were period (pre- or post-treatment), sex, session, and year. The microhabitat data collected from the Daubenmire plots were analyzed with Proc GLM (SAS 2002) to test for differences in means among areas of high use and no use by PMJM. PRELIMINARY RESULTS A comparison of microhabitat characteristics from a random sample of "high-use" and "no-use" areas indicated a greater (P < 0.0001) shrub canopy cover in "high-use" areas verses "no-use" areas (47.7% ± 29.8%, 12.6% ± 14.11 %, respectively). Further, "high-use" areas had greater basal cover (P = 0.013) and bare ground (P = 0.0459) and "no-use" areas contained a greater (P = 0.0331) abundance of forb canopy cover. We conducted a manipulation experiment where we constructed patches of artificial resources (food and cover) in areas without previous PMJM activity. PMJM were radio collared and located hourly before and after the addition of food and cover. The majority of PMJM movements were not influenced by the addition of resources in 2002. These results may be due to site fidelity or lack of exploratory movement to locate the additional resources LITERATURE CITED Bingham, B. B. and B. R. Noon. 1998. The use of core areas in comprehensive mitigation strategies. Conservation Biology 12:241-243. Burnham K. P. and D.R. Anderson. 2002. Model selection and multimodel inference. Second edition. Springer, New York, New York, USA. Daubenmire, R. 1959. A canopy-coverage method ofvegetational analysis. Northwest Science 33:4364. Harding, E., E. Crone, B. D. Elderd, J.M. Hoekstra, A. J. McKerrow, J. D. Perrine, J. Regetz, L. J. Rissler, A.G. Stanley, E. L. Walters and NCEAS Habitat Conservation Plan Working Group. 2001. The scientific foundations of habitat conservation plans: a quantitative assessment. Conservation Biology 15:488-500. Meaney, C. A. 2000. Monitoring for Preble's meadow jumping mice along South Boulder Creek and Four Ditches. Boulder, Colorado, USA. Report prepared for the Colorado Division of Wildlife. SAS Institute. 2002. SAS Version 8.2. SAS Institute, Cary, North Carolina, USA. Schorr, R. 2001 Meadow jumping mice (Zapus hudsonius preblei) on the U.S. Air Force Academy, El Paso County, Colorado, USA. Shenk, T. M. and M. Sivert. 1999. Movement patterns of Preble's meadow jumping mouse (Zapus hudsonius preblei) as they very across time and space. Annual Report to the Colorado Division of Wildlife. Fort Collins, Colorado, USA. U.S. Fish and Wildlife Service. 1999. Interim Survey Guidelines for Preble's meadow jumping mouse. U.S. Fish and Wildlife Service. Denver, Colorado, USA. �Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Project No. Work Package No. Colorado Task No. 2 : Cost Center 3430 : Mammals Research : Preble’s Meadow Jumping Mouse Conservation Effects of Resource Addition on Preble’s Meadow Jumping Mouse (Zapus hudsonius : preblei) Movement Patterns 0662 Federal Aid Project: N/A : Period Covered: July 1, 2003 - June 30, 2004 Author: Anne M. Trainor. Personnel: T. M. Shenk, K. R. Wilson, G. C. White All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT A thesis, entitled ‘Influence of resource supplementation on movements of Preble’s meadow jumping mouse (Zapus hudsonius preblei) and habitat use characteristics,’ was completed and submitted to Colorado State University in partial fulfillment of a Master of Science degree. The thesis is available from The Colorado Division of Wildlife Library or the Colorado State University Library. Included in this report is an abstract of the thesis. 1 �JOB PROGRESS REPORT INFLUENCE OF RESOURCE SUPPLEMENTATION ON MOVEMENTS OF PREBLE’S MEADOW JUMPING MOUSE (Zapus hudsonius preblei) AND HABITAT USE CHARACTERISTICS Anne M. Trainor ABSTRACT Riparian wetlands are complex ecosystems containing great species diversity that may easily be affected by anthropogenic disturbances. Preble’s meadow jumping mouse (Zapus hudsonius preblei) is a federally threatened species dependent upon riparian wetlands. It has been the subject of habitat management and conservation efforts involving restoration and mitigation projects along the eastern Front Range of Colorado and southeastern Wyoming. Although habitat improvements for Z. h. preblei are designed for multiple spatial scales, most knowledge about the species’ habitat requirements has been described at a broad landscape scale. In addition, few projects have directly evaluated the mouse’s response to restoration and mitigation projects. The first objective of this study was to determine how supplementation using artificial resources influences the spatial movement patterns of a Z. h. preblei population. Previous studies described Z. h. preblei use areas through live trapping. This study more precisely evaluated Z. h. preblei spatial use by applying radio telemetry within a riparian ecosystem. I conducted an experiment by constructing treatment plots of artificial resources (food and cover) in areas with no previous detections of Z. h. preblei during 3 prior years (1998-2000) of intensive monitoring. Z. h. preblei were radio collared and then located hourly during nightly activity periods before and after the addition of food and cover. The second objective of this study was to improve understanding about micro-habitat characteristics that Z. h. preblei use. During the resource supplementation experiment, Z. h. preblei response to treatment plots varied by year with only 1 of 13 radio-tagged individuals using supplemental resources during 2002 and 6 of 8 in 2003. The lower use in 2002 may have been due to drought conditions, which decreased available herbaceous cover and thus protection from predators. While treatment plot use increased in 2003, the overall use was relatively low when compared to natural, high-use areas. The mean proportion of treatment plot use in 2003 was = 5.9% (SE =1.4%, range = 0% to 12%). Limited use of treatment plots may have been due to site fidelity and minimal exploratory movements by Z. h. preblei or to elevated predation risk. A comparison of micro-habitat characteristics from random samples of high-use and no-use areas indicated that areas used intensely by Z. h. preblei were closer to the center of the creek bed and positively associated with shrub, grass, and woody debris cover. Distance to center of the creek bed, percent shrub cover, and grass cover had the greatest relative importance of the habitat variables modeled in describing high-use areas. High-use areas contained three times the percent of grass cover as forb cover. There was a greater proportion of wetland shrub and grass cover in high-use versus no-use cells. However, proportion of cover type (shrub or grass) did not vary greatly between high-use and no use cells. Within riparian wetlands, the identification of key micro-habitat components that are intensively used by Z. h. preblei could improve conservation and management programs. In addition, results from the resource supplementation experiment suggest that TP pˆ mitigation and restoration may not ensure use of areas by threatened and endangered species. Therefore, understanding how species respond to changes in 2 �areas where they currently live will require development of more efficient and effective mitigation projects, and monitoring by conservation biologists and wildlife managers will be essential. Prepared by _______________________ Anne M. Trainor, Colorado State University 3 � https://cpw.cvlcollections.org/files/original/b97e5be10a7fc3f0ebee9c561b49ff4a.pdf c30aa4a6d7abfdf4fcee933246c8d0ee PDF Text Text 29 Colorado Division of Wildlife Wild.life Research Annual Report July 2000 JOB PROGRESS REPORT State of Colorado Cost Center 3430 Project No. W-153-R-13 Mammals Research Program Lynx Reintroduction Work Package No. ____0-=-67"""'0'-------Task No. l Post-Release Monitoring of Reintroduced Lynx Period Covered: July l, 1999 - June 30, 2000 Author: Tanya M. Shenk Personnel: Gene Byrne, Rick Kahn, Dave Kenvin, Jim Olterman, Scott Wait, Whitey Wannamaker, Margaret Wild, Dave Younkin ABSTRACT In an effort to reestablish a viable population of lynx (Lynx canadensis) in Colorado, 41 lynx were reintroduced into southwestern Colorado in 1999 and an additional 55 lynx released in Spring 2000. Release protocols were evaluated by closely monitoring each lynx released in 1999 through radiotelemetry. Number of mortalities and causes of each mortality were documented. With this new infonnation, release protocols were modified in an effort to release each lynx with the highest probability of survival. Three different release protocols were used in 1999. Differences in release protocol included the length of time animals were kept in the Colorado holding facility and timing of the release. Percent mortality due to starvation within six months of release date decreased with each modification of release protocols (75% under Protocol 1, 11 % under Protocol 2, 0% under Protocol 3 except for female lynx released pregnant). Of the 55 lynx released in 2000, 41 were released in April (Protocol 2) and 14 were released in May (Protocol 3) following a minimum of three weeks in the Colorado holding facility. Of the total 96 lynx released, 63 are being followed on a regular basis within Colorado, six male lynx have not been located since October 1999, one lynx possibly slipped a collar, and 26 lynx are known to have died. Known mortality factors included starvation (7), gunshot (3), vehicle collision (3), trauma (2), predation (1), and disease (l). Cause of death could not be determined for nine mortalities. Initial dispersal movement patterns of the lynx released in 1999 were extremely variable. Dispersal habitat used by the lynx released in 1999 were also highly variable, from high elevation Engelmann spruce/subalpine fir to Nebraska agricultural lands. Lynx released in 2000 have remained closer to their release site and fewer have been observed using atypical lynx habitats. Through snow-tracking efforts (221 snow-tracking days) in 1999 and 2000 we have located 147 kills, 371 beds, and 132 scats. Of the 147 kills, 75% were of snowshoe hare (Lepus americanus), 23% were pine (red) squirrel (Tamiasciurus hudsonicus), and the remaining 2% were made up of other mammals and birds. We collected 132 scat samples that will be analyzed later for content. No reproduction has been documented to date, however whether or not lynx have bred and had litters at some point remains unknown. �30 �31 POST-RELEASE MONITORING OF REINTRODUCED LYNX Tanya M. Shenk P. N. OBJECTIVE I. Post release monitoring of lynx released in Colorado. SEGMENT OBJECTIVES FY99-00 2. 3. 4. 5. Estimate first year survival rates oflynx reintroduced to Colorado. Identify first year mortality factors of lynx reintroduced to Colorado. Describe first year movement patterns of lynx reintroduced to Colorado. Refine and prioritize needed research components to develop sound management strategies for lynx in Colorado. 6. Prepare a Federal Aid Job Progress Report. PERFORMANCE INDICATORS FY99-00 I. 2. 3. 4. 5. 6. 7. 8. Survival estimates of lynx reintroduced to Colorado. Summary of mortality factors oflynx reintroduced to Colorado. Description and analysis of habitats used by lynx reintroduced to Colorado. Description of movement patterns for lynx reintroduced to Colorado. Reproduction estimates of lynx reintroduced to Colorado. Evaluation and modification of release protocols for reintroducing lynx. Sites selected for second year release of lynx Refinement and prioritization of needed research components to develop sound management strategies for lynx in Colorado. 9. Report on first year release of lynx in Colorado: movement patterns, habitat use, survival and reproduction INTRODUCTION In an effort to reestablish a viable population of lynx (Lynx canadensis) to Colorado, 41 lynx were reintroduced into southwestern Colorado in the spring of 1999 and an additional 55 lynx were released in Spring 2000. Monitoring of these lynx is crucial to evaluating the progress of this lynx reintroduction effort. The monitoring program will also provide information and data critical to improving release techniques to ensure the highest probability of survival for each individual lynx released in future years of the Colorado effort, and perhaps in other reintroduction efforts. The post-release monitoring program for the reintroduced lynx has two primary goals. The first goal is to obtain regular locations of released lynx. From these locations we will be able to determine how many lynx remain in Colorado and their locations relative to each other. Given this information and knowing the sex of each individual we will be able to assess the feasability of these lynx to form a breeding core from which a viable population might be established. Also from these data we can describe general movement patterns and habitats used. The second primary goal of the monitoring program is to estimate survival of the reintroduced lynx and, where possible, determine cause of mortality of reintroduced lynx. �32 Additional goals of the post-release monitoring program for lynx reintroduced to the southern Rocky Mountains include refining descriptions of habitat use and movement patterns, determining food habits, and obtaining information on reproduction. When the lynx establish home ranges that encompass their preferred habitat, more emphasis will be placed on refining descriptions of movement patterns and habitat use. Lynx is currently a species listed as threatened under the Endangered Species Act (ESA) of 1973, as amended (16 U.S. C. 1531 et. seq.)(U. S. Fish and Wildlife Service 2000). As a listed species, information specific to the ecology of the lynx in its southern range such as habitats used, movement patterns, mortality factors, survival, and reproduction in Colorado will be needed to develop recovery goals and conservation strategies for this species specific to its southern range. Thus, an additional objective of the post-release monitoring program is to develop conservation strategies relevant to lynx in Colorado. OBJECTIVES The initial post-release monitoring of reintroduced lynx will emphasize five primary objectives: 1. Assess and modify release protocols to enure the highest probability of survival. 2. To obtain regular locations of released lynx to describe general movement patterns and habitats used by lynx. 3. Determine causes of mortality occurring in reintroduced lynx. 4. Estimate survival of lynx reintroduced to Colorado. 5. Estimate reproduction of reintroduced lynx. Three additional objectives will run concurrently or become active after lynx become established in an area that encompasses their movements. These objectives include: 6. Better refine descriptions of habitats used by reintroduced lynx. 7. Better refine descriptions of daily and overall movement patterns of reintroduced lynx. 8. Describe food habits and prey of reintroduced lynx. The data collected during the post-release monitoring will be analyzed to evaluate habitat use, movement patterns, reproduction and survival. These data will be used to further the knowledge about habitat requirements for this species in the southern Rocky Mountains. Thus, the final objective for the postrelease monitoring plan is to: 9. Refine habitat protection recommendations and conservation strategies based on information collected from released lynx. STUDY AREA Five areas throughout Colorado were evaluated as potential lynx habitat (Byrne 1998). Criteria investigated in these five areas for comparison were (1) relative snowshoe hare densities (Reed at al., unpublished data), (2) road density, (3) size of area, (4) juxtaposition of habitats within the area, (5) historical records oflynx observations, and (6) public issues. Based on results from this analysis, the San Juan Mountains of southwestern Colorado were selected as the.release area for reintroducing lynx. Ten release sites within the San Juan Mountains were selected based on land ownership and accessability during time ofrelease for the 41 animals released in 1999. Of the 55 lynx released in Spring 2000, 45 were released at Rio Grande Reservoir and ten lynx were released at three sites west of the Continental Divide. Based on current locations of the majority of the released lynx, the core research area remains in the southern San Juan Mountains, however lynx may need to be captured from areas of non-suitable habitat in Colorado and adjacent states. �33 METHODS Assessment ofRelease Protocols A total of 41 lynx were released in 1999 at selected areas in the San Juan Mountains of southwestern Colorado. Prior to release each lynx was examined and age, sex, and body condition determined. Each lynx was fitted with a TelonicsTM VHF radio-collar for post-release monitoring. The collars were also equipped with a mortality switch that activates if the collar remains motionless for a period of four hours or more. Specific release sites were selected based on land ownership and accessability during times of release. Lynx were transported from the holding facility to the release site in individual cages. Release site location was recorded in Universal Trans Mercator (UTM) coordinates and identification of all other lynx released at the same location, on the same day, was recorded. Behavior of the lynx on release and movement away from the release site was documented. Monitoring of the survival and mortality factors (see below) of each lynx was used to modify release protocols in 1999 in an attempt to release each lynx with the highest probability of survival. Release protocols for the 55 lynx released in 2000 were developed from survival, mortality factor, and movement pattern data obtained from lynx released in 1999. Documenting Movement Patterns To obtain regular locations of released lynx to determine general movement patterns and habitats used by reintroduced lynx a combination of satellite, aerial and ground radio-tracking were conducted. Locations and general habitat descriptions of each location were recorded and mapped for all locations. All 41 of the lynx released in 1999 were monitored from the air through radio-tracking. Frequent flights (three times a week) were critical during the initial post-release periods because of the greater likelihood of dispersal and mortality in reintroduced carnivores. Every effort was made to locate every lynx during each flight during this period. Sixty days from the date of the last re~ease, aerial locations of the radio-collared lynx were to be determined two times per week for the remainder of the life of the transmitters. Flights were also conducted three times per week, weather permitting, to locate lynx during the snow-tracking field season (December through April) to aid in the snow-tracking efforts. When possible at least one observer flew with the pilot to become familiar with the terrain, to operate the radio telemetry receiver, and to record the global positioning system (OPS) locations of the lynx. Generally,- the pilot circled a strong telemetry signal and then bisected the circle activating the OPS unit when approaching directly overhead. The date and time of the beginning and ending of the flight, the time each collar was located, the UTM coordinates for each animal located, general weather conditions, primary overstory vegetation type, and name of the personnel were recorded. All locations were entered into a database for mapping and data analysis. Fifty-one of the 55 lynx released in spring 2000 were fitted with SirtrackTM dual VHF/satellite transmitter collars. The remaining four lynx were fitted with TelonicsTM VHF collars identical to those used on lynx released in 1999. Each dual collar weighed 137-156 grams. The satellite component of each collar is programmed to be active for 12 hours per week. The 12-hour active periods are staggered throughout the week, with approximately seven collars being active each day of the week. Signals from the collars allow for locations of the animals to be made via Argos, NASA, and NOAA satellites. The location information was processed by ServiceArgos and distributed daily to the Colorado Division of Wildlife through e-mail messages. Both the VHF and satellite transmitter in the dual collar has a mortality switch which is triggered by four or more hours of stationarity. Determining Causes of Mortality To determine causes of mortality occurring in reintroduced lynx every effort was made to locate and retrieve carcasses of dead lynx as soon as possible. When a mortality signal (75 ppm vs 50 ppm for the TelonicsTM VHF transmitters, 20bpm vs 40bpm for the SirtrackTM VHF transmitters, 0 activity for �34 Sirtrack™ PIT) was heard during either satellite, aerial or ground surveys, the location (UTM coordinates) was recorded. Ground crews located and retrieved the carcasses. The immediate area was searched for evidence of other predators and the carcass photographed in place before removal. Additionally, the mortality site was described, habitat associations, and exact location were recorded. Any scat found near the dead lynx that appeared to be from the lynx was collected. All carcasses were transported immediately to the Colorado State University Veterinary Hospital for a post mortem exam. Lynx carcasses were not frozen but kept cool. If carcasses were already frozen due to field conditions, this was noted on the field form. The objectives of the post-mortem examination were to 1) determine the cause of death and document with evidence, 2) collect samples for a variety of research projects, and 3) archive samples for future reference (research or forensic). The gross necropsy and histology were performed by, or under the lead and direct supervision of a board certified veterinary pathologist. At least one research personnel from the Colorado Division of Wildlife involved with the lynx program was also present. In general, the protocol followed standard procedures used for thorough post-mortem examination and sample collection for histopathology and diagnostic testing. Some additional data/samples were routinely collected for research, forensics, and archiving. Other data/samples were collected based on the circumstances of the death (e.g., photographs, video, radiographs, bullet recovery, samples for toxicology or other diagnostic tests, etc.,). The CDOW retained all samples and carcass remains with the exception of tissues in formalin for histopathology, brain for rabies exam, feces for parasitology, external parasites for ID, and other diagnostic samples. Estimating Survival Survival rates of lynx reintroduced to Colorado will be estimated using the Kaplan-Meier method with staggered entries (Pollock et al. 1989). Documenting Habitat Use and Hunting Behavior More refined descriptions of habitats used by reintroduced lynx were obtained through snow-tracking of animals. Data were collected on habitats used, daybed and hunting bed locations, and travel corridors. Hunting and feeding behavior information was also collected by documenting prey taken, prey chases, relative abundance of prey (tracks and sightings), and use of carrion. Snow-tracking was conducted during February-May, 1999 and beginning again in November, 1999 through April 2000. Locations from the aerial-tracking were used to help ground-trackers "locate lynx tracks in the snow. One or more persons working together conducted the snow-tracking surveys. Snowmobiles, where permitted, were used to gain the closest possible access to the lynx tracks without disturbing the animal. From that point, snowshoes were used by the tracking team to reach the tracks. Once tracks are found, the ground crew back-tracked the animal. Back-tracking avoided the possibility of disturbing the lynx by moving away from the animal rather than toward the animal. However, monitoring of the lynx through radiotelemetry was also used to assure that ground crews stayed a sufficient distance away from the lynx in the event the lynx might double back on its tracks. If the lynx began to move in response to the observers, the observers retreated. If the lynx began to move and the movement did not appear to be a response to the observers, the crew continued to follow and record locations, habitats used, and behavioral information for as long as possible. Locations oflynx tracks were recorded using a Garmin XL12 GPS and 7.5° topographic map. Habitat descriptions included overstory and understory vegetation and seral stage. Locations and behavioral observations that could be interpreted from the tracks (e.g., chases, scent marking) were recorded. These data will be used for mapping and spatial analyses and analyzed to make inferences on how different habitats are used, frequency of use, daily movement patterns, hunting areas, daybed locations, den sites, and travel corridors. Data will also be used to document any changes in habitat use as animals begin to settle into a home range. �35 An attempt was made to locate tracks from all lynx. However, first priority was given to locating any animal that appeared to be consistently in the same location from aerial surveys. Such stationarity may indicate an injured, starving, or otherwise traumatized animal. Data on hunting behavior was collected by location of kills, food caches, chases, and through scat analysis. Prey from attempted and successful hunting attempts were identified by either tracks or prey remains. Information from scat analysis will also provide information on foods consumed. Scat samples were collected wherever found, recording location and individual lynx identification. Only part of the scat was collected, the remainder was left where found so as not to interfere with the possibility the scat was being used by the animal as a territory mark. Comparisons of food composition and percent occurrence will be made within and among individuals. Analyses of temporal, spatial, and individual differences will be conducted to provide information on feeding ecology of reintroduced lynx in the southern Rocky Mountains. Estimating Reproduction Reproductive status of all female lynx was determined prior to release through radiographs. All females known to be pregnant or thought to possibly be pregnant on release were monitored closely from their release through the following August to determine reproductive success. Females remaining within a limited area immediately after release through August were located and observed to look for accompanying kittens or a den site. Females that had been released in 1999 and were alive in spring 2000 were monitored for proximity to males during breeding season and for site fidelity to a given area during the denning period of May and June. Each female lynx from the 1999 releases were directly observed in summer 2000 over 3-5 different visits to look for accompanying kittens or evidence of denning. Locations of both males and females released in 1999 were evaluated during March and April 2000 to document proximity of males to females in an attempt to determine if breeding could have occurred. RESULTS Assessment ofRelease Protocols A total of 41 lynx were released in Colorado in 1999 under five different release protocols (Table I). The initial release protocol called for the immediate release of females once they passed veterinary inspection in Colorado. Males were to be held for a period of weeks until females established a territory, and then males were to be released near female territories. Four animals were released in early February, however, three of these died of starvation within six weeks of their release and the fourth was recaptured and returned to the holding facility where she recovered and was later re-released (Table 2). Reevaluation on the condition of animals released under the first protocol suggested that these animals may not have been in optimal physical shape when released. Therefore, a second release protocol was initiated whereby lynx were held at the Colorado holding facility for a minimum of three weeks and fed high quality diets to encourage weight gain. Most lynx gained considerable body weight while in captivity (Wild 1999). Nine lynx were released under this second protocol (Table 2). Of these nine lynx, one juvenile female died of starvation seven weeks after release. After the starvation death of the first lynx under the second protocol, a third release protocol was developed that called for releasing all subsequent lynx in the spring after a minimum stay in the holding facility of at least three weeks (Table I). A spring release would assure the lynx were released when prey was most abundant (i.e., young of the year would be most abundant and hibernating and migratory prey would be available). Twenty lynx were released under this protocol (Table 2). Additionally, six females were released under this third protocol that were known to be pregnant (Protocol 3P) and two that were possibly pregnant (3P?). No lynx reintroduced under Protocol 3 died of starvation within six months postrelease (Table 2). However, two of the six lynx released when pregnant died of starvation within six months post-release. �36 An assessment of the fates of each lynx under all five release protocols used in 1999 led to release protocols for lynx released in 2000. Release protocols 2 and 3 resulted in the fewest post-release (up to six months after release date) starvation mortalities (Table 2). The common element in both protocols was increased captivity time in the Colorado holding facility. The single starvation mortality for lynx released under Protocol 2 in 1999 was also the only juvenile released under that protocol and the only animal released in February (the other eight Protocol 2 lynx were released in March 1999). Thus, all lynx released in 2000 were released under either Protocol 2 or 3 but not before April I. Because of the high percentage of starvation mortalities in females pregnant on release (Table 2), we also attempted to avoid reintroducing lynx that were known to be pregnant. This was best accomplished by trying to have animals captured for the reintroduction effort in Canada and Alaska prior to their breeding season. Movement Patterns Through extensive aerial and satellite tracking, we continue to search and locate 63 of the 70 lynx with collars on and assumed to be alive (one lynx has presumably slipped her collar). We have 1206 satellite locations for 49 of the 51 lynx fitted with dual collars (2 satellite collars never worked after the lynx were released) and 1122 aerial VHF locations for all 96 reintroduced lynx. Six males from the 1999 releases have not been found since at least 1 October 1999. Possible reasons for not locating these six males include (I) long distance dispersal, beyond the areas currently being searched, (2) radio failure, or (3) destruction of the radio (e.g., run over by car). We continue to search for all missing lynx during both aerial and ground searches. Last known locations for each of the 70 lynx assumed to be alive are presented in Figure I. Initial dispersal movement patterns of the lynx released in 1999 were extremely variable. Dispersal habitat used by lynx released in 1999 has been highly variable, from high elevation Engelmann spruce/Subalpine fir to Nebraska agricultural lands. However, numerous travel corridors have been used repeatedly by more than one lynx, possibly suggesting route selection based on olfactory cues. Dispersal movement patterns of lynx released in 2000 were much less than those observed by lynx released in 1999. Most of 2000 releases have remained within an area encompassed by 100 km radius circle from the release locations. Most movement away from this core area has been to the north (Figure I). We currently have six lynx using areas near Interstate 70. Survival and Mortality Factors - Of the 96 lynx released, 26 mortalities have been recorded to date (Table 3). From the 1999 releases (41 animals) we have had 22 known mortalities (6 from starvation, 8 unknown, 3 gunshot, 2 hit by car, 2 trauma, and 1 predation). We have six missing males. We are following 13 of the lynx from the 1999 releases on a regular basis. From the 2000 releases (55 animals) we have four known mortalities (1 hit by car, 1 disease, I starvation, 1 unknown) and one animal that possibly slipped her collar. We are following the remaining 50 animals on a regular basis. Of the total seven confirmed starvation deaths, three were associated with animals released in less than ideal body condition and two were lynx less than one year old. Percent mortality due to starvation decreased with each modification of release protocols (75% under Protocol 1, 11 % under Protocol 2, 0% under Protocol 3). Necropsy results for lynx BC00F3, a female released on April 2, 2000 near Creede, Colorado indicated she died from pneumonic plague. The lynx was in fairly good condition, there was some abdominal fat, no muscle wasting, and the bone marrow had fat in it. The only gross lesion was an acute fibrinous pneumonia (i.e., lung infection of short duration). The lynx had probably only been sick a few days before it died. The carcass was recovered near her release site. Plague was diagnosed by flourescent antibody test and isolation of Yersinia pestis from lung and spleen samples. �37 Recaptures Three lynx have been recaptured and subsequently re-released since their initial release. Lynx BC99F6 was released in 1999 under Protocol l. Her behavior and incidental sightings by the public suggested the lynx was in poor condition. We trapped her using a TomahawkTM live trap baited with rabbit. She was recaptured the first night (March 25, 1999) we set the trap. On capture, we found she was severely emaciated. We anesthetized her with Telezol (2 mg/kg) and returned her to the Colorado holding facility. She was rehabilitated through diet. The lynx gained weight steadily and was re-released on May 28, 1999. She was hit by a car on Interstate 70 on July 19, 1999. Necropsy results indicated she was in excellent body condition at her time of death. Lynx AK.99M9 was released on May 12, 1999 and recaptured on March 24, 2000. Field observations by the lynx monitoring crew suggested that the lynx was severely emaciated. Live-trapping the lynx failed, so the lynx was darted with Telazol (3 mg/kg) using a Dan-Inject CO2 pistol. Physical examination revealed severe emaciation (6 kg). The lynx was returned to the Colorado holding facility and rehabilitated through diet. The lynx gained weight steadily and was re-released on May 3, 2000. Lynx AK.99F2 was released on May 7, 1999 and recaptured on April 18, 2000. Field observations by the lynx monitoring crew suggested that the lynx was emaciated. She was live-trapped with a TomohawkTM live trap with one nights effort. On capture, we found she was emaciated. We anesthetized her with Telezol (2 mg/kg) and returned her to the Colorado holding facility. She was rehabilitated through diet. The lynx gained weight steadily and was re-released on May 22, 2000. Habitat Use and Hunting Behavior February 1999-May 1999 Through snow-tracking, we were able to document habitat use, daily movement patterns, and hunting behavior of the earlier released lynx. Snow-tracking of lynx began shortly after the first release, Feb. 6, and continued until May 15, 1999. Although we tried to continue beyond May 15, efforts beyond this date did not yield any information because of either the lack of snow in the areas where the lynx were, or the snow conditions were too difficult to track in (hard, crusty, patchy). Because the majority (28) of the lynx were first released under Protocol 3, after May 6, the snow-tracking effort focused on the 13 lynx released prior to this date, under Release Protocols 1 and 2. Approximately 114 km of lynx tracks were followed. These tracks were from 11 different lynx, with kilometers tracked for any individual varying from 1 to 31 kilometers (Table 4). Two lynx (one female and one male) from Release Protocols 1 and 2 were never snow-tracked because we were either not able to locate the animals or because when we did locate them we could not readily access where they were. Daybeds and hunting beds were each located for eight of the lynx. Prey chases or kills were found for four lynx, scat samples were collected from five lynx, and possibly from a sixth. From the kills found and from initial examination of the scat samples, the lynx fed on snowshoe hare (Lepus americanus), pine (red) squirrel (Tamiasciurus hudsonicus), and waterfowl. All the snow-tracking effort was conducted on nine lynx released under Protocols 1 and 2. Any lynx released under Protocol 3 were released too late to track. November 1999 -April 2000 Ground crews tracked 13 of the lynx released in 1999 during this period (Table 4). Two other lynx were being located during this time but were not in snow. A total of 139 kills or chases were located, 75% were snowshoe hare, 23% were pine (red) squirrel, and the remaining 2% were made up of other mammals and birds. We collected 115 scat samples that will be analyzed for content. Lynx released in 2000 were released too late to snow track. �38 Reproduction Six lynx released under Protocol 3 in 1999 were known to be pregnant (Table 1, Release Protocol 3P). Two other females may have been pregnant, the radiographs were suggestive but inconclusive (Table 1, Release Protocol 3P?). Three of the six lynx known to have been pregnant on release in 1999 died within two months after release. Two starved and one was killed on the road (Table 2). Long distance movements and lack of stationarity in the movement patterns of the other three lynx known to have been pregnant on release in 1999 suggests these females did not have young with them by July 1999. Of the two females that might have been pregnant, movement patterns were not suggestive of a female rearing young. It is not known if any other females bred and/or had young once released, however no females snow-tracked November 1999 through April 2000 had young with them. From radiographs taken of the 35 females released in 2000, one female was known to be pregnant and three were possibly pregnant. Movement patterns suggest that none of these females have kittens with them as of July 2000. There were seven females released in 1999 that were alive during the Spring 2000 breeding season. All seven females were in close (< 5 km) proximity to a male during the breeding season and could have bred. The seven females were monitored closely for stationary movement patterns, indicative of denning, from May-July 2000. Ground trackers also walked in on all seven females for visual observations on a minimum of three occasions and two females were visited on five occasions. No kittens were observed. However, the question of whether they successfully bred or had kittens at some point in 2000 is unknown. One of these females has since died and three others have made movements of over 100 km. Although we are confident none of the six live females have kittens at this time, for further confirmation we will snow-track each of these females as soon as they are in areas with fresh snow to check for kitten tracks. Beginning in March 2000 both male and female lynx began to exhibit extensive movements (> 100 km) away from areas they had used throughout the winter. For example, female (AK.99F3) moved from the area near Grizzly Gulch she used throughout the winter to the Wolf Creek Pass area, a straight line distance of approximately 255km (Figure 3). Male YK99M3 moved from the area near the Climax mine which he had used throughout the winter to Taylor Mesa, a straight line distance of approximately 270km (Figure 4). Such movements by both females and males put them in close(< 5 km) proximity to a lynx of the opposite sex. Two isolated males did not move during March or April and thus were not in close proximity to a known female during reeding season. This was a male that had used the area in and adjacent to the northwest comer of Rocky Mountain National Park and a male that used the area around Cuchara, Colorado throughout the winter. DISCUSSION Monitoring of lynx reintroduced to southwestern Colorado is crucial to evaluating the progress of the lynx reintroduction. Monitoring of these released lynx provides information and data necessary for improving release techniques to ensure the highest probability of survival for each individual lynx released in future years, and perhaps in other areas. Lynx is currently a species listed as threatened under the ESA. Information collected on the progress of the lynx reintroduction program, including habitats used, movement patterns, mortality factors, survival, and reproduction, could also be used to help develop recovery goals and conservation strategies for this species specific to its southern range. Three release protocols were used in the reintroduction of lynx to Colorado in 1999. Release protocols were modified as new information became available from monitoring the released lynx through radio-telemetry and snow-tracking. Each modification of the release protocols decreased the percent of animals dying from starvation. The primary element in later, more successful release protocols was an increased time in captivity at the Colorado holding facility. Increasing the amount of time lynx were held in �39 the Colorado holding facility provided each lynx with an opportunity to increase body weight and acclimate to the climate, elevation, and local conditions of the environment they would be released into. Although most lynx were housed in individual pens, with a few sharing a pen with one other lynx, the holding facility also allowed the lynx to hear and smell each other throughout this acclimation period. Such contact may have provided time for social interactions to occur. Such social interactions. may improve the likelihood these animals could form a breeding population. Post-release monitoring provided preliminary information on habitat use specific to Colorado that might later be used to refine habitat protection and management recommendations specific to Colorado. However, caution must be used in interpreting the information collected to date on habitats used by the introduced lynx. The aerial locations and snow-tracking results do provide some information but may also reflect behavior of displaced animals. General observations to note may be repeated use by multiple lynx of certain travel corridors and lack of use of tundra areas for any length of time. Both these habitat use characteristics have been noted for naturally occurring lynx populations. Preliminary data collected on kills suggests the reintroduced lynx are feeding on their preferred prey species, snowshoe hare and pine (red) squirrel in similar proportions as those reported for northen lynx during lows in the snowshoe hare cycle (Aubry et al., 1999). Caution must be used in interpreting the proportion of identified kills. Such a proportion ignores other food items that are consumed in their entirety. Through snow-tracking we have evidence that lynx are mousing and several of the fresh carcasses have yielded small mammals in the gut on necropsy. Nearly all the scat samples collected have been found through snow-tracking efforts and thus are representative of winter diet only. However, the summer diet of lynx has been documented to include less snowshoe hare and more alternative prey than in winter (Mowat et al., 1999). The extreme movements observed by both females and males in March and April 2000 may have been related to breeding behavior. March and April are the natural breeding periods for northern lynx (Tumlison 1987). We do not know if any of the females bred or had kittens but we are fairly sure that no female has kittens at this time. With only seven females from the 1999 releases in the wild in spring 2000 it was not unexpected that there might not be successful reproduction in 2000. During the summer of 2000, some lynx that were released in 1999 and had been faithful to a given area have made large movements away from these areas. Extensive summer movements away from areas used throughout the rest of the year have been documented by native lynx in Wyoming and Montana (Squires and Laurion 1999). Proposed monitoring and research include continued aerial radiotelemetry to document current locations and movement patterns, documentation of mortalities and causes of death, use of snow-tracking to document habitat use and hunting behavior, and further assessment of snowshoe hare densities in the state. The habitats used by the lynx will continue to be identified, mapped, and analyzed. These data will be used to further the knowledge about habitat requirements and preferences for this species in the southern Rocky Mountains. This information will be used to identify other blocks of potential habitat located throughout the Southern Rocky Mountains and evaluate conflicts that might jeopardize the recovery of lynx in Colorado. If conflicts are identified, such information can be used to develop conservation strategies and recommend land management strategies to mitigate them. ACKNOWLEDGMENTS The Colorado lynx reintroduction program and post-release monitoring is a large project involving many people. John Mumma, former director of the Colorado Division of Wildlife was instrumental in the implementation of the program. Rick Kahn of the CDOW is the program leader. Many CDOW biologists, researchers, wildlife managers and other personnel are involved in the program and or have advised us in the development of the monitoring protion of the program including Bill Andree, Tom Beck, Gene Byrne, Bruce Gill, Dave Kenvin, Todd Malmsbury, Jim Olterman, Dale Reed, John Seidel, Scott Wait, Margaret �40 Wild. The Lynx Advisory Team members from outside the CDOW include Steve Buskirk, Jeff Copeland, Dave Kenny, Steve King, John Krebs, Brian Miller, Gary Patton, Jerry Mastel, Kim Poole, Rob Ramey, Rich Reading, John Weaver, and Mike Wunder. We thank Susan and Herman Dieterich of the Frisco Creek Wildlife Rehabilitation Center for the care and maintenance of the lynx while being held in Colorado. The aerial post-release monitoring has been conducted by state pilots including Dell Dhabolt, Jim Olterman, Matt Secor, Whitey Wannamaker, and Dave Younkin. Ground field crew members include Bob Dickman, Chris Parmater, Jake Powell, and Jennifer Zahratka. Jon Kindler and Anne Trainor of CDOW were most helpful in preparation of the maps. Funding has been provided by Vail Associates, Turner Foundation, Great Outdoors Colorado (GOCO), and the Colorado Division of Wildlife. LITERATURE CITED Aubry, K. B., G. M. Koehler, and J. R. Squires. 1999. Ecology of Canada lynx in southern boreal forests. in Ecology and Conservation of Lynx in the United States. General Technical Report for U.S. D. A. Rocky Mountain Research Station. University Press of Colorado. Byrne, G. I 998. Core area release site selection and considerations for a Canada lynx reintroduction in Colorado. Report for the Colorado Division of Wildlife. Mowat, G., B. G. Slough, and S. Boutin. 1996. Lynx recruitment during a snowshoe hare population peak and decline in southwest Yukon. Journal of Wildlife Management 60:441-452. Mowat, G. and B. G. Slough. 1998. Some observations on the natural history and behaviour of the Canada lynx, Lynx canadensis. Canadian Field Naturalist 112: 32-36. Mowat, G., K. G. Poole, and M. O'Donoghue. 1999. Ecology oflynx in northern Canada and Alaska. in Ecology and Conservation of Lynx in the United States. General Technical Report for U.S. D. A. Rocky Mountain Research Station. University Press of Colorado. Nava, J. 1970. The reproductive biology of the Alaska lynx. M.S. Thesis University of Alaska, Fairbanks. Pollock, K. H., S. R. Winterstein, C. M. Bunck, and P. D. Curtis. 1989. Survival analysis in telemetry studies: the staggered entry design. Journal of wildlife management 53: 7-15. Poole, K. G., G. Mowat, and B. G. Slough. 1993. Chemical immobilization oflynx. Wildlife Society Bulletin 21:136-140. Seidel, J., B. Andree, S. Berlinger, K. Buell, G. Byrne, B. Gill, D. Kenvin, and D. Reed. 1998. Draft strategy for the conservation and reestablishment of lynx and wolverine in the southern Rocky Mountains. Report for the Colorado Division of Wildlife. Slough, B. G. 1999. Characteristics of Canada lynx, Lynx canadensis, maternal dens and denning habitat. Canadian Field Naturalist 113:605-608. Squires, J. R. and T. Laurion. 1999. Lynx home range and movements in Montana and Wyoming: preliminary results. in Ecology and Conservation of Lynx in the United States. General Technical Report for U.S. D. A. Rocky Mountain Research Station. University Press of Colorado. Tumlison, R. 1987. Mammalian Species: Fe/is lynx. American Society of Mammalogists. U.S. Fish and Wildlife Service. 2000. Endangered and threatened wildlife and plants: final rule to list the contiguous United States distinct population segment of the Canada lynx as a threatened species. Federal Register 63, Number 58. Wild, M.A. 1999. Lynx veterinary services and diagnostics. Job Progress Report for the Colorado Division of Wildlife. Fort Collins, Colorado. �41 Table 1. Release protocols for l)'TIX released in southwestern Colorado in 1999. Protocol 1 Description Release females as soon as they pass veterinary inspection in Colorado. Release males once females appear to have settled into an area. 2 Release males or females after they have been held in Colorado holding facility for a minimum of 3 weeks. During this holding period, the lynx were fed high quality diets to encourage weight gain, assuring each l)'TIX would be released in optimal physical condition. Such a minimal holding period also provided an opportunity for the l)'TIX to acclimate to the climate, elevation, and local conditions of the environment they would be released into. Although most l)'TIX were housed in individual pens, with a few sharing a pen with one other l)'TIX, the holding facility allowed the l)'TIX to hear and smell each other throughout this acclimation period. Such contact may also have provided time for social interactions to occur. 3 All l)'TIX to be kept in the holding facility for not only the minimal three week period but until spring. A spring release would assure the l)'TIX were released when prey was most abundant (i.e., young of the year would be most abundant and hibernating prey would be available). Coupled with the minimum holding period of three weeks, these lynx would also be released when in optimal physical condition and after a period of acclimation to their new surroundings. 3P Pregnant females released under Protocol 3. 3P? Possibly pregnant females released under Protocol 3. Table 2. Summary of number of l)'TIX released under each release protocol and numbers of lynx mortalities six months post-release for lynx released into southwestern Colorado in 1999 and four months post-release for l)'TIX released in 2000. 1999 2000 Protocol 1 2 3 3P 3P? Total Number released Mortalities 6 months post-release (n, %) Number released Mortalities 4 months post-release (n, %) Female Male Starvation Other Female Male Starvation Other 3 3 8 6 2 22 1 6 12 3, 75% 1, 11% 0, 00/4 2,33% 0, 00/4 6, 14% 0,0% 0,0% 3, 15% 2,33% 0, 00/4 5, 12% 0 25 6 l 3 35 0 16 4 1,2% l, 100/4 0, 00/4 0,0% l, 1% 3, 7% 0,0% 0,0% 0,0% 3, 5% 19 20 �42 Table 3. Release and mortality information for lynx released into southwestern Colorado in 1999 and 2000. Mortality Infonnation Release Infonnation Animal ID Sex Age Date Site Protocol Date Cause of death BC99Ml M 8mo 2/4/99 Goose Creek 1 2/24/99 staivation BC99F9 BC99F7 BC99F8 AK99F4 AK99M23 BC99F6 AK99Fl7 AK99F8 AK99Fl8 AK99Fl0 BC99M2 AK99F27 AK99M6 AK99Fl5 YK99F4 AK99Mll BC00F3 YKOOM5 YK99F3 YK99M6 AKOOF4 AK99Fl3 YKOOF17 BC99Ml0 AK99F25 YKOOF6 F F F F M F F F F F M F M F F M F M F M F F F M F F 2+ 3+ 9mo 1-2 1-2 2+ 2-3 5+ 1-2 l0mo 4+ lOmo 5 2-3 4-5 2-3 1 lOmos 2 3 lOmos lOmo 1 3-4 lOmo 2 2/3/99 2/3/99 2/20/99 5/7/99 5/14/99 2/4/99 5/10/99 5/10/99 5/14/99 5/12/99 3/19/99 5/14/99 5/13/99 5/14/99 5/13/99 5/12/99 4/2/00 4/2/00 5/10/99 5/13/99 5/22/00 5/12/99 4/17/00 3/19/99 5/7/99 4/2/00 Goose Creek Goose Creek Red Mtn Creek Sand Bench Love Lake Goose Creek First Fork First Fork Love Lake Lemon Res Red Mtn Creek Love Lake Vallecito Res Love Lake Vallecito Res Lemon Res Goose Creek Beaver Meadows First Fork Vallecito Res Rio Grande Res Lemon Res Rio Grande Res Red Mtn Creek Sand Bench Rio Grande Res 1 1 2 3p 3 1 3p 3p 3 3p 2 3 3 3 3 3 2 2 3 3 3 3 2 2 3 2 2/26/99 3/16/99 4/10/99 6/13/99 6/18/99 7/19/99 7/22/99 7/30/99 8/25/99 9/13/99 10/20/99 10/31/99 11/16/99 11/24/99 1/25/00 1/29/00 5/24/00 5/25/00 6/7/00 6/19/00 6/19/00 6/22/00 7/29/00 8/2/00 8/10/00 8/17/00 staivation staivation staivation staivation shot hit by car hit by car staivation trawna, emaciation wucnown. not staivation wucnown, not staivation shot shot blunt trauma predation, emaciation wucnown pneumonic plague staivation wucnown. not staivation wucnown slipped collar? wucnown wucnown. not staivation wucnown wucnown. not staivation hit br car Table 4. Habitat use and hunting behavior as described by summarizing kills, mousing activity, territory marks, hunting beds, day beds, and chases for each lynx tracked. Total number of days tracked to date and number of scat samples collected are also summarized. Data presented here are for the 1999-2000 snowtracking field season. Tracking Period No. of lynx tracked Kills Beds Scats Tracking Days Feb 99 - May 99 11 8 71 17 84 Nov 99 - Apr 00 13 139 300 115 137 �. -- _J --- ---- . - I --- -- -- - - _/ .,l_ .. ,--, - -, WYOMING • I, ---r '- 7 I i-r-~';'l---.....L i I --··t~ - 7, ., __ !_ ,J ' -- ·•·-. ,--.,_I • /t£WMEXICO • i I I /-------/ j -- ! s Figure 1. Most recent locations, as of August 17, 2000, of the 70 lynx known or assumed to be alive from the 96 lynx reintroduced to southwestern Colorado in 1999 and 2000. Black triangles indicate last known VHF location, black circles indicate last known satellite location. Black lines are Colorado highways, grey lines are county boundaries. Each location is identified with the animal code. �! __ L _ _ ___ : j -1 l ! ! l- - -- J - . ------ ~ i NEBRA~KA WfOMING i ---~r· '' I ,I :~. 't! -- -- -~ --- I I -- -- - -1 -------- - - --/~EW::o I i ? i • ; I 'I l ? <._ \ i Figure 2. Locations of all 26 known lynx mortalities from the 96 lynx reintroduced to southwestern Colorado in 1999 and 2000. Different symbols indicate different causes of death: starvation (e), hit by car(+), predation(*), disease (A), gunshot(*), and unknown(?). Dark lines within the Colorado border are highways, grey lines are county lines. �NEBRASKA WYOMING I~ NEW MEXICO ------- ·-·-;------ .,.r---.L~ s Figure 3. Movements of lynx female AK.99F3 from her release to August 2000. Smallest circles are oldest locations with circles increasing in size as date becomes more recent. Largest circle is most recent location. Black lines are Colorado highways, grey lines are county boundaries. �NEBRASKA WYOMING ·1-----(~~~- -~\ f- NEW MEXICO Figure 4. Movements of lynx male YK99M3 from his release to August 2000. Smallest circles are oldest locations with circles increasing in size as date becomes more recent. Largest circle is most recent location. Black lines are Colorado highways, grey lines are county boundaries. �7 Colorado Division of Wildlife Wildlife Research Report July 2001 and July 2002 PROGRESS REPORT State of_._ _ _ _ _----=Ca.ao=lo=r=a=do.:.,__ _ _ __ Division of Wildlife - Mammals Research Work Package No._ _---"-0-=-67'-'0.....__ _ _ __ Lynx Conservation Task No. ---------=-1_ _ _ _ _ __ Post-Release Monitoring of Lynx Reintroduced to Colorado Period Covered: January 1, 2001-December 31, 2001 Author: Tanya M. Shenk, Ph. D. Personnel: A. B. Franklin, L. Gephert, R. Kahn, A. Keith, D. Kenvin, G. Miller, J. Olterman, M. Secor, C. Wagner, S. Wait, G. C. White, D. Younkin Interim Report - Preliminary Results This work continues, and precise analysis ofdata has yet to be accomplished. Manipulation or interpretation of these data beyond that contained in this report should be labeled as such and is discouraged. ABSTRACT In an effort to establish a viable population of lynx (Lynx canadensis) in Colorado 96 lynx were reintroduced into southwestern Colorado in 1999 and 2000. Release protocols were evaluated by monitoring released individuals through radiotelemetry. Numbers of mortalities and causes of death were documented and this information used to modify subsequent release protocols in an effort to attain the highest probability of survival for released lynx. In general, release protocols were modified by increasing length of time lynx were kept at the Colorado holding facility, delaying time of release to spring, and releasing non-pregnant females. Mortality due to starvation decreased as earlier protocols were modified. A suite of hypotheses was developed to model early survival and factors that may have influenced survival, including sex, age on capture, pregnancy, time spent in the Colorado holding facility, and release time. Models were evaluated using AICc model selection and model averaging used to estimate survival rates. There have been 39 confirmed deaths. Human-caused mortality factors such as gunshot and vehicle collision are the highest cause of death for lynx> 8 months post-release. Locations of each lynx were collected through aerial- or satellite-tracking to document movement patterns. Initial dispersal movement patterns and distances traveled by lynx released in 1999 were highly variable and • more extreme than movements oflynx released in 2000. Movement patterns suggest lynx are pairing in March, but successful reproduction has not been do~umented to date. Snow-tracking results indicate the primary winter prey are snowshoe hare (Lepus americanus) and red squirrel (Tamiasciurus hudsonicus), with waterfowl and other mammals and birds forming a minor part of the winter diet. Site-scale habitat data collected from snow-tracking efforts indicate Engelmann spruce (Picea engelmannii) and subalpine fir (Abies lasiocarpa) are the most common forest stands used by lynx in southwestern Colorado. There is a seasonal trend in use of willows (Salix spp.) with use peaking in November and being at its lowest in May and June. ��9 Post-Release Monitoring of Lynx Reintroduced to Colorado Annual Progress Report for the U.S. Fish and Wildlife Service December 2001 Interim Report - Preliminary Results This work continues, and precise analysis ofdata has yet to be accomplished. Manipulation or interpretation of these data beyond that contained in this report should be labeled as such, and is discouraged. Tanya M. Shenk Mammals Research Colorado Division of Wildlife Abstract In an effort to establish a viable population oflynx (Lynx canadensis) in Colorado 96 lynx were reintroduced into southwestern Colorado in 1999 and 2000. Release protocols were evaluated by monitoring released individuals through radiotelemetry. Numbers of mortalities and causes of death were documented and this information used to modify subsequent release protocols in an effort to attain the highest probability of survival for released lynx. In general, release protocols were modified by increasing length of time lynx were kept at the Colorado holding facility, delaying time ofrelease to spring, and releasing non-pregnant females. Mortality due to starvation decreased as earlier protocols were modified. A suite of hypotheses was developed to model early survival and factors that may have influenced survival, including sex, age on capture, pregnancy, time spent in the Colorado holding facility, and release time. Models were evaluated using AICc model selection and model averaging used to estimate survival rates. There have been 39 confirmed deaths. Human-caused mortality factors such as gunshot and vehicle collision are the highest cause of death for lynx >8 months post-release. Locations of each lynx were collected through aerial- or satellite-tracking to document movement patterns. Initial dispersal movement patterns and distances traveled by lynx released in 1999 were highly variable and more extreme than movements of lynx released in 2000. Movement patterns suggest lynx are pairing in March, but successful reproduction has not been documented to date. Snow-tracking results indicate the primary winter prey are snowshoe hare (Lepus americanus) and red squirrel (Tamiasciurus hudsonicus), with waterfowl and other mammals and birds forming a minor part of the winter diet. Site-scale habitat data collected from snow-tracking efforts indicate Engelmann spruce (Picea engelmannii) and subalpine fir (Abies lasiocarpa) are the most common forest stands used by lynx in southwestern Colorado. There is a seasonal trend in use of willows (Salix spp.) with use peaking in November and being at its lowest in May and June. Introduction In an effort to establish a viable population of lynx (Lynx canadensis) in Colorado (Seidel et al. 1998), 41 lynx were reintroduced into southwestern Colorado in the winter and spring of 1999 and an additional 55 lynx were released in April and May of 2000. Post-release monitoring of these lynx is crucial to evaluating the progress of this reintroduction effort. The monitoring program also provides information and data critical for improving release techniques to ensure the highest probability of survival for each individual lynx released in the Colorado effort, and perhaps in other reintroduction efforts. The post-release monitoring program for the reintroduced lynx has 2 primary goals. The first goal is to determine how many lynx remain in Colorado and their locations relative to each other. Given this information and knowing the sex of each individual we can assess whether these lynx can form a breeding core from which a viable population might be established. From these data we can also describe general movement patterns and habitats used. The second primary goal of the monitoring �program is to estimate survival of the reintroduced lynx and, where possible, determine cause of mortality of reintroduced lynx. Such information will help in assessing and modifying release protocols and management of lynx once they have been released. . Additional goals of the post-release monitoring program for lynx reintroduced to th~ southern Rocky Mountains include refining descriptions of habitat use and movement patterns, determining hunting habits, and obtaining information on reproduction. When the lynx establish home ranges that encompass their preferred habitat, more emphasis will be placed on refining descriptions of movement patterns and habitat use. Lynx is listed as threatened under the Endangered Species Act (ESA) of 1973, as amended ( 16 U.S. C. 1531 et. seq.) (U.S. Fish and Wildlife Service 2000). As a listed species, information specific to the ecology of the lynx in its southern range such as habitats used, movement patterns, mortality factors, survival, and reproduction in Colorado will be needed to develop recovery goals and conservation strategies for this species specific to its southern Rocky Mountain range. Thus, an additional objective of the post-release monitoring program is to develop conservation strategies relevant to lynx in Colorado Objectives The initial post-release monitoring of reintroduced lynx will emphasize five primary objectives: 1. Assess and modify release protocols to enure the highest probability of survival for each lynx released. 2. Obtain regular locations ofreleased lynx to describe general movement patterns and habitats used by lynx. 3. Determine causes of mortality in reintroduced lynx. 4. Estimate survival of lynx reintroduced to Colorado. 5. Estimate reproduction oflynx reintroduced to Colorado. Three additional objectives will be emphasized after lynx display site fidelity to an area: 6. Refine descriptions of habitats used by reintroduced lynx. 7. Refine descriptions of daily and overall movement patterns of reintroduced lynx. 8. Describe hunting habits and prey of reintroduced lynx. Information gained to achieve these objectives will form a basis for the development of lynx conservation strategies in the southern Rocky Mountains. Study Area Five areas throughout Colorado were evaluated as potential lynx habitat (Byrne 1998). Criteria investigated in these 5 areas for comparison were (1) relative snowshoe hare densities (Reed at al., unpublished data), (2) road density, (3) size of area, (4)juxtaposition ofhapitats within the area, (5) historical records of fynx observations, and (6) puolic issues: Based on results from this analysis, the San Juan Mountains of southwestern Colorado were selected as the release area for reintroducing lynx. Ten release sites within the San Juan Mountains were selected based on land ownership and accessability during time of release for the 41 animals released in 1999. Of the 55 lynx released in spring 2000, 45 were released at Rio Grande Reservoir and 10 lynx were released at 3 sites west of the Continental Divide. Based on current locations of the majority of the released lynx, the core research area remains in the southern San Juan Mountains. Methods Reintroduction Effort A total of 96 lynx were released at selected areas in the San Juan Mountains of southwestern Colorado (Table 1). Estimated age, sex and body condition were ascertained and recorded for each lynx prior to release (see Wild 1999). Specific release sites were selected based on land ownership and accessibility during times of release. Lynx were transported from the holding facility to the release site in cages (usually!, occasionally 2 lynx per cage). Release site location was recorded in Universal �11 Transverse Mercator (UTM) coordinates and identification of all other lynx released at the same location, on the same day, was recorded. Behavior of the lynx on release and movement away from the release site were documented. Table 1. Colorado lynx reintroduction effort. Assessment of Release Protocols Year Females Males TOTAL In 1999, lynx were released under 5 41 1999 22 19 different release protocols (Table 2). Protocol 1 55 2000 35 20 called for the immediate release of females once TOTAL 57 39 96 they passed veterinary inspection in Colorado. Males were to be held for a period of weeks u_ntil females established a territory, and then males were to be released near female territories. Four lynx were released under this protocol with poor survival. Protocol 2 was developed whereby lynx were held at the Colorado holding facility for a minimum of 3 weeks and fed high quality diets to encourage weight gain. Nine lynx were released under Protocol 2. After a starvation death under Protocol 2, Protocol 3 was developed, requiring the 3-week minimum holding time and high-quality feeding of Protocol 2 plus a release date no earlier than May 1. A spring release would assure that lynx were released when prey was most abundant (i.e., young of the year would be most abundant and hibernating and migratory prey would be available). Twenty lynx were released under Protocol 3. Additionally, 6 females were released under Protocol 3 that were known to be pregnant (Protocol 3P) and 2 that were possibly pregnant (Protocol 3P?). An assessment of the fates of each lynx under all 5 release protocols used in 1999 led to release protocols for lynx released in 2000. Release protocols 2 and 3 resulted in the fewest post-release (up to 8 months after release date) starvation mortalities. The common element in both protocols was increased captivity time in the Colorado hold~ng facility. The single starvation mortality for lynx released under Protocol 2 in 1999 was also the only juvenile released under that protocol and the only animal released in February (the other 8 Protocol 2 lynx were released in March 1999). Thus, all lynx released in 2000 were released under either Protocol 2 or 3 but not before April 1. Because of the high percentage of starvation mortalities in females pregnant on release, we also attempted to avoid reintroducing lynx that were known to be pregnant. This was best accomplished by trying to have animals captured for the reintroduction effort in Canada prior to their breeding season. Table 2. Release protocols for lynx released in southwestern Colorado in 1999 and 2000. Protocol Description 1 Release females as soon as they pass veterinary inspection in Colorado. Release males once females appear to have settled into an area. 2 Release males or females after they have been held in Colorado holding facility for a minimum of 3 weeks and fed a high quality diet. 3 Release males or females after they have been held in Colorado holding facility for a minimum of 3 weeks, fed a high quality diet, and released no earlier than May 1. 3P Pregnant females released under Protocol 3. 3P? Possibly pregnant females released under Protocol 3. To evaluate the efficacy of the changes in release protocols we developed a series of a priori hypotheses concerning factors that affected lynx survival up to 8 months post-release. These factors included (1) the timing ofrelease (winter vs spring), (2) age oflynx released ( adults vs. kittens), (3) sex oflynx released, (4) whether or not females were released while pregnant and the interaction of pregnancy and age of the female (adult vs. kitten), and (5) the duration of holding time in the Colorado facility. A series of 11 models were developed using various combinations of these factors. We used �12 AICc (Burnham and Anderson 1998) as the model selection criterion to select the model that best explained the data. Movement Patterns To determine general movement patterns and habitats used by reintroduced lynx, regular locations of released lynx were collected through a combination of aerial, satellite and ground radiotracking. Locations and general habitat descriptions at each location were recorded and mapped. Frequent flights (at least 2 times per week) were critical during the initial post-release periods because of the greater likelihood of dispersal and mortality in reintroduced carnivores during this period. Every effort was made to locate every lynx each flight during this period. All 41 of the lynx released in the winter and spring of 1999 were fitted with Telonics™ VHF radio-collars, equipped with a mortality switch that activates if the collar remains motionless for 4 hours or more. Fifty-one of the 55 lynx released in the spring 2000 were fitted with Sirtrack™ dual satelliteNHF radio-collars (the other 4 lynx were fitted with Telonics™ VHF collars). These collars also had a mortality indicator switch that operated on both the satellite and VHF mode. The satellite component of each collar was programmed to be active for 12 hours per week. The 12-hour active periods were staggered throughout the week, with approximately 7 collars being active each day of the week. Signals from the collars allowed for locations of the animals to be made via Argos, NASA, and NOAA satellites. The location information was processed by ServiceArgos and distributed to the CDOW through e-mail messages. Survival and Mortality Factors When a mortality signal (75 ppm vs. 50 ppm for the Telonics™ VHF transmitters, 20 bpm vs. 40 bpm for the Sirtrack™ VHF transmitters, 0 activity for Sirtrack™ PTT) was heard during either satellite, aerial or ground surveys, the location (UTM coordinates) was recorded. Ground crews then located and retrieved the carcass as soon as possible. The immediate area was searched for evidence of other predators and the carcass photographed in place before removal. Additionally, the mortality site was described, habitat associations, and exact location were recorded. Any scat found near the dead lynx that appeared to be from the lynx was collected. All carcasses were transported immediately to the Colorado State University Veterinary Hospital for a post mortem exam to 1) determine the cause of death and document with evidence, 2) collect samples for a variety ofresearch projects, and 3) archive samples for future reference (research or forensic). The gross necropsy and histology were performed by, or under the lead and direct supervision of a board certified veterinary pathologist. At least one research personnel from the Colorado Division of Wildlife involved with the lynx program was also present. The protocol followed standard procedures used for thorough post-mortem examination and sample collection for histopathology and diagnostic testing (see Shenk 1999 for details). Some additional data/samples were routinely collected for research, forensics, and archiving. Other data/samples were collected based on the circumstances of the death (e.g., photographs, video, radiographs, bullet recovery, samples for toxicology or other diagnostic tests, etc.). The CDOW retained all samples and carcass remains with the exception of tissues in formalin for histopathology, brain for rabies exam, feces for parasitology, external parasites for ID, and other diagnostic samples. Survival rates oflynx reintroduced to Colorado were estimated using the Kaplan-Meier method with staggered entries (Pollock et al. 1989) in Program MARK (White and Burnham 1999). Recaptures Recaptures were attempted on lynx that were either in poor body condition or need to have their radio collars replaced. Methods ofrecapture included trapping using a Tomahawk™ live trap baited with a rabbit, and darting lynx with Telazol (3 mg/kg) using a Dan-Inject CO2 pistol (modified from Poole et al. 1993 as recommended by M.Wild, DVM). Hounds trained to pursue felids were also used to tree lynx for capture. Treed lynx were immobilized with Telazol or medetomidine (0.09mg/kg) and �21 JOB PROGRESS REPORT Stateof _ _ _ _ _____,C=o=l=or=a=d=o_ _ _ __ Division of Wildlife - Mammals Research Work Package No. _06~7_0_ _ _ _ _ _ __ Lynx Conservation Task No. _ _ _ _ _~1_ _ _ _ _ _ __ Post-Release Monitoring of Lynx Reintroduced to Colorado Period Covered: July 1, 2002 - June 30, 2003 Author: Tanya M. Shenk Personnel: R. Dickman, R. Kahn, A. Keith, G. Miller, C. Wagner, S. Wait, D. Younkin Interim Report - Preliminary Results This work continues, and precise analysis ofdata has yet to be accomplished. Manipulation or interpretation of these data beyond that contained in this report should be labeled as such and is discouraged ABSTRACT Reproduction is critical to the success of any reintroduction effort if a self-sustaining, viable population is the ultimate goal of the conservation effort. As of winter 2002-2003, no reproduction had been documented from lynx reintroduced to Colorado beginning in winter 1999. However, the low density of lynx present in Colorado by winter 2002-2003 limited the ability to answer the question of whether Colorado is suitable to sustain a viable lynx population because either insufficient habitat or lynx at too low a density to achieve reproductive success could have resulted in the lack of reproduction. Following an analysis of possible management options, it was decided that an augmentation of this reintroduction effort was necessary to eliminate an ambiguous result if successful reproduction had not occurred under densities such as exist in winter 2002-03. The reintroduction effort was augmented with 33 additional animals, released within the Core Area in April 2003, to increase lynx density so that the question of whether lynx can sustain viable populations in Colorado could be more definitively addressed. Based on dispersal patterns of lynx released in 2000, the second cohort, it was hypothesized that lynx released in the Core Area would show the necessary site fidelity to increase lynx densities to enhance the probability of successful reproduction. The first lynx kittens documented to be born to lynx reintroduced to Colorado were found on May 21, 2003. A total of 6 dens and 16 kittens were found in 2003. From results to date it can be concluded that CDOW has developed release protocols that ensure high initial post-release survival, and on an individual level lynx have demonstrated they can survive long-term in areas of Colorado. It had also been documented that reintroduced lynx could exhibit site fidelity, engage in breeding behavior and produce kittens. What is yet to be demonstrated is whether Colorado conditions can support the recruitment necessary to offset annual mortality for a population to sustain itself. Monitoring ofreintroduced lynx will continue in an effort to document such viability. �22 Post-Release Monitoring of Lynx (Lynx canadensis) Reintroduced to Colorado Tanya M. Shenk Mammals Research Colorado Division of Wildlife INTRODUCTION The Canada lynx (Lynx canadensis) occurs throughout the boreal forests of northern North America. Colorado represents the southern-most historical distribution of lynx, where the species occupied the higher elevation, montane forests in the state. Little was known about the population dynamics or habitat use of this species in their southern distribution. Lynx were extirpated or reduced to a few animals in the state by the late l 970's. Given the isolation of Colorado to the nearest northern populations, the Colorado Division of Wildlife (CDOW) considered reintroduction as the only option to attempt to reestablish the species in the state. A key question to be asked when considering the re-establishment of any species is, "What is different now from when they disappeared?" For lynx, the causative factor(s) of their extirpation may never be known. Many of the hypothesized factors, however, have changed substantially since the early and mid-l 900's. For example, widespread predator poisoning no longer occurs; conservation of wildlife habitat is now given much stronger consideration in public land management decisions; trapping and hunting are more strictly regulated and regulations enforced; and in some areas, at least, the passage of time has allowed the landscape to recover from abuses of the past, perhaps to a state that is more conducive to lynx survival. It must be acknowledged, however, that there may be other detrimental factors operating now that did not exist previously. In particular, increased human density and development have occurred in some areas and exotic diseases such as plague have been introduced in Colorado. The uncertainty surrounding the cause of the extirpation of lynx and the effects of current conditions in Colorado on lynx makes it impossible to predict with confidence whether Colorado has sufficient habitat to sustain viable population(s) oflynx. In order to perform the best test of this question the CDOW led a cooperative effort to reintroduce wild-trapped lynx from Canada and Alaska into southwestern Colorado beginning in 1999. It was hoped the effort would clarify whether or not Colorado is or is not suitable for sustaining viable lynx populations, provided the fate of the released animals could be determined. The goal of the Colorado lynx reintroduction program is to establish a viable population of lynx in this state. Evaluation of incremental achievements necessary for establishing viable populations is an interim method of assessing if the reintroduction effort is progressing towards success. There are seven critical criteria for achieving a viable population: ( 1) development of release protocols that lead to a high initial post-release survival of reintroduced animals, (2) long-term survival of lynx in Colorado, (3) development of site fidelity by the lynx to areas supporting good habitat in densities sufficient to breed, (4) reintroduced lynx must breed, (5) breeding must lead to reproduction of surviving kittens (6) lynx born in Colorado must reach breeding age and reproduce successfully, and (7) recruitment must be equal to or greater than mortality. Prior to the reintroduction, it was hypothesized that a minimum of 100 animals would need to be released for a fair evaluation of the suitable/unsuitable question. In 1999 and 2000, 96 lynx (57 females, 39 males) were released into the San Juan Mountains of southwestern Colorado. The 1999 cohort of 41 individuals scattered widely, and suffered a first year mortality of 17 (41 %) lynx (Shenk 2001). The 2000 cohort of 55 animals, being released into areas already occupied (although sparsely) by the previous year's animals, were more sedentary, and experienced a first year mortality of 10 (18%) lynx. Humancaused mortalities due to vehicle collision, gunshot, and the mortalities where only a cut collar was found �23 comprise the greatest known cause of mortality for all the reintroduced lynx (31 %). Mortalities due to starvation (23%) were minimized with the improved release protocols. To date, only 2 of the 55 lynx released in 2000 died of starvation. However, the improved survival of reintroduced lynx provided only partial evidence that Colorado could sustain a viable population of the species. As of winter 2003, no successful reproduction had been documented. This lack of reproduction resulted in an increased emphasis on the question of whether or not Colorado could provide sufficient habitat to sustain a selfsustaining population of lynx. Two options existed to address the problem of answering the suitable/unsuitable question. The first was to continue to monitor the existing animals for recruitment, with the understanding that the probability of detection would decrease rapidly as radio-collars failed, and the probability of successful pairing might further decrease with lowered densities due to natural mortality. Possible outcomes include 1) the animals currently out there would eventually reproduce with sufficient success to establish a viable population of lynx, 2) the animals currently out there would reproduce although not in sufficient numbers to offset mortality or 3) the animals currently out there would fail to reproduce. The primary reasons for outcomes 2 and 3 are either that Colorado does not have sufficient habitat to support viable populations of lynx or there were too few lynx released to achieve sufficient successful reproduction. Thus, the question of whether or not Colorado can support viable population(s) oflynx would remain arguable. A second option would be to supplement the existing lynx by re-introducing additional lynx over multiple years into the Core Area to attain a density approaching that of established populations of lynx. The possible outcomes could be any of those listed for the first option. The difference, however, would be that the low-density explanation for failure to establish a viable population would be difficult to support. Thus, CDOW could more definitively address the question of the suitability of Colorado for lynx populations. An analysis of these two options was conducted to determine the best management strategy to pursue to enhance the ability to assess the outcome. An update of the post-release monitoring program was also conducted. OBJECTIVES The initial post-release monitoring ofreintroduced lynx will emphasize 5 primary objectives: 1. Assess and modify release protocols to enure the highest probability of survival for each lynx released. 2. Obtain regular locations ofreleased lynx to describe general movement patterns and habitats used by lynx. 3. Determine causes of mortality in reintroduced lynx. 4. Estimate survival oflynx reintroduced to Colorado. 5. Estimate reproduction oflynx reintroduced to Colorado. Three additional objectives will be emphasized after lynx display site fidelity to an area: 6. Refine descriptions of habitats used by reintroduced lynx. 7. Refine descriptions of daily and overall movement patterns of reintroduced lynx. 8. Describe hunting habits and prey ofreintroduced lynx. Information gained to achieve these objectives will form a basis for the development of l y ~ oervation "'-. strategies in the southern Rocky Mountains. Lastly, an analysis was conducted to evaluat tow --------7/ management options for assessing Colorado's suitability for sustaining a viable lynx populat10 . �24 METHODS Augmentation An analysis of two management options was conducted to determine the best management strategy to pursue to enhance the ability to assess whether Colorado provided suitable habiat for a viable, self-sustaining population of lynx. In response to the completed analysis, the current reintroduction effort was augmented with additional animals, released within the Core Area, to increase lynx density so that the question of whether lynx can sustain viable populations in Colorado could be more definitively addressed. These new releases were conducted under the protocols found to maximize survival (see Shenk 1999). Based on dispersal patterns of lynx released in 2000, the second cohort, it was hypothesized that lynx released in the Core Area would show the necessary site fidelity to increase lynx densities to enhance the probability of successful reproduction. Movement Patterns To determine general movement patterns and habitat used by reintroduced lynx, regular locations of released lynx were collected through a combination of aerial, satellite and ground radio-tracking. Locations and general habitat descriptions at each location were recorded and mapped. Frequent flights (at least 2 times per week) were critical during the initial post-release periods because of the greater likelihood of dispersal and mortality in reintroduced carnivores during this period. Every effort was made to locate all lynx each flight during this period. All lynx released in the winter and spring of 1999 were fitted with Telonics™ VHF radio-collars, equipped with a mortality switch that activates if the collar remains motionless for 4 hours or more. Fiftyone of the 55 lynx released in the spring 2000 were fitted with Sirtrack™ dual satellite/VHF radio-collars (the other 4 lynx were fitted with Telonics™ VHF collars). All 33 lynx released in 2003 were fitted with Sirtrack™ dual satellite/VHF radio-collars. These collars also had a mortality indicator switch that operated on both the satellite and VHF mode. The satellite component of each collar was programmed to be active for 12 hours per week. The 12-hour active periods were staggered throughout the week, with approximately 7 collars being active each day of the week. Signals from the collars allowed for locations of the animals to be made via Argos, NASA, and NOAA satellites. The location information was processed by ServiceArgos and distributed to the CDOW through e-mail messages. Survival and Mortality Factors When a mortality signal (75 ppm vs. 50 ppm for the Telonics™ VHF transmitters, 20 bpm vs. 40 bpm for the Sirtrack™ VHF transmitters, 0 activity for Sirtrack™ PTT) was heard during either satellite, aerial or ground surveys, the location (UTM coordinates) was recorded. Ground crews then located and retrieved the carcass as soon as possible. The immediate area was searched for evidence of other predators and the carcass photographed in place before removal. Additionally, the mortality site was described, habitat associations, and exact location were recorded. Any scat found near the dead lynx that appeared to be from the lynx was collected. All carcasses were transported immediately to the Colorado State University Veterinary Hospital for a post mortem exam to 1) determine the cause of death and document with evidence, 2) collect samples for a variety of research projects, and 3) archive samples for future reference (research or forensic). The gross necropsy and histology were performed by, or under the lead and direct supervision of a board certified veterinary pathologist. At least one research personnel from the CDOW involved with the lynx program was also present. The protocol followed standard procedures used for thorough post-mortem examination and sample collection for histopathology and diagnostic testing (see Shenk �25 1999 for details). Some additional data/samples were routinely collected for research, forensics, and archiving. Other data/samples were collected based on the circumstances of the death (e.g., photographs, video, radiographs, bullet recovery, samples for toxicology or other diagnostic tests, etc.). The CDOW retained all samples and carcass remains with the exception of tissues in formalin for histopathology, brain for rabies exam, feces for parasitology, external parasites for ID, and other diagnostic samples. Reproduction Females were monitored for proximity to males during breeding season and for site fidelity to a given area during the denning period of May and June. Each female that exhibited stationary movement patterns in May or June 2003 was observed to look for accompanying kittens. If kittens were found at a den site they were weighed, sexed and photographed. Each kitten was uniquely marked by inserting a sterile passive integrated transponder (PIT, Biomark, Inc., Boise, Idaho, USA) tag subcutaneously between the shoulder blades. Time spent at the den was minimized to ensure the least amount of disturbance to the female and the kittens. Weight, PIT-tag number, sex and any distinguishing characteristics of each kitten was recorded. Den site location was recorded as Universal Transmercator (UTM) Coordinates. Other data to be recorded include general vegetation characteristics, elevation, weather, field personnel, time at the den, and behavioral responses of the kittens and female. RESULTS Rationale for Augmentation Thirty-six reintroduced lynx were known to be in the Core Area in winter 2002-2003, which is approximately 10,000 mi 2 . Thus, the lowest possible density oflynx in the Core Area was approximately 2 lynx/ 500 mi2, an area slightly larger than Rocky Mountain National Park. The highest density of reintroduced lynx in the Core Area was approximately 3 lynx/ 500 mi2, if all the missing lynx at that time were currently there but not being detected due to faulty radio collars. If additional naturally occurring lynx were in the area these densities could have been even higher. Lynx densities reported for natural populations occurring in northern habitats range from< 13 lynx/ 500 mi2 during snowshoe hare lows to 104-259 lynx/ 500 mi2 in years of peak hare densities in mature forests. The densities of lynx reported for populations in the north during the low in the hare cycle may not represent the lowest densities at which lynx could exist and maintain viable populations. At these lows, northern lynx still reproduce although at a much lower rate then when the hare density is higher. This low reproductive rate could be related to poor body condition, low lynx densities, or a combination of both. What can be assumed is that lynx occurring at these low densities are able to rebound and achieve higher densities. Given that reintroduced lynx in Colorado are in good body condition, CDOW may only need to increase densities to achieve reproductive rates that would sustain a viable population of lynx. Densities of lynx reported for their northern range reflect densities where lynx habitat is more uniform and consistent than in Colorado. In Colorado, although the Core Area is described as 10,000 mi2, lynx are not using the Core Area uniformly but rather are dispersed in patches throughout the Core Area. Therefore the densities calculated for lynx in Colorado are not directly comparable to those estimated from the north. It is difficult, however, to estimate an appropriate correction factor for Colorado densities to make them comparable to those reported for northern populations. Therefore, the number of lynx needed to augment the current population to achieve a density of 13 lynx/ 500 mi2 under several combinations of current density and percent of the Core Area that has suitable habitat was estimated (Table 1). �26 Although this analysis required numerous assumptions, an augmentation effort of at least 150 animals (no more than 50 per year) is a minimum target for achieving densities oflynx conducive to successful reproduction and recruitment. Once minimum densities have been achieved, additional releases should continue over four to five years to maintain the minimum densities considered necessary for successful reproduction. Monitoring of the lynx population throughout the augmentation will be critical and should be conducted rigorously. The target density of 13 lynx I 500 mi2 is based on the lowest densities documented for northern populations. However, lynx may be able to rebound from lower densities. Thus, through monitoring CDOW should estimate at what densities reproduction occurs, and at what densities successful recruitment of animals occurs. This may happen at densities lower than low lynx densities estimated for the north. Table 1. Estimates of current densities of reintroduced lynx in the Core Area under various combinations of number of lynx and percent suitable habitat. Calculations of how many lynx would be needed under these conditions to achieve densities similar to the lowest densities reported for northern populations are presented and the number of additional lynx needed to achieve this density. Minimum Density Lynx needed to achieve additional lynx lynx/ 500 mi2 13 lynx/ 500 mi2 Density Assumptions needed 1 No.of lynx % Area suitable mimmum 100% 1.8 260 224 maximum 100% 2.6 260 208 mimmum 75% 2.4 195 159 maximum 75% 3.5 195 143 mmimum 50% 3.6 130 94 maximum 50% 5.2 130 78 Assumes no mortality. Augmentation Based on the adoption of the augmentation management strategy, 33 lynx were released in April 2003, bringing the total number of lynx reintroduced to Colorado to 129 (Table 2). The 33 lynx reintroduced in 2003 had been captured in Quebec, Manitoba and British Columbia. These new releases were conducted under the protocols found to maximize survival (see Shenk 2001). All 33 lynx were released in the Core Area of southwestern Colorado. Each lynx was released with a dual VHF /satellite radio collar so that the lynx can be monitor for movement and mortality. Estimated age, sex and body condition were ascertained and recorded for each lynx prior to release (see Wild 1999). Specific release sites were selected based on land ownership and accessibility during times of release. Lynx were transported from the holding facility to the release site in cages (usually 1, occasionally 2 lynx per cage). Release site location was recorded in Universal Transverse Mercator (UTM) coordinates and identification of all other lynx released at the same location, on the same day, was recorded. Behavior of the lynx on release and movement away from the-release site were documented. �27 Table 2. Colorado lynx reintroduction effort as of June 30, 2004. Females Males TOTAL Year 22 19 41 1999 35 20 55 2000 17 16 33 2003 74 55 129 TOTAL Reproduction Nine pairs of lynx were documented during the 2003 breeding season (March and April). In May and June 2003, 6 dens and a total of 16 kittens were found in the lynx core research area in southwestern Colorado (Table 3). At all dens the females appeared in excellent condition, as did the kittens. The kittens weighed from 270-500 grams. Lynx kittens weigh approximately 200 grams at birth and do not open their eyes until they are 10-17 days old. Dens were found when field crews walked in on females that exhibited virtually no movement for at least 10 days from both aerial and ground telemetry. Table 3. Reproduction information for summer 2003. Kittens Date Den Female Release Year Found Females 2000 BCO0F8 5/21/03 ? 2000 BC00F19 5/26/03 l 2000 YK00F16 6/19/03 l YK99Fl 1999 6/10/03 2 YK00F19 2000 6/11/03 YK00Fl0 2000 5/31/03 2 TOTAL 7 Males ? l 1 2 2 7 Total 2 2 2 3 3 4 16 The dens were scattered throughout the Core Area, with no dens found outside the Core Area. All the dens were in Engelmann spruce/subalpine fir forests in areas of extensive downfall. Elevations ranged from 3240-3557 m (10,630 - 11,670 feet). Field crews weighed, photographed, and PIT-tagged the kittens. Field crews also took hair samples from the kittens for genetic work in an attempt to confirm paternity. Kittens were processed as quickly as possible (11-32 minutes) to minimize the time the kittens were without their mother. While working with the kittens the females remained nearby, often making themselves visible to us. The females generally continued a low growling vocalization the entire time personnel were at the den. In all cases, the female returned to the den site once field crews left the area. Locations •The 2003 releases have remained in the Core Area with the exception of 2 lynx that went briefly to New Mexico but subsequently returned to Colorado. Most lynx continue to use terrain within the Core Area: New Mexico north to Gunnison, west as far as Taylor Mesa and east to Monarch Pass. There are some lynx north of Gunnison up to the 170 corridor and in the :Caylor Park area. No lynx are known to be north of 170 at this time. Mortalities Of the total 129 lynx released in 1999, 2000 and 2003 there are 46 known mortalities. Of these 46 mortalities, 25 are from the 1999 releases, 20 are from the 2000 releases, and 1 is from the 2003 releases. Causes of death are listed in Table 4. �28 Table 4. Causes of death for lynx released into southwestern Colorado in 1999, 2000 and 2003. 2000 2003 2003 2000 2000 1999 1999 Male Female Male Female Unk Male Female Cause of Death 1 6 1 Starvation 1 3 Hit by Vehicle 2 1 1 3 1 1 Shot Probable Predation 1 3 Plague Unknown: Human Caused 1 2 1 Probable Shot Probable Hit by Vehicle 2 Unknown: Not Starvation 1 2 1 2 4 3 1 Unknown 1 12 Total Mortalities 8 17 7 1 1 Total 9 6 6 1 3 4 2 4 11 46 Current Status At this time, CDOW is tracking 61 of the 83 lynx still possibly alive. A lynx is listed as missing if a signal has not been heard from the animal for at least 1 year. There are 21 lynx that CDOW has not heard signals on since at least June 30, 2002 (fable 5). One of these missing lynx cannot be identified but was hit by a truck in New Mexico, thus only 20 are truly missing. Possible reasons for not locating these missing lynx include (1) long distance dispersal, beyond the areas currently being searched, (2) radio failure, or (3) destruction of the radio (e.g., run over by car). CDOW continues to search for all missing lynx during both aerial and ground searches. Expanded flights outside the research area during the summer and fall months may yield locating these missing lynx. Two of the lynx released in 2000 have probably slipped their collars. Thus, CDOW has tracked 61 individual lynx since at least June 30, 2002. Table 5. Status oflynx reintroduced to Colorado as of June 30, 2003. Females Males TOTALS Unk.nown 74 55 129 Released 29 46 Known Dead 16 1 Possible Alive 45 39 83 7 14 Missing 21 (1 is unknown mortality) 1-2 Slipped Collar 1 1? 37 24 Tracking 61 �29 DISCUSSION The low density oflynx present in Colorado in winter 2002-2003 limited the ability to answer to the question of whether Colorado has sufficient suitable habitat to sustain a viable lynx population. At that time, the lack of successful reproduction may have reflected either insufficient habitat or lynx at too low a density to achieve reproductive success. It was decided that an augmentation of this reintroduction effort was necessary to eliminate an ambiguous result if successful reproduction had not occurred under densities existing in winter 2002-03. In order to maintain densities equal to those in areas that have maintained breeding populations the CDOW would need to reintroduce 50 lynx per year for the next three years and augment the population with an additional I 0-12 lynx for years 4 through 6. The reintroduction effort was augmented with 33 additional animals, released within the Core Area in April 2003, to increase lynx density so that the question of whether lynx can sustain viable populations in Colorado could be more definitively addressed. Based on dispersal patterns of lynx released in 2000, the second cohort, it was assumed lynx released in the Core Area would show the necessary site fidelity to increase lynx densities to enhance the probability of successful reproduction. The first lynx kittens documented to be born to lynx reintroduced to Colorado were found on May 21, 2003. A total of 6 dens and 16 kittens were found in 2003. While this is a milestone CDOW has been hoping to achieve, live births are the first step towards recruitment. Recruitment into a population would require these kittens to survive through their first year oflife and produce offspring of their own. To achieve a viable population of lynx, enough kittens need to be recruited into the population to offset the mortality that occurs in that year and hopefully even add more so that the population can grow. Although den sites will not be visited again until fall 2003, so as not to disturb the female and kittens further, the female's movement patterns will be monitored through aerial telemetry. During fall 2003, demales with kittens will be observed through walk-ins to try to count the number of kittens still with her. Alternatively, the females will be snow-tracked once there is sufficient snowfall on the ground to document the presence and number of kittens. Kittens typically stay with their mothers until they are I 0 months old. The Colorado lynx reintroduction effort has overcome most obstacles encmm.tered so far. From results to date it can be concluded that the CDOW has developed release protocols that ensure high initial post-release survival (Shenk 2001), and on an individual level lynx have demonstrated they can survive long-term in areas of Colorado. It had also been documented that reintroduced lynx could exhibit site fidelity, engage in breeding behavior and produce kittens. What is yet to be demonstrated is whether Colorado conditions can support the recruitment necessary to more-than-offset annual mortality for a population to sustain itself. Monitoring of reintroduced lynx will continue in an effort to document such viability. LITERATURE CITED Shenk, T. M. 1999. Program narrative: Post-release monitoring ofreintroduced lynx (Lynx canadensis) to Colorado. Report for the Colorado Division of Wildlife. Shenk, T. M. 200 I. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report. Colorado Division of Wildlife. Wild, M.A. 1999. Lynx veterinary services and diagnostics. Job Progress Report for the Colorado Division of Wildlife. Fort Collins, Colorado. �Colorado Division of Wildlife Wildlife Research Report July 2003- June 2004 JOB PROGRESS REPORT State of Colorado Project No. Work Package No. 0670 Task No. 1 : : : : Federal Aid Project: : N/A Cost Center 3430 Mammals Research Lynx Conservation Post-Release Monitoring of Lynx Reintroduced to Colorado Period Covered: July 1, 2003 - June 30, 2004 Author: Tanya M. Shenk Personnel: R. Dickman, L. Gepfert, R. Kahn, A. Keith, G. Merrill, G. Miller, C. Wagner, S. Wait, S. Waters, L. Wolfe, D. Younkin All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT In an effort to establish a viable population of lynx (Lynx canadensis) in Colorado, a reintroduction effort was initiated in 1999. The reintroduction effort was augmented with the release of 37 additional animals in April 2004, bringing the total to 166 lynx reintroduced to southwestern Colorado. Each lynx is released with dual satellite and VHF radio transmitters to allow intensive monitoring of animals after release. Through documentation of lynx mortalities and causes of death, human-caused mortality factors such as gunshot and vehicle collision are currently the highest source of mortality for reintroduced lynx. Locations of each lynx were collected through aerial- or satellite-tracking to document movement patterns. Most lynx remain in the southwestern quarter of Colorado. Reproduction was first documented during the 2003 reproduction season. A second successful breeding season was documented in 2004 with 11 dens and 30 kittens found as of June 30, 2004. Snow-tracking results indicate the primary winter prey species are snowshoe hare (Lepus americanus) and red squirrel (Tamiasciurus hudsonicus), with other mammals and birds forming a minor part of the winter diet. Site-scale habitat data collected from snowtracking efforts indicate Engelmann spruce (Picea engelmannii) and subalpine fir (Abies lasiocarpa) are the most common forest stands used by lynx in southwestern Colorado. From results to date it can be concluded that CDOW has developed release protocols that ensure high initial post-release survival, and on an individual level lynx have demonstrated they can survive long-term in areas of Colorado. It has also been documented that reintroduced lynx could exhibit site fidelity, engage in breeding behavior and produce kittens. What is yet to be demonstrated is whether Colorado conditions can support the recruitment necessary to offset annual mortality for a population to sustain itself. Monitoring of reintroduced lynx will continue in an effort to document such viability. 5 �JOB PROGRESS REPORT POST-RELEASE MONITORING OF LYNX (Lynx canadensis) REINTRODUCED TO COLORADO Tanya M. Shenk SEGMENT OBJECTIVES The initial post-release monitoring of reintroduced lynx will emphasize 5 primary objectives: 1. Assess and modify release protocols to ensure the highest probability of survival for each lynx released. 2. Obtain regular locations of released lynx to describe general movement patterns and habitats used by lynx. 3. Determine causes of mortality in reintroduced lynx. 4. Estimate survival of lynx reintroduced to Colorado. 5. Estimate reproduction of lynx reintroduced to Colorado. Three additional objectives will be emphasized after lynx display site fidelity to an area: 6. Refine descriptions of habitats used by reintroduced lynx. 7. Refine descriptions of daily and overall movement patterns of reintroduced lynx. 8. Describe hunting habits and prey of reintroduced lynx. Information gained to achieve these objectives will form a basis for the development of lynx conservation strategies in the southern Rocky Mountains. INTRODUCTION The Canada lynx occurs throughout the boreal forests of northern North America. Colorado represents the southern-most historical distribution of lynx, where the species occupied the higher elevation, montane forests in the state. Little was known about the population dynamics or habitat use of this species in their southern distribution. Lynx were extirpated or reduced to a few animals in the state by the late 1970’s. Given the isolation of Colorado to the nearest northern populations, the Colorado Division of Wildlife (CDOW) considered reintroduction as the only option to attempt to reestablish the species in the state. A reintroduction effort was begun in 1999. To date, 166 wild-caught lynx from Alaska and Canada have been released in southwestern Colorado. The goal of the Colorado lynx reintroduction program is to establish a self-sustaining, viable population of lynx in this state. Evaluation of incremental achievements necessary for establishing viable populations is an interim method of assessing if the reintroduction effort is progressing towards success. There are seven critical criteria for achieving a viable population: (1) development of release protocols that lead to a high initial post-release survival of reintroduced animals, (2) long-term survival of lynx in Colorado, (3) development of site fidelity by the lynx to areas supporting good habitat in densities sufficient to breed, (4) reintroduced lynx must breed, (5) breeding must lead to reproduction of surviving kittens (6) lynx born in Colorado must reach breeding age and reproduce successfully, and (7) recruitment must be equal to or greater than mortality. The post-release monitoring program for the reintroduced lynx has 2 primary goals. The first goal is to determine how many lynx remain in Colorado and their locations relative to each other. Given this information and knowing the sex of each individual, we can assess whether these lynx can form a breeding core from which a viable population might be established. From these data we can also describe general movement patterns and habitats used. The second primary goal of the monitoring program is to 6 �estimate survival of the reintroduced lynx and, where possible, determine cause of mortality of reintroduced lynx. Such information will help in assessing and modifying release protocols and management of lynx once they have been released. Additional goals of the post-release monitoring program for lynx reintroduced to the southern Rocky Mountains include refining descriptions of habitat use and movement patterns, determining hunting habits, and obtaining information on reproduction. When the lynx establish home ranges that encompass their preferred habitat, more emphasis will be placed on refining descriptions of movement patterns and habitat use. Lynx is listed as threatened under the Endangered Species Act (ESA) of 1973, as amended (16 U. S. C. 1531 et. seq.)(U. S. Fish and Wildlife Service 2000). As a listed species, an additional objective of the post-release monitoring program is to develop conservation strategies relevant to lynx in Colorado. Therefore, information specific to the ecology of the lynx in its southern Rocky Mountain range such as habitats used, movement patterns, mortality factors, survival, and reproduction in Colorado is needed. METHODS Reintroduction Effort All 2004 lynx releases were conducted under the protocols found to maximize survival (see Shenk 2001). Estimated age, sex and body condition were ascertained and recorded for each lynx prior to release (see Wild 1999). Specific release sites were selected based on land ownership and accessibility during times of release. Lynx were transported from the holding facility to the release site in individual cages. Release site location was recorded in Universal Transverse Mercator (UTM) coordinates and identification of all lynx released at the same location, on the same day, was recorded. Behavior of the lynx on release and movement away from the release site were documented. Movement Patterns All lynx released in spring 2004 were fitted with SirtrackTM dual satellite/VHF radio-collars. These collars have a mortality indicator switch that operated on both the satellite and VHF mode. The satellite component of each collar was programmed to be active for 12 hours per week. The 12-hour active periods were staggered throughout the week. Signals from the collars allowed for locations of the animals to be made via Argos, NASA, and NOAA satellites. The location information was processed by ServiceArgos and distributed to the CDOW through e-mail messages. To determine general movement patterns of reintroduced lynx, regular locations of released lynx were collected through a combination of aerial, satellite and ground radio-tracking. Locations and general habitat descriptions at each location were recorded and mapped. Survival and Mortality Factors When a mortality signal (75 ppm vs. 50 ppm for the Telonics™ VHF transmitters, 20 bpm vs. 40 bpm for the Sirtrack™ VHF transmitters, 0 activity for Sirtrack™ PTT) was heard during either satellite, aerial or ground surveys, the location (UTM coordinates) was recorded. Ground crews then located and retrieved the carcass as soon as possible. The immediate area was searched for evidence of other predators and the carcass photographed in place before removal. Additionally, the mortality site was described and habitat associations and exact location were recorded. Any scat found near the dead lynx that appeared to be from the lynx was collected. All carcasses were transported immediately to the Colorado State University Veterinary Hospital for a post mortem exam to 1) determine the cause of death and document with evidence, 2) collect samples for a variety of research projects, and 3) archive samples for future reference (research or 7 �forensic). The gross necropsy and histology were performed by, or under the lead and direct supervision of a board certified veterinary pathologist. At least one research personnel from the CDOW involved with the lynx program was also present. The protocol followed standard procedures used for thorough post-mortem examination and sample collection for histopathology and diagnostic testing (see Shenk 1999 for details). Some additional data/samples were routinely collected for research, forensics, and archiving. Other data/samples were collected based on the circumstances of the death (e.g., photographs, video, radiographs, bullet recovery, samples for toxicology or other diagnostic tests, etc.). The CDOW retained all samples and carcass remains with the exception of tissues in formalin for histopathology, brain for rabies exam, feces for parasitology, external parasites for ID, and other diagnostic samples. Reproduction Females were monitored for proximity to males during breeding season and for site fidelity to a given area during the denning period of May and June. Each female that exhibited stationary movement patterns in May or June 2004 was observed to look for accompanying kittens. Kittens found at den sites were weighed, sexed and photographed. Each kitten was uniquely marked by inserting a sterile passive integrated transponder (PIT, Biomark, Inc., Boise, Idaho, USA) tag subcutaneously between the shoulder blades. Time spent at the den was minimized to ensure the least amount of disturbance to the female and the kittens. Weight, PIT-tag number, sex and any distinguishing characteristics of each kitten was also recorded. Den site location was recorded as UTM coordinates. General vegetation characteristics, elevation, weather, field personnel, time at the den, and behavioral responses of the kittens and female were also recorded. Diet Winter diet of reintroduced lynx was estimated by documenting successful kills through snowtracking. Prey species from failed and successful hunting attempts were identified by either tracks or remains. Scat analysis also provided information on foods consumed. Scat samples were collected wherever found and labeled with location and individual lynx identification. Only part of the scat was collected; the remainder was left in place in the event that the scat was being used by the animal as a territory mark. RESULTS Reintroduction Effort Based on the adoption of the approved augmentation management strategy (Shenk 2003), 37 lynx (17 females and 20 males) were released in April 2004, bringing the total number of lynx reintroduced to Colorado to 166 (Table 1). Lynx for the 2004 augmentation were captured in Quebec and British Columbia. All 37 lynx were released at previously used release sites in southwestern Colorado. Table 1. Colorado lynx reintroduction effort as of June 30, 2004. Year Females Males TOTAL 1999 22 19 41 2000 35 20 55 2003 17 16 33 2004 17 20 37 TOTAL 91 75 166 8 �Movement Patterns Most of the lynx released in 2004 have remained in the southwestern quarter of Colorado, with the exception of 2 lynx that went briefly to New Mexico but subsequently returned to Colorado. The majority of surviving lynx from the entire reintroduction effort continue to use areas from New Mexico north to Gunnison, west as far as Taylor Mesa and east to Monarch Pass. There are some lynx north of Gunnison up to the I70 corridor and in the Taylor Park area. No lynx are known to be north of I70 at this time. Mortalities Of the total 166 adult lynx released in 1999, 2000, 2003 and 2004 we have 56 known mortalities. Of these 56 mortalities, 25 are from the 1999 releases, 23 are from the 2000 releases, 4 are from the 2003 releases, and 4 are from the 2004 releases. Causes of death are listed in Table 2. Of the 16 kittens known to have been born in Colorado in 2003, we have 7 confirmed mortalities and 3 possible mortalities. Table 2. Causes of death for adult lynx released into southwestern Colorado in 1999, 2000, 2003 and 2004. 1999 Total Cause of death 2000 2003 2004 M F M F U M F M F Starvation 1 Hit by Vehicle Shot 6 1 2 3 Probable Predation 1 1 3 1 9 1 1 1 1 7 7 1 1 Plague 4 4 Unknown: Human Caused 1 1 Probable Shot 1 Probable Hit by Vehicle 2 1 4 2 Unknown: Not Starvation 1 2 Unknown 2 1 2 1 4 1 4 2 1 Human Caused Total 1 14 1 8 17 7 15 1 3 1 6 2 1 2 56 As of June 30, 2004, CDOW was actively tracking 85 of the 110 lynx still possibly alive (Table 3). Of the remaining 25 remaining lynx possibly alive, 24 were ‘missing’ as of June 30, 2004 (Table 3). A lynx was listed as missing if a signal has not been heard from the animal for at least 1 year. One of these missing lynx is the unknown mortality, thus only 23 are truly missing. Possible reasons for not locating these missing lynx include (1) long distance dispersal, beyond the areas currently being searched, (2) radio failure, or (3) destruction of the radio (e.g., run over by car). CDOW continues to search for all missing lynx during both aerial and ground searches. Two of the lynx released in 2000 are thought to have slipped their collars. Thus, CDOW tracked 85 individual lynx since at least June 30, 2003. 9 �Table 3. Status of adult lynx reintroduced to Colorado as of June 30, 2004. Females Males Unknown TOTALS Released 91 75 Known Dead 35 20 Possible Alive 56 55 110 Missing 9 15 23 (1 is unknown mortality) Slipped Collar 1 1? 1-2 Tracking 46 39 85 166 1 56 Reproduction Of the 6 females that had kittens in 2003, 1 died and 2 had collars that shut off prior to denning season in 2004. Of the 3 that could be monitored during the 2004 denning season, 1 had a litter of 2 kittens (YK00F10), 1 did not have kittens (BC00F08) and it is highly probable the third female (YK99F1) has kittens with her based on her movement patterns. We are still trying to document her status. The 2004 dens were scattered throughout Colorado and 1 den was found in southern Wyoming. Most of the dens were in Engelmann spruce/subalpine fir forests in areas of extensive downfall. Elevations at the den sites ranged from 2652-3560 m (8701 - 11,680 feet). We weighed, photographed, and PIT-tagged the kittens and recorded sex. We also took hair samples from the kittens for genetic work in an attempt to confirm paternity. We processed the kittens as quickly as possible (15-35 minutes) to minimize the time the kittens were without their mother. Four of the females would not leave the den until we reached out to pick up a kitten. While we were working with the kittens the females remained nearby, often remaining visible to us. The females generally continued a low growling vocalization the entire time we were at the den. In all cases, the female returned to the den site once we left the area. Four of the 6 females that we know had kittens in summer 2003 were still with kittens at the end of April 2004. Two of those females still had 2 kittens with them at that time. Visual observations in February 2004 of one female with 2 kittens indicated all 3 were in good body condition. Snow-trackers documented at least 1 snowshoe hare kill by a kitten in winter 2003-04. The mortality of the female YK00F16 and her 1 kitten from plague was not due to poor habitat or prey conditions, and thus we might assume she would have raised the 1 kitten to this stage as well. Three probable kitten deaths from female YK00F19 were from 1 litter that most likely failed very early. Through snow-tracking an unknown female (no radio frequency heard in the area of the tracks) we also documented 1-2 additional kittens born spring 2003 and still alive in winter 2004. Of the 16 kittens we found in summer 2003, we documented the following by April 2004: 6 confirmed alive, 7 confirmed dead, and 3 some evidence dead (Table 4). Although we tried, we were not able to capture any of the 6 surviving kittens to fit them with radio collars. Unless we capture or find any of these kittens from other methods we will not know their fate beyond this 10 months of survival. 10 �Table 4. Known reproduction for summer 2003 and subsequent kitten fates by April 2004. Female Release Year Date Den Found Females Males Total BC00F8 2000 5/21/2003 ? ? 2 Kittens Known Alive in April 2004 1 BC00F19 2000 5/26/2003 1 1 2 1 1 YK00F16 2000 6/19/2003 1 1 2 0 2 YK99F1 1999 6/10/2003 2 1 3 2 1 YK00F19 2000 6/11/2003 1 2 3 ? 3? YK00F10 2000 5/31/2003 2 2 4 2 2 16 6 7, 3? Kittens Born TOTAL Kittens Known Dead in April 2004 1 In spring 2004 we had 26 females from the releases in 1999, 2000 and 2003 that had active radio collars. We documented 18 possible mating pairs of lynx during breeding season. We defined a possible mating pair as any male and female documented within at least 1 km of each other in breeding season through either flight data or snow-tracking data. All 4 of the females that had kittens with them through winter 2003-04 bred again this spring, 2 with the same male they successfully bred with last spring. During May-June 2004 we found 11 dens and a total of 30 kittens (Table 5). Dens were found when we walked in on females that exhibited virtually no movement for at least 10 days from both aerial and ground telemetry. At all dens the females appeared in excellent condition, as did the kittens. The kittens weighed from 250-770 grams. Lynx kittens weigh approximately 200 grams at birth and do not open their eyes until they are 10-17 days old. Three of the 11 females with kittens were from the 2003 releases (Table 5). Table 5. Lynx reproduction documented in 2004. Kittens Born Female Release Year Date Den Found Females Males Total YK00F2 2000 5/28/2004 3 1 4 AK00F2 2000 5/31/2004 2 1 3 YK00F1 2000 6/1/2004 3 YK00F15 2000 6/4/2004 1 2 3 BC00F14 2000 6/7/2004 1 2 3 BC00F18 2000 6/10/2004 4 YK00F10 2000 6/17/2004 1 BC03F02 2003 6/25/2004 BC03F10 2003 6/26/2004 BC03F09 2003 YK00F7 2000 TOTAL 3 4 1 2 2 2 1 1 2 6/29/2004 1 1 2 6/30/2004 1 1 2 18 12 30 11 �Diet Winter diet of lynx was documented through detection of kills found through snow-tracking. In each winter, the most common prey item was snowshoe hare (Lepus americanus), followed by red squirrel (Tamiasciurus hudsonicus) (Table 6). Table 6. Number of kills found each winter field season through snow-tracking of lynx and percent composition of kills of the three primary prey species. Field Season 1999 1999-2000 2000-2001 2001-2002 2002-2003 2003-2004 n 9 81 88 54 65 37 Snowshoe Hare 55.56 67.46 67.41 90.74 90.77 67.56 Prey (%) Red Squirrel Cottontail 22.22 0 19.27 1.20 19.10 8.99 5.56 0 6.15 0 27.03 2.70 Other 22.22 12.05 4.49 3.70 3.08 2.70 DISCUSSION In an effort to establish a viable population of lynx in Colorado, a reintroduction effort was initiated in 1999. The reintroduction effort was augmented with the release of 37 additional animals in April 2004, bringing the total to 166 lynx reintroduced to southwestern Colorado. Locations of each lynx were collected through aerial- or satellite-tracking to document movement patterns and to detect mortalities. Most lynx remain in the southwestern quarter of Colorado. Humancaused mortality factors such as gunshot and vehicle collision are currently the highest causes of death. Reproduction was first documented from the 2003 reproduction season. A second successful breeding season was documented in 2004 with 11 dens and 30 kittens found as of June 30, 2004. Live births are the first step towards recruitment. Recruitment into a population would require these kittens to survive through their first year of life and produce offspring of their own. To achieve a viable population of lynx, enough kittens need to be recruited into the population to offset the mortality that occurs in that year and hopefully even add more so that the population can grow. Snow-tracking of released lynx provided preliminary information on hunting behavior by documenting location of kills, food caches, chases, and diet composition estimated through scat analysis. Snow-tracking results indicate the primary winter prey species are snowshoe hare and red squirrel, with other mammals and birds forming a minor part of the winter diet. Site-scale habitat data collected from snow-tracking efforts indicate Engelmann spruce and subalpine fir are the most common forest stands used by lynx in southwestern Colorado. From results to date it can be concluded that CDOW has developed release protocols that ensure high initial post-release survival, and on an individual level lynx have demonstrated they can survive long-term in areas of Colorado. It has also been documented that reintroduced lynx could exhibit site fidelity, engage in breeding behavior and produce kittens. What is yet to be demonstrated is whether Colorado conditions can support the recruitment necessary to offset annual mortality for a population to sustain itself. Monitoring of reintroduced lynx will continue in an effort to document such viability. 12 �LITURATURE CITED Shenk, T. M. 1999. Program narrative: Post-release monitoring of reintroduced lynx (Lynx canadensis) to Colorado. Job Progress Report for the Colorado Division of Wildlife. Fort Collins, Colorado. __________ 2001. Post-release monitoring of lynx reintroduced to Colorado. Job Progress Report for the Colorado Division of Wildlife. Fort Collins, Colorado. __________ 2003. Post-release monitoring of lynx reintroduced to Colorado. Job Progress Report for the Colorado Division of Wildlife. Fort Collins, Colorado. U. S. Fish and Wildlife Service. 2000. Endangered and threatened wildlife and plants: final rule to list the contiguous United States distinct population segment of the Canada lynx as a threatened species. Federal Register 65, Number 58. Wild, M. A. 1999. Lynx veterinary services and diagnostics. Job Progress Report for the Colorado Division of Wildlife. Fort Collins, Colorado. Prepared by _______________________________ Tanya M. Shenk, Wildlife Researcher 13 �Colorado Division of Wildlife July 2004 – June 2005 WILDLIFE RESEARCH REPORT State of Cost Center Work Package Task No. Federal Aid Project: Colorado 3430 0670 1 : : : : N/A : Division of Wildlife Mammals Research Lynx Conservation Post-Release Monitoring of Lynx Reintroduced to Colorado Period Covered: July 1, 2004- June 30, 2005 Author: T. M. Shenk Personnel: L. Baeten, B. Diamond, R. Dickman, S. Dieterich, D. Freddy, L. Gepfert, R. Kahn, A. Keith, G. Merrill, G. Miller, J. Stewart, C. Wagner, S. Wait, S. Waters, L. Wolfe, D. Younkin All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT In an effort to establish a viable population of lynx (Lynx canadensis) in Colorado, the Colorado Division of Wildlife (CDOW) initiated a reintroduction effort in 1997 with the first lynx released in February 1999. A total of 166 lynx were released from 1999-2004 and an augmentation of 38 additional animals (20 males:18 females) was completed in 2005 resulting in a total of 204 lynx reintroduced to southwestern Colorado. Each lynx was released with dual satellite and VHF radio transmitters to allow intensive monitoring of animals after release. Locations of each lynx were collected through aerial- or satellite-tracking to document movement patterns. Most lynx remain in the southwestern quarter of Colorado. Through documentation of lynx mortalities and causes of death, human-caused mortality factors, such as gunshot and vehicle collision, are currently the highest source of mortality for reintroduced lynx. Reproduction was first documented during the 2003 reproduction season with 6 dens and 16 kittens found. A second successful breeding season was documented in 2004 with 30 kittens found at 11 dens and an addition 9 kittens found after denning season. In 2005, 46 kittens were found at 16 dens with an additional den located but not visited for safety reasons. Data collected from snowtracking indicate the primary winter prey species are snowshoe hare (Lepus americanus) and red squirrel (Tamiasciurus hudsonicus), with other mammals and birds forming a minor part of the winter diet. Sitescale habitat data collected from snow-tracking efforts indicate Engelmann spruce (Picea engelmannii) and subalpine fir (Abies lasiocarpa) are the most common forest stands used by lynx in southwestern Colorado. Results to date have demonstrated that CDOW has developed release protocols that ensure high initial post-release survival, and on an individual level, lynx have demonstrated an ability to survive long-term in areas of Colorado. Reintroduced lynx have also exhibited site fidelity, engaged in breeding behavior and produced kittens. What is yet to be demonstrated is whether conditions in Colorado can support the recruitment necessary to offset annual mortality for a population to remain viable for several generations of lynx. Monitoring of reintroduced lynx will continue in an effort to document such viability. 1 �WILDLIFE RESEARCH REPORT POST RELEASE MONITORING OF LYNX (LYNX CANADENSIS) REINTRODUCED TO COLORADO TANYA M. SHENK P. N. OBJECTIVE The initial post-release monitoring of lynx reintroduced into Colorado will emphasize 5 primary objectives: 1. Assess and modify release protocols to ensure the highest probability of survival for each lynx released. 2. Obtain regular locations of released lynx to describe general movement patterns and habitats used by lynx. 3. Determine causes of mortality in reintroduced lynx. 4. Estimate survival of lynx reintroduced to Colorado. 5. Estimate reproduction of lynx reintroduced to Colorado. Three additional objectives will be emphasized after lynx display site fidelity to an area: 6. Refine descriptions of habitats used by reintroduced lynx. 7. Refine descriptions of daily and overall movement patterns of reintroduced lynx. 8. Describe hunting habits and prey of reintroduced lynx. Information gained to achieve these objectives will form a basis for the development of lynx conservation strategies in the southern Rocky Mountains. SEGMENT OBJECTIVES 1. Release additional adult lynx captured in Canada in southwestern Colorado during spring 2005. 2. Complete winter 2004-05 field data collection on lynx habitat use, hunting behavior, diet, mortalities, and movement patterns. 3. Complete winter 2004-05 lynx trapping field season to collar Colorado born lynx and re-collar adult lynx. 4. Complete spring 2005 field data on lynx reproduction. 5. Summarize and analyze data and publish information as Progress Reports, peer-reviewed manuscripts for appropriate scientific journals, or CDOW technical publications. INTRODUCTION The Canada lynx occurs throughout the boreal forests of northern North America. Colorado represents the southern-most historical distribution of lynx, where the species occupied the higher elevation, montane forests in the state. Little was known about the population dynamics or habitat use of this species in their southern distribution. Lynx were extirpated or reduced to a few animals in the state by the late 1970’s due, most likely, to predator control efforts such as poisoning and trapping. Given the isolation of Colorado to the nearest northern populations, the CDOW considered reintroduction as the only option to attempt to reestablish the species in the state. A reintroduction effort was begun in 1997, with the first lynx released in Colorado in 1999. To date, 204 wild-caught lynx from Alaska and Canada have been released in southwestern Colorado. The goal of the Colorado lynx reintroduction program is to establish a self-sustaining, viable population of 2 �lynx in this state. Evaluation of incremental achievements necessary for establishing viable populations is an interim method of assessing if the reintroduction effort is progressing towards success. There are 7 critical criteria for achieving a viable population: 1) development of release protocols that lead to a high initial post-release survival of reintroduced animals, 2) long-term survival of lynx in Colorado, 3) development of site fidelity by the lynx to areas supporting good habitat in densities sufficient to breed, 4) reintroduced lynx must breed, 5) breeding must lead to reproduction of surviving kittens 6) lynx born in Colorado must reach breeding age and reproduce successfully, and 7) recruitment must be equal to or greater than mortality. The post-release monitoring program for the reintroduced lynx has 2 primary goals. The first goal is to determine how many lynx remain in Colorado and their locations relative to each other. Given this information and knowing the sex of each individual, we can assess whether these lynx can form a breeding core from which a viable population might be established. From these data we can also describe general movement patterns and habitat use. The second primary goal of the monitoring program is to estimate survival of the reintroduced lynx and, where possible, determine causes of mortality for reintroduced lynx. Such information will help in assessing and modifying release protocols and management of lynx once they have been released to ensure their highest probability of survival. Additional goals of the post-release monitoring program for lynx reintroduced to the southern Rocky Mountains included refining descriptions of habitat use and movement patterns and describing successful hunting habitat once lynx established home ranges that encompassed their preferred habitat. Specific objectives for the site-scale habitat data collection include: 1) describe and quantify site-scale habitat use by lynx reintroduced to Colorado, 2) compare site-scale habitat use among types of sites (e.g., kills vs. long-duration beds), and 3) compare habitat features at successful and unsuccessful snowshoe hare chases. The program will also investigate the ecology of snowshoe hare in Colorado. Documenting reproduction is critical to the success of the program and lynx are monitored intensively to document breeding, births, survival and recruitment of lynx born in Colorado. Site-scale habitat descriptions of den sites are also collected and compared to other sites used by lynx. Lynx is listed as threatened under the Endangered Species Act (ESA) of 1973, as amended (16 U. S. C. 1531 et. seq.)(U. S. Fish and Wildlife Service 2000). Colorado is included in the federal listing as lynx habitat. Thus, an additional objective of the post-release monitoring program is to develop conservation strategies relevant to lynx in Colorado. To develop these conservation strategies, information specific to the ecology of the lynx in its southern Rocky Mountain range, such as habitat use, movement patterns, mortality factors, survival, and reproduction in Colorado is needed. STUDY AREA Southwestern Colorado is characterized by wide plateaus, river valleys, and rugged mountains that reach elevations over 4200 m. Engelmann spruce-subalpine fir is the most widely distributed coniferous forest type at elevations most typically used by lynx. The Core Research Area is defined as areas bounded by the New Mexico state line to the south, Taylor Mesa to the west and Monarch Pass on the north and east and > 2900 meters in elevation. METHODS REINTRODUCTION Effort All 2005 lynx releases were conducted under the protocols found to maximize survival (see Shenk 2001). Estimated age, sex and body condition were ascertained and recorded for each lynx prior to 3 �release (see Wild 1999). Specific release sites were those used in earlier years of the project and were selected based on land ownership and accessibility during times of release (Byrne 1998). Lynx were transported from the Frisco Creek Wildlife Rehabilitation Center, where they were held from their time of arrival in Colorado, to their release site in individual cages. Release site location was recorded in Universal Transverse Mercator (UTM) coordinates and identification of all lynx released at the same location, on the same day, was recorded. Behavior of the lynx on release and movement away from the release site were documented. Distribution and Movement Patterns All lynx released in 1999 were fitted with TelonicsTM radio-collars. All lynx released since 1999, with the exception of 5 males released in spring 2000, were fitted with SirtrackTM dual satellite/VHF radio-collars. These collars have a mortality indicator switch that operated on both the satellite and VHF mode. The satellite component of each collar was programmed to be active for 12 hours per week. The 12-hour active periods for individual collars were staggered throughout the week. Signals from the collars allowed for locations of the animals to be made via Argos, NASA, and NOAA satellites. The location information was processed by ServiceArgos and distributed to the CDOW through e-mail messages. To determine general movement patterns of reintroduced lynx, regular locations of released lynx were collected through a combination of aerial, satellite and ground radio-tracking. Locations were recorded in UTM coordinates and general habitat descriptions for each ground and aerial location were recorded. Survival and Mortality Factors When a mortality signal (75 beats per minute [bpm] vs. 50 bpm for the Telonics™ VHF transmitters, 20 bpm vs. 40 bpm for the Sirtrack™ VHF transmitters, 0 activity for Sirtrack™ PTT) was heard during either satellite, aerial or ground surveys, the location (UTM coordinates) was recorded. Ground crews then located and retrieved the carcass as soon as possible. The immediate area was searched for evidence of other predators and the carcass photographed in place before removal. Additionally, the mortality site was described and habitat associations and exact location were recorded. Any scat found near the dead lynx that appeared to be from the lynx was collected. All carcasses were transported to the Colorado State University Veterinary Teaching Hospital (CSUVTH) for a post mortem exam to 1) determine the cause of death and document with evidence, 2) collect samples for a variety of research projects, and 3) archive samples for future reference (research or forensic). The gross necropsy and histology were performed by, or under the lead and direct supervision of a board certified veterinary pathologist. At least one research personnel from the CDOW involved with the lynx program was also present. The protocol followed standard procedures used for thorough post-mortem examination and sample collection for histopathology and diagnostic testing (see Shenk 1999 for details). Some additional data/samples were routinely collected for research, forensics, and archiving. Other data/samples were collected based on the circumstances of the death (e.g., photographs, video, radiographs, bullet recovery, samples for toxicology or other diagnostic tests, etc.). From 1999–2004 the CDOW retained all samples and carcass remains with the exception of tissues in formalin for histopathology, brain for rabies exam, feces for parasitology, external parasites for ID, and other diagnostic samples. Since 2005 carcasses are disposed of at the CSUVTH with the exception of the lower canine, fecal samples, stomach content samples and tissue or bone marrow samples to be delivered by CDOW to the Center for Disease control for plague testing. The lower canine is sent to Matson Labs (Missoula, Montana) for aging and the fecal and stomach content samples are evaluated for diet. 4 �Reproduction Females were monitored for proximity to males during each breeding season. We defined a possible mating pair as any male and female documented within at least 1 km of each other in breeding season through either flight data or snow-tracking data. Females were then monitored for site fidelity to a given area during each denning period of May and June. Each female that exhibited stationary movement patterns in May or June were closely monitored to locate possible dens. Dens were found when field crews walked in on females that exhibited virtually no movement for at least 10 days from both aerial and ground telemetry. Kittens found at den sites were weighed, sexed and photographed. Each kitten was uniquely marked by inserting a sterile passive integrated transponder (PIT, Biomark, Inc., Boise, Idaho, USA) tag subcutaneously between the shoulder blades. Time spent at the den was minimized to ensure the least amount of disturbance to the female and the kittens. Weight, PIT-tag number, sex and any distinguishing characteristics of each kitten was also recorded. Beginning in 2005, blood and saliva samples were collected for genetic identification. During the den site visits, den site location was recorded as UTM coordinates. General vegetation characteristics, elevation, weather, field personnel, time at the den, and behavioral responses of the kittens and female were also recorded. Once the females moved the kittens from the natal den area, den sites were visited again and site-specific habitat data were collected (see Habitat Use section below). Recaptures Recaptures were attempted for either lynx that were in poor body condition or lynx that needed to have their radio-collars replaced due to failed or failing batteries or to radio-collar kittens born in Colorado once they reached at least 10-months of age when they were nearly adult size. Methods of recapture included 1) trapping using a Tomahawk™ live trap baited with a rabbit and visual and scent lures, 2) calling in and darting lynx using a Dan-Inject CO2 rifle, 3) custom box-traps modified from those designed by other lynx researchers (Kolbe et al. 2003) and 4) hounds trained to pursue felids were also used to tree lynx and then the lynx was darted while treed. Lynx were immobilized either with Telazol (3 mg/kg; modified from Poole et al. 1993 as recommended by M. Wild, DVM) or medetomidine (0.09mg/kg) and ketamine (3 mg/kg; as recommended by L. Wolfe, DVM)) administered intramuscularly (IM) with either an extendible pole-syringe or a pressurized syringe-dart fired from a Dan-Inject air rifle. Immobilized lynx were monitored continuously for decreased respiration or hypothermia. If a lynx exhibited decreased respiration 2mg/kg of Dopram was administered under the tongue; if respiration was severely decreased, the animal was ventilated with a resuscitation bag. If medetomidine/ketamine were the immobilization drugs, the antagonist Atipamezole hydrochloride (Antisedan) was administered. Hypothermic (body temperature < 95o F) animals were warmed with hand warmers and blankets. While immobilized, lynx were fitted with replacement SirtrackTM VHF/satellite collar and blood and hair samples were collected. Once an animal was processed, recovery was expedited by injecting the equivalent amount of the antagonist Antisedan IM as the amount of medetomidine given, if medetomodine/ketemine was used for immobilization. Lynx were then monitored while confined in the box-trap until they were sufficiently recovered to move safely on their own. No antagonist is available for Telezol so lynx anesthetized with this drug were monitored until the animal recovered on its own in the box-trap and then released. If captured and in poor body condition lynx were anesthetized with Telezol (2 mg/kg) and returned to the Frisco Creek Wildlife Rehabilitation Center for treatment. HABITAT USE Gross habitat use was documented by recording canopy vegetation at aerial locations. More refined descriptions of habitat use by reintroduced lynx were obtained through following lynx tracks in 5 �the snow (i.e., snow-tracking) and site-scale habitat data collection conducted at sites found through this method to be used by lynx. Snow-tracking Locations from aerial- and satellite-tracking were used to help ground-trackers locate lynx tracks in snow. Snowmobiles, where permitted, were used to gain the closest possible access to the lynx tracks without disturbing the animal. From that point, the tracking team used snowshoes to access tracks. Once tracks were found, the ground crew back- or forward-tracked the animal if it was far enough away not to be disturbed. Back-tracking generally avoided the possibility of disturbing the lynx by moving away from the animal rather than towards the animal. However, monitoring of the lynx through radio-telemetry was used to assure that the ground crew was staying a sufficient distance away from the lynx in the event the lynx might double back on its tracks. Radio-telemetry was also used in forward-tracking to make sure the team did not disturb the animal. If it appeared the lynx began to move in response to the observers, the observers stopped following the tracks. If the lynx began to move and the movement did not appear to be a response to the observers, the ground crew continued following the track. An attempt was made in Season 1 (February-May 1999) and Season 2 (December 1999-April 2000) to snow-track each lynx. In Season 3 (December 2000-April 2001), we attempted to snow-track all lynx within the Core Research Area. In tracking Season 4 (December 2001-April 2002), Season 5 (December 2002-April 2003), Season 6 (December 2003-April 2004) and Season 7 (December 2004April 2005) we attempted to track all accessible lynx in the Core Research Area and some lynx north of the Core Research Area. Ground crews were instructed to track lynx only where it was safe to travel. Restrictions to safe travel included avalanche danger and extremely rugged terrain. Ground crews worked in pairs and were fully equipped for winter back-country survival. Data Collection For each day of tracking the date, lynx being tracked, slope, aspect, UTM coordinates, elevation, general habitat description, and summary of the days tracking were recorded. Aspect was defined as the direction of 'downhill' or 'fall line' on a slope. This is the direction along the ground in a dihedral angle between the horizontal and the plane of the ground surface. Units were compass degrees. Slope was defined as the dihedral angle between the horizontal and the plane of the ground surface (e.g., 45"). Once a track was located there were 2 types of 'sites' that were encountered. Site I areas needed documentation but either did not reflect areas lynx selected for specific habitat features, or were sites that occurred too frequently to measure each in detail. These sites included the start and end of the track being followed, the location of scat, and short-duration beds defined as being small in size (approximating an area a lynx would crouch), and with little ice formed in the bed indicating little time spent there. Site II areas included areas that might reflect specific habitat features lynx selected for and included locations where the following were found: kills, start of chases, territory marks (e.g., spray sites, buried scat, scat placed on prominent locations), long-duration beds (encompasses an area where a lynx would have lain for an extended period, iced bottom), and road crossing (both sides of road). In addition, habitat plots were conducted along lynx travel routes if no other sites sampled in last hour. At each of the 2 types of sites the date, lynx tracked, slope, aspect, forest structure class, UTM coordinates, and elevation were recorded. Forest structure classes included grass/forb, shrub/seedling, sapling/pole, mature, and old growth as defined in Table 1. For Site I areas, the only additional data that was collected was identification of what the site was used for (e.g., short-duration bed), and a brief description of the site. Habitat plots (see below) were conducted at Site II areas. 6 �Description of the Habitat Plot The habitat plot consisted of a 12 m x 12 m square defined by a series of 25 points placed in 5 rows of 5 with the center point being on the object that defined the site (e.g., a kill)(Figure 1). Each point was 3 m apart. The 12 m x 12 m sampling square exceeded the minimum requirement of 0.01 ha. recommended by Curtis (1959) for sampling trees. Measurements taken at each of the 25 points included: 1. Snow depth - measured vertically by an avalanche probe marked in cm. 2. Understory - measured from top of snow to 150 cm above snow in a column of 3-cm radius around the avalanche probe. Because understory measurements were influenced by vegetation outside the perimeter of the 25 sampling points (12 m x 12 m) the area used for estimating undersory cover was 15 m by 15 m. At each point, crews recorded all shrubs, trees and coarse woody debris (CWD) that fell within this column and was visible above the snow. Crews also recorded number of branches of each species that fell within the column at 3 different height categories (0-0.5 m, 0.51-1.0 m, 1.01-1.5 m). 3. Overstory: measured at 150 cm above snow with a sighting tube. The tube was made of PVC pipe, with a curved viewing end and a crosshair made of wire on the opposite end. The sighting tube was attached to the avalanche probe used to measure snow depth. Species that hit the crosshair were recorded at each of the 25 points in the vegetation plot. Ganey and Block (1994) found this method of measuring canopy cover (with 20 sample points per plot; Laymon 1988) provided greater precision among observers. 4. Species composition: all the different species of tree or shrub that hit the crosshair of the sighting tube at each of the 25 points were recorded. 5. Tree composition of the vegetation plot was recorded by species and diameter at breast height (DBH). Snow depth was used in conjunction with this recorded DBH to estimate true DBH. Within the 12 m x 12 m square all conifers and deciduous trees were recorded by DBH size class (A = 0-6 in, B = 6.1-12 in, C = 12.1 -18 in, D = 18.1-24 in, E = > 24 in). Area for the tree composition analysis was 12 m x 12 m. Understory was estimated as: 1) percent occurrence within the vegetation plot (number of points with understory/total number of points surveyed) and 2) mean percent occurrence and variance by species and height category over the total points sampled within the vegetation plot. Overstory was estimated as percent occurrence over the vegetation plot (number of points with overstory/total number of points surveyed). DIET AND HUNTING BEHAVIOR Winter diet of reintroduced lynx was estimated by documenting successful kills through snowtracking. Prey species from failed and successful hunting attempts were identified by either tracks or remains. Scat analysis also provided information on foods consumed. Scat samples were collected wherever found and labeled with location and individual lynx identification. Only part of the scat was collected (approximately 75%); the remainder was left in place in the event that the scat was being used by the animal as a territory mark. Site-scale habitat data collected for successful and unsuccessful snowshoe hare kills were compared. RESULTS REINTRODUCTION Effort From 1999 through 2004 166 lynx were reintroduced into southwestern Colorado. An additional 37 lynx were released in April 2005 (17 females and 20 males), one female was released in June 2005. This brings the total number of lynx released in Colorado to 204 (Table 2). These lynx released in 2005 were captured in Quebec, British Columbia and Manitoba. All lynx were released in the Core Research 7 �Area of southwestern Colorado at or near previously used release sites in southwestern Colorado. Lynx were released with dual VHF/satellite radio collars so they can be monitored for movement and mortality. The CDOW plans to release up to 15 lynx annually from 2006-2008. Distribution and Movement Patterns A total of 7421 aerial VHF locations for all 204 reintroduced lynx have been collected to date. An additional 14,788 satellite locations have been collected. Most lynx released remained in the southwestern quarter of Colorado. The majority of surviving lynx from the entire reintroduction effort continue to use areas from New Mexico north to Gunnison, west as far as Taylor Mesa and east to Monarch Pass. Most movements away from the Core Research Area were to the north. Numerous travel corridors have been used repeatedly by more than one lynx. These travel corridors include the Cochetopa Hills area for northerly movements, the Rio Grande Reservoir-SilvertonLizardhead Pass for movements to the west, and southerly movements down the east side of Wolf Creek Pass to the southeast through the Conejos River Valley. Lynx appear to remain faithful to an area during winter months, and exhibit more extensive movements away from these areas in the summer. Such movement patterns have also been documented by native lynx in Wyoming and Montana (Squires and Laurion 1999). Survival and Mortality Factors Of the total 204 adult lynx released from 1999-2005 there are 66 known mortalities. Of these 66 mortalities, 26 are from the 1999 releases, 24 are from the 2000 releases, 5 are from the 2003 releases, 8 are from the 2004 releases, and 3 are from the 2005 releases. Causes of death are listed in Table 3. Starvation was a significant cause of mortality in the first year of releases only. Mortalities occurred throughout the areas through which lynx moved. As of June 30, 2005, CDOW was actively tracking 110 of the 138 lynx still possibly alive. There are 29 lynx that we have not heard signals on since at least June 30, 2004 and these animals are classified as ‘missing’ (Table 4). One of these missing lynx is a mortality of unknown identity, thus only 28 are truly missing. Possible reasons for not locating these missing lynx include 1) long distance dispersal, beyond the areas currently being searched, 2) radio failure, or 3) destruction of the radio (e.g., run over by car). CDOW continues to search for all missing lynx during both aerial and ground searches. Two of the missing lynx released in 2000 are thought to have slipped their collars. Reproduction 2003.-- Nine pairs of lynx were documented during the 2003 breeding season (March and April). In May and June, 6 dens and a total of 16 kittens were found in the lynx Core Research Area in southwestern Colorado (Table 5). At all dens the females appeared in excellent condition, as did the kittens. The kittens weighed from 270-500 grams. Lynx kittens weigh approximately 200 grams at birth and do not open their eyes until they are 10-17 days old. The dens were scattered throughout the Core Research Area, with no dens found outside the core area. All the dens were in Engelmann spruce/subalpine fir forests in areas of extensive downfall. Elevations ranged from 3240-3557 m. Field crews weighed, photographed, PIT-tagged the kittens and . took hair samples from the kittens for genetic work in an attempt to confirm paternity. Kittens were processed as quickly as possible (11-32 minutes) to minimize the time the kittens were without their mother. While working with the kittens the females remained nearby, often making themselves visible to the field crews. The females generally continued a low growling vocalization the entire time personnel were at the den. In all cases, the female returned to the den site once field crews left the area. 8 �Four of the 6 females that we know had kittens in summer 2003 were still with kittens at the end of April 2004. Two of those females still had 2 kittens with them at that time. Visual observations in February 2004 of one female with 2 kittens indicated all 3 were in good body condition. The mortality of female YK00F16 and her 1 kitten in October 2003 from plague was not due to poor habitat or prey conditions, and thus we might assume she would have raised the 1 kitten to this stage as well. Three probable kitten deaths from female YK00F19 were from 1 litter that most likely failed very early. Through snow-tracking in winter 2003-04 an unknown female (no radio frequency heard in the area of the tracks) we also documented 1-2 additional kittens born spring 2003 and still alive in winter 2004. Of the 16 kittens we found in summer 2003, we documented the following by April 2004: 6 confirmed alive, 7 confirmed dead, and 3 some evidence dead. Although we tried, we were not able to capture any of the 6 surviving kittens to fit them with radio-collars. 2004.-- In Spring 2004, 26 females from the releases in 1999, 2000 and 2003 had active radiocollars. Of these, we documented 18 possible mating pairs of lynx during breeding season. All 4 of the females that had kittens with them through winter 2003-04 bred again spring 2004, 2 with the same male they successfully bred with spring 2003. During May-June 2004 we found 11 dens and a total of 30 kittens (Table 6). At all dens the females appeared in excellent condition, as did the kittens. The kittens weighed from 250-770 grams. Three of the 11 females with kittens were from the 2003 releases (Table 6). Three additional litters were documented after denning season through either observation of a female lynx with kittens or snow-tracking females with kittens that were not one of the 11 females found on dens. From the size of the kittens they would have been born during the normal denning season in May or June. Nine additional kittens were observed from these litters for a total of 39 known kittens born in 2004. Two of these additional litters were documented from direct follow-ups to sighting made by the public and reported to CDOW. Two females that had kittens in 2003 and reared at least part of their litters through March 2004, bred and had kittens again in 2004. Two of the litters documented by direct observation or snow-tracking are from females whose collars no longer work. Seven kittens born in 2004 were captured at 10-months of age and fitted with dual satellite/VHF collars. All 7 are alive and currently being monitored. 2005.-- In spring 2005 we had 34 females from the releases in 1999, 2000, 2003 and 2004 that had active radio-collars. We documented 23 possible mating pairs of lynx during breeding season. During May-June 2005 we visited 16 dens and found a total of 46 kittens (Table 7). At all dens the females appeared in excellent condition, as did the kittens. An additional female had a den we were not able to get to during May or June due to high water. Female BC03F03 was hit and killed on I70 on 5/19/2005. She had 2 fetuses in her uterus, so would have contributed to reproduction this year had she lived. We weighed, photographed, PIT-tagged the kittens and recorded sex. We also took blood samples from the kittens for genetic work in an attempt to confirm paternity. While we were working with the kittens the females remained nearby, often remaining visible to us. The females generally continued a low growling vocalization the entire time we were at the den. In all cases, the female returned to the den site once we left the area. All of the 2005 dens were scattered throughout the high elevation areas of Colorado, south of Interstate 70. Most of the dens were in Engelmann spruce/subalpine fir forests in areas of extensive downfall. Elevations ranged from 3117-3586 m. We weighed, photographed, and PIT-tagged the kittens, recorded sex and took hair samples from the kittens for genetic work in an attempt to confirm paternity. Four of the females would not leave the den until we reached out to pick up a kitten. While we were working with the kittens the females remained nearby, often remaining visible to us. The females 9 �generally continued a low growling vocalization the entire time we were at the den. In all cases, the female returned to the den site once we left the area. One female, YK00F10 has had litters 3 years in a row. In 2003 she had 4 kittens and raised 2 through the winter. In 2004 she had 2 kittens and raised both through the winter, this year she had 4 kittens again. She has had all 3 litters in the same general area and has had the same mate for 3 years. Eight additional females had a second litter in Colorado this year. Three females from the 2004 releases had litters in 2005. This is the second year in a row we had females released the prior spring, find a territory and a mate within a year and produced live young. In reproduction season 2004 we had 3 females released in spring 2003 that did the same thing. Of those 3, 2 successfully raised at least part of their litters through winter 2005. Den Sites.--A total of 33 dens have been found. All of the dens except one have been scattered throughout the high elevation areas of Colorado, south of I-70. One den was found in southeastern Wyoming, near the Colorado border. Dens were located on steep ( x slope = 29o), north-facing, high elevation ( x = 3347 m) slopes (Figure 2). The dens were typically in Engelmann spruce/subalpine fir forests in areas of extensive downfall (Figures 3, 4, 5). Recaptures Two adult lynx were captured in 2001 for collar replacement. One lynx was captured in a tomahawk live-trap, the other was treed by hounds and then anesthetized using a jab pole. Five adult lynx were captured in 2002; 3 were treed by hounds and 2 were captured in padded leghold traps. In 2004, 1 lynx was captured with a Belisle snare and 6 other adult lynx were captured in box-traps. Trapping effort was substantially increased in winter and spring 2005 and 12 adult lynx were captured and re-collared. In addition, 7 kittens born in Colorado in 2004 were also captured and collared. All lynx captured in 2005 were caught in box-traps. All captured lynx were fitted with new Sirtrack TM dual VHF/satellite collars. Six adult lynx were captured from March 1999-June 30, 2005 because they were in poor body condition. Five of these lynx were successfully treated at the Frisco Creek Rehabilitation Center and rereleased in the Core Research Area. One lynx, BC00F7, died from starvation and hypothermia. Two lynx were captured because they were in atypical habitat outside the state of Colorado. They were held at Frisco Creek Rehabilitation Center for a minimum of 3 weeks and re-released in the Core Research Area in Colorado. Prior to release these lynx were fitted with new Sirtrack TM dual VHF/satellite collars. HABITAT USE Landscape-scale daytime habitat use was documented from 7421 aerial locations of lynx collected from February 1999-June 30, 2005. Throughout the year Engelmann spruce / subalpine fir was the dominant cover used by lynx. A mix of Engelmann spruce, subalpine fir and aspen (Populus tremuloides) was the second most common cover type used throughout the year. Various riparian and riparian-mix areas were the third most common cover type where lynx were found during the daytime flights. Use of Engelmann spruce-subalpine fir forests and Engelmann spruce-subalpine fir-aspen forests was similar throughout the year. There was a trend in increased use of riparian areas beginning in July, peaking in November, and dropping off December through June. Site-scale habitat data collected from snow-tracking efforts indicate Engelmann spruce and subalpine fir were also the most common forest stands used by lynx for all activities during winter in southwestern Colorado. Comparisons were made among sites used for long beds, dens, travel and where they made kills. Little difference in aspect, mean slope and mean elevation were detected for 3 of the 4 site types including long beds, travel and kills where lynx typically use gentler slopes ( x = 15.7o ) at an mean elevation of 3173 m, and varying aspects with a slight preference for north-facing slopes (Figure 2). 10 �Den sites however, were located at higher elevations ( x = 3347 m), steeper slopes ( x = 29o) and more commonly on north-facing slopes (Figure 2). Mean percent total overstory was higher for long bed and kill sites than travel or den sites (Figure 3). Engelmann spruce provided a mean of 35.87% overstory for kills and long beds, with travel sites averaging 28% and den sites having the lowest mean percent overstory of 23% (Figure 3). Mean percent subalpine fir or aspen overstory did not vary across use sites (Figure 3). Willow overstory was highly variable and no dens were located in willow overstory. A total of 1841 site-scale habitat plots were completed in winter from December 2002 through April 2005. The most common understory species at all 3 height categories above the snow (low = 00.5m, medium = 0.51 - 1.0 m, high = 1.1 - 1.5 m) was Engelmann spruce, subalpine fir, willow (Salix spp.) and aspen (Figure 4). Various other species such as Ponderosa pine (Pinus ponderosa), lodgepole pine (Pinus contorta), cottonwood (Populus sargentii), birch (Betula spp.) and others were also found in less than 5% of the habitat plots. If present, willow provided the greatest percent cover within a plot followed by Engelmann spruce, subalpine fir, aspen and coarse woody debris for long beds, kills and travel sites. Areas documented in willow used by lynx are typically on the edge of willow thickets as tracks are quickly lost within the thicket. Den sites had significantly higher percent understory cover for all three height categories. Understory at den sites was primarily made up of coarse woody debris (Figure 3). The most common tree species documented in the site-scale habitat plots was Engelmann spruce Figure 5). Subalpine fir and aspen were also present in >35% of the plots. Most habitat plots were vegetated with trees of DBH < 6" (Figure 5). As DBH increased, percent occurrence decreased within the plot. Although decreasing in abundance as size increased, most lynx use sites had trees in each of the DBH categories, indicating mature forest stands except for dens. Den sites had a broad spectrum of Engelmann spruce tree sizes, including > 18” but no large subalpine fir or aspen trees. While Engelmann spruce and subalpine fir occurred in similar densities for kills, long beds and travel sites, den sites had twice the density of subalpine firs found at all other sites (Figure 5). DIET AND HUNTING BEHAVIOR Winter diet of lynx was documented through detection of kills found through snow-tracking. Prey species from failed and successful hunting attempts were identified by either tracks or remains. Scat analysis also provided information on foods consumed. A total of 400 kills were located from February 1999-April 2005. We collected 671 scat samples from February 1999-April 2004 that will be analyzed for content. In each winter, the most common prey item was snowshoe hare, followed by red squirrel (Table 8). A comparison of percent overstory for successful and unsuccessful snowshoe hare chases indicated lynx were more successful at sites with slightly higher percent overstory, if the overstory species were Englemann spruce, subalpine fir or willow. Lynx were slightly less successful in areas of greater aspen overstory (Figure 6). This trend was repeated for percent understory at all 3 height categories (Figure 7) except that higher aspen understory improved hunting success. Higher density of Engelmann spruce and subalpine fir increased hunting success while increased aspen density decreased hunting success (Figure 8). DISCUSSION In an effort to establish a viable population of lynx in Colorado, CDOW initiated a reintroduction effort in 1997 with the first lynx released in winter 1999. From 1999 through spring 2004, 166 lynx were released in the Core Research Area. The reintroduction effort was augmented with the release of 37 11 �additional animals in April 2005 and 1 in June 2005, bringing the total to 204 lynx reintroduced to southwestern Colorado. Locations of each lynx were collected through aerial- or satellite-tracking to document movement patterns and to detect mortalities. Most lynx remain in the southwestern quarter of Colorado. Dispersal movement patterns for lynx released in 2000 and subsequent years were similar to those of lynx released in 1999. However, more animals released in 2000 and subsequent years remained within the Core Research Area than those released in 1999. This increased site fidelity may have been due to the presence of con-specifics in the area on release. Numerous travel corridors have been used repeatedly by more than 1 lynx. These travel corridors include the Cochetopa Hills area for northerly movements, the Rio Grande Reservoir-Silverton-Lizardhead Pass for movements to the west, and southerly movements down the east side of Wolf Creek Pass to the southeast to the Conejos River Valley. Lynx appear to remain faithful to an area during winter months, and exhibit more extensive movements away from these areas in the summer. Most lynx currently being tracked are within the Core Research Area. During the summer months, lynx were documented to make extensive movements away from their winter use areas. Extensive summer movements away from areas used throughout the rest of the year have been documented in native lynx in Wyoming and Montana (Squires and Laurion 1999). Human-caused mortality factors such as gunshot and vehicle collision are currently the highest causes of death. Reproduction is critical to achieving a self-sustaining viable population of lynx in Colorado. Reproduction was first documented from the 2003 reproduction season and again in 2004 and 2005. Additional reproduction is likely to have occurred in females we are no longer tracking, and from Colorado born lynx that have not been collared. The dens we find are more representative of the minimum number of litters and kittens in a reproduction season. Live-births and over-winter survival of kittens are the first steps towards recruitment into the breeding population defined as when these Colorado-born lynx will produce offspring of their own. To achieve a viable population of lynx, enough kittens need to be recruited into the population to offset the mortality that occurs in that year and hopefully even exceed the mortality rate for an increasing population. Mowat et al. (1999) suggest lynx and snowshoe hare select similar habitats except that hares select more dense stands than lynx. Very dense understory limits hunting success of the lynx and provides refugia for hares. Given the high proportion of snowshoe hare in the lynx diet in Colorado, we might then assume the habitats used by reintroduced lynx also depict areas where snowshoes hare are abundant and available for capture by lynx in Colorado. From both aerial locations taken throughout the year and from the site-scale habitat data collected in winter, the most common areas used by lynx are in stands of Engelmann spruce and subalpine fir. This is in contrast to adjacent areas of Ponderosa pine, pinyon juniper, aspen and oakbrush. The lack of lodgepole pine in the areas used by the lynx may be more reflective of the limited amount of lodgepole pine in southwestern Colorado, the Core Research Area, rather than avoidance of this tree species. Hodges (1999) summarized habitats used by snowshoe hare from 15 studies as areas of dense understory cover from shrubs, stands that are densely stocked, and stands at ages where branches have more lateral cover. Species composition and stand age appears to be less correlated with hare habitat use than is understory structure (Hodges 1999). The stands need to be old enough to provide dense cover and browse for the hares and cover for the lynx. In winter, the cover/browse needs to be tall enough to still provide browse and cover in average snow depths. Hares also use riparian areas and mature forests with understory. Site-scale habitat use documented for lynx in Colorado indicate lynx are most commonly using areas with Engelmann spruce understory present from the snow line to at least 1.5 m above the snow. The mean percent understory cover within the habitat plots is typically less than 15% regardless of understory species. However, if the understory species is willow, percent understory cover is typically 12 �double that, with mean number of shrubs per plot approximately 80, far greater than for any other understory species. In winter, hares browse on small diameter woody stems (<0.25"), bark and needles. In summer, hares shift their diet to include forbs, grasses, and other succulents as well as continuing to browse on woody stems. This shift in diet may express itself in seasonal shifts in habitat use, using more or denser coniferous cover in winter than in summer. The increased use of riparian areas by lynx in Colorado from July to November may reflect a seasonal shift in hare habitat use in Colorado. Major (1989) suggested lynx hunted the edge of dense riparian willow stands. The use of these edge habitats may allow lynx to hunt hares that live in habitats normally too dense to hunt effectively. The use of riparian areas and riparian-Engelmann spruce-subalpine fir and riparian-aspen mixes documented in Colorado may stem from a similar hunting strategy. However, too little is known about habitat use by hares in Colorado to test this hypothesis at this time. Lynx also require sufficient denning habitat. Denning habitat has been described by Koehler (1990) and Mowat et al. (1999) as areas having dense downed trees, roots, or dense live vegetation. We found this to be in true in Colorado as well. In addition, the dens used by reintroduced lynx were at high elevation, steep north-facing slopes. All females that were documented with kittens denned in areas within their winter-use area. Snow-tracking of released lynx provided information on hunting behavior and diet through documentation of kills, food caches, chases, and diet composition estimated through prey remains. Snowtracking results indicate the primary winter prey species are snowshoe hare and red squirrel, with other mammals and birds forming a minor part of the winter diet. In winter, lynx reintroduced to Colorado appear to be feeding on their preferred prey species, snowshoe hare and red squirrel in similar proportions as those reported for northern lynx during lows in the snowshoe hare cycle (Aubry et al., 1999). Caution must be used in interpreting the proportion of identified kills. Such a proportion ignores other food items that are consumed in their entirety and thus are biased towards larger prey and may not accurately represent the proportion of smaller prey items, such as microtines, in lynx winter diet. Through snowtracking we have evidence that lynx are mousing and several of the fresh carcasses have yielded small mammals in the gut on necropsy. The summer diet of lynx has been documented to include less snowshoe hare and more alternative prey than in winter (Mowat et al., 1999). All evidence suggests reintroduced lynx are finding adequate food resources. SUMMARY From results to date it can be concluded that CDOW has developed release protocols that ensure high initial post-release survival, and on an individual level lynx have demonstrated they can survive long-term in areas of Colorado. It has also been documented that reintroduced lynx could exhibit site fidelity, engage in breeding behavior and produce kittens. What is yet to be demonstrated is whether current conditions in Colorado can support the recruitment necessary to offset annual mortality for a population to sustain itself. Monitoring of reintroduced lynx will continue in an effort to document such viability. 13 �ACKNOWLEDGEMENTS The lynx reintroduction program involves the efforts of literally hundreds of people across North America, in Canada and the U. S. Any attempt to properly acknowledge all the people who played a role in this effort is at risk of missing many people. The following list should be considered to be incomplete. CDOW CLAWS Team (1998-2001): Bill Andree, Tom Beck, Gene Byrne, Bruce Gill, Mike Grode, Rick Kahn (Program Leader), Dave Kenvin, Todd Malmsbury, Jim Olterman, Dale Reed, John Seidel, Scott Wait, Margaret Wild. CDOW: John Mumma (Director 1996-2000), Conrad Albert, Jerry Apker, Laurie Baeten, Cary Carron, Don Crane, Larry DeClaire, Phil Ehrlich, Lee Flores, Delana Friedrich, Dave Gallegos, Juanita Garcia, Drayton Harrison, Jon Kindler, Ann Mangusso, Jerrie McKee, Melody Miller, Mike Miller, Kirk Navo, Robin Olterman, Jerry Pacheo, Mike Reid, Ellen Salem, Eric Schaller, Mike Sherman, Jennie Slater, Steve Steinert, Kip Stransky, Suzanne Tracey, Anne Trainor, Brad Weinmeister, Nancy Wild, Perry Will, Lisa Wolfe, Brent Woodward, Kelly Woods, Kevin Wright. Lynx Advisory Team (1998-2001): Steve Buskirk, Jeff Copeland, Dave Kenny, John Krebs, Brian Miller (Co-leader), Mike Phillips, Kim Poole, Rich Reading (Co-leader), Rob Ramey, John Weaver. U. S. Forest Service: Kit Buell, Joan Friedlander, Dale Gomez, Jerry Mastel, John Squires, Fred Wahl, Nancy Warren. U. S. Fish and Wildlife Service: Lee Carlson, Gary Patton (1998-2000), Kurt Broderdorp. State Agencies: Gary Koehler (Washington). National Park Service: Steve King. Colorado State University: Alan B. Franklin, Gary C. White. Colorado Natural Heritage Program: Rob Schorr, Mike Wunder. Alaska: ADF&G: Cathie Harms, Mark Mcnay, Dan Reed (Regional Manager), Wayne Reglin (Director), Ken Taylor (Assist. Director), Ken Whitten, Randy Zarnke, Other:Ron Perkins (trapper), Dr. Cort Zachel (veterinarian). British Columbia: Dr. Gary Armstrong (veterinarian), Mike Badry (government), Paul Blackwell (trapper coordinator), Trappers: Dennis Brown, Ken Graham, Tom Sbo, Terry Stocks, Ron Teppema, Matt Ounpuu. Yukon: Government: Arthur Hoole (Director), Harvey Jessup, Brian Pelchat, Helen Slama, Trappers: Roger Alfred, Ron Chamber, Raymond Craft, Lance Goodwin, Jerry Kruse, Elizabeth Hofer, Jurg Hofer, Guenther Mueller (YK Trapper’s Association), Ken Reeder, Rene Rivard (Trapper coordinator), Russ Rose, Gilbert Tulk, Dave Young. Alberta: Al Cook. Northwest Territories: Albert Bourque, Robert Mulders (Furbearer Biologist), Doug Steward (Director NWT Renewable Res.), Fort Providence Native People. Quebec: Luc Farrell, Pierre Fornier. Colorado Holding Facility: Herman and Susan Dieterich, Loree Harvey, Rachel Riling. Pilots: Dell Dhabolt, Larry Gepfert, Al Keith, Jim Olterman, Matt Secor, Whitey Wannamaker, Steve Waters, Dave Younkin. Field Crews (1999-2005): Steve Abele, Brandon Barr, Bryce Bateman, Todd Bayless, Ryan Besser, Mandi Brandt, Brad Buckley. Patrick Burke, Paula Capece, Stacey Ciancone, Doug Clark, John DePue, Shana Dunkley, Tim Hanks, Matt Holmes, Andy Jennings, Susan Johnson, Paul Keenlance, Tony Lavictoire, Clay Miller, Denny Morris, Kieran O’Donovan, Gene Orth, Chris Parmater, Jake Powell, Jeremy Rockweit, Josh Smith, Adam Strong, Dave Unger, David Waltz, Andy Wastell, Lyle Willmarth, Leslie Witter, Kei Yasuda, Jennifer Zahratka. Research Associates: Bob Dickman, Grant Merrill. Data Analysts: Karin Eichhoff, Joanne Stewart, Anne Trainor. Data Entry: Charlie Blackburn, Patrick Burke, Rebecca Grote, Angela Hill, Mindy Paulek. Photographs: Tom Beck, Bruce Gill, Mary Lloyd, Rich Reading, Rick Thompson. Funding: CDOW, Great Outdoors Colorado (GOCO), Turner Foundation, U.S.D.A. Forest Service, Vail Associates. 14 �LITERATURE CITED AUBRY, K. B., G. M. KOEHLER, J. R. SQUIRES. 1999. Ecology of Canada lynx in southern boreal forests. Pages 373-396 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. BYRNE, G. 1998. Core area release site selection and considerations for a Canada lynx reintroduction in Colorado. Report for the Colorado Division of Wildlife. CURTIS, J. T. 1959. The vegetation of Wisconsin. University of Wisconsin Pres, Madison. GANEY, J. L. AND W. M. BLOCK. 1994. A comparison of two techniques for measuring canopy closure. Western Journal of Applied Forestry 9:1: 21-23. HODGES, K. E. 1999. Ecology of snowshoe hares in southern boreal and montane forests. Pages 163206 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. KOEHLER, G. M. 1990. Population and habitat characteristics of lynx and snowshoe hares in north central Washington. Canadian Journal of Zoology 68:845-851. KOLBE, J. A,, J. R. SQUIRES, T. W. PARKER. 2003. An effective box trap for capturing lynx. Journal of Wildlife Management 31:980-985. LAYMON, S. A. 1988. The ecology of the spotted owl in the central Sierra Nevada, California. PhD Dissertation University of California, Berkeley, California. MAJOR, A. R. 1989. Lynx, Lynx canadensis canadensis (Kerr) predation patterns and habitat use in the Yukon Territory, Canada. M. S. Thesis, State University of New York, Syracuse. MOWAT, G., K. G. POOLE, AND M. O’DONOGHUE. 1999. Ecology of lynx in northern Canada and Alaska. Pages 265-306 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. POOLE, K. G., G. MOWAT, AND B. G. SLOUGH. 1993. Chemical immobilization of lynx. Wildlife Society Bulletin 21:136-140. SHENK, T. M. 1999. Program narrative: Post-release monitoring of reintroduced lynx (Lynx canadensis) to Colorado. Report for the Colorado Division of Wildlife. __________. 2001. Post-release monitoring of lynx reintroduced to Colorado. Job Progress Report for the Colorado Division of Wildlife. Fort Collins, Colorado. SQUIRES, J. R. AND T. LAURION. 1999. Lynx home range and movements in Montana and Wyoming: preliminary results. Pages 337-349 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University Press of Colorado, Boulder, Colorado. U. S. FISH AND WILDLIFE SERVICE. 2000. Endangered and threatened wildlife and plants: final rule to list the contiguous United States distinct population segment of the Canada lynx as a threatened species. Federal Register 65, Number 58. WILD, M. A. 1999. Lynx veterinary services and diagnostics. Job Progress Report for the Colorado Division of Wildlife. Fort Collins, Colorado. Prepared by _______________________________ Tanya M. Shenk, Wildlife Researcher 15 �Table 1. Definitions of forest structure classes used to describe habitat sites (Thomas 1979). Forest Structure Class Definition Grass/forb The grass/forb stage is created naturally by a catastrophic event, such as wildfire, and is typified by the near complete absence of snags, litter or down material in the aspen and ponderosa pine types, or vice versa in the lodgepole or subalpine forest types. Shrub/seedling The shrub/seedling stage occurs when tree seedlings or shrubs grow up to 2.5 cm at diameter breast height (DBH), either naturally or artificially through planting. Sapling/pole The sapling/pole stage is a young stage where tree DBH's range from 2.5-17.5 cm with tree heights ranging 1.8-13.5 m. These trees are 5-100 years of age, depending on species and site condition. Mature The mature stage occurs when tree diameters reach a relatively large size (25-50 cm) and the trees are usually 90 or more years old. Forest stands begin to experience accelerated mortality from disease and insects. Old-growth The old-growth stage occurs when a mature stand is at advanced age (100 years for aspen or 200 years for spruce), is very slow growing, and has advanced degrees of disease, insects, snags, and down, dead material. An exception to this occurs in ponderosa pine and aspen types where these old-growth stands typically experience low densities of down dead material or snags. Table 2. Lynx released in Colorado from February 1999 through June 30, 2005. Year Females Males TOTAL 1999 22 19 41 2000 35 20 55 2003 17 16 33 2004 17 20 37 2005 18 20 38 TOTAL 109 95 204 16 �Table 3. Causes of death for adult lynx released into southwestern Colorado in 1999-2005 as of June30, 2005. Number of Mortalities Cause of Death Unknown 22 9 Starvation Hit by Vehicle 9 Shot 8 Probable Shot 6 Plague 4 Probable Predation 2 2 Probable Hit by Vehicle Other Human Caused 2 Illness 1 Territorial Dispute 1 Total Mortalities 66 Table 4. Status of adult lynx reintroduced to Colorado as of June 30, 2005. Females Males Unknown Released 109 95 Known Dead 40 25 1 Possible Alive 69 70 Missing 16 13 Tracking 53 57 a 1 is unknown mortality TOTALS 204 66 138 28a 110 Table 5. Lynx reproduction documented in 2003. Female BC00F8 BC00F19 YK00F16 YK99F1 YK00F19 YK00F10 Release Year 2000 2000 2000 1999 2000 2000 Date Den Found 5/21/03 5/26/03 6/19/03 6/10/03 6/11/03 5/31/03 TOTAL Females ? 1 1 2 1 2 7 17 Number of Kittens Males ? 2 1 2 1 2 1 3 2 3 2 4 7 16 Total �Table 6. Lynx reproduction documented in 2004. Female ID YK00F2 AK00F2 YK00F1 YK00F15 BC00F14 BC00F18 YK00F10 BC03F02 BC03F10 BC03F09 YK00F7 YK99F1 Unknown Unknown TOTAL Release Year 2000 2000 2000 2000 2000 2000 2000 2003 2003 2003 2000 1999 Previous Litter Date Den Found 5/28/2004 5/31/2004 6/1/2004 6/4/2004 6/7/2004 6/10/2004 6/17/2004 6/25/2004 6/26/2004 6/29/2004 6/30/2004 Date Kittens Found Dec 2004 Sept 2004 Feb 2005 Number of Kittens Females Males Total 3 1 4 2 1 3 3 3 1 2 3 1 2 3 4 4 1 1 2 2 2 2 2 1 1 2 1 1 2 2 4 3 19 11 39 Table 7. Lynx reproduction documented in 2005. Female ID AK00F02 YK00F15 YK00F10 YK00F11 YK00F01 YK00F07 BC00F18 BC03F02 BC03F01 QU03F06 QU03F04 QU03F07 BC03F09 BC03F10 BC04F01 BC04F03 BC04F05 TOTAL Release year 2000 2000 2000 2000 2000 2000 2000 2003 2003 2003 2003 2003 2003 2003 2004 2004 2004 Previous Litters 2004 2004 2003, 2004 2004 2004 2004 2004 2004 2004 Date Den Found 5/21/2005 5/28/2005 6/1/2005 6/9/2005 6/10/2005 6/14/2005 6/24/2005 5/25/2005 5/27/2005 6/5/2005 6/14/2005 6/16/2005 6/27/2005 6/11/2005 6/19/2005 6/23/2005 Total 3 2 4 2 3 3 2 2 4 3 3 4 2 ? 3 3 3 46 Number of Kittens Males Females 2 1 1 1 2 2 2 2 1 1 2 1 1 1 1 2 2 3 1 2 3 1 1 1 2 3 25 1 3 21 Table 8. Number of kills found each winter field season through snow-tracking of lynx and percent composition of kills of the three primary prey species. Field Season 1999 1999-2000 2000-2001 2001-2002 2002-2003 2003-2004 2004-2005 n 9 83 89 54 65 37 78 Snowshoe Hare 55.56 67.47 67.42 90.74 90.77 67.57 83.33 Prey (%) Red Squirrel Cottontail 22.22 0 19.28 1.20 19.10 8.99 5.56 0 6.15 0 27.03 2.70 10.26 0 18 Other 22.22 12.05 4.49 3.70 3.08 2.70 6.41 �·•-•--• •-• • • e •• •--•--•. 6 1 r 1 l l i 1 l 2• • • 1 l EB 3• 1 l l 4. l l se--•--•--•--• 12meters 15meters Figure 1. Design of site-scale habitat plot sampling plot. Each of the 25 points are 3 meters apart (the first 6 points are labeled 1-6). The object that triggered a habitat plot (e.g., kill ) is the center point, depicted by the hollow circle. The actual pints encompass a 12 m x 12 m square but the understory and overstory data collected are influenced by vegetation occurring within a 15 m x 15 m square. DEN SITES LONG BED SITES KILL SITES n = 33 n = 458 n = 216 x Elevation = 3347 m x Elevation = 3179 m x SE = 33 m Slope = 29° SE = 2° x SE= 9 m Slope = 14° SE = 0.5° TRAVEL SITES n = 441 x Elevatio n= 3145 m x Elevation = 3194 m SE = 14 m SE = 9 m x Slope = 16° x Slope = 17° SE = 0.8° SE = 0.6° Figure 2. Frequency of aspect, mean elevation and SE and mean slope and SE for 4 lynx use sites; dens, long beds, kills and travel. 19 �80 70 60 ...., 50 IC: 8 40 V P.. 30 20 10 0 ES SF AS \XII Total Cover Tree Species I ■ Long Beds ■ Kills □ Travel □ Den Sitesl Figure 3. Mean percent overstory by tree species Engelmann spruce (ES), subalpine fir (SF), aspen (AS), willow (WI) and total cover for 4 different lynx use sites: long beds, kill sites, travel and den sites. Confidence intervals (95%) are depicted by error bars. 10-r----------------~ ~ ES SF C\JI/D AS WI Total 60 Co~ lO ~ ~-r-----------------, V nu1I i ~1~B·~ p.... 1".S ES SF Den Sites 40 ITT nl Cll'ID lo ll'/1 WI 20 10 T,tl.C ..., 'I' otu "' " Understory Species CWD 30 Most Common Understory Species ES= Engelmann spruce SF = Subalpine fir \l7I= Willow AS= Aspen LO = Lodgepole pine C\YD = Coarse woody debris C ov~r Figure 4. Mean percent understory by tree species Engelmann spruce (ES), subalpine fir (SF), coarse woody debris (CWD), aspen (AS), willow (WI) and total cover for 4 different lynx use sites: long beds, kill sites, travel and den sites. 20 �1000 ~ - - ~ - - - ~ 800 Kills 600 ES SF SF ES AS AS Tree Species 1 ■ o-6 □ 6-12 12-18 □ 18-24 □ >24 I Figure 5. Mean tree density by species Engelmann spruce (ES), subalpine fir (SF) and aspen (AS) and dbh size class for 4 different lynx use sites. 80 70 ...u 60 > 0 50 _µ 40 lJ t:: u ... 30 u u p_, 20 10 0 SF AS WI Total Cover Tree Species I■ SuocessfuJ. Chases ~ Unsuocessful Chases I Figure 6. Mean percent overstory by tree species Engelmann spruce (ES), subalpine fir (SF), aspen (AS), willow (WI) and total cover for successful and unsuccessful snowshoe hare chases. Confidence intervals (95%) are depicted by error bars. 21 �50 45 H 40 G) i> 35 0 u.w 30 25 ~ G) u 20 H G) 15 p., 10 5 0 ES SF C\XJD T o tal Cover \XII T ree Species ■ 0 - 0. 5 m S IE 0 - 0.5 m U ■ 0.5 - 1.0 m S IE 0.5 - 1. 0 m U □ 1.0 - 1.5 m S ll1! 10 - 1.5 m U Figure 7. Mean percent understory by tree species Engelmann spruce (ES), subalpine fir (SF), aspen (AS), willow (WI) and total cover for 3 different understory height categories for successful and unsuccessful snowshoe hare chases. Confidence intervals (95%) are depicted by error bars. 1200 I 0-6 SC ir:il 0-6 UC 1000 .... V ~ u - ■ 6-12 SC 800 ir:il 6-12 UC V ::r:: □ 12-18 SC 600 ~ 12-18 UC r.n V V .... 400 ■ 18-24 SC E--< ir:il 18-24 UC 200 I >24 SC ir:il >24 UC 0 ES SF AS Tree Species Figure 8. Mean tree density by species Engelmann spruce (ES), subalpine fir (SF) and aspen (AS) and 5 DBH size classes for successful chases (SC) and unsuccessful chases (UC) of snowshoe hare. 22 �Colorado Division of Wildlife July 2005 - June 2006 WILDLIFE RESEARCH REPORT State of Cost Center Work Package Task No. Colorado 3430 0670 1 Federal Aid Project: N/A : Division of Wildlife : Mammals Research : Lynx Conservation : Post-Release Monitoring of Lynx Reintroduced to Colorado : Period Covered: July 1, 2005 - June 30, 2006 Author: T. M. Shenk Personnel: L. Baeten, B. Diamond, R. Dickman, D. Freddy, L. Gepfert, J. Ivan, R. Kahn, A. Keith, G. Merrill, T. Spraker, S. Wait, S. Waters, L. Wolfe, D. Younkin All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT In an effort to establish a viable population of lynx (Lynx canadensis) in Colorado, the Colorado Division of Wildlife (CDOW) initiated a reintroduction effort in 1997 with the first lynx released in February 1999. From 1999-2005, 204 lynx were released in Colorado. Fourteen additional animals (8 males: 6 females) were released in spring 2006 resulting in a total of 218 lynx reintroduced to southwestern Colorado. We documented survival, movement patterns, reproduction, and habitat-use through aerial (n = 8680) and satellite (n = 18, 963) tracking. Most lynx remained near the core release area in southwestern Colorado. From 1999-2006, there were 80 mortalities of released adult lynx. Approximately 31.3% were human-induced which were attributed to collisions with vehicles or gunshot. Malnutrition and disease/illness accounted for 21.3% of the deaths while 32.5% of the deaths were from unknown causes. Reproductive females had the smallest 90% utilization distribution home ranges ( x = 75.2 km2, SE = 15.9 km2 ), followed by attending males ( x = 102.5 km2, SE = 39.7 km2) and nonreproductive animals ( x = 653.8 km2, SE = 145.4 km2). Reproduction was first documented in 2003 with subsequent successful reproduction in 2004 and 2005. Four dens with 11 kittens were found in 2006. Lynx CO04F07, a female lynx born in Colorado in 2004 was the mother of one of these litters which documented the first recruitment of Colorado-born lynx into the Colorado breeding population. From snow-tracking, the primary winter prey species (n = 426) were snowshoe hare (Lepus americanus, annual x = 75.1%, SE = 5.17) and red squirrel (Tamiasciurus hudsonicus, annual x = 15.3%, SE = 3.09); other mammals and birds formed a minor part of the winter diet. Mature Engelmann spruce (Picea engelmannii)-subalpine fir (Abies lasiocarpa) forest stands with 42-65% canopy cover and 15-20% conifer understory cover were the most commonly used areas in southwestern Colorado. Little difference in aspect (slight preference for north-facing slopes), slope ( x = 15.7°) or elevation ( x = 3173 m) were detected for long beds, travel and kill sites (n = 1841). Den sites (n = 37) however, were located at higher 1 �elevations ( x = 3354 m, SE = 31 m) on steeper ( x = 30°, SE = 2°) and more commonly north-facing slopes with a dense understory of coarse woody debris. A study to evaluate snowshoe hare densities, demography and seasonal movement patterns among small and medium tree-sized lodgepole pine stands and mature spruce/fir stands was initiated in 2005 and will continue through 2009. Results to date have demonstrated that CDOW has developed release protocols that ensure high initial post-release survival followed by high long-term survival, site fidelity, reproduction and recruitment of Colorado-born lynx into the Colorado breeding population. What is yet to be demonstrated is whether Colorado can support sufficient recruitment to offset annual mortality for a viable lynx population over time. Monitoring continues in an effort to document such viability. 2 �WILDLIFE RESEARCH REPORT POST RELEASE MONITORING OF LYNX (LYNX CANADENSIS) REINTRODUCED TO COLORADO TANYA M. SHENK P. N. OBJECTIVE The initial post-release monitoring of lynx reintroduced into Colorado will emphasize 5 primary objectives: 1. Assess and modify release protocols to ensure the highest probability of survival for each lynx released. 2. Obtain regular locations of released lynx to describe general movement patterns and habitats used by lynx. 3. Determine causes of mortality in reintroduced lynx. 4. Estimate survival of lynx reintroduced to Colorado. 5. Estimate reproduction of lynx reintroduced to Colorado. Three additional objectives will be emphasized after lynx display site fidelity to an area: 6. Refine descriptions of habitats used by reintroduced lynx. 7. Refine descriptions of daily and overall movement patterns of reintroduced lynx. 8. Describe hunting habits and prey of reintroduced lynx. Information gained to achieve these objectives will form a basis for the development of lynx conservation strategies in the southern Rocky Mountains. SEGMENT OBJECTIVES 1. Release additional adult lynx captured in Canada in southwestern Colorado during spring 2006. 2. Complete winter 2005-06 field data collection on lynx habitat use, hunting behavior, diet, mortalities, and movement patterns. 3. Complete winter 2005-06 lynx trapping field season to collar Colorado born lynx and re-collar adult lynx. 4. Complete spring 2006 field data on lynx reproduction. 5. Summarize and analyze data and publish information as Progress Reports, peer-reviewed manuscripts for appropriate scientific journals, or CDOW technical publications. 6. Complete a study plan to evaluate snowshoe hare densities, demography and seasonal movement patterns among small and medium tree-sized lodgepole pine stands and mature spruce/fir stands. INTRODUCTION The Canada lynx occurs throughout the boreal forests of northern North America. Colorado represents the southern-most historical distribution of lynx, where the species occupied the higher elevation, montane forests in the state. Little was known about the population dynamics or habitat use of this species in their southern distribution. Lynx were extirpated or reduced to a few animals in the state by the late 1970’s due, most likely, to predator control efforts such as poisoning and trapping. Given the isolation of Colorado to the nearest northern populations, the CDOW considered reintroduction as the only option to attempt to reestablish the species in the state. 3 �A reintroduction effort was begun in 1997, with the first lynx released in Colorado in 1999. To date, 218 wild-caught lynx from Alaska and Canada have been released in southwestern Colorado. The goal of the Colorado lynx reintroduction program is to establish a self-sustaining, viable population of lynx in this state. Evaluation of incremental achievements necessary for establishing viable populations is an interim method of assessing if the reintroduction effort is progressing towards success. There are 7 critical criteria for achieving a viable population: 1) development of release protocols that lead to a high initial post-release survival of reintroduced animals, 2) long-term survival of lynx in Colorado, 3) development of site fidelity by the lynx to areas supporting good habitat in densities sufficient to breed, 4) reintroduced lynx must breed, 5) breeding must lead to reproduction of surviving kittens 6) lynx born in Colorado must reach breeding age and reproduce successfully, and 7) recruitment must equal or be greater than mortality over an extended period of time. The post-release monitoring program for the reintroduced lynx has 2 primary goals. The first goal is to determine how many lynx remain in Colorado and their locations relative to each other. Given this information and knowing the sex of each individual, we can assess whether these lynx can form a breeding core from which a viable population might be established. From these data we can also describe general movement patterns and habitat use. The second primary goal of the monitoring program is to estimate survival of the reintroduced lynx and, where possible, determine causes of mortality for reintroduced lynx. Such information will help in assessing and modifying release protocols and management of lynx once they have been released to ensure their highest probability of survival. Additional goals of the post-release monitoring program for lynx reintroduced to the southern Rocky Mountains included refining descriptions of habitat use and movement patterns and describing successful hunting habitat once lynx established home ranges that encompassed their preferred habitat. Specific objectives for the site-scale habitat data collection include: 1) describe and quantify site-scale habitat use by lynx reintroduced to Colorado, 2) compare site-scale habitat use among types of sites (e.g., kills vs. long-duration beds), and 3) compare habitat features at successful and unsuccessful snowshoe hare chases. Documenting reproduction is critical to the success of the program and lynx are monitored intensively to document breeding, births, survival and recruitment of lynx born in Colorado. Site-scale habitat descriptions of den sites are also collected and compared to other sites used by lynx. The program will also investigate the ecology of snowshoe hare in Colorado. A study comparing snowshoe hare densities among mature stands of Engelmann spruce (Picea engelmannii)/subalpine fir (Abies lasiocarpa), lodgepole pine (Pinus contorta) and Ponderosa pine (Pinus ponderosa) was completed in 2004 with highest hare densities found in Engelmann spruce/subalpine fir stands and no hares found in Ponderosa pine stands. A study to evaluate the importance of young, regenerating lodgepole pine and mature Engelmann spruce/subalpine fir stands in Colorado by examining density and demography of snowshoe hares that reside in each was initiated in 2005 and will continue through 2009. Lynx is listed as threatened under the Endangered Species Act (ESA) of 1973, as amended (16 U. S. C. 1531 et. seq.)(U. S. Fish and Wildlife Service 2000). Colorado is included in the federal listing as lynx habitat. Thus, an additional objective of the post-release monitoring program is to develop conservation strategies relevant to lynx in Colorado. To develop these conservation strategies, information specific to the ecology of the lynx in its southern Rocky Mountain range, such as habitat use, movement patterns, mortality factors, survival, and reproduction in Colorado is needed. 4 �STUDY AREA Southwestern Colorado is characterized by wide plateaus, river valleys, and rugged mountains that reach elevations over 4200 m. Engelmann spruce-subalpine fir is the most widely distributed coniferous forest type at elevations most typically used by lynx. The Core Release Area is defined as areas bounded by the New Mexico state line to the south, Taylor Mesa to the west and Monarch Pass on the north and east and > 2900 m in elevation (Figure 1). The lynx-established core area is roughly bounded by areas used by lynx in the Taylor Park/ Collegiate Peak areas in central Colorado and includes areas of continuous use by lynx, including areas used during breeding and denning (Figure 1). METHODS REINTRODUCTION Effort All 2006 lynx releases were conducted under the protocols found to maximize survival (see Shenk 2001). Estimated age, sex and body condition were ascertained and recorded for each lynx prior to release (see Wild 1999). Specific release sites were those used in earlier years of the project and were selected based on land ownership and accessibility during times of release (Byrne 1998). Lynx were transported from the Frisco Creek Wildlife Rehabilitation Center, where they were held from their time of arrival in Colorado, to their release site in individual cages. Release site location was recorded in Universal Transverse Mercator (UTM) coordinates and identification of all lynx released at the same location, on the same day, was recorded. Behavior of the lynx on release and movement away from the release site were documented. Distribution and Movement Patterns All lynx released in 1999 were fitted with TelonicsTM radio-collars. All lynx released since 1999, with the exception of 5 males released in spring 2000, were fitted with SirtrackTM dual satellite/VHF radio-collars. These collars have a mortality indicator switch that operated on both the satellite and VHF mode. The satellite component of each collar was programmed to be active for 12 hours per week. The 12-hour active periods for individual collars were staggered throughout the week. Signals from the collars allowed for locations of the animals to be made via Argos, NASA, and NOAA satellites. The location information was processed by ServiceArgos and distributed to the CDOW through e-mail messages. To determine general movement patterns of reintroduced lynx, regular locations of released lynx were collected through a combination of aerial, satellite and ground radio-tracking. Locations were recorded in UTM coordinates and general habitat descriptions for each ground and aerial location were recorded. Home Range Annual home ranges were calculated as a 95% utilization distribution using a kernel home-range estimator for each lynx we had at least 30 locations for within a year. A year was defined as March 15 – March 14 of the following year. Locations used in the analyses were collected from September 1999 – January 2006 and all locations obtained for an individual during the first six months after its release were eliminated from any home range analyses as it was assumed movements of lynx initially post-release may not be representative of normal habitat use. Locations were obtained either through aerial VHF surveys or locations or the midpoint (ArcView Movement Extension) of all high quality (accuracy rating of 01km) satellite locations obtained within a single 24-hour period. All locations used within a single home range analysis were taken a minimum of 24 hours apart. 5 �Home range estimates were classified as being for a reproductive or non-reproductive animal. A reproductive female was defined as one that had kittens with her; a reproductive male was defined as a male whose movement patterns overlapped that of a reproductive female. If a litter was lost within the defined year a home range described for a reproductive animal were estimated using only locations obtained while the kittens were still with the female. Survival Survival was estimated as ragged telemetry data using the nest survival models in Program MARK (White and Burnham 1999). Mortality Factors When a mortality signal (75 beats per minute [bpm] vs. 50 bpm for the Telonics™ VHF transmitters, 20 bpm vs. 40 bpm for the Sirtrack™ VHF transmitters, 0 activity for Sirtrack™ PTT) was heard during either satellite, aerial or ground surveys, the location (UTM coordinates) was recorded. Ground crews then located and retrieved the carcass as soon as possible. The immediate area was searched for evidence of other predators and the carcass photographed in place before removal. Additionally, the mortality site was described and habitat associations and exact location were recorded. Any scat found near the dead lynx that appeared to be from the lynx was collected. All carcasses were transported to the Colorado State University Veterinary Teaching Hospital (CSUVTH) for a post mortem exam to 1) determine the cause of death and document with evidence, 2) collect samples for a variety of research projects, and 3) archive samples for future reference (research or forensic). The gross necropsy and histology were performed by, or under the lead and direct supervision of a board certified veterinary pathologist. At least one research personnel from the CDOW involved with the lynx program was also present. The protocol followed standard procedures used for thorough post-mortem examination and sample collection for histopathology and diagnostic testing (see Shenk 1999 for details). Some additional data/samples were routinely collected for research, forensics, and archiving. Other data/samples were collected based on the circumstances of the death (e.g., photographs, video, radiographs, bullet recovery, samples for toxicology or other diagnostic tests, etc.). From 1999–2004 the CDOW retained all samples and carcass remains with the exception of tissues in formalin for histopathology, brain for rabies exam, feces for parasitology, external parasites for ID, and other diagnostic samples. Since 2005 carcasses are disposed of at the CSUVTH with the exception of the lower canine, fecal samples, stomach content samples and tissue or bone marrow samples to be delivered by CDOW to the Center for Disease control for plague testing. The lower canine, from all carcasses, is sent to Matson Labs (Missoula, Montana) for aging and the fecal and stomach content samples are evaluated for diet. Reproduction Females were monitored for proximity to males during each breeding season. We defined a possible mating pair as any male and female documented within at least 1 km of each other in breeding season through either flight data or snow-tracking data. Females were then monitored for site fidelity to a given area during each denning period of May and June. Each female that exhibited stationary movement patterns in May or June were closely monitored to locate possible dens. Dens were found when field crews walked in on females that exhibited virtually no movement for at least 10 days from both aerial and ground telemetry. Kittens found at den sites were weighed, sexed and photographed. Each kitten was uniquely marked by inserting a sterile passive integrated transponder (PIT, Biomark, Inc., Boise, Idaho, USA) tag subcutaneously between the shoulder blades. Time spent at the den was minimized to ensure the least amount of disturbance to the female and the kittens. Weight, PIT-tag number, sex and any distinguishing 6 �characteristics of each kitten was also recorded. Beginning in 2005, blood and saliva samples were collected and archived for genetic identification. During the den site visits, den site location was recorded as UTM coordinates. General vegetation characteristics, elevation, weather, field personnel, time at the den, and behavioral responses of the kittens and female were also recorded. Once the females moved the kittens from the natal den area, den sites were visited again and site-specific habitat data were collected (see Habitat Use section below). Captures Captures were attempted for either lynx that were in poor body condition or lynx that needed to have their radio-collars replaced due to failed or failing batteries or to radio-collar kittens born in Colorado once they reached at least 10-months of age when they were nearly adult size. Methods of recapture included 1) trapping using a Tomahawk™ live trap baited with a rabbit and visual and scent lures, 2) calling in and darting lynx using a Dan-Inject CO2 rifle, 3) custom box-traps modified from those designed by other lynx researchers (Kolbe et al. 2003) and 4) hounds trained to pursue felids were also used to tree lynx and then the lynx was darted while treed. Lynx were immobilized either with Telazol (3 mg/kg; modified from Poole et al. 1993 as recommended by M. Wild, DVM) or medetomidine (0.09mg/kg) and ketamine (3 mg/kg; as recommended by L. Wolfe, DVM)) administered intramuscularly (IM) with either an extendible pole-syringe or a pressurized syringe-dart fired from a Dan-Inject air rifle. Immobilized lynx were monitored continuously for decreased respiration or hypothermia. If a lynx exhibited decreased respiration 2mg/kg of Dopram was administered under the tongue; if respiration was severely decreased, the animal was ventilated with a resuscitation bag. If medetomidine/ketamine were the immobilization drugs, the antagonist Atipamezole hydrochloride (Antisedan) was administered. Hypothermic (body temperature < 95o F) animals were warmed with hand warmers and blankets. While immobilized, lynx were fitted with replacement SirtrackTM VHF/satellite collar and blood and hair samples were collected. Once an animal was processed, recovery was expedited by injecting the equivalent amount of the antagonist Antisedan IM as the amount of medetomidine given, if medetomodine/ketemine was used for immobilization. Lynx were then monitored while confined in the box-trap until they were sufficiently recovered to move safely on their own. No antagonist is available for Telezol so lynx anesthetized with this drug were monitored until the animal recovered on its own in the box-trap and then released. If captured and in poor body condition, lynx were anesthetized with either Telezol (2 mg/kg) or medetomodine/ketemine and returned to the Frisco Creek Wildlife Rehabilitation Center for treatment. HABITAT USE Gross habitat use was documented by recording canopy vegetation at aerial locations. More refined descriptions of habitat use by reintroduced lynx were obtained through following lynx tracks in the snow (i.e., snow-tracking) and site-scale habitat data collection conducted at sites found through this method to be used by lynx. Snow-tracking Locations from aerial- and satellite-tracking were used to help ground-trackers locate lynx tracks in snow. Snowmobiles, where permitted, were used to gain the closest possible access to the lynx tracks without disturbing the animal. From that point, the tracking team used snowshoes to access tracks. Once tracks were found, the ground crew back- or forward-tracked the animal if it was far enough away not to be disturbed. Back-tracking generally avoided the possibility of disturbing the lynx by moving away from the animal rather than towards the animal. However, monitoring of the lynx through radio-telemetry was used to assure that the ground crew was staying a sufficient distance away from the lynx in the event the lynx might double back on its tracks. Radio-telemetry was also used in forward-tracking to make sure 7 �the team did not disturb the animal. If it appeared the lynx began to move in response to the observers, the observers stopped following the tracks. If the lynx began to move and the movement did not appear to be a response to the observers, the ground crew continued following the track. An attempt was made in Season 1 (February-May 1999) and Season 2 (December 1999-April 2000) to snow-track each lynx. In Season 3 (December 2000-April 2001), we attempted to snow-track all lynx within the Core Release Area. In tracking Season 4 (December 2001-April 2002), Season 5 (December 2002-April 2003), Season 6 (December 2003-April 2004), Season 7 (December 2004-April 2005) and Season 8 (December 2005-March 2006) we attempted to track all accessible lynx in the Core Release Area and some lynx north of the Core Release Area. Ground crews were instructed to track lynx only where it was safe to travel. Restrictions to safe travel included avalanche danger and extremely rugged terrain. Ground crews worked in pairs and were fully equipped for winter back-country survival. Data Collection For each day of tracking the date, lynx being tracked, slope, aspect, UTM coordinates, elevation, general habitat description, and summary of the days tracking were recorded. Aspect was defined as the direction of 'downhill' or 'fall line' on a slope. This is the direction along the ground in a dihedral angle between the horizontal and the plane of the ground surface. Units were compass degrees. Slope was defined as the dihedral angle between the horizontal and the plane of the ground surface (e.g., 45°). Once a track was located there were 2 types of 'sites' that were encountered. Site I areas needed documentation but either did not reflect areas lynx selected for specific habitat features, or were sites that occurred too frequently to measure each in detail. These sites included the start and end of the track being followed, the location of scat, and short-duration beds defined as being small in size (approximating an area a lynx would crouch), and with little ice formed in the bed indicating little time spent there. Site II areas included areas that might reflect specific habitat features lynx selected for and included locations where the following were found: kills, start of chases, territory marks (e.g., spray sites, buried scat, scat placed on prominent locations), long-duration beds (encompasses an area where a lynx would have lain for an extended period, iced bottom), and road crossing (both sides of road). In addition, habitat plots were conducted along lynx travel routes if no other sites were sampled in the last hour. At each of the 2 types of sites the date, lynx tracked, slope, aspect, forest structure class, UTM coordinates, and elevation were recorded. Forest structure classes included grass/forb, shrub/seedling, sapling/pole, mature, and old growth as defined in Table 1. For Site I areas, the only additional data that was collected was identification of what the site was used for (e.g., short-duration bed), and a brief description of the site. Habitat plots (see below) were conducted at Site II areas. Description of the Habitat Plot The habitat plot consisted of a 12 m x 12 m square defined by a series of 25 points placed in 5 rows of 5 with the center point being on the object that defined the site (e.g., a kill)(Figure 2). Each point was 3 m apart. The 12 m x 12 m sampling square exceeded the minimum requirement of 0.01 ha. recommended by Curtis (1959) for sampling trees. Measurements taken at each of the 25 points included: 1. Snow depth - measured vertically by an avalanche probe marked in cm. 2. Understory - measured from top of snow to 150 cm above snow in a column of 3-cm radius around the avalanche probe. Because understory measurements were influenced by vegetation outside the perimeter of the 25 sampling points (12 m x 12 m) the area used for estimating undersory cover was 15 m by 15 m. At each point, crews recorded all shrubs, trees and coarse woody debris (CWD) that fell within this column and was visible above the snow. Crews also recorded number of branches of each species that fell within the column at 3 different height categories (0-0.5 m, 0.51-1.0 m, 1.01-1.5 m). 8 �3. 4. 5. Overstory: measured at 150 cm above snow with a sighting tube. The tube was made of PVC pipe, with a curved viewing end and a crosshair made of wire on the opposite end. The sighting tube was attached to the avalanche probe used to measure snow depth. Species that hit the crosshair were recorded at each of the 25 points in the vegetation plot. Ganey and Block (1994) found this method of measuring canopy cover (with 20 sample points per plot; Laymon 1988) provided greater precision among observers. Species composition: all the different species of tree or shrub that hit the crosshair of the sighting tube at each of the 25 points were recorded. Tree composition of the vegetation plot was recorded by species and diameter at breast height (DBH). Snow depth was used in conjunction with this recorded DBH to estimate true DBH. Within the 12 m x 12 m square all conifers and deciduous trees were recorded by DBH size class (A = 0-6 in, B = 6.1-12 in, C = 12.1 -18 in, D = 18.1-24 in, E = > 24 in). Area for the tree composition analysis was 12 m x 12 m. Understory was estimated as: 1) percent occurrence within the vegetation plot (number of points with understory/total number of points surveyed) and 2) mean percent occurrence and variance by species and height category over the total points sampled within the vegetation plot. Overstory was estimated as percent occurrence over the vegetation plot (number of points with overstory/total number of points surveyed). DIET AND HUNTING BEHAVIOR Winter diet of reintroduced lynx was estimated by documenting successful kills through snowtracking. Prey species from failed and successful hunting attempts were identified by either tracks or remains. Scat analysis also provided information on foods consumed. Scat samples were collected wherever found and labeled with location and individual lynx identification. Only part of the scat was collected (approximately 75%); the remainder was left in place in the event that the scat was being used by the animal as a territory mark. Site-scale habitat data collected for successful and unsuccessful snowshoe hare kills were compared. SNOWSHOE HARE ECOLOGY A study plan was designed to evaluate the importance of young, regenerating lodgepole pine (Pinus contorta) and mature Engelmann spruce / subalpine fir stands in Colorado by examining density and demography of snowshoe hares that reside in each. Specifically, the study was designed to evaluate small and medium lodgepole pine stands and large spruce/fir stands where the classes “small”, “medium”, and “large” refer to the diameter at breast height (dbh) of overstory trees as defined in the United States Forest Service R2VEG Database (small = 2.54−12.69 cm dbh, medium = 12.70−22.85 cm, and large = 22.86−40.64 cm dbh; J. Varner, United States Forest Service, personal communication). The study design was also developed to identify which of the numerous density-estimation procedures available perform accurately and consistently using an innovative, telemetry augmentation approach as a baseline. Movement patterns and seasonal use of deciduous cover types such as riparian willow will be assessed. Finally, the study was designed to further expound on the relationship between density, demography, and stand type by examining how snowshoe hare density and demographic rates vary with specific vegetation, physical, and landscape characteristics of a stand. 9 �RESULTS REINTRODUCTION Effort From 1999 through 2005 204 lynx were reintroduced into southwestern Colorado. An additional 14 lynx were released in April 2006 (6 females: 8 males), bringing the total number of lynx released in Colorado to 218 (Table 2). Lynx released in 2006 were captured in British Columbia and Yukon. These 14 lynx were released in the Core Release Area of southwestern Colorado at or near previously used release sites in southwestern Colorado. Lynx were released with dual VHF/satellite radio collars so they could be monitored for movement, reproduction and survival. The CDOW does not plan to release any additional lynx in 2007. Distribution and Movement Patterns A total of 8680 aerial VHF locations for all 218 reintroduced lynx have been collected to date (June 30, 2006). An additional 18,963 satellite locations have been collected. Most lynx released in 2006 remained in southwestern Colorado. The majority of surviving lynx from the entire reintroduction effort continue to use high elevation (> 2900 m), forested areas from New Mexico north to Gunnison, west as far as Taylor Mesa and east to Monarch Pass. Most movements away from the Core Release Area were to the north. Numerous travel corridors have been used repeatedly by more than one lynx. These travel corridors include the Cochetopa Hills area for northerly movements, the Rio Grande Reservoir-SilvertonLizardhead Pass for movements to the west, and southerly movements down the east side of Wolf Creek Pass to the southeast through the Conejos River Valley. Lynx appear to remain faithful to an area during winter months, and exhibit more extensive movements away from these areas in the summer. Such movement patterns have also been documented by native lynx in Wyoming and Montana (Squires and Laurion 1999). Home Range Reproductive females had the smallest 90% utilization distribution annual home ranges ( x = 75.2 km2, SE = 15.9 km2, n = 19), followed by attending males ( x = 102.5 km2, SE = 39.7 km2, n = 4). Nonreproductive females had the largest annual home ranges ( x = 703.9 km2, SE = 29.8 km2, n = 32) followed by non-reproductive males ( x = 387.0 km2, SE = 73.5 km2, n = 6). Combining all nonreproductive animals yielded a mean annual home range of 653.8 km2 (SE = 145.4 km2, n = 38). Survival Initial survival rate estimates for reintroduced lynx were completed, however, further analyses need to be conducted before estimates will be presented. As of June 30, 2006, CDOW was actively tracking 95 of the 138 lynx still possibly alive. There are 43 lynx that we have not heard signals on since at least June 30, 2005 and these animals are classified as ‘missing’ (Table 3). One of these missing lynx is a mortality of unknown identity, thus only 42 are truly missing. Possible reasons for not locating these missing lynx include 1) long distance dispersal, beyond the areas currently being searched, 2) radio failure, or 3) destruction of the radio (e.g., run over by car). CDOW continues to search for all missing lynx during both aerial and ground searches. Two of the missing lynx released in 2000 are thought to have slipped their collars. Mortality Factors Of the total 218 adult lynx released from 1999-2006 there are 80 known mortalities as of June 30, 2006. Causes of death are listed in Table 4. Starvation was a significant cause of mortality in the first 10 �year of releases only. Mortalities occurred throughout the areas through which lynx moved. Approximately 31.3% were human-induced which were attributed to collisions with vehicles or gunshot. Malnutrition and disease/illness accounted for 21.3% of the deaths while 32.5% of the deaths were from unknown causes (Table 4). Reproduction 2003.-- Nine pairs of lynx were documented during the 2003 breeding season (March and April) from the 17 females we were monitoring. In May and June, 6 dens and a total of 16 kittens were found in the lynx Core Release Area in southwestern Colorado (Table 5, Figure 1). At all dens the females appeared in excellent condition, as did the kittens. The kittens weighed from 270-500 grams. Lynx kittens weigh approximately 200 grams at birth and do not open their eyes until they are 10-17 days old. The dens were scattered throughout the Core Release Area, with no dens found outside the core area. All the dens were in Engelmann spruce/subalpine fir forests in areas of extensive downfall. Elevations ranged from 3240-3557 m. Field crews weighed, photographed, PIT-tagged the kittens and took hair samples from the kittens for genetic work in an attempt to confirm paternity. Kittens were processed as quickly as possible (11-32 minutes) to minimize the time the kittens were without their mother. While working with the kittens the females remained nearby, often making themselves visible to the field crews. The females generally continued a low growling vocalization the entire time personnel were at the den. In all cases, the female returned to the den site once field crews left the area. Four of the 6 females that we know had kittens in summer 2003 were still with kittens at the end of April 2004. Two of those females still had 2 kittens with them at that time. Visual observations in February 2004 of one female with 2 kittens indicated all 3 were in good body condition. The mortality of female YK00F16 and her 1 kitten in October 2003 from plague was not due to poor habitat or prey conditions, and thus we might assume she would have raised the 1 kitten to this stage as well. Three probable kitten deaths from female YK00F19 were from 1 litter that most likely failed very early. Through snow-tracking in winter 2003-04 an unknown female (no radio frequency heard in the area of the tracks) we also documented 1-2 additional kittens born spring 2003 and still alive in winter 2004. Of the 16 kittens we found in summer 2003, we documented the following by April 2004: 6 confirmed alive, 7 confirmed dead, and 3 some evidence dead. Although we tried, we were not able to capture any of the 6 surviving kittens to fit them with radio-collars for subsequent monitoring. 2004.-- In Spring 2004, 26 females from the releases in 1999, 2000 and 2003 had active radiocollars. Of these, we documented 18 possible mating pairs of lynx during breeding season. All 4 of the females that had kittens with them through winter 2003-04 bred again spring 2004; 2 with the same male they successfully bred with spring 2003. During May-June 2004 we found 11 dens and a total of 30 kittens (Table 6). At all dens the females appeared in excellent condition, as did the kittens. The kittens weighed from 250-770 grams. Three of the 11 females with kittens were from the 2003 releases (Table 6). Three additional litters were documented after denning season through either observation of a female lynx with kittens or snow-tracking females with kittens that were not one of the 11 females found on dens. From the size of the kittens they would have been born during the normal denning season in May or June. Nine additional kittens were observed from these litters for a total of 39 known kittens born in 2004. Two of these additional litters were documented from direct follow-ups to sighting made by the public and reported to CDOW. Two females that had kittens in 2003 and reared at least part of their litters through March 2004, bred and had kittens again in 2004. Two of the litters documented by direct observation or snow-tracking are from females whose collars were no longer functioning. Seven kittens born in 2004 were captured at approximately 10-months of age and fitted with dual satellite/VHF collars. Six of the 7 were still alive 11 �and being monitored as of June 30, 2006. The cut collar of one kitten CO04M15 was left at the Silverton Post Office on October 25, 2005. We assume this lynx is dead. 2005.-- In spring 2005 we had 40 females from the releases in 1999, 2000, 2003 and 2004 that had active radio-collars. We documented 23 possible mating pairs of lynx during breeding season. During May-June 2005 we visited 16 dens and found a total of 46 kittens (Table 7). At all dens the females appeared in excellent condition, as did the kittens. An additional female (BC03F10) had a den we were not able to get to during May or June due to high water during spring run-off. Female BC03F03 was hit and killed on I-70 on 5/19/2005. She had 2 fetuses in her uterus, so would have contributed to reproduction this year had she lived. We weighed, photographed, PIT-tagged the kittens and recorded sex. We also took blood samples from the kittens for genetic work in an attempt to confirm paternity. While we were working with the kittens the females remained nearby, often remaining visible to us. The females generally continued a low growling vocalization the entire time we were at the den. In all cases, the female returned to the den site once we left the area. All of the 2005 dens were scattered throughout the high elevation areas of Colorado, south of I70. Most of the dens were in Engelmann spruce/subalpine fir forests in areas of extensive downfall. Elevations ranged from 3117-3586 m. We weighed, photographed, and PIT-tagged the kittens, recorded sex and took hair samples from the kittens for genetic work in an attempt to confirm paternity. Four of the females would not leave the den until we reached out to pick up a kitten. While we were working with the kittens the females remained nearby, often remaining visible to us. The females generally continued a low growling vocalization the entire time we were at the den. In all cases, the female returned to the den site once we left the area. One female, YK00F10 has had litters 3 years in a row. In 2003 she had 4 kittens and raised 2 through the winter. In 2004 she had 2 kittens and raised both through the winter, in 2005 she had 4 kittens again. She has had all 3 litters in the same general area and has had the same mate for 3 years. Eight additional females had their second litter in Colorado in 2005. Three females from the 2004 releases had litters in 2005. Year 2005 was the second consecutive year that we had females released the prior spring, find a territory and a mate within a year and produced live young. In reproduction season 2004 we had 3 females released in spring 2003 that also produced live young the next year. Of those 3, 2 successfully raised at least part of their litters through winter 2005. Seven kittens born in 2005 were captured at approximately 10-months of age and fitted with dual satellite/VHF collars. One of the 7 was still alive and being monitored as of June 30, 2006. 2006.--In spring 2006, 42 females were being monitored. We found 4 dens in May and June 2006 with 11 kittens total (Table 8). Lynx CO04F07, a female lynx born in Colorado in 2004, was the mother of one of these litters which documented the first recruitment of Colorado-born lynx into the Colorado breeding population. The percent of tracked females found with litters in 2006 was lower (0.095) than in the 3 previous years (0.413, SE = 0.032, Table 9). However, all demographic and habitat characteristics measured at the 4 dens that were found in 2006 were comparable to all other dens found (Table 9). Mean number of kittens per litter from 2003-2006 was 2.78 (SE = 0.05) and sex ratio of females to males was equal ( x = 1.14, SE = 0.14). 12 �Den Sites.--A total of 37 dens have been found from 2003-2006. All of the dens except one have been scattered throughout the high elevation areas of Colorado, south of I-70. In 2004, 1 den was found in southeastern Wyoming, near the Colorado border. Dens were located on steep ( x slope = 30o , SE=2o), north-facing, high elevation ( x = 3354 m, SE = 31 m) slopes (Figure 3). The dens were typically in Engelmann spruce/subalpine fir forests in areas of extensive downfall of coarse woody debris (Figures 4, 5, 6). All dens were located within the winter use areas used by the females. Captures Two adult lynx were captured in 2001 for collar replacement. One lynx was captured in a tomahawk live-trap, the other was treed by hounds and then anesthetized using a jab pole. Five adult lynx were captured in 2002; 3 were treed by hounds and 2 were captured in padded leghold traps. In 2004, 1 lynx was captured with a Belisle snare and 6 other adult lynx were captured in box-traps. Trapping effort was substantially increased in winter and spring 2005 and 12 adult lynx were captured and re-collared. Eight reintroduced lynx were captured in winter and spring 2006. All lynx captured in 2005 and 2006 were caught in box-traps. All captured lynx were fitted with new Sirtrack TM dual VHF/satellite collars. Seven adult lynx were captured from March 1999-June 30, 2006 because they were in poor body condition. Five of these lynx were successfully treated at the Frisco Creek Rehabilitation Center and rereleased in the Core Release Area. One lynx, BC00F7, died from starvation and hypothermia. Lynx QU04M07 died on February 5, 2006 at the rehabilitation center. Necropsy results documented starvation as the cause of death that was precipitated by hydrocephalus and bronchopneumonia (unpublished data T. Spraker, CSUVTH). Two lynx were captured because they were in atypical habitat outside the state of Colorado. They were held at Frisco Creek Rehabilitation Center for a minimum of 3 weeks and rereleased in the Core Release Area in Colorado. Prior to release these lynx were fitted with new Sirtrack TM dual VHF/satellite collars. In addition, 14 Colorado-born kittens were captured and collared at approximately 10-months of age. Seven 2004-born kittens were collared in spring 2005, and 7 2005-born kittens collared in spring 2006. HABITAT USE Landscape-scale daytime habitat use was documented from 7421 aerial locations of lynx collected from February 1999-June 30, 2005. Throughout the year Engelmann spruce / subalpine fir was the dominant cover used by lynx. A mix of Engelmann spruce, subalpine fir and aspen (Populus tremuloides) was the second most common cover type used throughout the year. Various riparian and riparian-mix areas were the third most common cover type where lynx were found during the daytime flights. Use of Engelmann spruce-subalpine fir forests and Engelmann spruce-subalpine fir-aspen forests was similar throughout the year. There was a trend in increased use of riparian areas beginning in July, peaking in November, and dropping off December through June. Site-scale habitat data collected from snow-tracking efforts indicate Engelmann spruce and subalpine fir were also the most common forest stands used by lynx for all activities during winter in southwestern Colorado. Comparisons were made among sites used for long beds, dens, travel and where they made kills. Little difference in aspect, mean slope and mean elevation were detected for 3 of the 4 site types including long beds, travel and kills where lynx typically use gentler slopes ( x = 15.7o ) at an mean elevation of 3173 m, and varying aspects with a slight preference for north-facing slopes (Figure 3). Mean percent total overstory was higher for long bed and kill sites than travel or den sites (Figure 4). Engelmann spruce provided a mean of 35.87% overstory for kills and long beds, with travel sites averaging 28% and den sites having the lowest mean percent overstory of 23% (Figure 4). Mean percent 13 �subalpine fir or aspen overstory did not vary across use sites (Figure 4). Willow overstory was highly variable and no dens were located in willow overstory. A total of 1841 site-scale habitat plots were completed in winter from December 2002 through April 2005. The most common understory species at all 3 height categories above the snow (low = 00.5m, medium = 0.51 - 1.0 m, high = 1.1 - 1.5 m) was Engelmann spruce, subalpine fir, willow (Salix spp.) and aspen (Figure 5). Various other species such as Ponderosa pine (Pinus ponderosa), lodgepole pine (Pinus contorta), cottonwood (Populus sargentii), birch (Betula spp.) and others were also found in less than 5% of the habitat plots. If present, willow provided the greatest percent cover within a plot followed by Engelmann spruce, subalpine fir, aspen and coarse woody debris for long beds, kills and travel sites. Areas documented in willow used by lynx are typically on the edge of willow thickets as tracks are quickly lost within the thicket. Den sites had significantly higher percent understory cover for all three height categories. Understory at den sites was primarily made up of coarse woody debris (Figure 5). The most common tree species documented in the site-scale habitat plots was Engelmann spruce Figure 6). Subalpine fir and aspen were also present in >35% of the plots. Most habitat plots were vegetated with trees of DBH < 6" (Figure 6). As DBH increased, percent occurrence decreased within the plot. Although decreasing in abundance as size increased, most lynx use sites had trees in each of the DBH categories, indicating mature forest stands except for dens. Den sites had a broad spectrum of Engelmann spruce tree sizes, including > 18” but no large subalpine fir or aspen trees. While Engelmann spruce and subalpine fir occurred in similar densities for kills, long beds and travel sites, den sites had twice the density of subalpine firs found at all other sites (Figure 6). DIET AND HUNTING BEHAVIOR Winter diet of lynx was documented through detection of kills found through snow-tracking. Prey species from failed and successful hunting attempts were identified by either tracks or remains. Scat analysis also provided information on foods consumed. A total of 400 kills were located from February 1999-April 2005. We collected 671 scat samples from February 1999-April 2004 that will be analyzed for content. In each winter, the most common prey item was snowshoe hare, followed by red squirrel (Table 10). A comparison of percent overstory for successful and unsuccessful snowshoe hare chases indicated lynx were more successful at sites with slightly higher percent overstory, if the overstory species were Englemann spruce, subalpine fir or willow. Lynx were slightly less successful in areas of greater aspen overstory (Figure 7). This trend was repeated for percent understory at all 3 height categories (Figure 8) except that higher aspen understory improved hunting success. Higher density of Engelmann spruce and subalpine fir increased hunting success while increased aspen density decreased hunting success (Figure 9). SNOWSHOE HARE ECOLOGY A study plan was completed to evaluate snowshoe hare densities, demography and seasonal movement patterns among small and medium tree-sized lodgepole pine stands and mature spruce/fir stands (Appendix I). DISCUSSION In an effort to establish a viable population of lynx in Colorado, CDOW initiated a reintroduction effort in 1997 with the first lynx released in winter 1999. From 1999 through spring 2005, 204 lynx were 14 �released in the Core Release Area. The reintroduction effort was augmented with the release of 14 additional animals in April 2006, bringing the total to 218 lynx reintroduced to southwestern Colorado. Locations of each lynx were collected through aerial- or satellite-tracking to document movement patterns and to detect mortalities. Most lynx remain in the high elevation, forested areas in southwestern Colorado. Dispersal movement patterns for lynx released in 2000 and subsequent years were similar to those of lynx released in 1999. However, more animals released in 2000 and subsequent years remained within the Core Release Area than those released in 1999. This increased site fidelity may have been due to the presence of con-specifics in the area on release. Numerous travel corridors have been used repeatedly by more than 1 lynx. These travel corridors include the Cochetopa Hills area for northerly movements, the Rio Grande Reservoir-Silverton-Lizardhead Pass for movements to the west, and southerly movements down the east side of Wolf Creek Pass to the southeast to the Conejos River Valley. Lynx appear to remain faithful to an area during winter months, and exhibit more extensive movements away from these areas in the summer. Most lynx currently being tracked are within the Core Release Area. During the summer months, lynx were documented to make extensive movements away from their winter use areas. Extensive summer movements away from areas used throughout the rest of the year have been documented in native lynx in Wyoming and Montana (Squires and Laurion 1999). Humancaused mortality factors such as gunshot and vehicle collision are currently the highest causes of death. Reproduction is critical to achieving a self-sustaining viable population of lynx in Colorado. Reproduction was first documented from the 2003 reproduction season and again in 2004, 2005 and 2006. Reproduction in 2006 included a Colorado-born female giving birth to 2 kittens, documenting the first recruitment of Colorado-born lynx into the Colorado breeding population. Additional reproduction is likely to have occurred in all years from females we are no longer tracking, and from Colorado-born lynx that have not been collared. The dens we find are more representative of the minimum number of litters and kittens in a reproduction season. To achieve a viable population of lynx, enough kittens need to be recruited into the population to offset the mortality that occurs in that year and hopefully even exceed the mortality rate for an increasing population. Mowat et al. (1999) suggest lynx and snowshoe hare select similar habitats except that hares select more dense stands than lynx. Very dense understory limits hunting success of the lynx and provides refugia for hares. Given the high proportion of snowshoe hare in the lynx diet in Colorado, we might then assume the habitats used by reintroduced lynx also depict areas where snowshoes hare are abundant and available for capture by lynx in Colorado. From both aerial locations taken throughout the year and from the site-scale habitat data collected in winter, the most common areas used by lynx are in stands of Engelmann spruce and subalpine fir. This is in contrast to adjacent areas of Ponderosa pine, pinyon juniper, aspen and oakbrush. The lack of lodgepole pine in the areas used by the lynx may be more reflective of the limited amount of lodgepole pine in southwestern Colorado, the Core Release Area, rather than avoidance of this tree species. Hodges (1999) summarized habitats used by snowshoe hare from 15 studies as areas of dense understory cover from shrubs, stands that are densely stocked, and stands at ages where branches have more lateral cover. Species composition and stand age appears to be less correlated with hare habitat use than is understory structure (Hodges 1999). The stands need to be old enough to provide dense cover and browse for the hares and cover for the lynx. In winter, the cover/browse needs to be tall enough to still provide browse and cover in average snow depths. Hares also use riparian areas and mature forests with understory. Site-scale habitat use documented for lynx in Colorado indicate lynx are most commonly using areas with Engelmann spruce understory present from the snow line to at least 1.5 m above the snow. The mean percent understory cover within the habitat plots is typically less than 15% regardless of understory species. However, if the understory species is willow, percent understory cover is typically 15 �double that, with mean number of shrubs per plot approximately 80, far greater than for any other understory species. In winter, hares browse on small diameter woody stems (<0.25"), bark and needles. In summer, hares shift their diet to include forbs, grasses, and other succulents as well as continuing to browse on woody stems. This shift in diet may express itself in seasonal shifts in habitat use, using more or denser coniferous cover in winter than in summer. The increased use of riparian areas by lynx in Colorado from July to November may reflect a seasonal shift in hare habitat use in Colorado. Major (1989) suggested lynx hunted the edge of dense riparian willow stands. The use of these edge habitats may allow lynx to hunt hares that live in habitats normally too dense to hunt effectively. The use of riparian areas and riparian-Engelmann spruce-subalpine fir and riparian-aspen mixes documented in Colorado may stem from a similar hunting strategy. However, too little is known about habitat use by hares in Colorado to test this hypothesis at this time. Lynx also require sufficient denning habitat. Denning habitat has been described by Koehler (1990) and Mowat et al. (1999) as areas having dense downed trees, roots, or dense live vegetation. We found this to be in true in Colorado as well. In addition, the dens used by reintroduced lynx were at high elevations and on steep north-facing slopes. All females that were documented with kittens denned in areas within their winter-use area. Snow-tracking of released lynx provided information on hunting behavior and diet through documentation of kills, food caches, chases, and diet composition estimated through prey remains. Snowtracking results indicate the primary winter prey species are snowshoe hare and red squirrel, with other mammals and birds forming a minor part of the winter diet. In winter, lynx reintroduced to Colorado appear to be feeding on their preferred prey species, snowshoe hare and red squirrel in similar proportions as those reported for northern lynx during lows in the snowshoe hare cycle (Aubry et al., 1999). Caution must be used in interpreting the proportion of identified kills. Such a proportion ignores other food items that are consumed in their entirety and thus are biased towards larger prey and may not accurately represent the proportion of smaller prey items, such as microtines, in lynx winter diet. Through snowtracking we have evidence that lynx are mousing and several of the fresh carcasses have yielded small mammals in the gut on necropsy. The summer diet of lynx has been documented to include less snowshoe hare and more alternative prey than in winter (Mowat et al., 1999). All evidence suggests reintroduced lynx are finding adequate food resources. SUMMARY From results to date it can be concluded that CDOW has developed release protocols that ensure high initial post-release survival, and on an individual level lynx have demonstrated they can survive long-term in areas of Colorado. It has also been documented that reintroduced lynx could exhibit site fidelity, engage in breeding behavior and produce kittens that are recruited into the Colorado breeding population. What is yet to be demonstrated is whether current conditions in Colorado can support the recruitment necessary to offset annual mortality for a population to sustain itself. Monitoring of reintroduced lynx will continue in an effort to document such viability. ACKNOWLEDGEMENTS The lynx reintroduction program involves the efforts of literally hundreds of people across North America, in Canada and the U. S. Any attempt to properly acknowledge all the people who played a role in this effort is at risk of missing many people. The following list should be considered to be incomplete. 16 �CDOW CLAWS Team (1998-2001): Bill Andree, Tom Beck, Gene Byrne, Bruce Gill, Mike Grode, Rick Kahn (Program Leader), Dave Kenvin, Todd Malmsbury, Jim Olterman, Dale Reed, John Seidel, Scott Wait, Margaret Wild. CDOW: John Mumma (Director 1996-2000), Bruce McCloskey (Director 2001-present), Conrad Albert, Jerry Apker, Laurie Baeten, Cary Carron, Don Crane, Larry DeClaire, Phil Ehrlich, Lee Flores, Delana Friedrich, Dave Gallegos, Juanita Garcia, Drayton Harrison, Jon Kindler, Ann Mangusso, Jerrie McKee, Melody Miller, Mike Miller, Kirk Navo, Robin Olterman, Jerry Pacheo, Mike Reid, Ellen Salem, Eric Schaller, Mike Sherman, Jennie Slater, Steve Steinert, Kip Stransky, Suzanne Tracey, Anne Trainor, Brad Weinmeister, Nancy Wild, Perry Will, Lisa Wolfe, Brent Woodward, Kelly Woods, Kevin Wright. Lynx Advisory Team (1998-2001): Steve Buskirk, Jeff Copeland, Dave Kenny, John Krebs, Brian Miller (Co-leader), Mike Phillips, Kim Poole, Rich Reading (Co-leader), Rob Ramey, John Weaver. U. S. Forest Service: Kit Buell, Joan Friedlander, Dale Gomez, Jerry Mastel, John Squires, Fred Wahl, Nancy Warren. U. S. Fish and Wildlife Service: Lee Carlson, Gary Patton (1998-2000), Kurt Broderdorp. State Agencies: Gary Koehler (Washington). National Park Service: Steve King. Colorado State University: Alan B. Franklin, Gary C. White. Colorado Natural Heritage Program: Rob Schorr, Mike Wunder. Alaska: ADF&G: Cathie Harms, Mark Mcnay, Dan Reed (Regional Manager), Wayne Reglin (Director), Ken Taylor (Assist. Director), Ken Whitten, Randy Zarnke, Other:Ron Perkins (trapper), Dr. Cort Zachel (veterinarian). British Columbia: Dr. Gary Armstrong (veterinarian), Mike Badry (government), Paul Blackwell (trapper coordinator), Trappers: Dennis Brown, Ken Graham, Tom Sbo, Terry Stocks, Ron Teppema, Matt Ounpuu. Yukon: Government: Arthur Hoole (Director), Harvey Jessup, Brian Pelchat, Helen Slama, Trappers: Roger Alfred, Ron Chamber, Raymond Craft, Lance Goodwin, Jerry Kruse, Elizabeth Hofer, Jurg Hofer, Guenther Mueller (YK Trapper’s Association), Ken Reeder, Rene Rivard (Trapper coordinator), Russ Rose, Gilbert Tulk, Dave Young. Alberta: Al Cook. Northwest Territories: Albert Bourque, Robert Mulders (Furbearer Biologist), Doug Steward (Director NWT Renewable Res.), Fort Providence Native People. Quebec: Luc Farrell, Pierre Fornier. Colorado Holding Facility: Herman and Susan Dieterich, Loree Harvey, Rachel Riling. Pilots: Dell Dhabolt, Larry Gepfert, Al Keith, Jim Olterman, Matt Secor, Whitey Wannamaker, Steve Waters, Dave Younkin. Field Crews (1999-2006): Steve Abele, Brandon Barr, Bryce Bateman, Todd Bayless, Nathan Berg, Ryan Besser, Mandi Brandt, Brad Buckley. Patrick Burke, Braden Burkholder, Paula Capece, Stacey Ciancone, Doug Clark, John DePue, Shana Dunkley, Tim Hanks, Dan Haskell, Matt Holmes, Andy Jennings, Susan Johnson, Paul Keenlance, Patrick Kolar, Tony Lavictoire, Clay Miller, Denny Morris, Kieran O’Donovan, Gene Orth, Chris Parmater, Jake Powell, Jeremy Rockweit, Jenny Shrum, Josh Smith, Heather Stricker, Adam Strong, Dave Unger, David Waltz, Andy Wastell, Lyle Willmarth, Leslie Witter, Kei Yasuda, Jennifer Zahratka. Research Associates: Bob Dickman, Grant Merrill. Data Analysts: Karin Eichhoff, Joanne Stewart, Anne Trainor. Data Entry: Charlie Blackburn, Patrick Burke, Rebecca Grote, Angela Hill, Mindy Paulek. Photographs: Tom Beck, Bruce Gill, Mary Lloyd, Rich Reading, Rick Thompson. Funding: CDOW, Great Outdoors Colorado (GOCO), Turner Foundation, U.S.D.A. Forest Service, Vail Associates. LITERATURE CITED AUBRY, K. B., G. M. KOEHLER, J. R. SQUIRES. 1999. Ecology of Canada lynx in southern boreal forests. Pages 373-396 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. BYRNE, G. 1998. Core area release site selection and considerations for a Canada lynx reintroduction in Colorado. Report for the Colorado Division of Wildlife. CURTIS, J. T. 1959. The vegetation of Wisconsin. University of Wisconsin Pres, Madison. GANEY, J. L. AND W. M. BLOCK. 1994. A comparison of two techniques for measuring canopy closure. Western Journal of Applied Forestry 9:1: 21-23. 17 �HODGES, K. E. 1999. Ecology of snowshoe hares in southern boreal and montane forests. Pages 163206 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. KOEHLER, G. M. 1990. Population and habitat characteristics of lynx and snowshoe hares in north central Washington. Canadian Journal of Zoology 68:845-851. KOLBE, J. A., J. R. SQUIRES, T. W. PARKER. 2003. An effective box trap for capturing lynx. Journal of Wildlife Management 31:980-985. LAYMON, S. A. 1988. The ecology of the spotted owl in the central Sierra Nevada, California. PhD Dissertation University of California, Berkeley, California. MAJOR, A. R. 1989. Lynx, Lynx canadensis canadensis (Kerr) predation patterns and habitat use in the Yukon Territory, Canada. M. S. Thesis, State University of New York, Syracuse. MOWAT, G., K. G. POOLE, AND M. O’DONOGHUE. 1999. Ecology of lynx in northern Canada and Alaska. Pages 265-306 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. POOLE, K. G., G. MOWAT, AND B. G. SLOUGH. 1993. Chemical immobilization of lynx. Wildlife Society Bulletin 21:136-140. SHENK, T. M. 1999. Program narrative: Post-release monitoring of reintroduced lynx (Lynx canadensis) to Colorado. Report for the Colorado Division of Wildlife. __________. 2001. Post-release monitoring of lynx reintroduced to Colorado. Job Progress Report for the Colorado Division of Wildlife. Fort Collins, Colorado. SQUIRES, J. R. AND T. LAURION. 1999. Lynx home range and movements in Montana and Wyoming: preliminary results. Pages 337-349 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University Press of Colorado, Boulder, Colorado. U. S. FISH AND WILDLIFE SERVICE. 2000. Endangered and threatened wildlife and plants: final rule to list the contiguous United States distinct population segment of the Canada lynx as a threatened species. Federal Register 65, Number 58. WHITE, G.C. AND K. P. BURNHAM. 1999. Program MARK: Survival estimation from populations of marked animals. Bird Study 46 Supplement, 120-138. WILD, M. A. 1999. Lynx veterinary services and diagnostics. Job Progress Report for the Colorado Division of Wildlife. Fort Collins, Colorado. Prepared by _______________________________ Tanya M. Shenk, Wildlife Researcher 18 �Table 1. Definitions of forest structure classes used to describe habitat sites (Thomas 1979). Forest Structure Class Definition Grass/forb The grass/forb stage is created naturally by a catastrophic event, such as wildfire, and is typified by the near complete absence of snags, litter or down material in the aspen and ponderosa pine types, or vice versa in the lodgepole or subalpine forest types. Shrub/seedling The shrub/seedling stage occurs when tree seedlings or shrubs grow up to 2.5 cm at diameter breast height (DBH), either naturally or artificially through planting. Sapling/pole The sapling/pole stage is a young stage where tree DBH's range from 2.517.5 cm with tree heights ranging 1.8-13.5 m. These trees are 5-100 years of age, depending on species and site condition. Mature The mature stage occurs when tree diameters reach a relatively large size (25-50 cm) and the trees are usually 90 or more years old. Forest stands begin to experience accelerated mortality from disease and insects. Old-growth The old-growth stage occurs when a mature stand is at advanced age (100 years for aspen or 200 years for spruce), is very slow growing, and has advanced degrees of disease, insects, snags, and down, dead material. An exception to this occurs in ponderosa pine and aspen types where these old-growth stands typically experience low densities of down dead material or snags. Table 2. Lynx released in Colorado from February 1999 through June 30, 2006. Year Females Males TOTAL 1999 22 19 41 2000 35 20 55 2003 17 16 33 2004 17 20 37 2005 18 20 38 2006 6 8 14 TOTAL 115 103 218 Table 3. Status of adult lynx reintroduced to Colorado as of June 30, 2006. Females Released Known Dead Possible Alive Missing Tracking a 1 is unknown mortality 115 46 69 20 49 Males 103 33 70 24 46 19 Unknown 1 TOTALS 218 80 138 43a 95 �Table 4. Causes of death for lynx released into southwestern Colorado from 1999-2006 as of June30, 2006. Cause of Death Unknown Hit by Vehicle Starvation Shot Other Trauma Probable Shot Plague Predation Probable Predation Illness Total Mortalities Number of Mortalities 26 11 10 9 7 5 5 3 2 2 80 Table 5. Lynx reproduction documented in 2003. Female BC00F8 BC00F19 YK00F16 YK99F1 YK00F19 YK00F10 Release Year 2000 2000 2000 1999 2000 2000 Date Den Found 5/21/03 5/26/03 6/19/03 6/10/03 6/11/03 5/31/03 TOTAL Females ? 1 1 2 1 2 7 Number of Kittens Males ? 1 1 1 2 2 7 Total 2 2 2 3 3 4 16 Table 6. Lynx reproduction documented in 2004. Female ID YK00F2 AK00F2 YK00F1 YK00F15 BC00F14 BC00F18 YK00F10 BC03F02 BC03F10 BC03F09 YK00F7 YK99F1 Unknown Unknown TOTAL Release Year 2000 2000 2000 2000 2000 2000 2000 2003 2003 2003 2000 1999 Previous Litters Date Den Found 5/28/2004 5/31/2004 6/1/2004 6/4/2004 6/7/2004 6/10/2004 6/17/2004 6/25/2004 6/26/2004 6/29/2004 6/30/2004 6/2004 Date Kittens Found Dec 2004 Sept 2004 Feb 2005 20 Number of Kittens Females Males Total 3 1 4 2 1 3 3 3 1 2 3 1 2 3 4 4 1 1 2 2 2 2 2 1 1 2 1 1 2 2 4 3 19 11 39 �Table 7. Lynx reproduction documented in 2005. Female ID AK00F02 YK00F15 YK00F10 YK00F11 YK00F01 YK00F07 BC00F18 BC03F02 BC03F01 QU03F06 QU03F04 QU03F07 BC03F09 BC03F10 BC04F01 BC04F03 BC04F05 BC04F04 TOTAL Release Year 2000 2000 2000 2000 2000 2000 2000 2003 2003 2003 2003 2003 2003 2003 2004 2004 2004 2004 Previous Litters 2004 2004 2003, 2004 2004 2004 2004 2004 2004 2004 Date Den Found 5/21/2005 5/28/2005 6/1/2005 6/9/2005 6/10/2005 6/14/2005 6/24/2005 5/25/2005 5/27/2005 6/5/2005 6/14/2005 6/16/2005 6/27/2005 6/2005 6/11/2005 6/19/2005 6/23/2005 Date Kittens Found 12/20/2005 12/10/2005 Number of Kittens Males Females Total 2 1 3 1 1 2 2 2 4 2 2 2 1 3 1 2 3 1 1 2 1 1 2 2 2 4 3 3 1 2 3 3 1 4 1 1 2 2 2 1 3 1 3 4 3 3 1 1 26 22 50 Table 8. Lynx reproduction in 2006. Female ID AK00F15 AK00F05 BC03F10 CO04F07 TOTAL Release Year 2000 2000 2003 Year Born in Colorado Previous Litters 2004, 2005 2004 2004, 2005 Date Den Found 5/21/2006 6/7/2006 6/9/2006 6/17/2006 2004 Number of Kittens Males Females Total 1 3 4 1 2 3 1 1 2 2 2 5 6 11 Table 9. Lynx reproduction summary statistics for 2003-2006. Additional Litters Found in Winter Mean # Kittens/Litter 16 0.462 2 0.425 1 2.67 (SE = 0.33 2.83 (SE = 0.24) 2.88 (SE = 0.18) 2.75 (SE = 0.47) 2.78 (SE = 0.05) 11 Year # Females Tracked # Dens Found in May/June % Tracked Females with Kittens 2003 17 6 0.353 2004 26 11 2005 40 17 Mean 2003-05 2006 Mean 2003-06 Total Kittens Found 39 50 Sex Ratio M/F 1.0 1.5 0.8 0.413 (SE =0.032) 42 4 0.095 0.334 (SE = 0.083) 21 TOTAL 116 1.2 1.14 (SE = 0.14) �Table 10. Number of kills found each winter field season through snow-tracking of lynx and percent composition of kills of the three primary prey species. Field Season 1999 1999-2000 2000-2001 2001-2002 2002-2003 2003-2004 2004-2005 n 9 83 89 54 65 37 78 Snowshoe Hare 55.56 67.47 67.42 90.74 90.77 67.57 83.33 Prey (%) Red Squirrel Cottontail 22.22 0 19.28 1.20 19.10 8.99 5.56 0 6.15 0 27.03 2.70 10.26 0 Other 22.22 12.05 4.49 3.70 3.08 2.70 6.41 Figure 1. Lynx are monitored throughout Colorado and by satellite throughout the western United States. The lynx core release area, where all lynx were released, is located in southwestern Colorado. A lynxestablished core use area has developed in the Taylor Park and Collegiate Peak area in central Colorado. 22 �: 6 : !1 • - • - • • - • l 1 l f , 1 12• • • • • i i f i l i , e EB e e 1• 1 1 l 1 l f : !4. •-•-• • 1 l 1 , !5•-•-•-•-• I - - - - - --- - --- -- - -- ---- - - - - -- ---- - -- - - - - - - - - - - l 12meters 15 meters Figure 2. Design of site-scale habitat plot sampling plot. Each of the 25 points are 3 meters apart (the first 6 points are labeled 1-6). The object that triggered a habitat plot (e.g., kill ) is the center point, depicted by the hollow circle. The actual pints encompass a 12 m x 12 m square but the understory and overstory data collected are influenced by vegetation occurring within a 15 m x 15 m square. TRAVEL n = 441 LONG BED n = 458 x Elevation= 3194 m SE = 9 m x Slope = 17° SE = 0.6° x Elevation= 3179 m SE = 9 m x Slope = 14° SE = 0 5° SSH KILL n = 238 DEN n = 37 x Elevation= 3145 m SE = 14 m x Slope = 16° SE = 0 8° x Elevation = 3354 SE= 31 m x Slope= 30° SE~. . . ' Figure 3. Frequency of aspect with mean vector and 95%confidence interval depicted as grey bars on graphs for 4 lynx use sites; dens, long beds, kills and travel as well as mean elevation and SE and mean slope and SE . 23 �80 70 60 'i:: 50 8 40 V p_. 30 20 10 0 ES SF AS \XII Total Cover Tree Species I ■ L ong Beds ■ Kills □ T ravel □ Den Sites I Figure 4. Mean percent overstory by tree species Engelmann spruce (ES), subalpine fir (SF), aspen (AS), willow (WI) and total cover for 4 different lynx use sites: long beds, kill sites, travel and den sites. Confidence intervals (95%) are depicted by error bars. 40 ~ - - - - - - - - - - - - - - ~ ~J; 01 ITT,CTl ~.rr1I 70 - r - - - - - - - - - - - - - - - - - - - , 60 Den Sites lO 40 ~o 20 10 Most Common Understory Species ES= Engelmann spruce SF = Subalpine fir 'w'.I = Willow AS = Aspen LO = Lodgepole pine C\YD = Coarse woody debri s Understory Species Figure 5. Mean percent understory by tree species Engelmann spruce (ES), subalpine fir (SF), coarse woody debris (CWD), aspen (AS), willow (WI), and total cover for 4 different lynx use sites: long beds, kill sites, travel, and den sites. 24 �Kills 800 60J (].) 40J ~ .su 20J (].) ~ SF ES SF ES AS AS Tree Species I■ o-6 0 6-12 12-18 018-24 0 > 241 Figure 6. Mean tree density by species Engelmann spruce (ES), subalpine fir (SF), and aspen (AS) and dbh size class for 4 different lynx use sites. 80 70 ...u 60 :,. 0 50 ~ 40 u C: ...uuu 30 p... 20 10 0 FS SF AS WI Total Cover: Tree Species I ■ Successful Chases i:r,!] Unsuccessful Chases I Figure 7. Mean percent overstory by tree species Engelmann spruce (ES), subalpine fir (SF), aspen (AS), willow (WI) and total cover for successful and unsuccessful snowshoe hare chases. Confidence intervals (95%) are depicted by error bars. 25 �50 45 H 40 (!) ~ 35 0 u.J-> 30 25 ~ (!) u 20 H (!) 15 p_, 10 5 0 ES SF C\X/D \XII AS Tree Species Tot:al Cover ■ 0 - 0. 5 m S rlol 0 - 0.5 m U ■ 0.5 - 1.0 m S rlol 0.5 - 1.0 m U □ 1.0 - 1.5 m S ~ 1.0 - 1.5 m U Figure 8. Mean percent understory by tree species Engelmann spruce (ES), subalpine fir (SF), apsen (AS), willow (WI), and total cover for 3 different understory height categories for successful and unsuccessful snowshoe hare chases. Confidence intervals (95%) are depicted by error bars. 1200 I 0-6 SC l!I 0-6 UC 1000 <l) 1 6-12 SC w Zlu 800 ri!I 6-12 UC <l) X 600 ,.,, ll 12-18 SC -....._ 1!112-18 UC <l) <l) w 400 118-24 SC E--< l!I 18-24 UC 200 I >24 SC ri!1 >24 UC 0 ES SF AS Tree Species Figure 9. Mean tree density by species Engelmann spruce (ES), subalpine fir (SF), and aspen (AS) and 5 dbh size classes for successful chases (SC) and unsuccessful chases (UC) of snowshoe hares. 26 �APPENDIX I PROGRAM NARRATIVE STUDY PLAN FOR MAMMALS RESEARCH FY 2005-06 – FY 2009-10 State of: Cost Center: Work Package: Task No.: Colorado 3430 0670 2 Federal Aid Project No.: : Division of Wildlife : Mammals Research : Lynx Conservation : Density, Demography, and Seasonal Movements of Snowshoe Hare in Colorado N/A : DENSITY, DEMOGRAPHY, AND SEASONAL MOVEMENTS OF SNOWSHOE HARES IN COLORADO Principal Investigators Jacob S. Ivan, Ph. D. Candidate, Colorado State University Tanya M. Shenk, Wildlife Researcher, Mammals Research, Colorado Division of Wildlife Cooperators Gary C. White, Professor, Fishery and Wildlife Biology, Colorado State University STUDY PLAN APPROVAL Prepared by: _____________________________ Date: __________________ Submitted by: ____________________________ Date: ___________________ Reviewed by: ____________________________ Date: ___________________ ____________________________ Date: ___________________ ____________________________ Date: ___________________ Reviewed by: ____________________________ Biometrician Date: ___________________ Approved by: ____________________________ Mammals Research Leader Date: ___________________ 27 �DENSITY, DEMOGRAPHY, AND SEASONAL MOVEMENTS OF SNOWSHOE HARES IN COLORADO NEED A program to reintroduce the threatened Canada lynx (Lynx canadensis) into Colorado was initiated in 1997. Since that time, 204 lynx have been released in the state, and an extensive effort to determine their movements, habitat use, reproductive success, and food habits has ensued (Shenk 2005). Analysis of scat collected from winter snow tracking indicates that snowshoe hares (Lepus americanus) comprise 65–90% of the winter diet of reintroduced lynx (T. Shenk, Colorado Division of Wildlife, unpublished data). Thus, as in the far north where the intimate relationship between lynx and snowshoe hares has captured the attention of ecologists for decades, it appears that the existence of lynx in Colorado and the success of the reintroduction effort may hinge on maintaining adequate and widespread populations of hares. Colorado represents the extreme southern range limit for both lynx and snowshoe hares (Hodges 2000). At this latitude, habitat for each species is less widespread and more fragmented compared to the continuous expanse of boreal forest at the heart of lynx and hare ranges. Neither exhibits dramatic cycles as occur farther north, and typical lynx (≤2−3 lynx/100km2; Aubry et al. 2000) and hare (≤1−2 hares/ha; Hodges 2000) densities in the southern part of their range correspond to cyclic lows form northern populations (2-30 lynx/100 km2, 1−16 hares/ha; Aubry et al. 2000, Hodges 2000, Hodges et al. 2001). Whereas extensive research on lynx-hare ecology has occurred in the boreal forests of Canada, literature regarding the ecology of these species in the southern portion of their range is relatively sparse. This scientific uncertainty is acknowledged in the “Canada Lynx Conservation Assessment and Strategy,” a formal agreement between federal agencies intended to provide a consistent approach to lynx conservation on public lands in the lower 48 states (Ruediger et al. 2000). In fact, one of the explicit guiding principles of this document is to “retain future options…until more conclusive information concerning lynx management is developed.” Thus, management recommendations in this agreement are decidedly conservative, especially with respect to timber management, and are applied broadly to cover all habitats thought to be of possible value to lynx and hare. This has caused controversy where recommendations conflict with competing resource management goals. Accurate identification and detailed description of lynx-hare habitat in the southern Rocky Mountains would permit more informed and refined management recommendations. A commonality throughout the snowshoe hare literature, regardless of geographic location, is that hares are associated with dense understory vegetation that provides both browse and protection from elements and predators (Wolfe et al. 1982, Litvaitis et al. 1985, Hodges 2000, Homyack et al. 2003, Miller 2005). In western mountains, this understory can be provided by relatively young conifer stands regenerating after stand-replacing fires or timber harvest (Sullivan and Sullivan 1988, Koehler 1990, Koehler 1990, Bull et al. 2005) as well as mature, uneven-aged stands (Beauvais 1997, Griffin 2004). Hares may also take advantage of seasonally abundant browse and cover provided by deciduous, open habitats (e.g., riparian willow [Salix spp.], aspen [Populus tremuloides]; Wolff 1980, Miller 2005). In drier portions of hare range, such as Colorado, regenerating stands can be relatively sparse, and hares may be more associated with mesic, late-seral forest and/or riparian areas than with young stands (Ruggiero et al. 2000). Numerous investigators have sought to determine the relative importance of these distinctly different habitat types with regards to snowshoe hare ecology. Most previous evaluations were based on hare density or abundance (Bull et al. 2005), indices to hare density and abundance (Wolfe et al. 1982, Koehler 1990, Beauvais 1997, Miller 2005), survival (Bull et al. 2005), and/or habitat use (Dolbeer and 28 �Clark 1975). Each of these approaches provides insight into hare ecology, but taken singly, none provide a complete picture and may even be misleading. For example, extensive use of a particular habitat type may not accurately reflect the fitness it imparts on individuals, and density can be high even in “sink” habitats (Van Horne 1983). A more informative approach would be to measure density, survival, and habitat use simultaneously in addition to recruitment and population growth rate through time. Griffin (2004) employed such an approach and found that summer hare densities were consistently highest in young, dense stands. However, he also noted that only dense mature stands held as many hares in winter as in summer. Furthermore hare survival seemed to be higher in dense mature stands, and only dense mature stands were predicted (by matrix projection) to impart a mean positive population growth rate on hares. Griffin’s (2004) study occurred in the relatively moist forests of Montana, which share many similarities but also many notable differences with Colorado forests including levels of fragmentation, species composition, elevation, and annual precipitation. Density estimation is a key component in assessing the value of a particular stand type and is the common currency by which hare populations are compared across time and space. However, it can be a difficult metric to estimate accurately. Density estimation based on capture-recapture methods is a welldeveloped field (Otis et al. 1978, White et al. 1982), but is often too costly and labor intensive to be implemented on scales necessary to effectively monitor density over a biologically meaningful area. Also, density can be difficult to assess from grid-trapping efforts because it is often unclear how much area was effectively sampled by the grid (Williams et al. 2002:314). Different approaches can produce density estimates that differ by an order of magnitude even when calculated from the same data (Zahratka 2004). Indices such as pellet plot counts and distance sampling of pellet groups can be used to estimate density, but each of these has limitations as well (Krebs et al. 1987, Eriksson 2006). Pellet plot counts are typically conducted by laying out numerous rectangular or circular plots along transect lines randomly placed within a study site. All pellets occurring within the plot are counted and removed on an annual basis. The mean number of pellets per plot is then inserted into a regression equation that gives an estimate of hare density (Krebs et al. 1987). Estimates from this technique correlate well with density estimates derived from simultaneous mark-recapture studies occurring in the same area (Krebs et al. 2001, Murray et al. 2002, Mills et al. 2005, Homyack et al. 2006). However, because fecal deposition rates can vary by season and diet, and because pellet decomposition rates can vary with altitude, climate, aspect, precipitation, and cover type, region-specific, stand-specific, and/or season-specific equations should be developed before this technique is employed for a given area and season (Krebs et al. 2001, Prugh and Krebs 2004, Murray et al. 2005). Density estimates vary with plot size and shape, requiring equations specific to these geometric considerations as well (McKelvey et al. 2002). Pellet counts tend to yield more precise and unbiased density estimates when plots are visited and cleared more than once per year (e.g., plots cleared in the fall and then counted in the spring to estimate winter density) because variability in deposition and decomposition rates is reduced (Homyack et al. 2006). However, this requires considerably more work and expense than an annual survey. Some studies have conducted pellet plot counts without first clearing plots (e.g., Bartmann and Byrne 2001). This saves time and money, but requires the ability to discern fresh (this year) pellets from old pellets, which can be difficult and is generally not a recommended approach (Prugh and Krebs 2004, Murray et al. 2005). Distance sampling is a well-developed method for estimating the density of objects in a given area (Buckland et al. 2001). In general, observers walk a pre-defined sampling transect and record each object of interest along with the perpendicular distance of that object from the transect line. This information is then used to develop a detection function which is in turn used to estimate density (Buckland et al. 2001). The method assumes all objects on the line are seen with certainty, objects are not double-counted, distance measures are accurate, and transect lines are located randomly within a study area (Buckland et al. 2001). Recently, distance sampling has been used to indirectly estimate hare density 29 �by first estimating the pellet group density of hares, then using fecal deposition and decomposition rates as a link back to hare density (Eriksson 2006). In general, distance sampling is more efficient than pellet plot counts as it does not require the tedious layout of hundreds of plots or counting individual pellets. This advantage is most recognizable in situations where pellet groups occur at low densities. Conversely, at extremely high densities, it may become difficult to distinguish pellet groups, and plots may be preferable (Marques et al. 2001). Regardless, distance sampling of pellet groups to estimate animal density also requires habitat and season specific decomposition and defecation rates, which can be difficult to obtain (Marques et al. 2001). For this project, I have chosen to provide land managers with information relating demographic rates, as well as density, to stand characteristics. Thus, I will use mark-recapture techniques as data from such an approach can provide information on both density and demography. I will address the “effective trapping area” issue using a new approach that augments mark-recapture data with telemetry locations of animals using the grid. The study outlined below is designed principally to evaluate the importance of young, regenerating lodgepole pine (Pinus contorta) and mature Engelmann spruce (Picea engelmannii)/ subalpine fir (Abies lasiocarpa) stands in Colorado by examining density and demography of snowshoe hares that reside in each (Figure 1). My hope is that information gathered from this research will be drawn upon as managers make routine decisions, leading to landscapes that include stands capable of supporting abundant populations of hares. I assume that if management agencies focus on providing habitat, hares will persist. Specifically, I will evaluate small and medium lodgepole pine stands and large spruce/fir stands where the classes “small”, “medium”, and “large” refer to the diameter at breast height (dbh) of overstory trees as defined in the United States Forest Service R2VEG Database (small = 2.54−12.69 cm dbh, medium = 12.70−22.85 cm, and large = 22.86−40.64 cm dbh; J. Varner, United States Forest Service, personal communication). To maximize comparability, I will choose lodgepole stands so that all are generating from harvest or all are regenerating following fire. I also intend to identify which of the numerous density-estimation procedures available perform accurately and consistently using an innovative, telemetry augmentation approach as a baseline. I will assess movement patterns and seasonal use of deciduous cover types such as riparian willow. Finally, I will further expound on the relationship between density, demography, and stand type by examining how snowshoe hare density and demographic rates vary with specific vegetation, physical, and landscape characteristics of a stand. Figure 1. Purported high quality snowshoe hare habitat in Colorado. From left to right: small lodgepole pine, medium lodgepole pine, and large Engelmann spruce/subalpine fir. 30 �OBJECTIVES 1) Compare telemetry-corrected estimates of density to those that would have been obtained from other commonly employed techniques used to convert population size estimated from a trapping grid to density (i.e., mean maximum distance moved, ½ mean maximum distance moved, ½ trap interval, nested grids, Program DENSITY). The purpose is to determine which common technique requiring less effort most consistently matches estimates from the intensive, telemetry-corrected approach. 2) Assess the relative value of the 3 stand types that purportedly provide high quality hare habitat by estimating and comparing survival, recruitment, finite population growth rate, and maximum (late summer) and minimum (late winter) snowshoe hare densities for each type. 3) Describe the timing, duration, and extent of broad-scale, seasonal movement patterns of snowshoe hares. 4) Relate specific vegetation, physical, and landscape characteristics of the 3 stand types to snowshoe hare density and demographics. APPROACH Hypotheses 1) In general, snowshoe hare density in Colorado will be relatively low (≤0.5 hares/ha) compared to densities reported in northern boreal forests, even immediately post-breeding when an influx of juveniles will bolster hare numbers. 2) Snowshoe hare density will be consistently highest in small lodgepole pine stands, followed by large spruce/fir and medium lodgepole pine, respectively. 3) Survival will generally be highest in mature (large) spruce/fir stands followed by small and medium lodgepole pine, respectively. 4) Finite population growth rate will be consistently at or above 1.0 in mature spruce/fir stands with survival contributing most significantly to the growth rate. Finite growth rates for the lodgepole pine stands will be more variable. 5) Snowshoe hares will significantly shift their home ranges to make use of abundant food and cover provided by riparian willow (and/or aspen) habitats in summer. 6) Snowshoe hare density, survival, and recruitment will be highly correlated with understory cover and stem density. Experimental Design/Procedures Variables.--The response variables of interest for this project include stand-specific snowshoe hare density (D), apparent survival (φ), recruitment (f), finite population growth rate (λ), and a metric of seasonal movement. Density is the number of hares per unit area and will be estimated using a variety of conventional techniques as well as a rigorous method that incorporates radio telemetry. The standspecific demographic parameters will be estimated primarily from capture-mark-recapture methods. As such, apparent survival is defined as the probability that a marked animal alive and in the population at time i survives and is in the population at time i + 1. Apparent survival encompasses losses due to both death and emigration. Recruitment is the number of new animals in the population at time i + 1 per animal in the population at time i. New recruits can arise from on-site reproduction as well as immigration. The finite population growth rate is the number of animals in a given age class at time i + 1 divided by the number present at time i. Shifts in home range will be assessed by comparing the seasonal proportion of telemetry locations in deciduous habitats using multi-response permutation procedures (MRPP; Zimmerman et al. 1985, White and Garrott 1990). Potential explanatory variables for snowshoe hare density, demographics, and movement include general species composition and structural stage of each stand in which response variables are measured. Additionally, stem density, horizontal cover, and canopy cover (to a lesser extent) are highly correlated 31 �with snowshoe hare abundance and habitat use (Wolfe et al. 1982, Litvaitis et al. 1985, Hodges 2000, Zahratka 2004, Miller 2005). Thus, I will further characterize vegetation in each stand by measuring stem density by size class (1-7 cm, 7.1-10 cm, and >10 cm), percent canopy cover, percent horizontal cover of understory and basal area. Basal area is an easily obtainable metric that may be correlated with the other variables and is recorded routinely during timber cruises, whereas the others are not. Thus, it might prove a useful link for biologists designing management strategies for snowshoe hare. Additionally, I will record physical covariates such as ambient temperature, precipitation, and snow depth at each stand during sampling periods as well as precipitation 1-3 years prior to sampling. Finally, I will calculate potentially important landscape metrics such as patch size and level of fragmentation. Location--.Identification of a suitable study area for this project and others that may follow is ongoing. The general study area must consist of an interspersion of young lodgepole pine and mature spruce/fir forest juxtaposed closely with open, seasonal habitats such as riparian willow. Within this general area, 3 sites will be selected such that 1) the 3 stand types of interest (small and medium lodgepole, large spruce/fir) occur within each site, 2) sites are close enough geographically to minimize differences due to climate, weather, and topography, but are far enough apart to be considered independent (e.g., 3 sites might occur in 3 different, but adjacent drainages), 3) each stand type within a site is adjacent to a riparian area, and 4) stand types of interest occur within 1 km of an access road (for logistical purposes). Such an arrangement often occurs in east-west drainages where spruce/fir grows on the north-facing slope, lodgepole pine covers the south-facing slope, and a riparian/willow area with road access separates the two (Figure 2). Additionally, sites must 1) include stands of suitable size and shape to admit a 16.5-ha trapping grid, 2) be consistent in their management history (i.e., all lodgepole pine stands in all sites must be either thinned or un-thinned, all regenerating after fire or all regenerating after harvest), and 3) be consistent in their intensity of use by lynx (core areas or not). I recently obtained the U.S. Forest Service R2VEG GIS database (newest, most detailed stand inventory information available statewide) and am currently working to objectively select a suite of potential study sites that satisfy the above-stated conditions. Depending on the number of potential sites within this suite, I will choose a small set of provisional study areas to ground-truth based on logistical considerations (e.g, housing, access). I will randomly select the final study sites from among those that appeared qualitatively suitable during ground-truthing. Prior to data collection I will more intensively sample the vegetation characteristics of the various stand types within the selected study sites to ensure that they represent intended conditions. Sampling.--All trapping and handling procedures will be approved by the Colorado State University Animal Care and Use Committee and filed with the Colorado Division of Wildlife. Snowshoe hares breed synchronously and generally exhibit 2 birth pulses in Colorado (although in some years, some individuals may have 3 litters), with the first pulse terminating approximately June 5−20 and the second approximately July 15–25 (Dolbeer 1972). To obtain a maximum density estimate, I will begin data collection on site 1 immediately following the second birth pulse in late July. Along with a crew of 5 technicians, I will deploy one 7 × 12 trapping grid (50-m spacing between traps; grid covers 16.5 ha) in each of the 3 stand types of interest following Griffin (2004) and Zahratka (2004). Grid locations and orientation will be chosen randomly within each stand subject to the logistical constraint that they must be within 1 km of a road. Traps will be deployed in all 3 stands in a single day. As traps are deployed, they will be locked open and “pre-baited” with apple slices and commercial rabbit chow. On days 2-4, the crew will continue pre-baiting, replacing apples and rabbit chow as necessary. The purpose of this extended pre-baiting is to maximize capture rates when trapping begins. This will minimize the number of trap-nights needed to capture the desired number of animals which in turn will minimize trappingrelated stress as well as the likelihood that American marten (Martes americana) will key into trap lines and prey on entrapped hares, as has occurred in previous studies (J. Zahratka, personal communication). During pilot work in winter 2005, I observed low but increasing capture rates (<0.20) during the first 3 32 �nights of trapping, with higher, more stable capture probabilities after 3 days (approximately 0.35–0.45). Thus 3 days of pre-baiting seems reasonable. Study Area Site 1 D Site 2 Site 3 _____A.._____ \ ( Summer Winter FY06-07 FY08-09 FY07-08 Figure 2. Experimental design for study of snowshoe hare density, demography, and movement. Within the study area, 3 sites, each consisting of 3 forest stand types (light to dark gray shades) and a riparian area (medium gray shade), will be sampled (dotted trapping grids) during late summer and late winter for 3 years. 33 �Traps will be set on the afternoon of the 4th day and checked early each morning and again in the evening on days 5–9. By checking traps in both morning and evening I prevent hares from being entrapped >13 hours, which should minimize capture stress. Based on Zahratka (2004) and personal experience, I anticipate capturing up to 10–15 individual hares per grid. A crew of 2 people will work together on each grid to check traps and process captures as quickly as possible. All captured hares will be coaxed out of the trap and into a dark handling bag by blowing quick shots of air on them from behind. Hares will remain in the handling bag, physically restrained with their eyes covered, for the entire handling process. Each individual will be aged, sexed, marked with a passive integrated transponder (PIT) tag and temporary ear mark (to track PIT tag retention), then released. Aging will consist of assigning each individual as either juvenile (<1 year old, <1000 g) or adult (≥1 year old, ≥1000 g) based on weight. This criterion is accurate through the end of September at which point juveniles are difficult to distinguish from adults (K. Hodges, University of British Columbia; P. Griffin, University of Montana, personal communication). After the first day of trapping, all captured hares will be scanned for a PIT tag prior to any handling and those already marked will be recorded and immediately released. Traps and bait will be completely removed from the grid on day 10. In addition to PIT tags and ear marks, I will radio collar up to 10 hares captured on each grid with a 28-g mortality-sensing transmitter (BioTrack, LTD) to facilitate unbiased density estimation as well as assessment of seasonal movements. I expect heterogeneity in snowshoe hare movements and use of the grid area, with potential bias surfacing due to location at which a hare is captured (e.g., hares captured on the edge of a grid may use the grid area differently than those captured at the center), and differential behavioral responses to trapping (e.g., young individuals may have lower capture probabilities and thus may be more likely to be captured on later occasions). To guard against the first potential bias, I will randomly select a starting trap location each morning and run the grid systematically from that point. Thus, the first several hares encountered (and collared) will be as likely to be from the inner part of the grid as from the edge. To protect against the second potential source of bias, I will refrain from deploying the final 3 collars until days 4 and 5 of the trapping session. Immediately following the removal of traps, the field crew will begin work locating each radiocollared hare 1–2 times per day for 10 days. Locations will be obtained by “homing” on a signal (Samuel and Fuller 1996, Griffin 2004) taking care not to push hares while approaching them. Technicians will record their location with hand-held GPS units (Garmin model 12XL) as soon as a visual of the collared hare is obtained or if the signal can be picked up by the receiver without an antenna. Using the same make and model collars, Griffin (2004) found that hares are usually within ~15m when the signal came be received without an antenna (Griffin 2004). I will test this assumption with my telemetry equipment over a variety of transmitter locations and orientations. Because hares are largely nocturnal (Keith 1964, Mech et al. 1966, Foresman and Pearson 1999), an effort will be made to conduct telemetry work at various times of the night (safety and logistics permitting) and day to gather a representative sample of locations for each hare. The crew will gather telemetry locations for radio-collared hares on site 1 for 8−9 days. Then the 10−day trapping procedure and 8 to 9−day telemetry work will be repeated on the 3 grids comprising site 2 (Figure 3). The cycle will be repeated once more for grids on site 3 (Figure 3). The entire process will be repeated during the following winter when densities should be at a minimum. In summary, for any given 9-week sampling period, I will collect data from 9 total grids, 1 grid in each of 3 habitat types (stand types) across 3 sites. Sampling will occur during 2 such 9-week periods each year − once in late summer and once in late winter – and will continue for 3 years (Figure 2). During the interim between intensive trapping and telemetry work, a single technician and myself will attempt to gather 1–2 telemetry locations/hare/month in order to keep closer tabs on these individuals, 34 �determine more precisely when mortality occurs, and retrieve collars from dead hares. Telemetry work will also occur during “pre-baiting” days to determine which hares are still alive and immediately available to be sampled by the grid during the ensuing trapping period. Jul Aug Sep 1 Jan 1 2 3 4 2 3 4 5 6 7 8 5 6 7 8 9 10 11 9 10 11 12 13 14 15 12 13 14 15 16 17 18 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 1 2 3 4 6 7 8 9 10 13 14 15 16 19 20 21 22 23 24 25 26 27 28 29 30 31 1 5 2 3 4 5 6 7 8 11 12 9 10 11 12 13 14 15 17 18 19 16 17 18 19 20 21 22 Feb 20 21 22 23 24 25 26 23 24 25 26 27 28 1 27 28 29 30 31 1 2 Mar 2 3 4 5 6 7 8 3 4 5 6 7 8 9 9 10 11 12 13 14 15 10 11 12 13 14 15 16 16 17 18 19 20 21 22 17 18 19 20 21 22 23 23 24 25 26 27 28 29 24 25 26 27 28 29 30 30 31 Figure 3. Approximate annual data collection schedule for trapping (�) and telemetry (�). Dates and weeks will change depending on calendar year and pay schedule. During telemetry work, the 6-person crew will be divided into 2 teams, only one of which will be working at any given time. Monthly locations on radio-collared hares will also be collected in the interim between the intensive sampling periods indicated here. Vegetation sampling at each stand will follow protocols established through previous snowshoe hare and lynx work in Colorado (Zahratka 2004, T. Shenk, Colorado Division of Wildlife, personal communication). Specifically, on each of the 9 live-trapping grids, I will lay out 5 × 5 grids (3-m spacing) of vegetation sampling points centered on 15 of the 84 trap locations (Figure 4). At each of the 25 vegetation sampling points, I will record: 1) distance to the nearest woody stem 1.0−7.0 cm, 7.1−10.0 cm, and >10.0 cm in diameter at heights of 0.1 m and 1.0 m above the ground (to capture both summer [0.1 m] and winter [1.0 m] stem density; Barbour et al. 1999), 2) horizontal cover in 0.5-m increments above the ground up to 2 m (Nudds 1977), and 3) canopy cover [present or absent] using a densitometer. Additionally, at the center of all 15 vegetation sampling grid points (i.e., at the trap location), I will measure basal area using an angle gauge. These measurements will be gathered once at the start of the project, unless conditions change due to disturbance such as fire. Temperature will be monitored hourly at each grid during the 6-week intensive sampling periods using data loggers. During winter sampling periods, snow depth measurements will be recorded daily at the same 15 trap locations used to quantify the vegetative attributes of that stand. 35 �0 0 0 0 0 0 0 0 • 0 • 0 • 0 0 0 0 0 0 0 0 0 • 0 • 0 • 0 0 0 0 0 0 0 l som □ □ □ □ □ □ □ □ □ □ □ □ ■ □ □ □ □ □ □ □ □ □ □ □ □ l 3m 0 • 0 • 0 0 0 0 0 0 0 0 0 • 0 • 0 • 0 0 0 0 0 0 0 0 0 • 0 • 0 • 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Figure 4. 15 trap locations (•) on 7 × 12 trapping grid where vegetation will be sampled by measuring stem density horizontal cover, and canopy cover at the 25 points on each 5 × 5 subgrid (inset). In addition, basal area will be measured at the trap location (�) on which each of the 15 subgrids are centered. Data Analysis Density.--I will assume that hare populations are demographically and geographically closed during the short 5-day mark-recapture sampling periods. To obtain a density estimate for each grid, I will use the Huggins closed capture model (Huggins 1989, 1991) in Program MARK (White and Burnham 1999) with some modifications. The basic Huggins estimator (no individual covariates) is based on the fact that if pj is the probability that a hare in the population will be captured (and marked) for the first time on trapping occasion j, then p * = 1 − (1 − p1 )...(1 − p5 ) is the probability that an individual is captured at least once during a 5-day trapping period (i.e., j = 1,…,5). Accordingly, the basic Huggins estimator for population size, N̂ , is Nˆ = M t +1 / p* where M t +1 is the total number of hares captured. The estimator can be re-written to allow each of the M t +1 individuals captured to have their own p*. In that case, Nˆ = M t +1 ∑1 / p . Presumably hares that reside near the edge of a grid encounter fewer traps and * i i =1 are less likely to be captured than hares residing near the center of a grid. To account for this, I will take advantage of the Huggins model with individual covariates to model p* by using the logit link function of program MARK to model pi* as a function of di, where di is distance from the edge of the grid for hare i based on mean capture coordinates. A naïve density estimate for each grid would then be Dˆ = Nˆ / A where A is the area of the grid. However, this gives full credit to all hares, even those whose home range only partially overlaps the grid, which results in a density estimate that is biased high. To correct for this bias, I will determine the proportion, ( ~ pk ), of telemetry locations for each of the k = 1,…,10 radiocollared hares that fall within the “naïve grid area.” By incorporating data from multiple grids, a logistic regression model will be developed to estimate p% i for all M t +1 animals captured on a grid based on 36 �distance from the edge of the grid for hare i (di). Replacing the numerator (i.e., 1) in the Huggins ⎛ M t +1 ⎞ ~ p / p ⎟ A. ⎜∑ ⎟ estimator with ( p% i ), gives a density estimate, Dˆ = ⎜ ⎝ i =1 i * i ⎠ The above-stated approach assumes that radio-collared hares neither gravitate toward nor avoid the former grid area after the 5 days of trapping, 10–20 locations per hare is enough to provide a reasonable representation of the proportion of time they spend on the grid, and their use of the grid area is representative of other hares that were captured but not collared (i.e., that the logistic regression model of p% i is a useful model). I contend that this type of estimate from grid-based trapping can be construed as a relatively unbiased estimate of density. Using these point estimates and their associated confidence intervals, I will compare hare density among seasons, years, and stand types. I will also compare these “true” density estimates to those that would have been obtained using other available methods such as ½ mean maximum distance moved (Wilson and Anderson 1985, Williams et al. 2002:314-315), full mean maximum distance moved (Parmenter et al. 2003), ½ trap interval (Parmenter et al. 2003), “nested grids” (White et al. 1982:120-131), and Program DENSITY (Efford et al. 2004). I Demography.--I will analyze mark-recapture data using Pradel temporal symmetry models (Pradel 1996, Nichols and Hines 2002) in a robust design framework (Williams et al. 2002:523-554), which will be available in Program MARK by summer 2006. Thus, I will treat summer and winter sampling occasions as primary periods, and the 5-day trapping sessions within each as secondary periods. The Pradel temporal symmetry models employ both forward and reverse-time evaluation of capture histories to provide estimates of apparent survival ( φ̂ ) and seniority ( γ̂ ). Apparent survival, φi, is the probability that a marked animal alive and in the population at time i survives and is in the population at time i + 1. The seniority parameter, γi , is the reverse-time analogue of survival. Reading backward through a capture history, it is the probability that a marked animal alive and in the population at time i was alive and in the sampled population at time i − 1. If N is the number of animals present in the population, N i φi ≈ N i +1γ i +1 and N i +1 / N i = φi / γ i +1 = λ i . Also, if fi is recruitment rate, or the number of recruits at time i + 1 per animal present at time i, then N i +1 = N i φi + N i f i . Rearranging and substituting into the previous equation gives f i = φi (1/ γ i − 1) . Thus, using Pradel models, one can estimate recruitment and finite population growth rate in addition to survival (Pradel 1996, Nichols and Hines 2002). I will use Akaike’s Information Criterion corrected for small sample size (AICc; Burnham and Anderson 1998) to determine whether models with time-dependent parameters or constant parameters are best supported by the data. I will derive estimates of the above-mentioned parameters from the best model or from model averaging. I anticipate pooling capture data across sites to obtain φˆ i , λˆ i , and fˆi for each stand type for each interval between primary sampling periods (5 estimates of each). I also anticipate simply estimating these parameters for “generic hares”, treating both juveniles and adults as a single group or age class. Given that juveniles are morphometrically indistinguishable from adults by their first fall of life (K. Hodges, University of British Columbia; P. Griffin, University of Montana, personal communication), adult and juvenile survival rates are similar (Griffin 2004), and there is little evidence for age-specific differences in pregnancy rates or litter size (Dolbeer 1972), this approach seems justified. However, if I happen to capture sufficient numbers of juveniles and adults, the design I have laid out here allows for treating the age classes separately. This, in turn, may permit me to decompose the contribution that fi makes to λi into the portion of that contribution due to on-site reproduction and that due to immigration (Nichols et al. 2000). Similarly, it may also be possible using my telemetry data to decompose apparent survival, φi , into emigration and mortality. Such fortuitous situations would facilitate the identification of source and sink habitats if they exist. 37 �Seasonal Movements.--I will assess whether snowshoe hares seasonally shift their home ranges using the multi-response permutation procedure (MRPP; Zimmerman et al. 1985, White and Garrott 1990:134-135). Under this approach, telemetry locations are grouped by season (summer and winter), and an MRPP statistic is calculated as the weighted average of the distance between all possible pairs of locations within groups compared to the average distance between all possible pairs ignoring groups. The null hypothesis is that the distribution of locations is the same for both groups (seasons). Sufficiently small values of the test statistic suggest that within group distances are smaller than distances measured ignoring groups, which is evidence against the null in favor of a group (seasonal) effect. P-values are obtained by calculating the percentile of the observed test statistic relative to all possible test statistics that could be computed by re-arranging the data into all possible groups of 2. The MRPP procedure is sensitive and can detect even small changes in use of an area (White and Garrott 1990:136). I propose a priori that changes in proportional use of deciduous habitats <0.10 in magnitude are unlikely to be biologically significant. Vegetation.--I will calculate mean stem density, horizontal cover, canopy cover, and basal area for each season−stand type as well as temperature, precipitation, snow depth information, and landscape metrics. These will be entered into the MARK design matrix as covariates to population size (~density) and survival in a random effects analysis. As such, I will be able to quantify the amount of variation in population size or survival that is due to differences in vegetation, landscape, or weather relative to the amount left to other causes. Sample size.--I conducted power analyses to determine the probability of discerning meaningful differences in density and survival for hares occupying different stand types. For density, I postulated that foraging lynx likely do not discriminate among stands that differ by only a few hares. However, it seems probable that if hare density in one stand is twice that of another, a lynx would choose the former given the opportunity. Thus, I conducted power calculations to determine the probability of distinguishing differences in densities between 2 stand types in which one had twice the density of hares as the second. Specifically, using the Huggins closed capture model (Huggins 1989, Huggins 1991) in Program MARK, I specified the number of hares (N) present in each of 2 groups (i.e., 2 stand types), allowed capture (p) and recapture (c) probabilities to vary with time but constrained them to be equal and the same for each group, then simulated this scenario 1000 times for a range of realistic capture probabilities. For each simulation I calculated a 95% confidence interval for the mean difference in N̂ between the 2 groups and determined the proportion of all simulations in which this confidence interval did not include zero. This proportion is the power, or probability of discerning a difference between the 2 groups when one actually exists. I compared 2-fold differences in density at the low (5 vs. 10 hares/grid) and high (15 vs. 30 hares/grid) end of the range of hare numbers and I expect to observe (Zahratka 2004). I also simulated the power to detect differences between 17 and 39 hares/grid, corresponding to recently published cut-points for low and high hare densities in the context of lynx conservation (Mills et al. 2005). Given capture/recapture probabilities I observed during winter 2005 (approximately 0.35–0.45), I expect to have reasonable power to detect 2-fold differences in density even if I encounter relatively few hares per grid (Figure 5). 38 �% Non-overlapping 95% CIs Density Power Analysis 100 90 80 70 60 50 40 30 20 10 0 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.50 0.55 0.60 0.45 capture/recapture probability ---+--- N=5 vs. N=10 N=15 vs. N=30 ___,,_ ---I N=17 vs. N=39 Figure 5. Power for distinguishing differences in snowshoe hare density between 2 habitat types when a difference actually exists. Gray area indicates the capture probability realized by the 3rd day of trapping during a pilot study in winter 2005. N indicates number of hares per grid, a range of roughly 0.1 (N = 5) to 0.7 hares/ha (N = 39). I conducted power analyses for survival in a similar manner using the Huggins estimator (Huggins 1989, Huggins 1991) in a robust design framework (Williams et al. 2002:524-556). For this analysis, I specified 3 primary periods (e.g., 3 years) with 5 secondary occasions for each. I established either 30 or 45 hares in each of 2 groups (i.e., pooled an expected 10-15 hares/grid across the 3 grids in a given habitat type), specified a different survival rate for each, and allowed p and c to vary with time but constrained them to be equal and the same for each group as before. I then specified a general model that assumed survival rates varied among groups and a second, reduced model that assumed survival rates were the same for each group. After 1000 simulations under a given scenario of hare numbers, capture probabilities, and survival rates, I conducted a likelihood ratio test between each pair of general and reduced models. As before, I used the proportion of significant tests as an estimate of power to detect differences in survival. I compared survival rates of 0.4 vs. 0.5, 0.3 vs. 0.5, and 0.2 vs. 0.5. These rates span the range of annual hare survival rates reported in the literature (Dolbeer 1972, Dolbeer and Clark 1975, Griffin 2004). Also, because each comparison is anchored at 0.5, these calculations provide a conservative estimate of power due to the nature of binomial probabilities. That is, I would be more likely to distinguish the difference between 0.1 and 0.2 than between 0.4 and 0.5 even though the difference in both cases is 0.1 because the sampling variance of the estimate for the same sample size is maximal at 0.5 and declines to 0 for survival rates of 0 or 1. Results indicate that I have ≥80% chance of discerning real differences in survival of ≥0.3 (Figure 6), but only 40-65% chance (depending on number of hares captured) of detecting a difference of 0.2, and very little chance of detecting differences smaller than 0.2. However, I plan to combine my telemetry data with my trapping data in the MARK Robust design model using separate groups for each data type. This should enhance my precision and power, thus making the prospect of detecting differences as small as 0.2 a possibility. 39 �Survival Power Analysis (N = 45) 100 Capture/Recapture Probability I-+- 0.2 vs. 0.5 --- 0.3 vs. 0.5 -------- _._..,,.- 0.60 0.10 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0 V / 0.55 10 0 ✓ 0.50 - 20 / - 0.45 ------------ / 30 ---- / 0.40 / 40 / / 0.35 20 L---"" 10 .---------- 50 / - 0.30 / - _,....... 70 60 / 0.25 V /' / / 90 80 0.20 70 60 50 40 30 ---- ____.,- 80 0.15 100 90 % Significant LR Tests % Significant LR Tests Survival Power Analysis (N = 30) Capture/Recapture Probability I-+- 0.2 vs. 0.5 --- 0.3 vs. 0.5 0.4 vs. 0.5 0.4 vs. 0.5 Figure 6. Power, or probability of distinguishing differences in snowshoe hare survival between 2 habitat types when differences actually exist. N = 30 (left) and N = 45 (right) correspond to reasonable estimates of the number of hares I expect to capture in each habitat type. Gray area indicates the capture probability realized by the 3rd day of trapping during a pilot study in winter 2005. To complete a power analysis for λ̂ requires running simulations of Pradel models in a robust design framework. This capability is not yet available in Program MARK, so such an analysis has not been completed. Sampling 15 vegetation plots per trapping grid provided reasonably precise characterizations of similar stands in similar locations during a previous study (Zahratka 2004). I trust this level of sampling will be adequate for the present study as well. If not, more plots can be established at a later date given that vegetative characteristics are unlikely to change appreciably over a few years. Project Schedule I will begin the first 9-week data collection period in mid July 2006. The first winter sampling period will begin in February 2007. Intensive sampling will occur across a total of 3 summer and 3 winter periods, with monthly telemetry work interspersed between the main sampling periods. All fieldwork will terminate with the winter 2009 sampling period. Analysis, write-up, and submission to journal outlets will occur during summer and Fall 2009. I plan to graduate during spring semester 2010. Personnel Jacob S. Ivan, Ph. D. student, Colorado State University will be the primary investigator responsible for the design and conduct of the study. Tanya M. Shenk, Mammals Research, Colorado Division of Wildlife, and Gary C. White, Professor, Colorado State University will serve as primary advisors. Also, as most lynx/hare habitat occurs on United States Forest Service (USFS) land, this project will require cooperation and coordination with USFS biologists and district rangers for permission and possibly logistical support (housing, campsites, trucks). As presented here, this project will require an estimated minimum of 3,600 person-hours/year (5 technicians, 720 hours) in technician labor to complete the intensive 9-week sampling periods as well as 360 person-hours/year of technician labor to run the monthly telemetry operation. Thus, completion of the 3-year project will require an estimated minimum of 11,880 person-hours in addition to time spent by the primary investigator, advisors, and cooperators. 40 �Estimated Annual Cost FY06-07 FY07-08 FY08-09 TFTE (5 techs, 360, $11.13/hr, 11.16% overhead)* $ 22,270 $ 22,830 $ 23,410 TFTE (1 tech, 360 hours, $11.13/hr, 11.16% overhead)** $ $ 4,565 $ 4,679 Personnel 4,454 Operating PURCHSERV (Ph.D. Stipend, tuition, minimal supplies)*** $ 27,500 $ 27,500 $ 27,500 SUPPLIES (bait, snowmobile repairs, handling supplies, etc.) $ $ 4,000 $ 4,000 EQUIPMENT (radio collars) $ 11,500 $ 11,500 $ 11,500 INSTTRAV $ 1,500 $ 1,500 $ 1,500 VEHICLE LEASE/MILEAGE (3 vehicles, 5 months/year)** $ 5,328 $ 5,328 $ 5,328 4,000 Travel TOTAL COST $76,552 $77,223 $77,917 TOTAL COST TO SSH BUDGET $43,724 $44,395 $45,089 *Assumes 2.5% cost-of-living wage increase/year **Telemetry work during interim between sampling periods ***Will be charged to budget centers other than lynx/snowshoe hare EXPECTED RESULTS/BENEFITS 1) Seasonal density estimates and associated variability will help establish where Colorado lies on the continuum of hare densities reported in the literature. Whether densities are relatively high or low, stable or highly variable, or drastically different or roughly equal among cover types could influence future land management decisions as well as decisions regarding the lynx reintroduction process. 2) Combined with Zahratka (2004) and future research, density estimates from this project may elucidate the degree to which hare populations fluctuate or cycle in Colorado, a phenomenon of interest to wildlife ecologists and managers. 3) Comparison of “known” densities to those obtained from other commonly used methods will inform future research and monitoring programs which techniques are likely to produce results that are most consistently in agreement with the intensively derived estimates reported from this project. This knowledge will also enhance interpretation of previously reported hare densities in Colorado and elsewhere. 4) Assessment of density, demographic parameters, and their variability among habitat types will help identify which type(s) consistently support(s) high hare numbers and productivity. The current, conservative approach to lynx/hare conservation is to treat all potential habitat as equally and highly valuable, although this has not been substantiated scientifically, especially in Colorado. This project should determine if the current approach is justified or if there is a disparity in the value of different habitat types relative to lynx-hare conservation. If the latter is true, those charged with managing forests may be allowed more flexibility to accommodate competing resource uses while maintaining lynx/hare habitat. 41 �5) Assessment of density and demographic parameters should help identify the general time period over which succession carries young, regenerating lodgepole pine stands into and then out of service as snowshoe hare habitat. It is apparent that stands in fresh clear cuts and mature lodgepole stands do not provide quality hare habitat (Zahratka 2004). The value of small and medium lodgepole stands to hares has not been quantified in Colorado and is of interest to resource managers. 6) Knowledge regarding the presence or absence of large-scale seasonal movements, and the extent to which this occurs will inform managers about the value of peripheral vegetation (other than conifer forest, such as riparian willow or aspen), will identify when and for how long peripheral vegetation is likely to be used, and will potentially identify other snowshoe hare management issues that have not received prior consideration. 7) A description and comparison of vegetation and landscape characteristics among the 3 stand types and their relationship to snowshoe hare demography and movement patterns should further aid land managers in creating and maintaining lynx/hare habitat. RELATED FEDERAL PROJECTS Given that the majority of lynx/hare habitat occurs on United States Forest Service land, this project will require cooperation with local ranger districts, regional biologists, and researchers within that agency. As soon as I have completed provisional study site selection, I will contact the appropriate collaborators to obtain permission, appropriate permits, etc. LITERATURE CITED AUBRY, K. B., G. M. KOEHLER, AND J. R. SQUIRES. 2000. Ecology of Canada lynx in southern boreal forests. Pages 373-396 in Ruggiero, L. F., K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation of lynx in the United States. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Collins, Colorado, USA. BARBOUR, M. G., J. H. BURK, W. E. PITTS, F. S. GILLIAM, AND M. W. SCHWARTZ. 1999. Terrestrial plant ecology, Addison Wesley Longman, Inc., Menlo Park, California, USA. BARTMANN, R. M. AND G. BYRNE. 2001. Analysis and critique of the 1998 snowshoe hare pellet survey. Colorado Division of Wildlife 20. BEAUVAIS, G. P. 1997. Mammals in fragmented forests in the Rocky Mountains: community structure, habitat selection, and individual fitness. Dissertation, University of Wyoming, Laramie, Wyoming, USA. BUCKLAND, S. T., D. R. ANDERSON, K. P. BURNHAM, J. L. LAAKE, D. L. BORCHERS, AND L. THOMAS. 2001. Introduction to distance sampling: estimating abundance of biological populations, Oxford University Press, Inc., New York, New York, USA. BULL, E. L., T. W. HEATER, A. A. CLARK, J. F. SHEPHERD, AND A. K. BLUMTON. 2005. Influence of precommercial thinning on snowshoe hares. U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station, General Technical Report PNW-RP-562. DOLBEER, R. A. AND W. R. CLARK. 1975. Population ecology of snowshoe hares in the central Rocky Mountains. Journal of Wildlife Management 39:535-549. DOLBEER, R. A. 1972. Population dynamics of snowshoe hare in Colorado. Dissertation, Colorado State University, Fort Collins, Colorado, USA. EFFORD, M. G., D. K. DAWSON, AND C. S. ROBBINS. 2004. DENSITY: software for analysing capturerecapture data from passive detector arrays. Animal Biodiversity and Conservation 27:1-12. ERIKSSON, H. M. 2006. Snowshoe hare densities in post-fire vegetation. Thesis, University of Alaska Fairbanks, Fairbanks, Alaska, USA. FORESMAN, K. R. AND D. E. PEARSON. 1999. Activity patterns of American martens, Martes americana, snowshoe hares, Lepus americanus, and red squirrels, Tamiasciurus hudsonicus, in westcentral 42 �Montana. The Canadian Field-Naturalist 113:386-389. GRIFFIN, P. C. 2004. Landscape ecology of snowshoe hares in Montana. Dissertation, University of Montana, Missoula, Montana, USA. HODGES, K. E. 2000. Ecology of snowshoe hares in southern boreal and montane forests. Pages 163-206 in Ruggiero, L. F., K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation of lynx in the United States. University Press of Colorado, Boulder, Colorado, USA. HODGES, K. E., C. J. KREBS, D. S. HIK, C. I. STEFAN, E. A. GILLIS, AND C. E. DOYLE. 2001. Snowshoe hare demography. Pages 141-178 in Krebs, C. J., S. Boutin, and R. Boonstra, editors. Ecosystem dynamics of the boreal forest: the Kluane project. Oxford University Press, New York, New York, USA. HOMYACK, J. A., D. J. HARRISON, AND W. B. KROHN. 2003. Effects of precommercial thinning on select wildlife species in northern Maine, with special emphasis on snowshoe hare. Maine Cooperative Fish and Wildlife Research Unit, Orono, Maine, USA. HOMYACK, J. A., D. J. HARRISON, J. A. LITVAITIS, AND W. B. KROHN. 2006. Quantifying densities of snowshoe hares in Maine using pellet plots. Wildlife Society Bulletin 34:74-80. HUGGINS, R. M. 1989. On the statistical analysis of capture-recapture experiments. Biometrika 76:133140. HUGGINS, R. M. 1991. Some practical aspects of a conditional likelihood approach to capture experiments. Biometrics 47:725-732. KEITH, L. B. 1964. Daily activity pattern of snowshoe hares. Journal of Mammalogy 45:626-627. KOEHLER, G. M. 1990. Snowshoe hare, Lepus americanus, use of forest successional stages and population changes during 1985-1989 in North-central Washington. Canadian Field-Naturalist 105:291-293. KOEHLER, G. M. 1990. Population and habitat characteristics of lynx and snowshoe hares in north central Washington. Canadian Journal of Zoology 68:845-851. KREBS, C. J., B. SCOTT GILBERT, S. BOUTIN, AND E. AL. R. BOONSTRA. 1987. Estimation of snowshoe hare population density from turd transects. Canadian Journal of Zoology 65:565-567. KREBS, C. J., R. BOONSTRA, V. NAMS, M. O'DONOGHUE, K. E. HODGES, AND S. BOUTIN. 2001. Estimating snowshoe hare population density from pellet plots: a further evaluation. Canadian Journal of Zoology 79:1-4. LITVAITIS, J. A., J. A. SHERBURNE, AND J. A. BISSONETTE. 1985. Influence of understory characteristics on snowshoe hare habitat use and density. Journal of Wildlife Management 49:866-873. MARQUES, F. F. C., S. T. BUCKLAND, D. GOFFIN, C. E. DIXON, D. L. BORCHERS, B. A. MAYLE, AND A. J. PEACE. 2001. Estimating deer abundance from line transect surveys of dung: sika deer in southern Scotland. Journal of Applied Ecology 38:349-363. MCKELVEY, K. S., G. W. MCDANIEL, L. S. MILLS, AND P. C. GRIFFIN. 2002. Effects of plot size and shape on pellet density estimates for snowshoe hares. Wildlife Society Bulletin 30:751-755. MECH, L. D., K. L. HEEZEN, AND D. B. SINIFF. 1966. Onset and cessation of activity in cottontail rabbit and snowshoe hares in relation to sunset and sunrise. Animal Behaviour 14:410-413. MILLER, M. A. 2005. Snowshoe hare habitat relationships in northwest Colorado. Thesis, Colorado State University, Fort Collins, Colorado, USA. MILLS, L. S., P. C. GRIFFIN, K. E. HODGES, K. MCKELVEY, L. RUGGIERO, AND T. ULIZIO. 2005. Pellet count indices compared to mark-recapture estimates for evaluating snowshoe hare density. Journal of Wildlife Management 69:1053-1062. MURRAY, D., E. ELLSWORTH, AND A. ZACK. 2005. Assessment of potential bias with snowshoe hare fecal pellet-plot counts. Journal of Wildlife Management 69:385-395. MURRAY, D. L., J. D. ROTH, E. ELLSWORTH, A. J. WIRSING, AND T. D. STEURY. 2002. Estimating lowdensity snowshoe hare populations using fecal pellet counts. Canadian Journal of Zoology 80:771-781. NICHOLS, J. D. AND J. E. HINES. 2002. Approaches for the direct estimation of lambda, and demographic 43 �contributions to lambda, using capture-recapture data. Journal of Applied Statistics 29:539-568. NICHOLS, J. D., J. E. HINES, J. D. LEBRETON, AND R. PRADEL. 2000. Estimation of contributions to population growth: a reverse-time capture-recapture approach. Ecology 81:3362-3376. NUDDS, T. D. 1977. Quantifying the vegetative structure of wildlife cover. Wildlife Society Bulletin 5:113-117. OTIS, D. L., K. P. BURNHAM, G. C. WHITE, AND D. R. ANDERSON. 1978. Statisical inference from capture data on closed animal populations. Wildlife Monographs 62: PARMENTER, R. R., T. L. YATES, D. R. ANDERSON, K. P. BURNHAM, J. L. DUNNUM, A. B. FRANKLIN, M. T. FRIGGENS, B. C. LUBOW, M. MILLER, G. S. OLSON, C. A. PARMENTER, J. POLLARD, E. REXSTAD, T. SHENK, T. R. STANLEY, AND G. C. WHITE. 2003. Small-mammal density estimation: a field comparison of grid-based vs. web-based density estimators. Ecological Monographs 73:126. PRADEL, R. 1996. Utilization of capture-mark-recapture for the study of recruitment and population growth rate. Biometrics 52:703-709. PRUGH, L. R. AND C. J. KREBS. 2004. Snowshoe hare pellet-decay rates and aging in different habitats. Wildlife Society Bulletin 32:386-393. RUEDIGER, B., J. CLAAR, S. GNIADEK, B. HOLT, LEWIS LYLE, S. MIGHTON, B. NANEY, G. PATTON , T. RINALDI, J. TRICK, A. VENDEHEY, F. WAHL, N. WARREN, D. WENGER, AND A. WILLIAMSON. 2000. Canada lynx conservation assessment and strategy. U.S. Department of Agriculture, Forest Service, U.S. Department of Interior, Fish and Wildlife Service, Bureau of Land Management, National Park Service R1-00-53 U.S. Department of Agriculture, Forest Service, U.S. Department of Interior, Fish and Wildlife Service, Bureau of Land Management, National Park Service, Missoula, Montana, USA. RUGGIERO, L. F., K. B. AUBRY, S. W. BUSKIRK, G. M. KOEHLER, C. J. KREBS, K. S. MCKELVEY, AND J. R. SQUIRES. 2000. The scientific basis for lynx conservation: qualified insights. Pages 443-454 in Ruggiero, L. F., K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation of lynx in the United States. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Collins, Colorado, USA. SAMUEL, M. D. AND M. R. FULLER. 1996. Wildlife radiotelemetry. Pages 370-418 in Bookhout, T. A., editors. Research and Management Techniques for Wildlife and Habitats. Allen Press, Inc., Lawrence, Kansas, USA. SHENK, T. M. 2005. General locations of lynx (Lynx canadensis) reintroduced to southwestern Colorado from February 4, 1999 through February 1, 2005. Colorado Division of Wildlife Colorado Division of Wildlife, Fort Collins, Colorado, USA. SULLIVAN, T. P. AND D. S. SULLIVAN. 1988. Influence of stand thinning on snowshoe hare population dynamics and feeding damage in lodgepole pine forest. Journal of Applied Ecology 25:791-805. VAN HORNE, B. 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47:893-901. WHITE, G. C., D. R. ANDERSON, K. P. BURNHAM, AND D. L. OTIS. 1982. Capture-recapture and removal methods for sampling closed populations. Los Alamos National Laboratory, Los Alamos, New Mexico, USA. WHITE, G. C. AND R. A. GARROTT. 1990. Analysis of wildlife radio-tracking data, Academic Press, San Diego, California, USA. WILLIAMS, B. K., J. D. NICHOLS, AND M. J. CONROY. 2002. Analysis and management of animal populations, Academic Press, San Diego, California, USA. WILSON, K. R. AND D. R. ANDERSON. 1985. Evaluation of two density estimators of small mammal population size. Journal of Mammalogy 66:13-21. WOLFE, M. L., N. V. DEBYLE, C. S. WINCHELL, AND T. R. MCCABE. 1982. Snowshoe hare cover relationships in northern Utah. Journal of Wildlife Management 46:662-670. WOLFF, J. O. 1980. The role of habitat patchiness in the population dynamics of snowshoe hares. Ecological Monographs 50:111-130. 44 �ZAHRATKA, J. L. 2004. The population and habitat ecology of snowshoe hares (Lepus americanus) in the southern Rocky Mountains. Thesis, The University of Wyoming, Laramie, Wyoming, USA. ZIMMERMAN, G. M., H. GOETZ, AND P. W. JR. MIELKE. 1985. Use of an improved statistical method for group comparisons to study effects of prairie fire. Ecology 66:606-611. 45 � https://cpw.cvlcollections.org/files/original/939c9c7f86d7cd43cc99d606aa5c6865.pdf a6b5b0747f7dd2d2165a1723826bb852 PDF Text Text Colorado Division of Wildlife July 2006 - June 2007 WILDLIFE RESEARCH REPORT State of Cost Center Work Package Task No. Colorado 3430 0670 1 Federal Aid Project: N/A : Division of Wildlife : Mammals Research : Lynx Conservation : Post-Release Monitoring of Lynx : Reintroduced to Colorado : Period Covered: July 1, 2006 - June 30, 2007 Author: T. M. Shenk Personnel: L. Baeten, B. Diamond, R. Dickman, D. Freddy, L. Gepfert, J. Ivan, R. Kahn, A. Keith, G. Merrill, M. Schuette, B. Smith, T. Spraker, S. Wait, S. Waters, L. Wolfe, D. Younkin All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT In an effort to establish a viable population of lynx (Lynx canadensis) in Colorado, the Colorado Division of Wildlife (CDOW) initiated a reintroduction effort in 1997 with the first lynx released in February 1999. From 1999-2007, 218 lynx were released in Colorado. We documented survival, movement patterns, reproduction, and landscape habitat-use through aerial (n = 9496) and satellite (n = 23,791) tracking. Most lynx remained near the core release area in southwestern Colorado. From 1999June 2007, there were 98 mortalities of released adult lynx. Approximately 30.6% were human-induced which were attributed to collisions with vehicles or gunshot. Starvation and disease/illness accounted for 19.4% of the deaths while 35.7% of the deaths were from unknown causes. Reproductive females had the smallest 90% utilization distribution home ranges ( x = 75.2 km2, SE = 15.9 km2 ), followed by attending males ( x = 102.5 km2, SE = 39.7 km2) and non-reproductive animals ( x = 653.8 km2, SE = 145.4 km2). Reproduction was first documented in 2003 with subsequent successful reproduction in 2004, 2005 and 2006. No dens were documented in 2007. From snow-tracking, the primary winter prey species (n = 506 kills) were snowshoe hare (Lepus americanus, annual x = 74.9%, SE = 4.6, n = 9) and red squirrel (Tamiasciurus hudsonicus, annual x = 16.5%, SE = 4.1, n = 9); other mammals and birds formed a minor part of the winter diet. Lynx use-density surfaces were generated to illustrate relative use of areas throughout Colorado and areas of use in New Mexico, Utah and Wyoming. Within the areas of high use in southwestern Colorado, site-scale habitat use, documented through snow-tracking, supports mature Engelmann spruce (Picea engelmannii)-subalpine fir (Abies lasiocarpa) forest stands with 42-65% canopy cover and 15-20% conifer understory cover as the most commonly used areas in southwestern Colorado. Little difference in aspect (slight preference for north-facing slopes), slope ( x = 15.7°) or elevation ( x = 3173 m) were detected for long beds, travel and kill sites (n = 1841). Den sites (n = 37) 1 �however, were located at higher elevations ( x = 3354 m, SE = 31 m) on steeper ( x = 30°, SE = 2°) and more commonly north-facing slopes with a dense understory of coarse woody debris. The first year of a study to evaluate snowshoe hare densities, demography and seasonal movement patterns among small and medium tree-sized lodgepole pine stands and mature spruce/fir stands was completed in 2006-2007 and will continue through 2009 (see Appendix I of this report). Results to date have demonstrated that CDOW has developed lynx release protocols that ensure high initial post-release survival followed by high long-term survival, site fidelity, reproduction and recruitment of Colorado-born lynx into the Colorado breeding population. What is yet to be demonstrated is whether Colorado can support sufficient recruitment to offset annual mortality for a viable lynx population over time. Monitoring continues in an effort to document such viability. 2 �WILDLIFE RESEARCH REPORT POST RELEASE MONITORING OF LYNX (LYNX CANADENSIS) REINTRODUCED TO COLORADO TANYA M. SHENK P. N. OBJECTIVE The initial post-release monitoring of Canada lynx (Lynx canadensis) reintroduced into Colorado will emphasize 5 primary objectives: 1. Assess and modify release protocols to ensure the highest probability of survival for each lynx released. 2. Obtain regular locations of released lynx to describe general movement patterns and habitats used by lynx. 3. Determine causes of mortality in reintroduced lynx. 4. Estimate survival of lynx reintroduced to Colorado. 5. Estimate reproduction of lynx reintroduced to Colorado. Three additional objectives will be emphasized after lynx display site fidelity to an area: 6. Refine descriptions of habitats used by reintroduced lynx. 7. Refine descriptions of daily and overall movement patterns of reintroduced lynx. 8. Describe hunting habits and prey of reintroduced lynx. Information gained to achieve these objectives will form a basis for the development of lynx conservation strategies in the southern Rocky Mountains. SEGMENT OBJECTIVES 1. Complete winter 2006-07 field data collection on lynx habitat use at the landscape scale, hunting behavior, diet, mortalities, and movement patterns. 2. Complete winter 2006-07 lynx trapping field season to collar Colorado born lynx and re-collar adult lynx. 3. Complete spring 2007 field data on lynx reproduction. 4. Summarize and analyze data and publish information as Progress Reports, peer-reviewed manuscripts for appropriate scientific journals, or CDOW technical publications. 5. Complete the first year of field work to evaluate snowshoe hare (Lepus americanus) densities, demography and seasonal movement patterns among small and medium tree-sized lodgepole pine stands and mature spruce/fir stands (see Appendix I). INTRODUCTION The Canada lynx occurs throughout the boreal forests of northern North America. Colorado represents the southern-most historical distribution of lynx, where the species occupied the higher elevation, montane forests in the state. Little was known about the population dynamics or habitat use of this species in their southern distribution. Lynx were extirpated or reduced to a few animals in the state by the late 1970’s due, most likely, to predator control efforts such as poisoning and trapping. Given the isolation of Colorado to the nearest northern populations, the CDOW considered reintroduction as the only option to attempt to reestablish the species in the state. 3 �A reintroduction effort was begun in 1997, with the first lynx released in Colorado in 1999. To date, 218 wild-caught lynx from Alaska and Canada have been released in southwestern Colorado. The goal of the Colorado lynx reintroduction program is to establish a self-sustaining, viable population of lynx in this state. Evaluation of incremental achievements necessary for establishing viable populations is an interim method of assessing if the reintroduction effort is progressing towards success. There are 7 critical criteria for achieving a viable population: 1) development of release protocols that lead to a high initial post-release survival of reintroduced animals, 2) long-term survival of lynx in Colorado, 3) development of site fidelity by the lynx to areas supporting good habitat in densities sufficient to breed, 4) reintroduced lynx must breed, 5) breeding must lead to reproduction of surviving kittens 6) lynx born in Colorado must reach breeding age and reproduce successfully, and 7) recruitment must equal or be greater than mortality over an extended period of time. The post-release monitoring program for the reintroduced lynx has 2 primary goals. The first goal is to determine how many lynx remain in Colorado and their locations relative to each other. Given this information and knowing the sex of each individual, we can assess whether these lynx can form a breeding core from which a viable population might be established. From these data we can also describe general movement patterns and habitat use. The second primary goal of the monitoring program is to estimate survival of the reintroduced lynx and, where possible, determine causes of mortality for reintroduced lynx. Such information will help in assessing and modifying release protocols and management of lynx once they have been released to ensure their highest probability of survival. Additional goals of the post-release monitoring program for lynx reintroduced to the southern Rocky Mountains included refining descriptions of habitat use and movement patterns and describing successful hunting habitat once lynx established home ranges that encompassed their preferred habitat. Specific objectives for the site-scale habitat data collection include: 1) describe and quantify site-scale habitat use by lynx reintroduced to Colorado, 2) compare site-scale habitat use among types of sites (e.g., kills vs. long-duration beds), and 3) compare habitat features at successful and unsuccessful snowshoe hare chases. Documenting reproduction is critical to the success of the program and lynx are monitored intensively to document breeding, births, survival and recruitment of lynx born in Colorado. Site-scale habitat descriptions of den sites are also collected and compared to other sites used by lynx. The program will also investigate the ecology of snowshoe hare in Colorado. A study comparing snowshoe hare densities among mature stands of Engelmann spruce (Picea engelmannii)/subalpine fir (Abies lasiocarpa), lodgepole pine (Pinus contorta) and Ponderosa pine (Pinus ponderosa) was completed in 2004 with highest hare densities found in Engelmann spruce/subalpine fir stands and no hares found in Ponderosa pine stands. A study to evaluate the importance of young, regenerating lodgepole pine and mature Engelmann spruce/subalpine fir stands in Colorado by examining density and demography of snowshoe hares that reside in each was initiated in 2005 and will continue through 2009 (see Appendix I). Lynx is listed as threatened under the Endangered Species Act (ESA) of 1973, as amended (16 U. S. C. 1531 et. seq.)(U. S. Fish and Wildlife Service 2000). Colorado is included in the federal listing as lynx habitat. Thus, an additional objective of the post-release monitoring program is to develop conservation strategies relevant to lynx in Colorado. To develop these conservation strategies, information specific to the ecology of the lynx in its southern Rocky Mountain range, such as habitat use, movement patterns, mortality factors, survival, and reproduction in Colorado is needed. 4 �STUDY AREA Southwestern Colorado is characterized by wide plateaus, river valleys, and rugged mountains that reach elevations over 4200 m. Engelmann spruce-subalpine fir is the most widely distributed coniferous forest type at elevations most typically used by lynx. The Core Release Area is defined as areas bounded by the New Mexico state line to the south, Taylor Mesa to the west and Monarch Pass on the north and east and > 2900 m in elevation (Figure 1). The lynx-established core area is roughly bounded by areas used by lynx in the Taylor Park/Collegiate Peak areas in central Colorado and includes areas of continuous use by lynx, including areas used during breeding and denning (Figure 1). METHODS REINTRODUCTION Effort All lynx releases were conducted under the protocols found to maximize survival (see Shenk 2001). Estimated age, sex and body condition were ascertained and recorded for each lynx prior to release (see Wild 1999). Specific release sites were those used in earlier years of the project and were selected based on land ownership and accessibility during times of release (Byrne 1998). Lynx were transported from the Frisco Creek Wildlife Rehabilitation Center, where they were held from their time of arrival in Colorado, to their release site in individual cages. Release site location was recorded in Universal Transverse Mercator (UTM) coordinates and identification of all lynx released at the same location, on the same day, was recorded. Behavior of the lynx on release and movement away from the release site were documented. Movement, Distribution and Relative Use of Areas by Lynx To monitor lynx movements and thus determine distribution and relative use of areas all released lynx were fitted with radio collars. All lynx released in 1999 were fitted with TelonicsTM radio-collars. All lynx released since 1999, with the exception of 5 males released in spring 2000, were fitted with SirtrackTM dual satellite/VHF radio-collars. These collars have a mortality indicator switch that operated on both the satellite and VHF mode. The satellite component of each collar was programmed to be active for 12 hours per week. The 12-hour active periods for individual collars were staggered throughout the week. Signals from the collars allowed for locations of the animals to be made via Argos, NASA, and NOAA satellites. The location information was processed by ServiceArgos and distributed to the CDOW through e-mail messages. Datasets.-- To determine recent (post-reintroduction) movement and distribution of lynx reintroduced, born or initially trapped in Colorado and relative use of areas by these lynx, regular locations of lynx were collected through a combination of aerial and satellite tracking. Locations were recorded and general habitat descriptions for each aerial location was recorded. The first dataset of lynx locations included all locations obtained from daytime flights conducted with a Cessna 185 or similar aircraft to locate lynx by their VHF collar transmitters (hereafter aerial locations). VHF transmitters have been used on lynx since the first lynx were released in February 1999. The second type of lynx location data was collected via satellite from the satellite collar transmitters placed on the lynx (hereafter satellite locations). Satellite transmitter collars were first used for lynx in April 2000. These satellite collars also contained a VHF transmitter which also allowed locating lynx from the air or ground. All locations were recorded in Universal Transverse Mercator (UTM) coordinates using the CONUS NAD27 datum. Flights to obtain lynx aerial locations were typically conducted on a weekly basis throughout most summer and winter months and twice a week during the den search field season (May 15 – June 30), depending on weather and availability of planes and pilots. Flights were typically concentrated in the high elevation (> 2700 m) southwest quadrant of Colorado which encompasses the core lynx release and 5 �research area (Figure 1). Flights during the den seasons were conducted to obtain locations on all female lynx within the state wearing an active VHF transmitter. VHF transmitters were outfitted with sufficient batteries to last 60 months. The satellite transmitters were designed to provide locations on a weekly basis with sufficient batteries to last for 18 months. Lynx may not be exhibiting typical behavior or habitat use within the first few months after their release in Colorado. Therefore, a subset of each of the aerial and satellite datasets was created that eliminated the first 180 days (approximately 6 months) of locations obtained for each lynx immediately after their initial release. As a result, the truncated aerial location dataset contained lynx locations from September 1999 through March 2007 while the truncated satellite location dataset began October 2000 and extended through March 2007. Accuracy of both aerial and satellite locations varied with the environmental conditions at the time the location was obtained. Accuracy of aerial locations was influenced by weather with accuracy ranging from 50 - 500 meters. Satellite location accuracy was also influenced by atmospheric conditions and position of the satellites. Satellite location accuracy ranged from 150 meters -10 km. Movement and Distribution.-- To document all known lynx locations maps were generated with all aerial and satellite locations displayed. Due to lynx movements outside of Colorado, particularly into the states of New Mexico, Utah and Wyoming we further evaluated lynx use throughout those three states, as well as the data would allow. All individual lynx located at least once in these 3 states (nontruncated datasets) were identified and tallied for each year. To document consistency and known use of these states after the initial effect of being reintroduced was minimized (i.e., 180 days post-release), each individual lynx located at least once in these states from the truncated datasets were identified and tallied. Relative Use.-- To document relative use of areas by lynx, 90% kernel use-density surfaces were calculated for truncated satellite and aerial lynx locations using the ArcGIS Spatial Analyst Kernel Density Tool. Due to differences in data collection frequency and accuracy between datasets, the truncated satellite and truncated aerial data were analyzed separately for generating the lynx use-density surfaces. These use-density surfaces fit a smoothly curved surface over each lynx location. The surface value was highest at the location of the point and diminished with increasing distance from the point. A fixed kernel was used with a smoothing parameter of 5 km, reaching 0 at the search radius distance from the point. Only a circular neighborhood was possible. The volume under the surface equaled the total value for the point. The use-density at each output GIS raster cell was calculated by adding the values of all the kernel surfaces from all the lynx point locations that overlaid each raster cell center. The kernel function was based on the quadratic kernel function described in Silverman (1986, p. 76, equation 4.5). The use-density surfaces were calculated at 100 m resolution. To enhance graphic displays of higher usedensity areas, density values representing single locations were not displayed. Home Range Annual home ranges were calculated as a 95% utilization distribution using a kernel home-range estimator for each lynx we had at least 30 locations for within a year. A year was defined as March 15 – March 14 of the following year. Locations used in the analyses were collected from September 1999 – January 2006 and all locations obtained for an individual during the first six months after its release were eliminated from any home range analyses as it was assumed movements of lynx initially post-release may not be representative of normal habitat use. Locations were obtained either through aerial VHF surveys or locations or the midpoint (ArcView Movement Extension) of all high quality (accuracy rating of 01km) satellite locations obtained within a single 24-hour period. All locations used within a single home range analysis were taken a minimum of 24 hours apart. 6 �Home range estimates were classified as being for a reproductive or non-reproductive animal. A reproductive female was defined as one that had kittens with her; a reproductive male was defined as a male whose movement patterns overlapped that of a reproductive female. If a litter was lost within the defined year a home range described for a reproductive animal were estimated using only locations obtained while the kittens were still with the female. Survival Survival was estimated as ragged telemetry data using the nest survival models in Program MARK (White and Burnham 1999). Mortality Factors When a mortality signal (75 beats per minute [bpm] vs. 50 bpm for the Telonics™ VHF transmitters, 20 bpm vs. 40 bpm for the Sirtrack™ VHF transmitters, 0 activity for Sirtrack™ PTT) was heard during either satellite, aerial or ground surveys, the location (UTM coordinates) was recorded. Ground crews then located and retrieved the carcass as soon as possible. The immediate area was searched for evidence of other predators and the carcass photographed in place before removal. Additionally, the mortality site was described and habitat associations and exact location were recorded. Any scat found near the dead lynx that appeared to be from the lynx was collected. All carcasses were transported to the Colorado State University Veterinary Teaching Hospital (CSUVTH) for a post mortem exam to 1) determine the cause of death and document with evidence, 2) collect samples for a variety of research projects, and 3) archive samples for future reference (research or forensic). The gross necropsy and histology were performed by, or under the lead and direct supervision of a board certified veterinary pathologist. At least one research personnel from the CDOW involved with the lynx program was also present. The protocol followed standard procedures used for thorough post-mortem examination and sample collection for histopathology and diagnostic testing (see Shenk 1999 for details). Some additional data/samples were routinely collected for research, forensics, and archiving. Other data/samples were collected based on the circumstances of the death (e.g., photographs, video, radiographs, bullet recovery, samples for toxicology or other diagnostic tests, etc.). From 1999–2004 the CDOW retained all samples and carcass remains with the exception of tissues in formalin for histopathology, brain for rabies exam, feces for parasitology, external parasites for ID, and other diagnostic samples. Since 2005 carcasses are disposed of at the CSUVTH with the exception of the lower canine, fecal samples, stomach content samples and tissue or bone marrow samples to be delivered by CDOW to the Center for Disease control for plague testing. The lower canine, from all carcasses, is sent to Matson Labs (Missoula, Montana) for aging and the fecal and stomach content samples are evaluated for diet. Reproduction Females were monitored for proximity to males during each breeding season. We defined a possible mating pair as any male and female documented within at least 1 km of each other in breeding season through either flight data or snow-tracking data. Females were then monitored for site fidelity to a given area during each denning period of May and June. Each female that exhibited stationary movement patterns in May or June were closely monitored to locate possible dens. Dens were found when field crews walked in on females that exhibited virtually no movement for at least 10 days from both aerial and ground telemetry. Kittens found at den sites were weighed, sexed and photographed. Each kitten was uniquely marked by inserting a sterile passive integrated transponder (PIT, Biomark, Inc., Boise, Idaho, USA) tag subcutaneously between the shoulder blades. Time spent at the den was minimized to ensure the least 7 �amount of disturbance to the female and the kittens. Weight, PIT-tag number, sex and any distinguishing characteristics of each kitten was also recorded. Beginning in 2005, blood and saliva samples were collected and archived for genetic identification. During the den site visits, den site location was recorded as UTM coordinates. General vegetation characteristics, elevation, weather, field personnel, time at the den, and behavioral responses of the kittens and female were also recorded. Once the females moved the kittens from the natal den area, den sites were visited again and site-specific habitat data were collected (see Habitat Use section below). Captures Captures were attempted for either lynx that were in poor body condition or lynx that needed to have their radio-collars replaced due to failed or failing batteries or to radio-collar kittens born in Colorado once they reached at least 10-months of age when they were nearly adult size. Methods of recapture included 1) trapping using a Tomahawk™ live trap baited with a rabbit and visual and scent lures, 2) calling in and darting lynx using a Dan-Inject CO2 rifle, 3) custom box-traps modified from those designed by other lynx researchers (Kolbe et al. 2003) and 4) hounds trained to pursue felids were also used to tree lynx and then the lynx was darted while treed. Lynx were immobilized either with Telazol (3 mg/kg; modified from Poole et al. 1993 as recommended by M. Wild, DVM) or medetomidine (0.09mg/kg) and ketamine (3 mg/kg; as recommended by L. Wolfe, DVM)) administered intramuscularly (IM) with either an extendible pole-syringe or a pressurized syringe-dart fired from a Dan-Inject air rifle. Immobilized lynx were monitored continuously for decreased respiration or hypothermia. If a lynx exhibited decreased respiration 2mg/kg of Dopram was administered under the tongue; if respiration was severely decreased, the animal was ventilated with a resuscitation bag. If medetomidine/ketamine were the immobilization drugs, the antagonist Atipamezole hydrochloride (Antisedan) was administered. Hypothermic (body temperature < 95o F) animals were warmed with hand warmers and blankets. While immobilized, lynx were fitted with replacement SirtrackTM VHF/satellite collar and blood and hair samples were collected. Once an animal was processed, recovery was expedited by injecting the equivalent amount of the antagonist Antisedan IM as the amount of medetomidine given, if medetomodine/ketemine was used for immobilization. Lynx were then monitored while confined in the box-trap until they were sufficiently recovered to move safely on their own. No antagonist is available for Telezol so lynx anesthetized with this drug were monitored until the animal recovered on its own in the box-trap and then released. If captured and in poor body condition, lynx were anesthetized with either Telezol (2 mg/kg) or medetomodine/ketemine and returned to the Frisco Creek Wildlife Rehabilitation Center for treatment. HABITAT USE Gross habitat use was documented by recording canopy vegetation at aerial locations. More refined descriptions of habitat use by reintroduced lynx were obtained through following lynx tracks in the snow (i.e., snow-tracking) and site-scale habitat data collection conducted at sites found through this method to be used by lynx. See Shenk (2006) for detailed methodologies. DIET AND HUNTING BEHAVIOR Winter diet of reintroduced lynx was estimated by documenting successful kills through snowtracking. Prey species from failed and successful hunting attempts were identified by either tracks or remains. Scat analysis also provided information on foods consumed. Scat samples were collected wherever found and labeled with location and individual lynx identification. Only part of the scat was collected (approximately 75%); the remainder was left in place in the event that the scat was being used by the animal as a territory mark. Site-scale habitat data collected for successful and unsuccessful snowshoe hare kills were compared. 8 �SNOWSHOE HARE ECOLOGY A study designed to evaluate the importance of young, regenerating lodgepole pine and mature Engelmann spruce / subalpine fir stands in Colorado by examining density and demography of snowshoe hares that reside in each was initiated in 2005. Specifically, the study was designed to evaluate small and medium lodgepole pine stands and large spruce/fir stands where the classes “small”, “medium”, and “large” refer to the diameter at breast height (dbh) of overstory trees as defined in the United States Forest Service R2VEG Database (small = 2.54−12.69 cm dbh, medium = 12.70−22.85 cm, and large = 22.86−40.64 cm dbh; J. Varner, United States Forest Service, personal communication). The study design was also developed to identify which of the numerous density-estimation procedures available perform accurately and consistently using an innovative, telemetry augmentation approach as a baseline. Movement patterns and seasonal use of deciduous cover types such as riparian willow were assessed. Finally, the study was designed to further expound on the relationship between density, demography, and stand-type by examining how snowshoe hare density and demographic rates vary with specific vegetation, physical, and landscape characteristics of a stand. RESULTS REINTRODUCTION Effort From 1999 through 2006, 218 lynx were reintroduced into southwestern Colorado (Table 1). No lynx were released in 2007. All lynx were released with either VHF or dual VHF/satellite radio collars so they could be monitored for movement, reproduction and survival. The CDOW does not plan to release any additional lynx in 2008. Movement Patterns and Distribution Numerous travel corridors were used repeatedly by more than one lynx. These travel corridors include the Cochetopa Hills area for northerly movements, the Rio Grande Reservoir-SilvertonLizardhead Pass for movements to the west, and southerly movements down the east side of Wolf Creek Pass to the southeast through the Conejos River Valley. Lynx appear to remain faithful to an area during winter months, and exhibit more extensive movements away from these areas in the summer. A total of 9496 aerial and 23791 satellite locations were obtained from the 218 reintroduced lynx, radio-collared Colorado kittens (n = 14) and unmarked lynx captured in Colorado (n = 2) as of June 30, 2007. The majority of these locations were in Colorado (Figure 2). Some reintroduced lynx dispersed outside of Colorado into Arizona, Idaho, Iowa, Kansas, Montana, Nebraska, Nevada, New Mexico, South Dakota, Utah and Wyoming (Figure 2). The majority of surviving lynx from the reintroduction effort currently continue to use high elevation (> 2900 m), forested terrain in an area bounded on the south by New Mexico north to Independence Pass, west as far as Taylor Mesa and east to Monarch Pass. Most movements away from the Core Release Area were to the north. Relative Use The lynx use-density surfaces resulting from the fixed kernel analyses provided relative probabilities of finding lynx in areas throughout their distribution. A single use-density surface was calculated separately for both the aerial (n = 8058) and satellite truncated datasets (n = 16240). Relative Use in Colorado.-- All 218 lynx released in Colorado, all radio-collared kittens and 2 captured unmarked adults were located at least once in Colorado. The majority of these lynx remained in Colorado. The use-density surfaces within Colorado were displayed separately for both the aerial (Figure 3) and satellite truncated datasets (Figure 4). Of the total locations available in the 9 �truncated datasets used to generate the use-density surfaces, 7953 of the aerial locations and 13,241 of the satellite locations were in Colorado. Aerial and satellite use-density surfaces indicated similar high usedensity areas. Satellite locations indicated broader spatial use by lynx because satellite collars provided more locations than flights. The use-density surface for lynx use in Colorado indicates two primary areas of use. The first is the Core Research Area (see Figure 1) and a secondary core centered in the Collegiate Peaks Wilderness (Figures 3 and 4). High use is also documented for 1) the area east of Dillon, on both the north and south sides of I70 and 2) the area north of Hwy 50 centered around Gunnison and then north to Crested Butte. These last 2 high use areas are smaller in extent than the 2 core areas. Relative Use in New Mexico.-- Combining the non-truncated aerial (n = 81) and satellite lynx location (n = 928) datasets, lynx used New Mexico consistently and with an increasing number of individuals from 1999 through 2006 (Table 2). Data for 2007 represents only a partial year and thus trend in numbers of individuals using New Mexico for 2007 cannot be made, however continued use of New Mexico into 2007 was documented Sixty lynx (37 females: 23 males) were found within New Mexico from February 1999 through March 2007 (Table 2). Excluding all aerial and satellite lynx locations collected in the first 180 days after release (truncated datasets; n = 61 aerial locations, n = 569 satellite locations), a total of 35 individual lynx (22 females: 13 males) were found within New Mexico from September 1999 through March 2007 (Table 3). The decrease in number of lynx frequenting New Mexico in 2001 through 2003 (Tables 2 and 3) was more likely due to fewer satellite collars functioning in those years rather than indicating less use of the area by lynx. The satellite transmitters placed on lynx in 2000 were failing and no new lynx were released or re-collared in 2001 and 2002. This decrease in satellite locations is present throughout the lynx distribution and is also reflected in the numbers presented below for Utah and Wyoming The use-density surface for lynx use in New Mexico indicates the primary area of use being located either immediately south of the Colorado border and south of the Conejos River Valley (an area of high use in Colorado) or east of Taos (Figure 5). The use-density surfaces throughout both Colorado and New Mexico are displayed so that lynx use within New Mexico can be directly compared to lynx use throughout Colorado (Figure 6). Relative Use in Utah.-- Combining the non-truncated aerial (n = 10) and satellite lynx location (n = 574) datasets, lynx used the analysis area consistently and with an increasing number of individuals from 1999 through 2006 (Table 4). Data for 2007 represents only a partial year and thus trend in numbers of individuals using the state for 2007 cannot be made, however continued use of Utah into 2007 was documented. Twenty-two lynx (7 females: 15 males) were found within Utah from February 1999 through March 2007 (Table 4). Excluding all aerial and satellite lynx locations collected in the first 180 days after release (truncated datasets; n = 7 aerial locations, n = 399 satellite locations), 17 individual lynx (6 females: 11 males) were found within Utah from September 1999 through March 2007 (Table 5). The use-density surface for lynx use in Utah indicates the primary area of use being located in the Uinta Mountains (Figure 7). The use-density surfaces throughout both Colorado and Utah are displayed so that lynx use within Utah can be directly compared to lynx use throughout Colorado (Figure 8). Relative Use in Wyoming.-- Combining the non-truncated aerial (n = 34) and satellite lynx location (n = 1780) datasets, lynx used the analysis area consistently and with an increasing number of individuals from 1999 through 2006 (Table 6). Data for 2007 represents only a partial year and thus trend in numbers of individuals using the state for 2007 cannot be made, however continued use of the Wyoming into 2007 was documented. Thirty-three lynx (14 females: 19 males) were found within 10 �Wyoming from February 1999 through March 2007 (Table 6). Excluding all aerial and satellite lynx locations collected in the first 180 days after release (truncated datasets; n = 28 aerial locations, n = 1533 satellite locations), 27 individual lynx (13 females: 14 males) were found within Wyoming from September 1999 through March 2007 (Table 7). The use-density surface for lynx use in Wyoming indicates the primary area of use being located either immediately north of the Colorado border in the Medicine Bow National Forest or in the northwest quadrant of the state including areas in Yellowstone and Teton National Parks and the Laramie Range (Figure 9). The use-density surfaces throughout both Colorado and Wyoming are displayed so that lynx use within Wyoming can be directly compared to lynx use throughout Colorado (Figure 10). Home Range Reproductive females had the smallest 90% utilization distribution annual home ranges ( x = 75.2 km2, SE = 15.9 km2, n = 19), followed by attending males ( x = 102.5 km2, SE = 39.7 km2, n = 4). Nonreproductive females had the largest annual home ranges ( x = 703.9 km2, SE = 29.8 km2, n = 32) followed by non-reproductive males ( x = 387.0 km2, SE = 73.5 km2, n = 6). Combining all nonreproductive animals yielded a mean annual home range of 653.8 km2 (SE = 145.4 km2, n = 38). Survival Initial survival rate estimates for reintroduced lynx were completed, however, further analyses need to be conducted before estimates will be presented. As of June 30, 2007, CDOW was actively monitoring/tracking 71 of the 120 lynx still possibly alive (Table 8). There are 50 lynx that we have not heard signals on since at least June 30, 2006 and these animals are classified as ‘missing’ (Table 8). One of these missing lynx is a mortality of unknown identity, thus only 49 are truly missing. Possible reasons for not locating these missing lynx include 1) long distance dispersal, beyond the areas currently being searched, 2) radio failure, or 3) destruction of the radio (e.g., run over by car). CDOW continues to search for all missing lynx during both aerial and ground searches. Two of the missing lynx released in 2000 are thought to have slipped their collars. Mortality Factors Of the total 218 adult lynx released, we have 98 known mortalities as of June 30, 2007 (Table 9). The primary known causes of death included 30.6% human-induced deaths which were confirmed or probably caused by collisions with vehicles or gunshot. Starvation and disease/illness accounted for 19.4% of the deaths; starvation was a significant cause of mortality in the first year of releases only. An additional 35.7% of known mortalities were from unknown causes. Mortalities occurred throughout the areas where lynx moved, including 13 in New Mexico, 4 in Wyoming and Nebraska, 3 in Utah and 1 each in Arizona, Kansas and Montana (Figure 2, Table 10). Reproduction Field crews weighed, photographed, PIT-tagged the kittens and took hair, blood and saliva samples from the kittens for genetic work in an attempt to confirm paternity. Lynx kittens weigh approximately 200 grams at birth and do not open their eyes until they are 10-17 days old. Kittens were processed as quickly as possible (11-32 minutes) to minimize the time the kittens were without their mother. While working with the kittens the females remained nearby, often making themselves visible to the field crews. The females generally continued a low growling vocalization the entire time personnel were at the den. In all cases, the female returned to the den site once field crews left the area. At all dens the females appeared in excellent condition, as did the kittens. 11 �2003.-- Nine pairs of lynx were documented during the 2003 breeding season (March and April) from the 17 females we were monitoring. In May and June, 6 dens and a total of 16 kittens were found in the lynx Core Release Area in southwestern Colorado (Table 11, Figure 1). The kittens weighed from 270-500 grams. The dens were scattered throughout the Core Release Area, with no dens found outside the core area. All the dens were in Engelmann spruce/subalpine fir forests in areas of extensive downfall. Elevations ranged from 3240-3557 m. Four of the 6 females that we know had kittens in summer 2003 were still with kittens at the end of April 2004. Two of those females still had 2 kittens with them at that time. Visual observations in February 2004 of one female with 2 kittens indicated all 3 were in good body condition. The mortality of female YK00F16 and her 1 kitten in October 2003 from plague was not due to poor habitat or prey conditions, and thus we might assume she would have raised the 1 kitten to this stage as well. Three probable kitten deaths from female YK00F19 were from 1 litter that most likely failed very early. Through snow-tracking in winter 2003-04 an unknown female (no radio frequency heard in the area of the tracks) we also documented 1-2 additional kittens born spring 2003 and still alive in winter 2004. Of the 16 kittens we found in summer 2003, we documented the following by April 2004: 6 confirmed alive, 7 confirmed dead, and 3 some evidence dead. Although we tried, we were not able to capture any of the 6 surviving kittens to fit them with radio-collars for subsequent monitoring. 2004.-- In Spring 2004, 26 females from the releases in 1999, 2000 and 2003 had active radiocollars. Of these, we documented 18 possible mating pairs of lynx during breeding season. All 4 of the females that had kittens with them through winter 2003-04 bred again spring 2004; 2 with the same male they successfully bred with spring 2003. During May-June 2004 we found 11 dens and a total of 30 kittens (Table 11). The kittens weighed from 250-770 grams. Three of the 11 females with kittens were from the 2003 releases. Three additional litters were documented after denning season through either observation of a female lynx with kittens or snow-tracking females with kittens that were not one of the 11 females found on dens. From the size of the kittens they would have been born during the normal denning season in May or June. Nine additional kittens were observed from these litters for a total of 39 known kittens born in 2004. Two of these additional litters were documented from direct follow-ups to sighting made by the public and reported to CDOW. Two females that had kittens in 2003 and reared at least part of their litters through March 2004, bred and had kittens again in 2004. Two of the litters documented by direct observation or snow-tracking are from females whose collars were no longer functioning. Seven kittens born in 2004 were captured at approximately 10-months of age and fitted with dual satellite/VHF collars. Six of the 7 were still alive and being monitored as of June 30, 2006. The cut collar of one kitten CO04M15 was left at the Silverton Post Office on October 25, 2005. We assume this lynx is dead. 2005.-- In spring 2005 we had 40 females from the releases in 1999, 2000, 2003 and 2004 that had active radio-collars. We documented 23 possible mating pairs of lynx during breeding season. During May-June 2005 we visited 16 dens and found a total of 46 kittens (Table 11). An additional female (BC03F10) had a den we were not able to get to during May or June due to high water during spring run-off. Female BC03F03 was hit and killed on I-70 on 5/19/2005. She had 2 fetuses in her uterus, so would have contributed to reproduction this year had she lived. All of the 2005 dens were scattered throughout the high elevation areas of Colorado, south of I70. Most of the dens were in Engelmann spruce/subalpine fir forests in areas of extensive downfall. Elevations ranged from 3117-3586 m. Four of the females would not leave the den until we reached out to pick up a kitten. 12 �One female, YK00F10 has had litters 3 years in a row. In 2003 she had 4 kittens and raised 2 through the winter. In 2004 she had 2 kittens and raised both through the winter, in 2005 she had 4 kittens again. She has had all 3 litters in the same general area and has had the same mate for 3 years. Eight additional females had their second litter in Colorado in 2005. Three females from the 2004 releases had litters in 2005. Year 2005 was the second consecutive year that we had females released the prior spring find a territory and a mate within a year and produced live young. In reproduction season 2004 we had 3 females released in spring 2003 that also produced live young the next year. Of those 3, 2 successfully raised at least part of their litters through winter 2005. Seven kittens born in 2005 were captured at approximately 10-months of age and fitted with dual satellite/VHF collars. One of the 7 was still alive and being monitored as of June 30, 2007. 2006.-- In spring 2006, 42 females were being monitored. We found 4 dens in May and June 2006 with 11 kittens total (Table 11). Lynx CO04F07, a female lynx born in Colorado in 2004, was the mother of one of these litters which documented the first recruitment of Colorado-born lynx into the Colorado breeding population. There were at least 2 surviving kittens as of spring 2007. We were unsuccessful in capturing these kittens for collar placement. The percent of tracked females found with litters in 2006 was lower (0.095) than in the 3 previous years (0.413, SE = 0.032, Table 11). However, all demographic and habitat characteristics measured at the 4 dens that were found in 2006 were comparable to all other dens found (Table 11). Mean number of kittens per litter from 2003-2006 was 2.78 (SE = 0.05) and sex ratio of females to males was equal ( x = 1.14, SE = 0.14). 2007.-- During May and June 2007 we monitored 34 females for reproduction (Table 11). No dens were found. Den Sites.-- A total of 37 dens have been found from 2003-2006. All of the dens except one have been scattered throughout the high elevation areas of Colorado, south of I-70. In 2004, 1 den was found in southeastern Wyoming, near the Colorado border. Dens were located on steep ( x slope = 30o , SE=2o), north-facing, high elevation ( x = 3354 m, SE = 31 m) slopes. The dens were typically in Engelmann spruce/subalpine fir forests in areas of extensive downfall of coarse woody debris (Shenk 2006). All dens were located within the winter use areas used by the females. Captures Two adult lynx were captured in 2001 for collar replacement. One lynx was captured in a tomahawk live-trap, the other was treed by hounds and then anesthetized using a jab pole. Five adult lynx were captured in 2002; 3 were treed by hounds and 2 were captured in padded leghold traps. In 2004, 1 lynx was captured with a Belisle snare and 6 adult lynx were captured in box-traps. Trapping effort was substantially increased in winter and spring 2005 and 12 adult lynx were captured and re-collared. Eight reintroduced lynx were captured in winter and spring 2006. In 2007, 11 reintroduced adult lynx were captured and re-collared. All lynx captured in Colorado from 2005-2007 were caught in box-traps. In addition, as part of the collaring trapping effort, 14 Colorado-born kittens were captured and collared at approximately 10-months of age. Seven 2004-born kittens were collared in spring 2005, and 7, 2005-born kittens were collared in spring 2006. We were not successful at capturing and collaring any kittens born in 2006 in winter 2006-07. We did however, capture 2 adults (approximate age 2 years old) in winter 2006-07 that had no PIT-tags or radio collars. We assume these 2 lynx were from litters born in Colorado that were never found at dens (i.e., why there were no PIT-tags). All lynx captured for collaring 13 �or re-collaring were fitted with new Sirtrack TM dual VHF/satellite collars and re-released at their capture locations. Seven adult lynx were captured from March 1999-June 30, 2007 because they were in poor body condition (Table 12). Five of these lynx were successfully treated at the Frisco Creek Rehabilitation Center and re-released in the Core Release Area. One lynx, BC00F7, died from starvation and hypothermia within 1 day of capture at the rehabilitation center. Lynx QU04M07 died 3 days after capture at the rehabilitation center. Necropsy results documented starvation as the cause of death that was precipitated by hydrocephalus and bronchopneumonia (unpublished data T. Spraker, CSUVTH). Seven lynx were captured (either by CDOW personnel or conservation personnel in other states) because they were in atypical habitat outside the state of Colorado (Table 12). They were held at Frisco Creek Rehabilitation Center for a minimum of 3 weeks, fitted with new Sirtrack TM dual VHF/satellite collars and re-released in the Core Release Area in Colorado. Five of these 7 lynx were still alive 6 months post-re-release but 3 had already dispersed out of Colorado and 2 stayed in Colorado through June 30, 2007. Two lynx died within 6 months of re-release: 1 died of starvation in Colorado and the other died of unknown causes in Nebraska. Two lynx captured out of state and re-released currently remain in Colorado. HABITAT USE Landscape-scale daytime habitat use was documented from 9496 aerial locations of lynx collected from February 1999-June 30, 2007. Throughout the year Engelmann spruce - subalpine fir was the dominant cover used by lynx. A mix of Engelmann spruce, subalpine fir and aspen (Populus tremuloides) was the second most common cover type used throughout the year. Various riparian and riparian-mix areas were the third most common cover type where lynx were found during the daytime flights. Use of Engelmann spruce-subalpine fir forests and Engelmann spruce-subalpine fir-aspen forests was similar throughout the year. There was a trend in increased use of riparian areas beginning in July, peaking in November, and dropping off December through June. Site-scale habitat data collected from snow-tracking efforts indicate Engelmann spruce and subalpine fir were also the most common forest stands used by lynx for all activities during winter in southwestern Colorado. Comparisons were made among sites used for long beds, dens, travel and where they made kills. Little difference in aspect, mean slope and mean elevation were detected for 3 of the 4 site types including long beds, travel and kills where lynx typically use gentler slopes ( x = 15.7o ) at an mean elevation of 3173 m, and varying aspects with a slight preference for north-facing slopes. See Shenk (2006) for more detailed analyses of habitat use. DIET AND HUNTING BEHAVIOR Winter diet of lynx was documented through detection of kills found through snow-tracking. Prey species from failed and successful hunting attempts were identified by either tracks or remains. Scat analysis also provided information on foods consumed. A total of 506 kills were located from February 1999-April 2007. We collected over 900 scat samples from February 1999-April 2007 that will be analyzed for content. In each winter, the most common prey item was snowshoe hare, followed by red squirrel (Tamiusciurus hudsonicus; Table 13). The percent of snowshoe hare kills found however, varied annually from a low of 55.56% in 1999 to a high of 90.77% in winter 2002-2003. SNOWSHOE HARE ECOLOGY The first year of a study to evaluate snowshoe hare densities, demography and seasonal movement patterns among small and medium tree-sized lodgepole pine stands and mature spruce/fir stands was completed and preliminary results presented (Appendix I). 14 �DISCUSSION In an effort to establish a viable population of lynx in Colorado, CDOW initiated a reintroduction effort in 1997 with the first lynx released in winter 1999. From 1999 through spring 2007, 218 lynx were released in the Core Release Area. Locations of each lynx were collected through aerial- or satellite-tracking to document movement patterns and to detect mortalities. Most lynx remain in the high elevation, forested areas in southwestern Colorado. The use-density surfaces for lynx use in Colorado indicate two primary areas of use. The first is the Core Research Area (see Figure 1) and a secondary core centered in the Collegiate Peaks Wilderness (Figures 3, 4). High use is also documented for 1) the area east of Dillon, on both the north and south sides of I70 and 2) the area north of Hwy 50 centered around Gunnison and then north to Crested Butte. These last 2 high use areas are smaller in extent than the 2 core areas. Dispersal movement patterns for lynx released in 2000 and subsequent years were similar to those of lynx released in 1999 (Shenk 2000). However, more animals released in 2000 and subsequent years remained within the Core Release Area than those released in 1999. This increased site fidelity may have been due to the presence of con-specifics in the area on release. Numerous travel corridors within Colorado have been used repeatedly by more than 1 lynx. These travel corridors include the Cochetopa Hills area for northerly movements, the Rio Grande Reservoir-Silverton-Lizardhead Pass for movements to the west, and southerly movements down the east side of Wolf Creek Pass to the southeast to the Conejos River Valley. Lynx appear to remain faithful to an area during winter months, and exhibit more extensive movements away from these areas in the summer. Reproductive females had the smallest 90% utilization distribution home ranges ( x = 75.2 km2, SE = 15.9 km2), followed by attending males ( x = 102.5 km2, SE = 39.7 km2) and non-reproductive animals ( x = 653.8 km2, SE = 145.4 km2). Most lynx currently being tracked are within the Core Release Area. During the summer months, lynx were documented to make extensive movements away from their winter use areas. Extensive summer movements away from areas used throughout the rest of the year have been documented in native lynx in Wyoming and Montana (Squires and Laurion 1999). Current data collection methods used for the Colorado lynx reintroduction program were not specifically designed to address the reintroduced lynx movements or use of areas in other states. In particular, the core research and release area were in Colorado. Therefore, the number of aerial locations obtained would be far fewer in other states than in Colorado which would bias low the number of lynx and intensity of lynx use documented outside the state. In contrast, obtaining satellite locations is not biased by the location of the lynx. Satellite locations are, however, biased by the shorter time the satellite transmitters function, approximately 18 months versus 60 months for the VHF transmitters used to obtain the aerial locations. However, data collected to meet objectives of the lynx reintroduction program were used to provide information to help address the question of lynx use outside of Colorado. Due to the rarity of flights conducted outside Colorado, only use-density surfaces generated from satellite locations were used to document relative lynx use of areas in New Mexico, Utah and Wyoming. New Mexico and Wyoming have been used continuously by lynx since the first year lynx were released in Colorado (1999) to the present (Tables 2, 6). Lynx reintroduced in Colorado were first documented in Utah in 2000 (Table 4) and are still being documented there to date. In addition, all levels of lynx use-density documented throughout Colorado are also represented in New Mexico, Utah and Wyoming from none to the highest level of use (Figures 5, 7, 9). One den was found in Wyoming. Although no reproduction has been documented in New Mexico or Utah to date, documenting areas of the 15 �highest intensity of use and the continuous presence of lynx within these states for over six years does suggest the potential for year-round residency of lynx and reproduction in those states. The use-density surface for lynx use in New Mexico indicates the primary areas of use being located immediately south of the Colorado border and south of the Conejos River Valley (an area of high use in Colorado) or east of Taos (Figure 5). In Utah, the primary area of use is located in the Uinta Mountains (Figure 7). Lynx use in Wyoming is focused in 2 primary areas, the Medicine Bow National Forest in south-central Wyoming and in the northwest quadrant of the state including areas in Yellowstone and Teton National Parks and the Laramie Range (Figure 9). From 1999-June 2007, there were 98 mortalities of released adult lynx. Human-caused mortality factors are currently the highest causes of death with approximately 30.6% attributed to collisions with vehicles or gunshot. Starvation and disease/illness accounted for 19.4% of the deaths while 35.7% of the deaths were from unknown causes. Lynx mortalities were documented throughout all areas lynx used, including 28 (28.6%) occurring in other states (Figure 2, Table10). Half of the out-of-state mortalities were documented in New Mexico. Reproduction is critical to achieving a self-sustaining viable population of lynx in Colorado. Reproduction was first documented from the 2003 reproduction season and again in 2004, 2005 and 2006. Lower reproduction occurred in 2006 (Table 11) but did include a Colorado-born female giving birth to 2 kittens, documenting the first recruitment of Colorado-born lynx into the Colorado breeding population. No reproduction was documented in 2007. The cause of the decreased reproduction in 2006 and 2007 is unknown. One possible explanation would be a decrease in prey abundance. Additional reproduction is likely to have occurred in all years from females we were no longer tracking, and from Colorado-born lynx that have not been collared. The dens we find are more representative of the minimum number of litters and kittens in a reproduction season. To achieve a viable population of lynx, enough kittens need to be recruited into the population to offset the mortality that occurs in that year and hopefully even exceed the mortality rate to achieve an increasing population. The use-density surfaces depict intensity of use by location. Why certain areas would be used more intensively than others should be explained by the quality of the habitat in those areas. Characteristics of areas used by lynx, as documented through aerial locations and snow-tracking of lynx in the Colorado core research area, include mature Engelmann spruce-subalpine fir forest stands with 4265% canopy cover and 15-20% conifer understory cover (Shenk 2006). Within these forest stand types, lynx appear to have a slight preference for north-facing, moderate slopes ( x = 15.7°) at high elevations ( x = 3173 m; Shenk 2006). Snow-tracking of released lynx also provided information on hunting behavior and diet through documentation of kills, food caches, chases, and diet composition estimated through prey remains. The primary winter prey species (n = 506) were snowshoe hare (Table 12) with an annual x = 74.9% (SE = 4.6, n = 9) and red squirrel (annual x = 16.5%, SE = 4.1, n = 9). Thus, areas of good habitat must also support populations of snowshoe hare and red squirrel. In winter, lynx reintroduced to Colorado appear to be feeding on their preferred prey species, snowshoe hare and red squirrel in similar proportions as those reported for northern lynx during lows in the snowshoe hare cycle (Aubry et al. 1999). Environmental conditions in the springs and summers of 2003 and 2006 resulted in high cone crops during their following winters based on field observations, resulting in increased red squirrel abundance. This may partially explain the higher percent of red squirrel kills, and thus a lower percent of snowshoe hare kills, found in winters 2003-04 and 2006-07 (Table 12). 16 �Caution must be used in interpreting the proportion of identified kills. Such a proportion ignores other food items that are consumed in their entirety and thus are biased towards larger prey and may not accurately represent the proportion of smaller prey items, such as microtines, in lynx winter diet. Through snow-tracking we have evidence that lynx are mousing and several of the fresh carcasses have yielded small mammals in the gut on necropsy. The summer diet of lynx has been documented to include less snowshoe hare and more alternative prey than in winter (Mowat et al., 1999). All evidence suggests reintroduced lynx are finding adequate food resources to survive. Mowat et al. (1999) suggest lynx and snowshoe hare select similar habitats except that hares select more dense stands than lynx. Very dense understory limits hunting success of the lynx and provides refugia for hares. Given the high proportion of snowshoe hare in the lynx diet in Colorado, we might then assume the habitats used by reintroduced lynx also depict areas where snowshoes hare are abundant and available for capture by lynx in Colorado. From both aerial locations taken throughout the year and from the site-scale habitat data collected in winter, the most common areas used by lynx are in stands of Engelmann spruce and subalpine fir. This is in contrast to adjacent areas of Ponderosa pine, pinyon juniper, aspen and oakbrush. The lack of lodgepole pine in the areas used by the lynx may be more reflective of the limited amount of lodgepole pine in southwestern Colorado, the Core Release Area, rather than avoidance of this tree species. Hodges (1999) summarized habitats used by snowshoe hare from 15 studies as areas of dense understory cover from shrubs, stands that are densely stocked, and stands at ages where branches have more lateral cover. Species composition and stand age appears to be less correlated with hare habitat use than is understory structure (Hodges 1999). The stands need to be old enough to provide dense cover and browse for the hares and cover for the lynx. In winter, the cover/browse needs to be tall enough to still provide browse and cover in average snow depths. Hares also use riparian areas and mature forests with understory. Site-scale habitat use documented for lynx in Colorado indicate lynx are most commonly using areas with Engelmann spruce understory present from the snow line to at least 1.5 m above the snow. The mean percent understory cover within the habitat plots is typically less than 15% regardless of understory species. However, if the understory species is willow, percent understory cover is typically double that, with mean number of shrubs per plot approximately 80, far greater than for any other understory species. In winter, hares browse on small diameter woody stems (<0.25"), bark and needles. In summer, hares shift their diet to include forbs, grasses, and other succulents as well as continuing to browse on woody stems. This shift in diet may express itself in seasonal shifts in habitat use, using more or denser coniferous cover in winter than in summer. The increased use of riparian areas by lynx in Colorado from July to November may reflect a seasonal shift in hare habitat use in Colorado. Major (1989) suggested lynx hunted the edge of dense riparian willow stands. The use of these edge habitats may allow lynx to hunt hares that live in habitats normally too dense to hunt effectively. The use of riparian areas and riparian-Engelmann spruce-subalpine fir and riparian-aspen mixes documented in Colorado may stem from a similar hunting strategy. However, too little is known about habitat use by hares in Colorado to test this hypothesis at this time. Lynx also require sufficient denning habitat. Denning habitat has been described by Koehler (1990) and Mowat et al. (1999) as areas having dense downed trees, roots, or dense live vegetation. We found this to be in true in Colorado as well (Shenk 2006). In addition, the dens used by reintroduced lynx were at high elevations and on steep north-facing slopes. All females that were documented with kittens denned in areas within their winter-use area. 17 �SUMMARY From results to date it can be concluded that CDOW developed release protocols that ensure high initial post-release survival of lynx, and on an individual level, lynx demonstrated they can survive longterm in areas of Colorado. We also documented that reintroduced lynx exhibited site fidelity, engaged in breeding behavior and produced kittens that were recruited into the Colorado breeding population. What is yet to be demonstrated is whether current conditions in Colorado can support the recruitment necessary to offset annual mortality in order to sustain the population. Monitoring of reintroduced lynx will continue in an effort to document such viability. ACKNOWLEDGEMENTS The lynx reintroduction program involves the efforts of literally hundreds of people across North America, in Canada and USA. Any attempt to properly acknowledge all the people who played a role in this effort is at risk of missing many people. The following list should be considered to be incomplete. CDOW CLAWS Team (1998-2001): Bill Andree, Tom Beck, Gene Byrne, Bruce Gill, Mike Grode, Rick Kahn (Program Leader), Dave Kenvin, Todd Malmsbury, Jim Olterman, Dale Reed, John Seidel, Scott Wait, Margaret Wild. CDOW: John Mumma (Director 1996-2000), Russell George (Director 2001-2003), Bruce McCloskey (Director 2004-2007), Conrad Albert, Jerry Apker, Laurie Baeten, Cary Carron, Don Crane, Larry DeClaire, Phil Ehrlich, Lee Flores, Delana Friedrich, Dave Gallegos, Juanita Garcia, Drayton Harrison, Jon Kindler, Ann Mangusso, Jerrie McKee, Gary Miller, Melody Miller, Mike Miller, Kirk Navo, Robin Olterman, Jerry Pacheo, Mike Reid, Tom Remington, Ellen Salem, Eric Schaller, Mike Sherman, Jennie Slater, Steve Steinert, Kip Stransky, Suzanne Tracey, Anne Trainor, Scott Wait, Brad Weinmeister, Nancy Wild, Perry Will, Lisa Wolfe, Brent Woodward, Kelly Woods, Kevin Wright. Lynx Advisory Team (1998-2001): Steve Buskirk, Jeff Copeland, Dave Kenny, John Krebs, Brian Miller (Co-Leader), Mike Phillips, Kim Poole, Rich Reading (Co-Leader), Rob Ramey, John Weaver. U. S. Forest Service: Kit Buell, Joan Friedlander, Dale Gomez, Jerry Mastel, John Squires, Fred Wahl, Nancy Warren. U. S. Fish and Wildlife Service: Lee Carlson, Gary Patton (1998-2000), Kurt Broderdorp. State Agencies: Alaska: ADF&G: Cathie Harms, Mark Mcnay, Dan Reed (Regional Manager), Wayne Reglin (Director), Ken Taylor (Assist. Director), Ken Whitten, Randy Zarnke, Other:Ron Perkins (trapper), Dr. Cort Zachel (veterinarian). Washington: Gary Koehler. National Park Service: Steve King. Colorado State University: Alan Franklin, Gary White. Colorado Natural Heritage Program: Rob Schorr, Mike Wunder. Canada: British Columbia: Dr. Gary Armstrong (veterinarian), Mike Badry (government), Paul Blackwell (trapper coordinator), Trappers: Dennis Brown, Ken Graham, Tom Sbo, Terry Stocks, Ron Teppema, Matt Ounpuu. Yukon: Government: Arthur Hoole (Director), Harvey Jessup, Brian Pelchat, Helen Slama, Trappers: Roger Alfred, Ron Chamber, Raymond Craft, Lance Goodwin, Jerry Kruse, Elizabeth Hofer, Jurg Hofer, Guenther Mueller (YK Trapper’s Association), Ken Reeder, Rene Rivard (Trapper coordinator), Russ Rose, Gilbert Tulk, Dave Young. Alberta: Al Cook. Northwest Territories: Albert Bourque, Robert Mulders (Furbearer Biologist), Doug Steward (Director NWT Renewable Res.), Fort Providence Native People. Quebec: Luc Farrell, Pierre Fornier. Colorado Holding Facility: Herman and Susan Dieterich, Kate Goshorn, Loree Harvey, Rachel Riling. Pilots: Dell Dhabolt, Larry Gepfert, Al Keith, Jim Olterman, Matt Secor, Brian Smith, Whitey Wannamaker, Steve Waters, Dave Younkin. 18 �Field Crews (1999-2007): Steve Abele, Brandon Barr, Bryce Bateman, Todd Bayless, Nathan Berg, Ryan Besser, Jessica Bolis, Mandi Brandt, Brad Buckley. Patrick Burke, Braden Burkholder, Paula Capece, Stacey Ciancone, Doug Clark, John DePue, Shana Dunkley, Tim Hanks, Carla Hanson, Dan Haskell, Nick Hatch, Matt Holmes, Andy Jennings, Susan Johnson, Paul Keenlance, Patrick Kolar, Tony Lavictoire, Jenny Lord, Clay Miller, Denny Morris, Kieran O’Donovan, Gene Orth, Chris Parmater, Jake Powell, Jeremy Rockweit, Jenny Shrum, Josh Smith, Heather Stricker, Adam Strong, Dave Unger, David Waltz, Andy Wastell, Mike Watrobka, Lyle Willmarth, Leslie Witter, Kei Yasuda, Jennifer Zahratka. Research Associates: Bob Dickman, Grant Merrill. Data Analysts: Karin Eichhoff, Joanne Stewart, Anne Trainor. Data Entry: Charlie Blackburn, Patrick Burke, Rebecca Grote, Angela Hill, Mindy Paulek. Mary Schuette and Dave Theobald provided assistance with the GIS analysis and M. Schuette generated the maps used in this report Photographs: Tom Beck, Bruce Gill, Mary Lloyd, Rich Reading, Rick Thompson. Funding: CDOW, Great Outdoors Colorado (GOCO), Turner Foundation, U.S.D.A. Forest Service, Vail Associates, Colorado Wildlife Heritage Foundation. LITERATURE CITED AUBRY, K. B., G. M. KOEHLER, J. R. SQUIRES. 1999. Ecology of Canada lynx in southern boreal forests. Pages 373-396 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. BYRNE, G. 1998. Core area release site selection and considerations for a Canada lynx reintroduction in Colorado. Report for the Colorado Division of Wildlife. CURTIS, J. T. 1959. The vegetation of Wisconsin. University of Wisconsin Pres, Madison. HODGES, K. E. 1999. Ecology of snowshoe hares in southern boreal and montane forests. Pages 163206 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. KOEHLER, G. M. 1990. Population and habitat characteristics of lynx and snowshoe hares in north central Washington. Canadian Journal of Zoology 68:845-851. KOLBE, J. A., J. R. SQUIRES, T. W. PARKER. 2003. An effective box trap for capturing lynx. Journal of Wildlife Management 31:980-985. LAYMON, S. A. 1988. The ecology of the spotted owl in the central Sierra Nevada, California. PhD Dissertation University of California, Berkeley, California. MAJOR, A. R. 1989. Lynx, Lynx canadensis canadensis (Kerr) predation patterns and habitat use in the Yukon Territory, Canada. M. S. Thesis, State University of New York, Syracuse. MOWAT, G., K. G. POOLE, AND M. O’DONOGHUE. 1999. Ecology of lynx in northern Canada and Alaska. Pages 265-306 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. POOLE, K. G., G. MOWAT, AND B. G. SLOUGH. 1993. Chemical immobilization of lynx. Wildlife Society Bulletin 21:136-140. SHENK, T. M. 1999. Program narrative Study Plan: Post-release monitoring of reintroduced lynx (Lynx canadensis) to Colorado. Colorado Division of Wildlife, Fort Collins, Colorado. __________. 2001. Post-release monitoring of lynx reintroduced to Colorado. Job Progress Report, Colorado Division of Wildlife, Fort Collins, Colorado. __________. 2006. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report July: 1-45. Colorado Division of Wildlife, Fort Collins, Colorado 19 �SILVERMAN, B.W. 1986. Density Estimation for Statistics and Data Analysis. Chapman and Hall. New York, New York, USA. SQUIRES, J. R. AND T. LAURION. 1999. Lynx home range and movements in Montana and Wyoming: preliminary results. Pages 337-349 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University Press of Colorado, Boulder, Colorado. U. S. FISH AND WILDLIFE SERVICE. 2000. Endangered and threatened wildlife and plants: final rule to list the contiguous United States distinct population segment of the Canada lynx as a threatened species. Federal Register 65, Number 58. WHITE, G.C. AND K. P. BURNHAM. 1999. Program MARK: Survival estimation from populations of marked animals. Bird Study 46 Supplement, 120-138. WILD, M. A. 1999. Lynx veterinary services and diagnostics. Job Progress Report for the Colorado Division of Wildlife. Fort Collins, Colorado. Prepared by ___________________________________ Tanya M. Shenk, Wildlife Researcher Table 1. Lynx released in Colorado from February 1999 through June 30, 2007. No lynx were released in 2001, 2002 or 2007. Year Females Males TOTAL 1999 22 19 41 2000 35 20 55 2003 17 16 33 2004 17 20 37 2005 18 20 38 2006 6 8 14 TOTAL 115 103 218 20 �Table 2. All individual lynx (n = 60) documented through either aerial or satellite locations (nontruncated datasets) by year in New Mexico from February 1999 – March 2007. Lynx ID AK99F10 AK99F13 AK99F17 AK99F3 AK99F5 AK99F8 AK99M11 AK99M26 AK99M9 BC99M4 YK99F3 YK99M3 YK99M6 YK99M7 AK00F2 AK00F5 BC00F10 BC00F14 BC00F6 BC00F8 BC00M04 BC00M11 BC00M4 YK00F11 YK00F2 YK00F4 YK00F7 BC03F03 BC03F04 BC03F06 BC03F08 BC03M02 BC03M05 BC03M08 QU03F01 QU03F04 QU03F07 BC04F02 BC04F03 BC04F05 BC04M02 BC04M13 QU04F05 QU04F08 QU04F09 QU04M02 QU04M04 BC05F04 BC05M02 QU05F03 QU05M01 QU05M05 YK05F01 YK05M01 BC06F05 BC06F07 BC06F09 BC06M12 YK06F01 YK06M01 Total Lynx 1999 X 2000 2001 2002 Year 2003 2004 2005 2006 2007 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X 8 11 2 3 9 21 17 17 X X X X X X X X 14 X 2 �Table 3. All individual lynx (n = 35) documented at least 180 days after their initial release (truncated datasets) through either aerial or satellite locations, by year in New Mexico from September 1999 – March 2007. Lynx ID AK99F13 AK99F3 AK99F5 AK99M11 AK99M9 BC99M4 YK99F3 YK99M6 AK00F2 AK00F5 BC00F14 BC00F8 BC00M04 BC00M11 BC00M4 YK00F11 YK00F2 YK00F4 YK00F7 BC03F03 BC03F06 BC03M02 BC03M08 QU03F04 QU03F07 BC04F02 BC04M13 QU04F05 QU04F08 QU04F09 QU05M05 YK05M01 BC06F07 BC06M12 YK06F01 Total Lynx 2000 X X 2001 2002 2003 Year 2004 X 2005 2006 2007 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X 6 2 3 2 12 22 X X X X X X 10 X X X X X X 11 X 2 �Table 4. All individual lynx (n = 22) documented through either aerial or satellite locations (nontruncated datasets) by year in Utah from February 1999 – March 2007. Lynx ID AK99F5 AK00F5 AK00M3 BC00M09 BC00M13 YK00F7 BC03F03 BC03M06 BC03M08 BC03M10 QU03F03 BC04M01 QU04M04 QU04M05 BC05M01 BC05M03 CO05F20 QU05F05 QU05M03 QU05M08 YK05M01 YK06M01 Total Lynx 2000 2001 2002 2003 Year 2004 2005 X 2006 2007 X X X X X X X X X X X X X X X X X X X X X 1 1 0 2 4 X X X X X 7 7 5 Table 5. All individual lynx (n = 17) documented at least 180 days after their initial release (truncated datasets) through either aerial or satellite locations, by year in Utah from September 1999 – March 2007. Year Lynx ID AK99F5 AK00F5 AK00M3 BC00M09 YK00F7 BC03M06 BC03M10 QU03F03 BC04M01 QU04M04 BC05M01 BC05M03 CO05F20 QU05F05 QU05M03 YK05M01 YK06M01 Total Lynx 2000 2001 2002 2003 2004 2005 X 2006 2007 X X X X X X X X X X X X X X X X 0 1 0 0 23 3 6 X X X X X 7 5 �Table 6. All individual lynx (n = 33) documented through either aerial or satellite locations (nontruncated datasets) by year in Wyoming from February 1999 – March 2007. Lynx ID AK99M6 BC00F14 BC00M13 YK00F11 BC03F03 BC03M02 BC03M06 BC03M09 QU03M01 BC04F02 BC04M01 BC04M08 BC04M13 CO04F10 CO04M05 CO04M06 QU04F01 QU04F02 QU04F07 QU04M04 QU04M05 BC05M03 BC05M08 MB05F01 MB05F02 MB05F03 QU05F04 QU05F05 QU05F08 QU05M08 YK05M03 BC06M10 BC06M13 Total Lynx 1999 X 2001 2003 X X Year 2004 2005 2006 2007 X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X X 1 1 2 16 24 14 X X X X X X X X X 16 X X X X X 5 �Table 7. All individual lynx (n = 27) documented at least 180 days after their initial release (truncated datasets) through aerial or satellite locations, by year in Wyoming from September 1999 – March 2007. Year Lynx ID BC00F14 BC00M13 YK00F11 BC03F03 BC03M02 BC03M06 BC03M09 QU03M01 BC04F02 BC04M08 BC04M13 CO04F10 CO04M05 CO04M06 QU04F01 QU04F02 QU04M04 QU04M05 BC05M03 MB05F01 MB05F02 MB05F03 QU05F04 QU05F05 QU05F08 QU05M08 BC06M13 Total Lynx 2001 2003 2004 X X X X 2005 2006 2007 X X X X X X X X X X X X X X X X X X X X X X X X X 1 2 14 11 X X X X X X X X X X X X X X 15 X X X X X 5 Table 8. Status of adult lynx reintroduced to Colorado as of June 30, 2007. Lynx Released Known Dead Possible Alive Missing Monitoring/tracking a 1 is unknown mortality Females 115 54 61 23 38 Males 103 43 60 27 33 Unknown 1 TOTALS 218 98 120 49a 71 Table 9. Causes of death for all lynx released into southwestern Colorado 1999-2006 as of June30, 2007. Cause of Death Unknown Gunshot Hit by Vehicle Starvation Other Trauma Plague Probable Gunshot Predation Probable Predation Illness Total Mortalities Mortalities In Colorado (%) 20 (57.1) 7 (53.8) 8 (66.7) 9 (90.0) 7 (87.5) 7 (100) 4 (80) 3 (100) 3 (100) 2 (100) 70 (71.4) Total (%) 35 (35.7) 13 (13.3) 12 (12.2) 10 (10.2) 8 (8.1) 7 (7.1) 5 (5.1) 3 (3.1) 3 (3.1) 2 (2.0) 98 25 Outside Colorado (%) 15 (42.9) 6 (46.2) 4 (33.3) 1 (10.0) 1 (12.5) 0 (0) 1 (20) 0 (0) 0 (0) 0 (0) 28 (28.6) �Table 10. Known lynx mortalities (n = 28) and causes of death documented by state outside of Colorado from February 1999 – June 30, 2007. Lynx ID AK99F8 Unknown AK99M11 YK99M06 AK99F13 YK00F04 BC99M04 QU05M01 QU04F05 QU03F07 BC00M04 YK06F01 BC03M08 BC06F07 AK99M06 AK99M01 QU05M08 MB05F02 BC00F14 QU04F07 BC06M10 QU04F02 AK00M03 QU05M03 YK06M01 YK99F01 YK00M03 YK05M03 State Date Mortality Recorded Cause of Death New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico Nebraska Nebraska Nebraska Nebraska Wyoming Wyoming Wyoming Wyoming Utah Utah Utah Arizona Kansas Montana 7/30/1999 2000 1/27/2000 6/19/2000 6/22/2000 4/20/2001 6/7/2002 8/22/2005 8/26/2005 9/15/2005 7/19/2006 10/19/2006 10/19/2006 1/8/2007 11/16/1999 1/11/2005 10/1/2006 2/13/2007 7/28/2004 9/21/2004 8/15/2006 3/14/2007 7/2/2001 10/26/2005 12/4/2006 9/15/2005 9/30/2005 11/8/2005 Starvation Hit by Vehicle Unknown Probable Gunshot Unknown Gunshot Gunshot Unknown Hit by Vehicle Unknown Unknown Unknown Unknown Gunshot Gunshot Snared (Other Trauma) Unknown Gunshot Unknown Unknown Vehicle Collision Unknown Unknown Unknown Unknown Gunshot Vehicle Collision Unknown Table 11. Lynx reproduction summary statistics for 2003-2007. No reproduction was documented from 1999-2002 or in 2007. Year Females Tracked Dens Found in May/June 2003 2004 2005 2006 2007 Total 17 26 40 42 34 6 11 17 4 0 Percent Tracked Females with Kittens 0.353 0.462 0.425 0.095 0.0 Additional Litters Found in Winter 2 1 26 Mean Kittens Per Litter (SE) 2.67 (0.33) 2.83 (0.24) 2.88 (0.18) 2.75 (0.47) Total Kittens Found 16 39 50 11 0 116 Sex Ratio M/F (SE) 1.0 1.5 0.8 1.2 1.14 (0.14) �Table 12. Lynx captured because they were in poor body condition or were in atypical habitat and their fates 6 months post re-release and as of June 30, 2007. Lynx ID BC99F6 Date of Capture 3/25/1999 State Where Captured Colorado Reason For Capture Poor body condition Date of Re-release 5/28/1999 Status 6 Months Post Re-release Dead AK99M9 3/24/2000 Colorado 5/3/2000 Missing AK99F2 4/18/2000 Colorado 5/22/2000 BC00F7 2/11/2001 Colorado Alive in Colorado Dead BC00M13 3/21/2001 Wyoming BC03M08 9/5/2003 Colorado QU04M07 2/2/2006 Colorado Poor body condition Poor body condition Poor body condition Poor body condition Poor body condition Poor body condition BC04M01 11/5/2004 Utah QU04F02 4/10/2005 Nebraska QU05M08 11/25/2005 Wyoming QU04M04 12/5/2006 Utah YK00F7 12/12/2006 Utah YK05M02 1/1/2007 Kansas BC04M08 1/22/2007 Wyoming N/A 4/24/2001 1/1/2004 N/A Alive in Colorado Alive in Colorado Dead Died 7/19/1999 in Colorado from vehicle collision Last located 5/3/2000, collar failure Last located 7/30/2003 in Colorado Died at Rehab Center on 2/12/2001 Last located 10/26/2004 in Colorado Died in New Mexico of unknown causes 10/19/06 Died at Rehab Center on 2/5/2006 from hydrocephalous and pneumonia Atypical habitat Atypical habitat 12/5/2004 Atypical habitat Atypical habitat Atypical habitat Atypical habitat Atypical habitat 4/18/2006 Dead 1/20/2007 1/20/2007 Dead in Colorado Alive in Utah Died 3/14/2007 in Wyoming (good habitat) of unknown causes Died of unknown causes in Nebraska 10/1/2006 Died of starvation in Colorado, found 3/19/07 In Utah as of 6/30/2007 2/2/2007 Alive in Iowa In Iowa as of 6/30/2007 2/15/2007 Alive in Colorado In Colorado as of 6/30/2007 5/7/2005 Alive in Colorado Alive in Wyoming Current Status In Colorado as of 6/30/2007 Table 13. Number of kills found each winter field season through snow-tracking of lynx and percent composition of kills of the three primary prey species. Field Season 1999 1999-2000 2000-2001 2001-2002 2002-2003 2003-2004 2004-2005 2005-2006 2006-2007 n 9 83 89 54 65 37 78 50 41 Snowshoe Hare 55.56 67.47 67.42 90.74 90.77 67.57 83.33 90.00 61.00 Prey (%) Cottontail Red Squirrel 22.22 0 19.28 1.20 19.10 8.99 5.56 0 6.15 0 27.03 2.70 10.26 0 0.08 0 39.0 0 27 Other 22.22 12.05 4.49 3.70 3.08 2.70 6.41 0.02 0 �Figure 1. Lynx are monitored throughout Colorado and by satellite throughout the western United States. The lynx core release area, where all lynx were released, is located in southwestern Colorado. A lynx-established core use area has developed in the Taylor Park and Collegiate Peak area in central Colorado. 28 �Figure 2. All documented lynx locations (non-truncated datasets) obtained from either aerial (yellow circles) or satellite (red circles) tracking from February 1999 through June 30, 2007. All known lynx mortality locations (n = 97) are displayed as stars. 29 �Figure 3. Use-density surface for lynx aerial locations (truncated dataset) in Colorado from September 1999-March 2007 30 �Figure 4. Use-density surface for lynx satellite locations (truncated dataset) in Colorado from September 1999-March 2007. 31 �Figure 5. Use-density surface for lynx satellite locations (truncated dataset) in New Mexico from September 1999-March 2007 32 �Figure 6. Use-density surface for lynx satellite locations (truncated dataset) in Colorado and New Mexico from September 1999-March 2007. 33 �Figure 7. Use-density surface for lynx satellite locations (truncated dataset) for Utah from September 1999-March 2007. 34 �Figure 8. Use-density surface for lynx satellite locations (truncated dataset) in Colorado and Utah from September 1999-March 2007. 35 �Figure 9. Use-density surface for lynx satellite locations (truncated dataset) in Wyoming from September 1999-March 2007. 36 �Figure 10. Use-density surface for lynx satellite locations (truncated dataset) in Colorado and Wyoming from September 1999-March 2007. 37 �APPENDIX I Colorado Division of Wildlife July 2006 - June 2007 WILDLIFE RESEARCH REPORT State of Cost Center Work Package Task No. Colorado 3430 0670 2 Federal Aid Project: N/A : Division of Wildlife : Mammals Research : Lynx Conservation : Density, Demography, and Seasonal Movements of Snowshoe Hare in Colorado : Period Covered: July 1, 2006- June 30, 2007 Author: J. S. Ivan, Ph. D. Candidate, Colorado State University Personnel: T. M. Shenk, CDOW and G. C. White of Colorado State University. All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT A program to reintroduce the threatened Canada lynx (Lynx canadensis) into Colorado was initiated in 1997 with the first lynx release in 1999. Analysis of scat collected from winter snow tracking indicated that snowshoe hares (Lepus americanus) comprised 65–90% of the winter diet of reintroduced lynx. Thus, existence of lynx in Colorado and success of the reintroduction effort hinge at least in part on maintaining adequate and widespread populations of hares. Beginning in July 2006, I initiated a study to assess the relative value of 3 forest stand types (mature [“large”] spruce/fir, sapling [“small”] lodgepole pine, pole-sized [“medium”] lodgepole pine) that purportedly provide high quality hare habitat in Colorado. Estimates and comparisons of survival, recruitment, finite population growth rate, and maximum (late summer) and minimum (late winter) snowshoe hare densities for each stand will provide the metrics for assessing value. Number of individuals captured, number of captures, and number of locations obtained per hare during the first year of the project appear adequate for attaining the objectives of this study. Some hare deaths due to capture myopathy (most likely cause) occurred during initial trapping periods in both the summer and winter sampling seasons. However, changes to the trapping protocol, trapping schedule, and bait provided seem to have alleviated the problem. Densities during summer were highest in small lodgepole stands (0.47 hares/ha, 95% CI: 0.41-0.54), followed by large spruce/fir (0.18 hares/ha, 95% CI: 0.12-0.25) and medium lodgepole (0.02 hares/ha, 95% CI: 0.01-0.03). During winter, densities in small lodgepole stands dropped and became more variable across replicates (0.18 hares/ha, 95% CI: 0.01-0.35). Medium lodgepole stands gained hares (0.07 hares/ha, 95% CI: 0.050.10). Spruce/fir stands remained at the same density as during summer (0.17 hares/ha, 95% CI: 0.110.23). 38 �WILDIFE RESEARCH REPORT DENSITY, DEMOGRAPHY, AND SEASONAL MOVEMENTS OF SNOWSHOE HARE IN COLORADO JACOB S. IVAN P. N. OBJECTIVE Assess the relative value of 3 forest stand types (old spruce/fir, sapling lodgepole, pole-sized lodgepole) that purportedly provide high quality snowshoe hare (Lepus americanus) habitat by estimating survival, recruitment, finite population growth rate, and maximum (late summer) and minimum (late winter) snowshoe hare densities for each type. SEGMENT OBJECTIVES 1. Complete mark-recapture work across all replicate stands during late summer (mid-July through midSeptember) and winter (mid-January through March). 2. Obtain daily telemetry locations on radio-tagged hares for 10 days immediately after capture periods, as well as monthly between primary trapping sessions. 3. Locate, retrieve, and refurbish radio tags as mortalities occur. 4. Summarize initial sampling efforts and provide initial density estimates for Wildlife Research Reports for Colorado Division of Wildlife (CDOW). INTRODUCTION A program to reintroduce the threatened Canada lynx (Lynx canadensis) into Colorado was initiated in 1997 with the first lynx released in 1999. Since that time, 204 lynx have been released in the state, and an extensive effort to determine their movements, habitat use, reproductive success, and food habits has ensued (Shenk 2006). Analysis of scat collected from winter snow tracking indicates that snowshoe hares (Lepus americanus) comprise 65–90% of the winter diet of reintroduced lynx (T. Shenk, Colorado Division of Wildlife, unpublished data). Thus, as in the far north where the intimate relationship between lynx and snowshoe hares has captured the attention of ecologists for decades, the existence of lynx in Colorado and success of the reintroduction effort may also hinge on maintaining adequate and widespread populations of hares. Colorado represents the extreme southern range limit for both lynx and snowshoe hares (Hodges 2000). At this latitude, habitat for each species is less widespread and more fragmented compared to the continuous expanse of boreal forest at the heart of lynx and hare ranges. Neither species exhibits dramatic cycles as occur farther north, and typical lynx (≤2−3 lynx/100km2; Aubry et al. 2000) and hare (≤1−2 hares/ha; Hodges 2000) densities in the southern part of their range correspond to cyclic lows form northern populations (2-30 lynx/100 km2, 1−16 hares/ha; Aubry et al. 2000, Hodges 2000, Hodges et al. 2001). Whereas extensive research on lynx-hare ecology has occurred in the boreal forests of Canada, literature regarding the ecology of these species in the southern portion of their range is relatively sparse. This scientific uncertainty is acknowledged in the “Canada Lynx Conservation Assessment and Strategy,” a formal agreement between federal agencies intended to provide a consistent approach to lynx conservation on public lands in the lower 48 states (Ruediger et al. 2000). In fact, one of the explicit guiding principles of this document is to “retain future options…until more conclusive information concerning lynx management is developed.” Thus, management recommendations in this agreement are 39 �decidedly conservative, especially with respect to timber management, and are applied broadly to cover all habitats thought to be of possible value to lynx and hare. This has caused controversy where recommendations conflict with competing resource management goals. Accurate identification and detailed description of lynx-hare habitat in the southern Rocky Mountains would permit more informed and refined management recommendations. A commonality throughout the snowshoe hare literature, regardless of geographic location, is that hares are associated with dense understory vegetation that provides both browse and protection from elements and predators (Wolfe et al. 1982, Litvaitis et al. 1985, Hodges 2000, Homyack et al. 2003, Miller 2005). In western mountains, this understory can be provided by relatively young conifer stands regenerating after stand-replacing fires or timber harvest (Sullivan and Sullivan 1988, Koehler 1990, Koehler 1990, Bull et al. 2005) as well as mature, uneven-aged stands (Beauvais 1997, Griffin 2004). Hares may also take advantage of seasonally abundant browse and cover provided by deciduous, open habitats (e.g., riparian willow [Salix spp.], aspen [Populus tremuloides]; Wolff 1980, Miller 2005). In drier portions of hare range, such as Colorado, regenerating stands can be relatively sparse, and hares may be more associated with mesic, late-seral forest and/or riparian areas than with young stands (Ruggiero et al. 2000). Numerous investigators have sought to determine the relative importance of these distinctly different habitat types with regards to snowshoe hare ecology. Most previous evaluations were based on hare density or abundance (Bull et al. 2005), indices to hare density and abundance (Wolfe et al. 1982, Koehler 1990, Beauvais 1997, Miller 2005), survival (Bull et al. 2005), and/or habitat use (Dolbeer and Clark 1975). Each of these approaches provides insight into hare ecology, but taken singly, none provide a complete picture and may even be misleading. For example, extensive use of a particular habitat type may not accurately reflect the fitness it imparts on individuals, and density can be high even in “sink” habitats (Van Horne 1983). A more informative approach would be to measure density, survival, and habitat use simultaneously in addition to recruitment and population growth rate through time. Griffin (2004) employed such an approach and found that summer hare densities were consistently highest in young, dense stands. However, he also noted that only dense mature stands held as many hares in winter as in summer. Furthermore hare survival seemed to be higher in dense mature stands, and only dense mature stands were predicted (by matrix projection) to impart a mean positive population growth rate on hares. Griffin’s (2004) study occurred in the relatively moist forests of Montana, which share many similarities but also many notable differences with Colorado forests including levels of fragmentation, species composition, elevation, and annual precipitation. Density estimation is a key component in assessing the value of a particular stand type and is the common currency by which hare populations are compared across time and space. However, density can be a difficult metric to estimate accurately. Density estimation based on capture-recapture methods is a well-developed field (Otis et al. 1978, White et al. 1982), but is often too costly and labor intensive to be implemented on scales necessary to effectively monitor density over a biologically meaningful area. Also, density can be difficult to assess from grid-trapping efforts because it is often unclear how much area was effectively sampled by the grid (Williams et al. 2002:314). Different approaches can produce density estimates that differ by an order of magnitude even when calculated from the same data (Zahratka 2004). Indices such as pellet plot counts and distance sampling of pellet groups can be used to estimate density, but each of these has limitations as well (Krebs et al. 1987, Eriksson 2006). Pellet plot counts are typically conducted by laying out numerous rectangular or circular plots along transect lines randomly placed within a study site. All pellets occurring within the plot are counted and removed on an annual basis. The mean number of pellets per plot is then inserted into a regression equation that gives an estimate of hare density (Krebs et al. 1987). Estimates from this technique correlate well with density estimates derived from simultaneous mark-recapture studies occurring in the 40 �same area (Krebs et al. 2001, Murray et al. 2002, Mills et al. 2005, Homyack et al. 2006). However, because fecal deposition rates can vary by season and diet, and because pellet decomposition rates can vary with altitude, climate, aspect, precipitation, and cover type, region-specific, stand-specific, and/or season-specific equations should be developed before this technique is employed for a given area and season (Krebs et al. 2001, Prugh and Krebs 2004, Murray et al. 2005). Density estimates vary with plot size and shape, requiring equations specific to these geometric considerations as well (McKelvey et al. 2002). Pellet counts tend to yield more precise and unbiased density estimates when plots are visited and cleared more than once per year (e.g., plots cleared in the fall and then counted in the spring to estimate winter density) because variability in deposition and decomposition rates is reduced (Homyack et al. 2006). However, this requires considerably more work and expense than an annual survey. Some studies have conducted pellet plot counts without first clearing plots (e.g., Bartmann and Byrne 2001). This saves time and money, but requires the ability to discern fresh (this year) pellets from old pellets, which can be difficult and is generally not a recommended approach (Prugh and Krebs 2004, Murray et al. 2005). Distance sampling is a well-developed method for estimating the density of objects in a given area (Buckland et al. 2001). In general, observers walk a pre-defined sampling transect and record each object of interest along with the perpendicular distance of that object from the transect line. This information is then used to develop a detection function which is in turn used to estimate density (Buckland et al. 2001). The method assumes all objects on the line are seen with certainty, objects are not double-counted, distance measures are accurate, and transect lines are located randomly within a study area (Buckland et al. 2001). Recently, distance sampling has been used to indirectly estimate hare density by first estimating the pellet group density of hares, then using fecal deposition and decomposition rates as a link back to hare density (Eriksson 2006). In general, distance sampling is more efficient than pellet plot counts as it does not require the tedious layout of hundreds of plots or counting individual pellets. This advantage is most recognizable in situations where pellet groups occur at low densities. Conversely, at extremely high densities, it may become difficult to distinguish pellet groups, and plots may be preferable (Marques et al. 2001). Regardless, distance sampling of pellet groups to estimate animal density also requires habitat and season specific decomposition and defecation rates, which can be difficult to obtain (Marques et al. 2001). For this project, I have chosen to provide land managers with information relating demographic rates, as well as density, to forest stand characteristics. Thus, I will use mark-recapture techniques as data from such an approach can provide information on both density and demography. I will address the “effective trapping area” issue using a new approach that augments mark-recapture data with telemetry locations of animals using the grid. The study outlined below is designed principally to evaluate the importance of young, regenerating lodgepole pine (Pinus contorta) and mature Engelmann spruce (Picea engelmannii)/ subalpine fir (Abies lasiocarpa) stands in Colorado by examining density and demography of snowshoe hares that reside in each (Figure 1). My hope is that information gathered from this research will be drawn upon as managers make routine decisions, leading to landscapes that include stands capable of supporting abundant populations of hares. I assume that if management agencies focus on providing habitat, hares will persist. Specifically, I will evaluate small and medium lodgepole pine stands and large spruce/fir stands where the classes “small”, “medium”, and “large” refer to the diameter at breast height (dbh) of overstory trees as defined in the United States Forest Service R2VEG Database (small = 2.54−12.69 cm dbh, medium = 12.70−22.85 cm, and large = 22.86−40.64 cm dbh; J. Varner, United States Forest Service, personal communication). I also intend to identify which of the numerous density-estimation procedures available perform accurately and consistently using an innovative, telemetry augmentation approach as a 41 �baseline. I will assess movement patterns and seasonal use of deciduous cover types such as riparian willow. Finally, I will further expound on the relationship between density, demography, and stand type by examining how snowshoe hare density and demographic rates vary with specific vegetation, physical, and landscape characteristics of a stand. Hypotheses 1) In general, snowshoe hare density in Colorado will be relatively low (≤0.5 hares/ha) compared to densities reported in northern boreal forests, even immediately post-breeding when an influx of juveniles will bolster hare numbers. 2) Snowshoe hare density will be consistently highest in small lodgepole pine stands, followed by large spruce/fir and medium lodgepole pine, respectively. 3) Survival will generally be highest in mature (large) spruce/fir stands followed by small and medium lodgepole pine, respectively. 4) Finite population growth rate will be consistently at or above 1.0 in mature spruce/fir stands with survival contributing most significantly to the growth rate. Finite growth rates for the lodgepole pine stands will be more variable. 5) Snowshoe hares will significantly shift their home ranges to make use of abundant food and cover provided by riparian willow (and/or aspen) habitats in summer. 6) Snowshoe hare density, survival, and recruitment will be highly correlated with understory cover and stem density. STUDY AREA The study area stretches from Taylor Park to Pitkin in central Colorado (Figure 2). Elevation ranges from 2700 m to 4000 m. Sagebrush (Artemisia spp.) dominates broad, low-lying valleys. Most montane areas are covered by even-aged, large-diameter lodgepole pine forests with sparse understory. Moist, north-facing slopes and areas near tree line are dominated by large-diameter Engelmann spruce/subalpine fir. Interspersed along streams and rivers are corridors of willow. Patches of aspen occur sporadically on southern exposures. This area was chosen over other potential study areas in the state because 1) it contained numerous examples of the 3 stand types of interest (more southern regions lack naturally occurring stands of lodgepole pine), 2) it was not subject to confounding effects of largescale mountain pine beetle outbreak as were more northern stands, and 3) an adequate number of radio frequencies were available to support a large study with hundreds of radio-tagged individuals. Within the study area I selected sample stands based on the following: Potential replicate stands were required to be 1) close enough geographically to minimize differences due to climate, weather, and topography, but are far enough apart to be considered independent, 2) adjacent to one or more riparian willow corridors, 3) within 1 km of an access road for logistical purposes, 4) of suitable size and shape to admit a 16.5-ha trapping grid, and 5) consistent in their management history (i.e., replicate lodgepole pine stands were clear-cut and/or thinned within 1-2 years of each other). I queried the U.S. Forest Service R2VEG GIS database using the criteria listed above to initially develop a suite of potential sample stands. I further narrowed this suite after obtaining updated standlevel information from local USFS personnel (Art Haines, Silviculturalist, USFS Gunnison Ranger District, personal communication). Finally, I ground-truthed potential stands and qualitatively assessed their representativeness and similarity to other potential replicates. Given the numerous constraints imposed, very few stands met all criteria. Thus, I was unable to randomly select sample stands from a population of suitable stands. Rather, I subjectively chose the “best” stands from among the handful that met my criteria. Small lodgepole stands rarely occur on the landscape in patches large enough to fit a full 42 �7 x 12 trapping grid. To accommodate this, I sampled 6 replicate small lodgepole stands (rather than 3) using 6 x 7 trapping grids (1/2 size). METHODS Experimental Design/Procedures Variables.--The response variables of interest for this project include stand-specific snowshoe hare density (D), apparent survival (φ), recruitment (f), finite population growth rate (λ), and a metric of seasonal movement. Density is the number of hares per unit area and will be estimated using a variety of conventional techniques as well as a rigorous method that incorporates radio telemetry. The standspecific demographic parameters will be estimated primarily from capture-mark-recapture methods. As such, apparent survival is defined as the probability that a marked animal alive and in the population at time i survives and is in the population at time i + 1. Apparent survival encompasses losses due to both death and emigration. Recruitment is the number of new animals in the population at time i + 1 per animal in the population at time i. New recruits can arise from on-site reproduction as well as immigration. The finite population growth rate is the number of animals in a given age class at time i + 1 divided by the number present at time i. Shifts in home range will be assessed by comparing the seasonal proportion of telemetry locations in deciduous habitats using multi-response permutation procedures (MRPP; Zimmerman et al. 1985, White and Garrott 1990). Potential explanatory variables for snowshoe hare density, demographics, and movement include general species composition and structural stage of each stand in which response variables are measured. Additionally, stem density, horizontal cover, and canopy cover (to a lesser extent) are highly correlated with snowshoe hare abundance and habitat use (Wolfe et al. 1982, Litvaitis et al. 1985, Hodges 2000, Zahratka 2004, Miller 2005). Thus, I will further characterize vegetation in each stand by measuring stem density by size class (1-7 cm, 7.1-10 cm, and >10 cm), percent canopy cover, percent horizontal cover of understory and basal area. Basal area is an easily obtainable metric that may be correlated with the other variables and is recorded routinely during timber cruises, whereas the others are not. Thus, it might prove a useful link for biologists designing management strategies for snowshoe hare. Additionally, I will record physical covariates such as ambient temperature, precipitation, and snow depth at each stand during sampling periods as well as precipitation 1-3 years prior to sampling. Finally, I will calculate potentially important landscape metrics such as patch size and level of fragmentation. Sampling.--All trapping and handling procedures have been approved by the Colorado State University Animal Care and Use Committee and filed with the Colorado Division of Wildlife. Snowshoe hares breed synchronously and generally exhibit 2 birth pulses in Colorado (although in some years, some individuals may have 3 litters), with the first pulse terminating approximately June 5−20 and the second approximately July 15–25 (Dolbeer 1972). To obtain a maximum density estimate, I began data collection on the first suite of sites immediately following the second birth pulse in late July. Along with a crew of 5 technicians, I deployed one 7 × 12 trapping grid (50-m spacing between traps; grid covers 16.5 ha) in the large spruce/fir and medium lodgepole stands within the first suite, along with 2 6 × 7 grids in 2 small lodgepole stands. Grid set up and trap deployment followed Griffin (2004) and Zahratka (2004). Grid locations and orientation within each stand were chosen subjectively to accommodate logistical constraints and to ensure that hares using the grid had ample opportunity to use adjacent riparian willow zones. Traps were deployed in all 4 stands in a single day. As traps are deployed, they were locked open and “pre-baited” with apple slices and commercial rabbit chow. During winter, hay cubes were added to traps as well (see Discussion). On days 2-4, the crew continued pre-baiting, replacing apples and rabbit chow as necessary. The purpose of this extended pre-baiting was to maximize capture rates when trapping began. This minimized the number of trap-nights needed to capture the desired number of animals which in turn minimized trapping-related stress as well as the likelihood that 43 �American marten (Martes americana) keyed into trap lines and preyed on entrapped hares, as has occurred in previous studies (J. Zahratka, personal communication). During pilot work in winter 2005, I observed low but increasing capture rates (<0.20) during the first 3 nights of trapping, with higher, more stable capture probabilities after 3 days (approximately 0.35–0.45). Thus 3 days of pre-baiting seems reasonable. Traps were set on the afternoon of the 4th day and checked early each morning and again in the evening on days 5–9. By checking traps in both morning and evening I prevent hares from being entrapped >13 hours, which should minimize capture stress. A crew of 2 people worked together on each grid to check traps and process captures as quickly as possible. All captured hares were coaxed out of the trap and into a dark handling bag by blowing quick shots of air on them from behind. Hares remained in the handling bag, physically restrained with their eyes covered, for the entire handling process. Each individual was aged, sexed, marked with a passive integrated transponder (PIT) tag and temporary ear mark (to track PIT tag retention), then released. Aging consisted of assigning each individual as either juvenile (<1 year old, <1000 g) or adult (≥1 year old, ≥1000 g) based on weight. This criterion is accurate through the end of September at which point juveniles are difficult to distinguish from adults (K. Hodges, University of British Columbia; P. Griffin, University of Montana, personal communication). After the first day of trapping, all captured hares were scanned for a PIT tag prior to any handling and those already marked were recorded and immediately released. Traps and bait were completely removed from the grid on day 10. In addition to PIT tags and ear marks, I radio collared up to 10 hares captured on each grid with a 28-g mortality-sensing transmitter (BioTrack, LTD) to facilitate unbiased density estimation as well as assessment of seasonal movements. I expected heterogeneity in snowshoe hare movements and use of the grid area, with potential bias surfacing due to location at which a hare is captured (e.g., hares captured on the edge of a grid may use the grid area differently than those captured at the center), and differential behavioral responses to trapping (e.g., young individuals may have lower capture probabilities and thus may be more likely to be captured on later occasions). To guard against the first potential bias, I randomly selected a starting trap location each morning and ran the grid systematically from that point. Thus, the first several hares encountered (and collared) were as likely to be from the inner part of the grid as from the edge. To protect against the second potential source of bias, I refrained from deploying the final 3 collars until days 4 and 5 of the trapping session. Immediately following the removal of traps, the field crew began work locating each radiocollared hare 1–2 times per day for 10 days. Most locations were obtained by triangulation from relatively close proximity, but some were obtained by “homing” on a signal (Samuel and Fuller 1996, Griffin 2004) taking care not to push hares while approaching them. Because hares are largely nocturnal (Keith 1964, Mech et al. 1966, Foresman and Pearson 1999), I made an effort to conduct telemetry work at various times of the night (safety and logistics permitting) and day to gather a representative sample of locations for each hare. The crew gathered telemetry locations for radio-collared hares on the initial sites for 8 to 10 days. Then the 10−day trapping procedure and 8 to 10−day telemetry work were repeated on the 3 grids comprising suite 2 (Figure 3). The cycle was repeated once more for grids in suite 3 (Figure 3). The entire process was repeated during the winter when densities should have been at a minimum. In summary, for any given 9-week sampling period, I collected data from 12 total grids, 1 spruce/fir, 1 medium lodgepole, and 2 small lodgepole across 3 replicates. Sampling will occur during 2 such 9-week periods each year − once in late summer and once in late winter – and will continue for 3 years. During the interim between intensive trapping and telemetry work, monthly telemetry checks were conducted from the air to track mortalities and facilitate retrieval of collars from dead hares. Telemetry 44 �work was also occur during “pre-baiting” days after the initial summer sampling session to determine which hares were still alive and immediately available to be sampled by the grid during the ensuing trapping period. Vegetation sampling at each stand will follow protocols established through previous snowshoe hare and lynx work in Colorado (Zahratka 2004, T. Shenk, Colorado Division of Wildlife, personal communication). Specifically, on each of the 12 live-trapping grids, I will lay out 5 × 5 grids (3-m spacing) of vegetation sampling points centered on 15 of the 84 trap locations (Figure 4; 9 points will be sampled on each of the ½-sized small lodgepole stands). At each of the 25 vegetation sampling points, I will record: 1) distance to the nearest woody stem 1.0−7.0 cm, 7.1−10.0 cm, and >10.0 cm in diameter at heights of 0.1 m and 1.0 m above the ground (to capture both summer [0.1 m] and winter [1.0 m] stem density; Barbour et al. 1999), 2) horizontal cover in 0.5-m increments above the ground up to 2 m (Nudds 1977), and 3) canopy cover [present or absent] using a densitometer. Additionally, at the center of all 15 vegetation sampling grid points (i.e., at the trap location), I will measure basal area using an angle gauge. These measurements will be gathered once at the start of the project, unless conditions change due to disturbance such as fire. Temperature will be monitored hourly at each grid during the 6-week intensive sampling periods using data loggers. During winter sampling periods, snow depth measurements will be recorded daily at the same 15 trap locations used to quantify the vegetative attributes of that stand. Data Analysis Density.--I assumed that hare populations were demographically and geographically closed during the short 5-day mark-recapture sampling periods. To obtain a density estimate for each grid, I used the Huggins closed capture model (Huggins 1989, 1991) in Program MARK (White and Burnham 1999) with some modifications. The basic Huggins estimator (no individual covariates) is based on the fact that if pj is the probability that a hare in the population is captured (and marked) for the first time on trapping occasion j, then p * = 1 − (1 − p1 )...(1 − p5 ) is the probability that an individual is captured at least once during a 5-day trapping period (i.e., j = 1,…,5). Accordingly, the basic Huggins estimator for population size, N̂ , is Nˆ = M t +1 / p* where M t +1 is the total number of hares captured. The estimator can be re-written to allow each of the M t +1 individuals captured to have their own p*. In that case, M t +1 Nˆ = ∑1 / pi* . Presumably hares that reside near the edge of a grid encounter fewer traps and are less i =1 likely to be captured than hares residing near the center of a grid. To account for this, I took advantage of the Huggins model with individual covariates to model p* by using the logit link function of program MARK to model pi* as a function of di, where di is distance from the edge of the grid for hare i based on mean capture coordinates. A naïve density estimate for each grid would then be Dˆ = Nˆ / A where A is the area of the grid. However, this gives full credit to all hares, even those whose home range only partially overlaps the grid, which results in a density estimate that is biased high. To correct for this bias, I determined the proportion, ( ~ pk ), of telemetry locations for each of the k = 1,…,10 radio-collared hares that fell within the “naïve grid area.” By incorporating data from multiple grids, a logistic regression model was developed to estimate p% i for all M t +1 animals captured on a grid based on distance from the edge of the grid for hare i (di). Replacing the numerator (i.e., 1) in the Huggins estimator with ( p% i ), gives ⎛ M t +1 ⎞ ~ p / p ⎟ A. ⎟ ⎜∑ a density estimate, Dˆ = ⎜ ⎝ i =1 i * i ⎠ The above-stated approach assumes that radio-collared hares neither gravitate toward nor avoid the former grid area after the 5 days of trapping, 10–20 locations per hare is enough to provide a 45 �reasonable representation of the proportion of time they spend on the grid, and their use of the grid area is representative of other hares that were captured but not collared (i.e., that the logistic regression model of p% i is a useful model). I contend that this type of estimate from grid-based trapping can be construed as a relatively unbiased estimate of density. Using these point estimates and their associated confidence intervals, I compared hare density among seasons and stand types. I will also compare these “true” density estimates to those that would have been obtained using other available methods such as ½ mean maximum distance moved (Wilson and Anderson 1985, Williams et al. 2002:314-315), full mean maximum distance moved (Parmenter et al. 2003), ½ trap interval (Parmenter et al. 2003), “nested grids” (White et al. 1982:120-131), and Program DENSITY (Efford et al. 2004). Demography.--I will analyze mark-recapture data using Pradel temporal symmetry models (Pradel 1996, Nichols and Hines 2002) in a robust design framework (Williams et al. 2002:523-554), which will be available in Program MARK by summer 2006. Thus, I will treat summer and winter sampling occasions as primary periods, and the 5-day trapping sessions within each as secondary periods. The Pradel temporal symmetry models employ both forward and reverse-time evaluation of capture histories to provide estimates of apparent survival ( φ̂ ) and seniority ( γ̂ ). Apparent survival, φi, is the probability that a marked animal alive and in the population at time i survives and is in the population at time i + 1. The seniority parameter, γi , is the reverse-time analogue of survival. Reading backward through a capture history, it is the probability that a marked animal alive and in the population at time i was alive and in the sampled population at time i − 1. If N is the number of animals present in the population, N i φi ≈ N i +1γ i +1 and N i +1 / N i = φi / γ i +1 = λ i . Also, if fi is recruitment rate, or the number of recruits at time i + 1 per animal present at time i, then N i +1 = N i φi + N i f i . Rearranging and substituting into the previous equation gives f i = φi (1/ γ i − 1) . Thus, using Pradel models, one can estimate recruitment and finite population growth rate in addition to survival (Pradel 1996, Nichols and Hines 2002). I will use Akaike’s Information Criterion corrected for small sample size (AICc; Burnham and Anderson 1998) to determine whether models with time-dependent parameters or constant parameters are best supported by the data. I will derive estimates of the above-mentioned parameters from the best model or from model averaging. I anticipate pooling capture data across sites to obtain φˆ i , λˆ i , and fˆi for each stand type for each interval between primary sampling periods (5 estimates of each). I also anticipate simply estimating these parameters for “generic hares”, treating both juveniles and adults as a single group or age class. Given that juveniles are morphometrically indistinguishable from adults by their first fall of life (K. Hodges, University of British Columbia; P. Griffin, University of Montana, personal communication), adult and juvenile survival rates are similar (Griffin 2004), and there is little evidence for age-specific differences in pregnancy rates or litter size (Dolbeer 1972), this approach seems justified. However, if I happen to capture sufficient numbers of juveniles and adults, the design I have laid out here allows for treating the age classes separately. This, in turn, may permit me to decompose the contribution that fi makes to λi into the portion of that contribution due to on-site reproduction and that due to immigration (Nichols et al. 2000). Similarly, it may also be possible using my telemetry data to decompose apparent survival, φi , into emigration and mortality. Such fortuitous situations would facilitate the identification of source and sink habitats if they exist. Seasonal Movements.--I will assess whether snowshoe hares seasonally shift their home ranges using the multi-response permutation procedure (MRPP; Zimmerman et al. 1985, White and Garrott 1990:134-135). Under this approach, telemetry locations are grouped by season (summer and winter), and an MRPP statistic is calculated as the weighted average of the distance between all possible pairs of locations within groups compared to the average distance between all possible pairs ignoring groups. The 46 �null hypothesis is that the distribution of locations is the same for both groups (seasons). Sufficiently small values of the test statistic suggest that within group distances are smaller than distances measured ignoring groups, which is evidence against the null in favor of a group (seasonal) effect. P-values are obtained by calculating the percentile of the observed test statistic relative to all possible test statistics that could be computed by re-arranging the data into all possible groups of 2. The MRPP procedure is sensitive and can detect even small changes in use of an area (White and Garrott 1990:136). I propose a priori that changes in proportional use of deciduous habitats <0.10 in magnitude are unlikely to be biologically significant. Vegetation.--I will calculate mean stem density, horizontal cover, canopy cover, and basal area for each season−stand type as well as temperature, precipitation, snow depth information, and landscape metrics. These will be entered into the MARK design matrix as covariates to population size (~density) and survival in a random effects analysis. As such, I will be able to quantify the amount of variation in population size or survival that is due to differences in vegetation, landscape, or weather relative to the amount left to other causes. Sample size.--I conducted power analyses to determine the probability of discerning meaningful differences in density and survival for hares occupying different stand types. For density, I postulated that foraging lynx likely do not discriminate among stands that differ by only a few hares. However, it seems probable that if hare density in one stand is twice that of another, a lynx would choose the former given the opportunity. Thus, I conducted power calculations to determine the probability of distinguishing differences in densities between 2 stand types in which one had twice the density of hares as the second. Specifically, using the Huggins closed capture model (Huggins 1989, Huggins 1991) in Program MARK, I specified the number of hares (N) present in each of 2 groups (i.e., 2 stand types), allowed capture (p) and recapture (c) probabilities to vary with time but constrained them to be equal and the same for each group, then simulated this scenario 1000 times for a range of realistic capture probabilities. For each simulation I calculated a 95% confidence interval for the mean difference in N̂ between the 2 groups and determined the proportion of all simulations in which this confidence interval did not include zero. This proportion is the power, or probability of discerning a difference between the 2 groups when one actually exists. I compared 2-fold differences in density at the low (5 vs. 10 hares/grid) and high (15 vs. 30 hares/grid) end of the range of hare numbers and I expect to observe (Zahratka 2004). I also simulated the power to detect differences between 17 and 39 hares/grid, corresponding to recently published cut-points for low and high hare densities in the context of lynx conservation (Mills et al. 2005). Given capture/recapture probabilities I observed during winter 2005 (approximately 0.35–0.45), I expect to have reasonable power to detect 2-fold differences in density even if I encounter relatively few hares per grid (Figure 5). I conducted power analyses for survival in a similar manner using the Huggins estimator (Huggins 1989, Huggins 1991) in a robust design framework (Williams et al. 2002:524-556). For this analysis, I specified 3 primary periods (e.g., 3 years) with 5 secondary occasions for each. I established either 30 or 45 hares in each of 2 groups (i.e., pooled an expected 10-15 hares/grid across the 3 grids in a given habitat type), specified a different survival rate for each, and allowed p and c to vary with time but constrained them to be equal and the same for each group as before. I then specified a general model that assumed survival rates varied among groups and a second, reduced model that assumed survival rates were the same for each group. After 1000 simulations under a given scenario of hare numbers, capture probabilities, and survival rates, I conducted a likelihood ratio test between each pair of general and reduced models. As before, I used the proportion of significant tests as an estimate of power to detect differences in survival. 47 �I compared survival rates of 0.4 vs. 0.5, 0.3 vs. 0.5, and 0.2 vs. 0.5. These rates span the range of annual hare survival rates reported in the literature (Dolbeer 1972, Dolbeer and Clark 1975, Griffin 2004). Also, because each comparison is anchored at 0.5, these calculations provide a conservative estimate of power due to the nature of binomial probabilities. That is, I would be more likely to distinguish the difference between 0.1 and 0.2 than between 0.4 and 0.5 even though the difference in both cases is 0.1 because the sampling variance of the estimate for the same sample size is maximal at 0.5 and declines to 0 for survival rates of 0 or 1. Results indicate that I have ≥80% chance of discerning real differences in survival of ≥0.3 (Figure 6), but only 40-65% chance (depending on number of hares captured) of detecting a difference of 0.2, and very little chance of detecting differences smaller than 0.2. However, I plan to combine my telemetry data with my trapping data in the MARK Robust design model using separate groups for each data type. This should enhance my precision and power, thus making the prospect of detecting differences as small as 0.2 a possibility. To complete a power analysis for λ̂ requires running simulations of Pradel models in a robust design framework. This capability is not yet available in Program MARK, so such an analysis has not been completed. Sampling 15 vegetation plots per trapping grid provided reasonably precise characterizations of similar stands in similar locations during a previous study (Zahratka 2004). I trust this level of sampling will be adequate for the present study as well. If not, more plots can be established at a later date given that vegetative characteristics are unlikely to change appreciably over a few years. RESULTS AND DISCUSSION Much of the analysis presented above is not possible or meaningful without several seasons of data, especially the survival, recruitment, and growth rate models. Below, I present a basic summary, relevant observations, and initial density estimates from the inaugural year of this project. I captured 75 hares 166 times during July-September 2006. I captured 99 hares 243 times during January-March 2007 (Table 1). Fourteen of these individuals were captured during both the summer and winter sampling sessions. During summer, I captured over twice as many individuals in small lodgepole stands as in spruce/fir. I captured only a few individuals in medium lodgepole stands. During winter, captures were more evenly distributed among the stands (Table 1). During the initial trapping session of the summer trapping period, 6 hares were captured, handled, and released (seemingly without harm) but were found dead in traps 1-3 days later. I collected the carcasses and submitted them for necropsy. Cause of death was attributed to capture myopathy, which is relatively common in lagomorphs (Laurie Baeten DVM, and Lisa Wolfe DVM, Colorado Division of Wildlife, personal communication). I subsequently altered my trapping protocol to further minimize both the amount of time a hare could be entrapped as well as the handling time at each capture. No trap deaths occurred during the remainder of the sampling season aside from 4 hares that succumbed to predation while inside traps. During the initial 2 trapping sessions of the winter trapping period, 6 more hares were captured, handled, and released multiple times, again with seemingly little adverse reaction, only to be found dead on a subsequent trapping occasion. Several more hares died during the 10-day telemetry session immediately following trapping. These “telemetry deaths” could have been due to natural causes, effects of capture, or a combination of both. Again, carcasses were submitted for necropsy, and again capture myopathy was cited as a potential cause of death. Further examination of the data indicated that hares trapped ≥3 days in a row were much more likely to die in a trap or during telemetry than other hares. Thus, I further modified the trapping protocol by locking traps open on day 3 of the 5-day trapping period so that hares could not be trapped more than 2 days in a row. Additionally, I began providing hay cubes 48 �in the traps as roughage to complement the high quality alfalfa pellets and apples. After implementing these changes, I did not observe any further trap-deaths or telemetry-deaths for the rest of the season. I averaged 9.9 and 6.3 locations per radio-tagged hare during the summer and winter sampling sessions, respectively (Table 2). Thus, “proportion of time on grid,” which is critical to my density estimation procedure, was based on relatively few points per individual for the first 2 sampling periods, and I was unable to attain my goal of 10-20 locations per individual. Following the winter field season, I conducted a series of simulations to examine the effects of sample size on precision of density estimates. I found that 1) the variability between hares (“proportion on grid” ranges from 0.00-1.00) overwhelms the variability within hares (i.e., the binomial variance for proportion of time on grid for any single individual, which decreases as number of locations increases), and 2) given a fixed effort, the variance of the density estimate is minimized by increasing the number of individuals collared as opposed to increasing the number of locations per individual. Thus, it is better to radio-tag more hares and get fewer locations than to tag fewer hares and get more locations. I will continue to deploy as many collars as possible, and will strive for 10-20 locations per individual, but the level of sampling achieved during the first 2 field seasons appears sufficient to detect the large differences in density that occur on the landscape. During summer, density estimates followed hypotheses 1) and 2) above. Specifically, hare density in small lodgepole stands was twice that observed in spruce/fir, which was more than twice that observed in medium lodgepole stands (Figure 7). However, even the relatively high density found in the small lodgepole stands was relatively low compared to densities that have been reported in other parts of hare range (Griffin 2004, Hodges 2000). However, different methods for computing density make this type of comparison difficult. During winter, hare densities remained the same in spruce/fir stands. Hare density in medium lodgpole stands more than doubled, although still remained relatively low compared to other stand types. Density in the small lodgepole stands dropped significantly compared to summer levels and was more variable among replicates. Hare density is likely driven by availability of food and cover. I submit that the interplay between food, cover, and snow depth provides a plausible explanation for the density patterns observed during the first year of this study. Spruce/fir stands probably provide adequate access to both necessities during both summer and winter due to their uneven-aged, multi-layered structure. Medium lodgepole stands, on the other hand, apparently provide very little forage/cover for hares during summer as the canopy in these stands is generally ≥1 meter off the ground. However, in winter, accumulated snow may bring that canopy back into reach for hares. Conversely, small lodgepole stands provided abundant food and cover during summer, but accumulated snow during winter brings hares closer to the crowns of the young trees, which then provide less cover. SUMMARY • The number of snowshoe hares captured, the number of captures, and the number of locations obtained per hare during the first year appeared adequate for attaining the objectives of this study. • Some deaths due to capture myopathy (most likely cause) occurred during initial trapping periods in each sampling season. Changes to the trapping protocol, trapping schedule, and bait provided seem to have alleviated the problem. • Snowshoe hare densities during summer were highest in small lodgepole stands, followed by large spruce/fir and medium lodgepole. During winter, densities in small lodgepole stands dropped and became more variable across replicates. Medium lodgepole stands gained hares. Spruce/fir stands remained at the same density as during summer. 49 �ACKNOWLEDGEMENTS Ken Wilson (CSU), Bill Romme (CSU), Paul Doherty (CSU), Dave Freddy (CDOW), Chad Bishop (CDOW), and Paul Lukacs (CDOW) provided helpful insight on the design of this study. We appreciate the invaluable logistical support provided by Mike Jackson (USFS), Jake Spritzer (USFS), Margie Michaels (CDOW), Gabriele Engler (CSU), Brandon Diamond (CDOW), Chris Parmeter (CDOW), Kathaleen Crane (CDOW), Lisa Wolfe (CDOW), and Laurie Baeten (CDOW). The following hardy individuals collected the hard-won data presented in this report: Braden Burkholder, Matt Cuzzocreo, Brian Gerber, Belita Marine, Adam Behney, Pete Lundberg, Katie Yale, Britta Schielke, Cory VanStratt, Mike Watrobka, Meredith Goss, Sidra Blake, Keith Rutz, Rob Saltmarsh, Jennie Sinclair, and Evan Wilson. Funding was provided by the Colorado Division of Wildlife. LITERATURE CITED AUBRY, K. B., G. M. KOEHLER, AND J. R. SQUIRES. 2000. Ecology of Canada lynx in southern boreal forests. Pages 373-396 in Ruggiero, L. F., K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation of lynx in the United States. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Collins, Colorado, USA. BARBOUR, M. G., J. H. BURK, W. E. PITTS, F. S. GILLIAM, AND M. W. SCHWARTZ. 1999. Terrestrial plant ecology, Addison Wesley Longman, Inc., Menlo Park, California, USA. BARTMANN, R. M. AND G. BYRNE. 2001. Analysis and critique of the 1998 snowshoe hare pellet survey. Colorado Division of Wildlife 20. BEAUVAIS, G. P. 1997. Mammals in fragmented forests in the Rocky Mountains: community structure, habitat selection, and individual fitness. Dissertation, University of Wyoming, Laramie, Wyoming, USA. BUCKLAND, S. T., D. R. ANDERSON, K. P. BURNHAM, J. L. LAAKE, D. L. BORCHERS, AND L. THOMAS. 2001. Introduction to distance sampling: estimating abundance of biological populations, Oxford University Press, Inc., New York, New York, USA. BULL, E. L., T. W. HEATER, A. A. CLARK, J. F. SHEPHERD, AND A. K. BLUMTON. 2005. Influence of precommercial thinning on snowshoe hares. U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station, General Technical Report PNW-RP-562. DOLBEER, R. A. AND W. R. CLARK. 1975. Population ecology of snowshoe hares in the central Rocky Mountains. Journal of Wildlife Management 39:535-549. DOLBEER, R. A. 1972. Population dynamics of snowshoe hare in Colorado. Dissertation, Colorado State University, Fort Collins, Colorado, USA. EFFORD, M. G., D. K. DAWSON, AND C. S. ROBBINS. 2004. DENSITY: software for analysing capturerecapture data from passive detector arrays. Animal Biodiversity and Conservation 27:1-12. ERIKSSON, H. M. 2006. Snowshoe hare densities in post-fire vegetation. Thesis, University of Alaska Fairbanks, Fairbanks, Alaska, USA. FORESMAN, K. R. AND D. E. PEARSON. 1999. Activity patterns of American martens, Martes americana, snowshoe hares, Lepus americanus, and red squirrels, Tamiasciurus hudsonicus, in westcentral Montana. The Canadian Field-Naturalist 113:386-389. GRIFFIN, P. C. 2004. Landscape ecology of snowshoe hares in Montana. Dissertation, University of Montana, Missoula, Montana, USA. HODGES, K. E. 2000. Ecology of snowshoe hares in southern boreal and montane forests. Pages 163-206 in Ruggiero, L. F., K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation of lynx in the United States. University Press of Colorado, Boulder, Colorado, USA. HODGES, K. E., C. J. KREBS, D. S. HIK, C. I. STEFAN, E. A. GILLIS, AND C. E. DOYLE. 2001. Snowshoe hare demography. Pages 141-178 in Krebs, C. J., S. Boutin, and R. Boonstra, editors. Ecosystem 50 �dynamics of the boreal forest: the Kluane project. Oxford University Press, New York, New York, USA. HOMYACK, J. A., D. J. HARRISON, AND W. B. KROHN. 2003. Effects of precommercial thinning on select wildlife species in northern Maine, with special emphasis on snowshoe hare. Maine Cooperative Fish and Wildlife Research Unit, Orono, Maine, USA. HOMYACK, J. A., D. J. HARRISON, J. A. LITVAITIS, AND W. B. KROHN. 2006. Quantifying densities of snowshoe hares in Maine using pellet plots. Wildlife Society Bulletin 34:74-80. HUGGINS, R. M. 1989. On the statistical analysis of capture-recapture experiments. Biometrika 76:133140. HUGGINS, R. M. 1991. Some practical aspects of a conditional likelihood approach to capture experiments. Biometrics 47:725-732. KEITH, L. B. 1964. Daily activity pattern of snowshoe hares. Journal of Mammalogy 45:626-627. KOEHLER, G. M. 1990. Snowshoe hare, Lepus americanus, use of forest successional stages and population changes during 1985-1989 in North-central Washington. Canadian Field-Naturalist 105:291-293. KOEHLER, G. M. 1990. Population and habitat characteristics of lynx and snowshoe hares in north central Washington. Canadian Journal of Zoology 68:845-851. KREBS, C. J., B. SCOTT GILBERT, S. BOUTIN, AND E. AL. R. BOONSTRA. 1987. Estimation of snowshoe hare population density from turd transects. Canadian Journal of Zoology 65:565-567. KREBS, C. J., R. BOONSTRA, V. NAMS, M. O'DONOGHUE, K. E. HODGES, AND S. BOUTIN. 2001. Estimating snowshoe hare population density from pellet plots: a further evaluation. Canadian Journal of Zoology 79:1-4. LITVAITIS, J. A., J. A. SHERBURNE, AND J. A. BISSONETTE. 1985. Influence of understory characteristics on snowshoe hare habitat use and density. Journal of Wildlife Management 49:866-873. MARQUES, F. F. C., S. T. BUCKLAND, D. GOFFIN, C. E. DIXON, D. L. BORCHERS, B. A. MAYLE, AND A. J. PEACE. 2001. Estimating deer abundance from line transect surveys of dung: sika deer in southern Scotland. Journal of Applied Ecology 38:349-363. MCKELVEY, K. S., G. W. MCDANIEL, L. S. MILLS, AND P. C. GRIFFIN. 2002. Effects of plot size and shape on pellet density estimates for snowshoe hares. Wildlife Society Bulletin 30:751-755. MECH, L. D., K. L. HEEZEN, AND D. B. SINIFF. 1966. Onset and cessation of activity in cottontail rabbit and snowshoe hares in relation to sunset and sunrise. Animal Behaviour 14:410-413. MILLER, M. A. 2005. Snowshoe hare habitat relationships in northwest Colorado. Thesis, Colorado State University, Fort Collins, Colorado, USA. MILLS, L. S., P. C. GRIFFIN, K. E. HODGES, K. MCKELVEY, L. RUGGIERO, AND T. ULIZIO. 2005. Pellet count indices compared to mark-recapture estimates for evaluating snowshoe hare density. Journal of Wildlife Management 69:1053-1062. MURRAY, D., E. ELLSWORTH, AND A. ZACK. 2005. Assessment of potential bias with snowshoe hare fecal pellet-plot counts. Journal of Wildlife Management 69:385-395. MURRAY, D. L., J. D. ROTH, E. ELLSWORTH, A. J. WIRSING, AND T. D. STEURY. 2002. Estimating lowdensity snowshoe hare populations using fecal pellet counts. Canadian Journal of Zoology 80:771-781. NICHOLS, J. D. AND J. E. HINES. 2002. Approaches for the direct estimation of lambda, and demographic contributions to lambda, using capture-recapture data. Journal of Applied Statistics 29:539-568. NICHOLS, J. D., J. E. HINES, J. D. LEBRETON, AND R. PRADEL. 2000. Estimation of contributions to population growth: a reverse-time capture-recapture approach. Ecology 81:3362-3376. NUDDS, T. D. 1977. Quantifying the vegetative structure of wildlife cover. Wildlife Society Bulletin 5:113-117. OTIS, D. L., K. P. BURNHAM, G. C. WHITE, AND D. R. ANDERSON. 1978. Statisical inference from capture data on closed animal populations. Wildlife Monographs 62: PARMENTER, R. R., T. L. YATES, D. R. ANDERSON, K. P. BURNHAM, J. L. DUNNUM, A. B. FRANKLIN, M. T. FRIGGENS, B. C. LUBOW, M. MILLER, G. S. OLSON, C. A. PARMENTER, J. POLLARD, E. 51 �REXSTAD, T. SHENK, T. R. STANLEY, AND G. C. WHITE. 2003. Small-mammal density estimation: a field comparison of grid-based vs. web-based density estimators. Ecological Monographs 73:126. PRADEL, R. 1996. Utilization of capture-mark-recapture for the study of recruitment and population growth rate. Biometrics 52:703-709. PRUGH, L. R. AND C. J. KREBS. 2004. Snowshoe hare pellet-decay rates and aging in different habitats. Wildlife Society Bulletin 32:386-393. RUEDIGER, B., J. CLAAR, S. GNIADEK, B. HOLT, LEWIS LYLE, S. MIGHTON, B. NANEY, G. PATTON , T. RINALDI, J. TRICK, A. VENDEHEY, F. WAHL, N. WARREN, D. WENGER, AND A. WILLIAMSON. 2000. Canada lynx conservation assessment and strategy. U.S. Department of Agriculture, Forest Service, U.S. Department of Interior, Fish and Wildlife Service, Bureau of Land Management, National Park Service R1-00-53 U.S. Department of Agriculture, Forest Service, U.S. Department of Interior, Fish and Wildlife Service, Bureau of Land Management, National Park Service, Missoula, Montana, USA. RUGGIERO, L. F., K. B. AUBRY, S. W. BUSKIRK, G. M. KOEHLER, C. J. KREBS, K. S. MCKELVEY, AND J. R. SQUIRES. 2000. The scientific basis for lynx conservation: qualified insights. Pages 443-454 in Ruggiero, L. F., K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation of lynx in the United States. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Collins, Colorado, USA. SAMUEL, M. D. AND M. R. FULLER. 1996. Wildlife radiotelemetry. Pages 370-418 in Bookhout, T. A., editors. Research and Management Techniques for Wildlife and Habitats. Allen Press, Inc., Lawrence, Kansas, USA. SHENK, T. M. 2005. General locations of lynx (Lynx canadensis) reintroduced to southwestern Colorado from February 4, 1999 through February 1, 2005. Colorado Division of Wildlife Colorado Division of Wildlife, Fort Collins, Colorado, USA. __________. 2006. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report July: 1-45. Colorado Division of Wildlife, Fort Collins, Colorado SULLIVAN, T. P. AND D. S. SULLIVAN. 1988. Influence of stand thinning on snowshoe hare population dynamics and feeding damage in lodgepole pine forest. Journal of Applied Ecology 25:791-805. VAN HORNE, B. 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47:893-901. WHITE, G. C., D. R. ANDERSON, K. P. BURNHAM, AND D. L. OTIS. 1982. Capture-recapture and removal methods for sampling closed populations. Los Alamos National Laboratory, Los Alamos, New Mexico, USA. WHITE, G. C. AND R. A. GARROTT. 1990. Analysis of wildlife radio-tracking data, Academic Press, San Diego, California, USA. WILLIAMS, B. K., J. D. NICHOLS, AND M. J. CONROY. 2002. Analysis and management of animal populations, Academic Press, San Diego, California, USA. WILSON, K. R. AND D. R. ANDERSON. 1985. Evaluation of two density estimators of small mammal population size. Journal of Mammalogy 66:13-21. WOLFE, M. L., N. V. DEBYLE, C. S. WINCHELL, AND T. R. MCCABE. 1982. Snowshoe hare cover relationships in northern Utah. Journal of Wildlife Management 46:662-670. WOLFF, J. O. 1980. The role of habitat patchiness in the population dynamics of snowshoe hares. Ecological Monographs 50:111-130. ZAHRATKA, J. L. 2004. The population and habitat ecology of snowshoe hares (Lepus americanus) in the southern Rocky Mountains. Thesis, The University of Wyoming, Laramie, Wyoming, USA. ZIMMERMAN, G. M., H. GOETZ, AND P. W. JR. MIELKE. 1985. Use of an improved statistical method for group comparisons to study effects of prairie fire. Ecology 66:606-611. Prepared by _________________________________________________ Jacob S. Ivan, Graduate Student, Colorado State University 52 �Table 1. Number of snowshoe hares (Lepus americanus) captured during 5-day trapping sessions conducted during July-September 2006 and January-March 2007 on 3 medium lodgepole, 3 spruce/fir, and 6 small lodgepole stands on the Gunnison National Forest, Taylor Park and Pitkin, Colorado. ____________________________________________________________________________________ Number of Hares Captured (Total Captures) ___________________________________________________________ Summer 2006 Winter 2007 Both Summer and Winter ____________________________________________________________________________________ Medium Lodgepole 3 24 2 Small Lodgepole 50 40 10 Spruce/Fir 22 35 2 ____________________________________________________________________________________ Table 2. Number of snowshoe hares (Lepus americanus) radio-collared and tracked during 10-day sessions immediately following 5-day trapping periods July-September 2006 and January-March 2007 on the Gunnison National Forest, Taylor Park and Pitkin, Colorado. ____________________________________________________________________________________ Summer 2006 Winter 2007 ____________________________________________________________________________________ Number of Hares Collared 41 79 Number of Locations 407 510 Number of Locations/Hare 9.9 6.5 ____________________________________________________________________________________ Figure 1. Purported high quality snowshoe hare habitat in Colorado. From left to right: small lodgepole pine, medium lodgepole pine, and large Engelmann spruce/subalpine fir. 53 �Figure 2. Study area near Taylor Park and Pitkin, Colorado including medium lodgepole (squares), small lodgepole (circles), and spruce/fir (triangles) stands selected for mark-recapture sampling. 54 �Jul Aug Sep 1 Jan 1 2 3 4 2 3 4 5 6 7 8 5 6 7 8 9 10 11 9 10 11 12 13 14 15 12 13 14 15 16 17 18 16 17 18 19 20 21 22 19 20 21 22 23 24 25 23 24 25 26 27 28 29 26 27 28 29 30 31 1 30 31 1 2 3 4 5 2 3 4 5 6 7 8 Feb 6 7 8 9 10 11 12 9 10 11 12 13 14 15 13 14 15 16 17 18 19 16 17 18 19 20 21 22 20 21 22 23 24 25 26 23 24 25 26 27 28 1 27 28 29 30 31 1 2 2 3 4 5 6 7 8 3 4 5 6 7 8 9 9 10 11 12 13 14 15 10 11 12 13 14 15 16 16 17 18 19 20 21 22 17 18 19 20 21 22 23 23 24 25 26 27 28 29 24 25 26 27 28 29 30 30 31 Mar Figure 3. Approximate annual data collection schedule for trapping (�) and telemetry (�). Dates and weeks will change depending on calendar year and pay schedule. During telemetry work, the 6-person crew will be divided into 2 teams, only one of which will be working at any given time. Monthly locations on radio-collared hares will also be collected in the interim between the intensive sampling periods indicated here. Figure 4. 15 trap locations (•) on 7 × 12 trapping grid where vegetation will be sampled by measuring stem density horizontal cover, and canopy cover at the 25 points on each 5 × 5 subgrid (inset). In addition, basal area will be measured at the trap location (�) on which each of the 15 subgrids are centered. 55 �Density Power Analysis % Non-overlapping 95% CIs 100 90 80 70 60 50 40 30 20 10 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0 capture/recapture probability N=5 vs. N=10 N=15 vs. N=30 N=17 vs. N=39 Figure 5. Power for distinguishing differences in snowshoe hare density between 2 habitat types when a difference actually exists. Gray area indicates the capture probability realized by the 3rd day of trapping during a pilot study in winter 2005. N indicates number of hares per grid, a range of roughly 0.1 (N = 5) to 0.7 hares/ha (N = 39). Survival Power Analysis (N = 45) 100 100 90 80 90 80 % Significant LR Tests 70 60 50 40 30 20 10 0 70 60 50 40 30 20 Capture/Recapture Probability 0.2 vs. 0.5 0.3 vs. 0.5 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.10 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 10 0 0.15 % Significant LR Tests Survival Power Analysis (N = 30) Capture/Recapture Probability 0.4 vs. 0.5 0.2 vs. 0.5 0.3 vs. 0.5 0.4 vs. 0.5 Figure 6. Power, or probability of distinguishing differences in snowshoe hare survival between 2 habitat types when differences actually exist. N = 30 (left) and N = 45 (right) correspond to reasonable estimates of the number of hares I expect to capture in each habitat type. Gray area indicates the capture probability realized by the 3rd day of trapping during a pilot study in winter 2005. 56 �Figure 7. Snowshoe hare density and 95% confidence intervals in 3 types of stands in central Colorado as determined by mark-recapture with telemetry augmentation, July-September 2006 and January-March 2007. 57 �Colorado Division of Wildlife July 2007- June 2008 WILDLIFE RESEARCH REPORT State of: Cost Center: Work Package: Task No.: Colorado 3430 0670 1 Federal Aid Project No. N/A : : : : : Division of Wildlife Mammals Research Lynx Conservation Post-Release Monitoring of Lynx Reintroduced to Colorado Period Covered: July 1, 2007 - June 30, 2008 Author: T. M. Shenk Personnel: L. Baeten, B. Diamond, R. Dickman, D. Freddy, L. Gepfert, J. Ivan, R. Kahn, A. Keith, G. Merrill, B. Smith, T. Spraker, S. Wait, S. Waters, L. Wolfe All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT In an effort to establish a viable population of Canada lynx (Lynx canadensis) in Colorado, the Colorado Division of Wildlife (CDOW) initiated a reintroduction effort in 1997 with the first lynx released in February 1999. From 1999-2007, 218 wild-caught lynx from Canada and Alaska were released in Colorado. We documented survival, movement patterns, reproduction, and landscape habitatuse through aerial (n = 10,935) and satellite (n = 26,082) tracking. Most lynx remained near the core release area in southwestern Colorado. From 1999-August 2008, there were 112 mortalities of released adult lynx. Approximately 30.4% were either human-induced or likely human-induced through either collisions with vehicles or gunshot. Starvation and disease/illness accounted for 18.8% of the deaths while 36.6% of the deaths were from unknown causes. Of these mortalities, 26.8% occurred outside of Colorado. Monthly mortality rate was lower inside the study area than outside, and slightly higher for male than for female lynx, although 95% confidence intervals for sexes overlapped. Mortality was higher immediately after release (first month = 0.0368 [SE = 0.0140] inside the study area, and 0.1012 [SE = 0.0359] outside the study area), and then decreased according to a quadratic trend over time. Reproductive females had the smallest 90% utilization distribution home ranges ( x = 75.2 km2, SE = 15.9 km2), followed by attending males ( x = 102.5 km2, SE = 39.7 km2) and non-reproductive animals ( x = 653.8 km2, SE = 145.4 km2). Reproduction was first documented in 2003 with subsequent successful reproduction in 2004, 2005 and 2006. No dens were documented in 2007 or 2008. From snow-tracking, the primary winter prey species (n = 548 kills) were snowshoe hare (Lepus americanus, annual x = 73.3%, SE = 4.7, n = 10) and red squirrel (Tamiasciurus hudsonicus, annual x = 18.2%, SE = 4.2, n = 10); other mammals and birds formed a minor part of the winter diet. Lynx use-density surfaces were generated to illustrate relative use of areas throughout Colorado. Within the areas of high use in southwestern Colorado, site-scale habitat use, documented through snow-tracking, supports mature 1 �Engelmann spruce (Picea engelmannii)-subalpine fir (Abies lasiocarpa) forest stands with 42-65% canopy cover and 15-20% conifer understory cover as the most commonly used areas in southwestern Colorado. Little difference in aspect (slight preference for north-facing slopes), slope ( x = 15.7°) or elevation ( x = 3173 m) were detected for long beds, travel and kill sites (n = 1841). Den sites (n = 37) however, were located at higher elevations ( x = 3354 m, SE = 31 m) on steeper ( x = 30°, SE = 2°) and more commonly north-facing slopes with a dense understory of coarse woody debris. Two years of a study to evaluate snowshoe hare densities, demography and seasonal movement patterns among small and medium tree-sized lodgepole pine stands and mature spruce/fir stands have been completed in 2006-2008 and will continue through 2009 (see Appendix I of this report). Results to date have demonstrated that CDOW has developed lynx release protocols that ensure high initial post-release survival followed by high long-term survival, site fidelity, reproduction and recruitment of Colorado-born lynx into the Colorado breeding population. What is yet to be demonstrated is whether Colorado can support sufficient recruitment to offset annual mortality for a viable lynx population over time. Monitoring continues in an effort to document such viability. 2 �WILDLIFE RESEARCH REPORT POST RELEASE MONITORING OF LYNX (LYNX CANADENSIS) REINTRODUCED TO COLORADO TANYA M. SHENK P. N. OBJECTIVE The initial post-release monitoring of Canada lynx (Lynx canadensis) reintroduced into Colorado will emphasize 5 primary objectives: 1. Assess and modify release protocols to ensure the highest probability of survival for each lynx released. 2. Obtain regular locations of released lynx to describe general movement patterns and habitats used by lynx. 3. Determine causes of mortality in reintroduced lynx. 4. Estimate survival of lynx reintroduced to Colorado. 5. Estimate reproduction of lynx reintroduced to Colorado. Three additional objectives will be emphasized after lynx display site fidelity to an area: 6. Refine descriptions of habitats used by reintroduced lynx. 7. Refine descriptions of daily and overall movement patterns of reintroduced lynx. 8. Describe hunting habits and prey of reintroduced lynx. Information gained to achieve these objectives will form a basis for the development of lynx conservation strategies in the southern Rocky Mountains. SEGMENT OBJECTIVES 1. Complete winter 2007-08 field data collection on lynx habitat use at the landscape scale, hunting behavior, diet, mortalities, and movement patterns. 2. Complete winter 2007-08 lynx trapping field season to collar Colorado born lynx and re-collar adult lynx. 3. Complete spring 2008 field data on lynx reproduction. 4. Summarize and analyze data and publish information as Progress Reports, peer-reviewed manuscripts for appropriate scientific journals, or CDOW technical publications. 5. Complete the second year of field work to evaluate snowshoe hare (Lepus americanus) densities, demography and seasonal movement patterns among small and medium tree-sized lodgepole pine stands and mature spruce/fir stands (see Appendix I). INTRODUCTION The Canada lynx occurs throughout the boreal forests of northern North America. Colorado represents the southern-most historical distribution of lynx, where the species occupied the higher elevation, montane forests in the state. Little was known about the population dynamics or habitat use of this species in their southern distribution. Lynx were extirpated or reduced to a few animals in the state by the late 1970‘s due, most likely, to predator control efforts such as poisoning and trapping. Given the isolation of Colorado to the nearest northern populations, the CDOW considered reintroduction as the only option to attempt to reestablish the species in the state. 3 �A reintroduction effort was begun in 1997, with the first lynx released in Colorado in 1999. To date, 218 wild-caught lynx from Alaska and Canada have been released in southwestern Colorado. The goal of the Colorado lynx reintroduction program is to establish a self-sustaining, viable population of lynx in this state. Evaluation of incremental achievements necessary for establishing viable populations is an interim method of assessing if the reintroduction effort is progressing towards success. There are 7 critical criteria for achieving a viable population: 1) development of release protocols that lead to a high initial post-release survival of reintroduced animals, 2) long-term survival of lynx in Colorado, 3) development of site fidelity by the lynx to areas supporting good habitat in densities sufficient to breed, 4) reintroduced lynx must breed, 5) breeding must lead to reproduction of surviving kittens 6) lynx born in Colorado must reach breeding age and reproduce successfully, and 7) recruitment must equal or be greater than mortality over an extended period of time. The post-release monitoring program for the reintroduced lynx has 2 primary goals. The first goal is to determine how many lynx remain in Colorado and their locations relative to each other. Given this information and knowing the sex of each individual, we can assess whether these lynx can form a breeding core from which a viable population might be established. From these data we can also describe general movement patterns and habitat use. The second primary goal of the monitoring program is to estimate survival of the reintroduced lynx and, where possible, determine causes of mortality for reintroduced lynx. Such information will help in assessing and modifying release protocols and management of lynx once they have been released to ensure their highest probability of survival. Additional goals of the post-release monitoring program for lynx reintroduced to the southern Rocky Mountains included refining descriptions of habitat use and movement patterns and describing successful hunting habitat once lynx established home ranges that encompassed their preferred habitat. Specific objectives for the site-scale habitat data collection include: 1) describe and quantify site-scale habitat use by lynx reintroduced to Colorado, 2) compare site-scale habitat use among types of sites (e.g., kills vs. long-duration beds), and 3) compare habitat features at successful and unsuccessful snowshoe hare chases. Documenting reproduction is critical to the success of the program and lynx are monitored intensively to document breeding, births, survival and recruitment of lynx born in Colorado. Site-scale habitat descriptions of den sites are also collected and compared to other sites used by lynx. The program will also investigate the ecology of snowshoe hare in Colorado. A study comparing snowshoe hare densities among mature stands of Engelmann spruce (Picea engelmannii)/subalpine fir (Abies lasiocarpa), lodgepole pine (Pinus contorta) and Ponderosa pine (Pinus ponderosa) was completed in 2004 with highest hare densities found in Engelmann spruce/subalpine fir stands and no hares found in Ponderosa pine stands. A study to evaluate the importance of young, regenerating lodgepole pine and mature Engelmann spruce/subalpine fir stands in Colorado by examining density and demography of snowshoe hares that reside in each was initiated in 2005 and will continue through 2009 (see Appendix I). Lynx is listed as threatened under the Endangered Species Act (ESA) of 1973, as amended (16 U. S. C. 1531 et. seq.)(U. S. Fish and Wildlife Service 2000). Colorado is included in the federal listing as lynx habitat. Thus, an additional objective of the post-release monitoring program is to develop conservation strategies relevant to lynx in Colorado. To develop these conservation strategies, information specific to the ecology of the lynx in its southern Rocky Mountain range, such as habitat use, movement patterns, mortality factors, survival, and reproduction in Colorado is needed. 4 �STUDY AREA Byrne (1998) evaluated five areas within Colorado as potential lynx habitat based on (1) relative snowshoe hare densities (Bartmann and Byrne 2001), (2) road density, (3) size of area, (4) juxtaposition of habitats within the area, (5) historical records of lynx observations, and (6) public issues. Based on results from this analysis, the San Juan Mountains of southwestern Colorado were selected as the core reintroduction area, and where all lynx were reintroduced. Wild Canada lynx captured in Alaska, British Columbia, Manitoba, Quebec and Yukon were transported to Colorado and held at The Frisco Creek Wildlife Rehabilitation Center located within the reintroduction area prior to release. Post-release monitoring efforts were focused in a 20,684 km2 study area which included the core reintroduction area, release sites and surrounding high elevation sites (> 2,591 m). The area encompassed the southwest quadrant of Colorado and was bounded on the south by New Mexico, on the west by Utah, on the north by interstate highway 70, and on the east by the Sangre de Cristo Mountains (Figure 1). Southwestern Colorado is characterized by wide plateaus, river valleys, and rugged mountains that reach elevations over 4,200 m. Engelmann spruce/subalpine fir is the most widely distributed coniferous forest type within the study area. The lynx-established core area is roughly bounded by areas used by lynx in the Taylor Park/Collegiate Peak areas in central Colorado and includes areas of continuous use by lynx, including areas used during breeding and denning (Figure 1). METHODS REINTRODUCTION Effort Wild Canada lynx were captured in Alaska, British Columbia, Manitoba, Quebec and Yukon and transported to Colorado where they were held at the Frisco Creek Wildlife Rehabilitation Center prior to release. All lynx releases were conducted under the protocols found to maximize survival (see Shenk 2001). Estimated age, sex and body condition were ascertained and recorded for each lynx prior to release (see Wild 1999). Lynx were transported from the rehabilitation facility to their release site in individual cages. Specific release site locations were recorded in Universal Transverse Mercator (UTM) coordinates and identification of all lynx released at the same location, on the same day, was recorded. Behavior of the lynx on release and movement away from the release site were documented. Movement, Distribution and Relative Use of Areas by Lynx To monitor lynx movements and thus determine distribution and relative use of areas all released lynx were fitted with radio collars. All lynx released in 1999 were fitted with TelonicsTM radio-collars. All lynx released since 1999, with the exception of 5 males released in spring 2000, were fitted with SirtrackTM dual satellite/VHF radio-collars. These collars have a mortality indicator switch that operated on both the satellite and VHF mode. The satellite component of each collar was programmed to be active for 12 hours per week. The 12-hour active periods for individual collars were staggered throughout the week. Signals from the collars allowed for locations of the animals to be made via Argos, NASA, and NOAA satellites. The location information was processed by ServiceArgos and distributed to the CDOW through e-mail messages. Datasets.-- To determine recent (post-reintroduction) movement and distribution of lynx reintroduced, born or initially trapped in Colorado and relative use of areas by these lynx, regular locations of lynx were collected through a combination of aerial and satellite tracking. Locations were recorded and general habitat descriptions for each aerial location was recorded. The first dataset of lynx locations included all locations obtained from daytime flights conducted with a Cessna 185 or similar aircraft to locate lynx by their VHF collar transmitters (hereafter aerial locations). VHF transmitters have been used on lynx since the first lynx were released in February 1999. The second type of lynx location 5 �data was collected via satellite from the satellite collar transmitters placed on the lynx (hereafter satellite locations). Satellite transmitter collars were first used for lynx in April 2000. These satellite collars also contained a VHF transmitter which also allowed locating lynx from the air or ground. All locations were recorded in Universal Transverse Mercator (UTM) coordinates using the CONUS NAD27 datum. Flights to obtain lynx aerial locations were typically conducted on a weekly basis throughout most summer and winter months and twice a week during the den search field season (May 15 – June 30), depending on weather and availability of planes and pilots. Flights were typically concentrated in the high elevation (> 2700 m) southwest quadrant of Colorado which encompasses the core lynx release and research area (Figure 1). Flights during the den seasons were conducted to obtain locations on all female lynx within the state wearing an active VHF transmitter. VHF transmitters were outfitted with sufficient batteries to last 60 months. The satellite transmitters were designed to provide locations on a weekly basis with sufficient batteries to last for 18 months. Lynx may not be exhibiting typical behavior or habitat use within the first few months after their release in Colorado. Therefore, a subset of each of the aerial and satellite datasets was created that eliminated the first 180 days (approximately 6 months) of locations obtained for each lynx immediately after their initial release. As a result, the truncated aerial location dataset contained lynx locations from September 1999 through March 2007 while the truncated satellite location dataset began October 2000 and extended through March 2007. Accuracy of both aerial and satellite locations varied with the environmental conditions at the time the location was obtained. Accuracy of aerial locations was influenced by weather with accuracy ranging from 50 - 500 meters. Satellite location accuracy was also influenced by atmospheric conditions and position of the satellites. Satellite location accuracy ranged from 150 meters -10 km. Movement and Distribution.-- To document all known lynx locations maps were generated with all aerial and satellite locations displayed. Due to lynx movements outside of Colorado, particularly into the states of New Mexico, Utah and Wyoming we further evaluated lynx use throughout those three states, as well as the data would allow. All individual lynx located at least once in these 3 states (nontruncated datasets) were identified and tallied for each year. To document consistency and known use of these states after the initial effect of being reintroduced was minimized (i.e., 180 days post-release), each individual lynx located at least once in these states from the truncated datasets were identified and tallied. Relative Use.-- To document relative use of areas by lynx, 90% kernel use-density surfaces were calculated for truncated satellite and aerial lynx locations using the ArcGIS Spatial Analyst Kernel Density Tool. Due to differences in data collection frequency and accuracy between datasets, the truncated satellite and truncated aerial data were analyzed separately for generating the lynx use-density surfaces. These use-density surfaces fit a smoothly curved surface over each lynx location. The surface value was highest at the location of the point and diminished with increasing distance from the point. A fixed kernel was used with a smoothing parameter of 5 km, reaching 0 at the search radius distance from the point. Only a circular neighborhood was possible. The volume under the surface equaled the total value for the point. The use-density at each output GIS raster cell was calculated by adding the values of all the kernel surfaces from all the lynx point locations that overlaid each raster cell center. The kernel function was based on the quadratic kernel function described in Silverman (1986, p. 76, equation 4.5). The use-density surfaces were calculated at 100 m resolution. To enhance graphic displays of higher usedensity areas, density values representing single locations were not displayed. 6 �Home Range Annual home ranges were calculated as a 95% utilization distribution using a kernel home-range estimator for each lynx we had at least 30 locations for within a year. A year was defined as March 15 – March 14 of the following year. Locations used in the analyses were collected from September 1999 – January 2006 and all locations obtained for an individual during the first six months after its release were eliminated from any home range analyses as it was assumed movements of lynx initially post-release may not be representative of normal habitat use. Locations were obtained either through aerial VHF surveys or locations or the midpoint (ArcView Movement Extension) of all high quality (accuracy rating of 01km) satellite locations obtained within a single 24-hour period. All locations used within a single home range analysis were taken a minimum of 24 hours apart. Home range estimates were classified as being for a reproductive or non-reproductive animal. A reproductive female was defined as one that had kittens with her; a reproductive male was defined as a male whose movement patterns overlapped that of a reproductive female. If a litter was lost within the defined year a home range described for a reproductive animal were estimated using only locations obtained while the kittens were still with the female. Survival Multi-state mark-recapture models were used to estimate monthly mortality rates and described in detail in Devineau et al. 2008 (in review). This approach accommodated missing data and allowed exploration of factors possibly affecting lynx survival such as sex, time spent in pre-release captivity, movement patterns, and origin. Mortality Factors When a mortality signal (75 beats per minute [bpm] vs. 50 bpm for the Telonics™ VHF transmitters, 20 bpm vs. 40 bpm for the Sirtrack™ VHF transmitters, 0 activity for Sirtrack™ PTT) was heard during either satellite, aerial or ground surveys, the location (UTM coordinates) was recorded. Ground crews then located and retrieved the carcass as soon as possible. The immediate area was searched for evidence of other predators and the carcass photographed in place before removal. Additionally, the mortality site was described and habitat associations and exact location were recorded. Any scat found near the dead lynx that appeared to be from the lynx was collected. All carcasses were transported to the Colorado State University Veterinary Teaching Hospital (CSUVTH) for a post mortem exam to 1) determine the cause of death and document with evidence, 2) collect samples for a variety of research projects, and 3) archive samples for future reference (research or forensic). The gross necropsy and histology were performed by, or under the lead and direct supervision of a board certified veterinary pathologist. At least one research personnel from the CDOW involved with the lynx program was also present. The protocol followed standard procedures used for thorough post-mortem examination and sample collection for histopathology and diagnostic testing (see Shenk 1999 for details). Some additional data/samples were routinely collected for research, forensics, and archiving. Other data/samples were collected based on the circumstances of the death (e.g., photographs, video, radiographs, bullet recovery, samples for toxicology or other diagnostic tests, etc.). From 1999–2004 the CDOW retained all samples and carcass remains with the exception of tissues in formalin for histopathology, brain for rabies exam, feces for parasitology, external parasites for ID, and other diagnostic samples. Since 2005 carcasses are disposed of at the CSUVTH with the exception of the lower canine, fecal samples, stomach content samples and tissue or bone marrow samples to be delivered by CDOW to the Center for Disease control for plague testing. The lower canine, from all carcasses, is sent to Matson Labs (Missoula, Montana) for aging and the fecal and stomach content samples are evaluated for diet. 7 �Reproduction Females were monitored for proximity to males during each breeding season. We defined a possible mating pair as any male and female documented within at least 1 km of each other in breeding season through either flight data or snow-tracking data. Females were then monitored for site fidelity to a given area during each denning period of May and June. Each female that exhibited stationary movement patterns in May or June were closely monitored to locate possible dens. Dens were found when field crews walked in on females that exhibited virtually no movement for at least 10 days from both aerial and ground telemetry. Kittens found at den sites were weighed, sexed and photographed. Each kitten was uniquely marked by inserting a sterile passive integrated transponder (PIT, Biomark, Inc., Boise, Idaho, USA) tag subcutaneously between the shoulder blades. Time spent at the den was minimized to ensure the least amount of disturbance to the female and the kittens. Weight, PIT-tag number, sex and any distinguishing characteristics of each kitten was also recorded. Beginning in 2005, blood and saliva samples were collected and archived for genetic identification. During the den site visits, den site location was recorded as UTM coordinates. General vegetation characteristics, elevation, weather, field personnel, time at the den, and behavioral responses of the kittens and female were also recorded. Once the females moved the kittens from the natal den area, den sites were visited again and site-specific habitat data were collected (see Habitat Use section below). Captures Captures were attempted for either lynx that were in poor body condition or lynx that needed to have their radio-collars replaced due to failed or failing batteries or to radio-collar kittens born in Colorado once they reached at least 10-months of age when they were nearly adult size. Methods of recapture included 1) trapping using a Tomahawk™ live trap baited with a rabbit and visual and scent lures, 2) calling in and darting lynx using a Dan-Inject CO2 rifle, 3) custom box-traps modified from those designed by other lynx researchers (Kolbe et al. 2003) and 4) hounds trained to pursue felids were also used to tree lynx and then the lynx was darted while treed. Lynx were immobilized either with Telazol (3 mg/kg; modified from Poole et al. 1993 as recommended by M. Wild, DVM) or medetomidine (0.09mg/kg) and ketamine (3 mg/kg; as recommended by L. Wolfe, DVM)) administered intramuscularly (IM) with either an extendible pole-syringe or a pressurized syringe-dart fired from a Dan-Inject air rifle. Immobilized lynx were monitored continuously for decreased respiration or hypothermia. If a lynx exhibited decreased respiration 2mg/kg of Dopram was administered under the tongue; if respiration was severely decreased, the animal was ventilated with a resuscitation bag. If medetomidine/ketamine were the immobilization drugs, the antagonist Atipamezole hydrochloride (Antisedan) was administered. Hypothermic (body temperature < 95o F) animals were warmed with hand warmers and blankets. While immobilized, lynx were fitted with replacement SirtrackTM VHF/satellite collar and blood and hair samples were collected. Once an animal was processed, recovery was expedited by injecting the equivalent amount of the antagonist Antisedan IM as the amount of medetomidine given, if medetomodine/ketemine was used for immobilization. Lynx were then monitored while confined in the box-trap until they were sufficiently recovered to move safely on their own. No antagonist is available for Telezol so lynx anesthetized with this drug were monitored until the animal recovered on its own in the box-trap and then released. If captured and in poor body condition, lynx were anesthetized with either Telezol (2 mg/kg) or medetomodine/ketemine and returned to the Frisco Creek Wildlife Rehabilitation Center for treatment. HABITAT USE Gross habitat use was documented by recording canopy vegetation at aerial locations. More 8 �refined descriptions of habitat use by reintroduced lynx were obtained through following lynx tracks in the snow (i.e., snow-tracking) and site-scale habitat data collection conducted at sites found through this method to be used by lynx. See Shenk (2006) for detailed methodologies. DIET AND HUNTING BEHAVIOR Winter diet of reintroduced lynx was estimated by documenting successful kills through snowtracking. Prey species from failed and successful hunting attempts were identified by either tracks or remains. Scat analysis also provided information on foods consumed. Scat samples were collected wherever found and labeled with location and individual lynx identification. Only part of the scat was collected (approximately 75%); the remainder was left in place in the event that the scat was being used by the animal as a territory mark. Site-scale habitat data collected for successful and unsuccessful snowshoe hare kills were compared. SNOWSHOE HARE ECOLOGY To further our understanding of snowshoe hare ecology in Colorado, a study was conducted comparing snowshoe hare densities among mature stands of Engelmann spruce/subalpine fir, lodgepole pine (Pinus contorta) and Ponderosa pine (Pinus ponderosa). The highest hare densities were found in Engelmann spruce/subalpine fir stands and no hares found in Ponderosa pine stands (Zahratka and Shenk 2008). A second study was initiated in 2005 to evaluate the importance of young, regenerating lodgepole pine and mature Engelmann spruce / subalpine fir stands in Colorado by examining density and demography of snowshoe hares that reside in each (Ivan 2005). Specifically, this study was designed to evaluate small and medium lodgepole pine stands and large spruce/fir stands where the classes ― small‖, ―medium‖, and ―l arge‖ refer to the diameter at breast height (dbh) of overstory trees as defined in the United States Forest Service R2VEG Database (small = 2.54 12.69 cm dbh, medium = 12.70 22.85 cm, and large = 22.86 40.64 cm dbh; J. Varner, United States Forest Service, personal communication). The study design was also developed to identify which of the numerous hare density-estimation procedures available perform accurately and consistently using an innovative, telemetry augmentation approach as a baseline. In addition, movement patterns and seasonal use of deciduous cover types such as riparian willow were assessed. Finally, the study was designed to further expound on the relationship between density, demography, and stand-type by examining how snowshoe hare density and demographic rates vary with specific vegetation, physical, and landscape characteristics of a stand. RESULTS REINTRODUCTION Effort From 1999 through 2006, 218 wild-caught lynx were reintroduced into southwestern Colorado (Table 1). No lynx were released in 2007 or 2008. All lynx were released with either VHF or dual VHF/satellite radio collars so they could be monitored for movement, reproduction and survival. The CDOW does not plan to release any additional lynx in 2009. Movement Patterns and Distribution Numerous travel corridors were used repeatedly by more than one lynx. These travel corridors include the Cochetopa Hills area for northerly movements, the Rio Grande Reservoir-SilvertonLizardhead Pass for movements to the west, and southerly movements down the east side of Wolf Creek Pass to the southeast through the Conejos River Valley. Lynx appear to remain faithful to an area during winter months, and exhibit more extensive movements away from these areas in the summer. 9 �A total of 10,935 aerial and 26,082 satellite locations were obtained from the 218 reintroduced lynx, radio-collared Colorado kittens (n = 14) and unmarked lynx captured in Colorado (n = 2) as of August 27, 2008. The majority of these locations were in Colorado (Figure 2). Some reintroduced lynx dispersed outside of Colorado into Arizona, Idaho, Iowa, Kansas, Montana, Nebraska, Nevada, New Mexico, South Dakota, Utah and Wyoming (Figure 2). The majority of surviving lynx from the reintroduction effort currently continue to use high elevation (> 2900 m), forested terrain in an area bounded on the south by New Mexico north to Independence Pass, west as far as Taylor Mesa and east to Monarch Pass. Most movements away from the Core Release Area were to the north. Relative Use The lynx use-density surfaces resulting from the fixed kernel analyses provided relative probabilities of finding lynx in areas throughout their distribution. A single use-density surface was calculated separately for both the aerial (n = 8058) and satellite truncated datasets (n = 16240). All 218 lynx released in Colorado, all radio-collared kittens and 2 captured unmarked adults were located at least once in Colorado. The majority of these lynx remained in Colorado. The use-density surfaces within Colorado were displayed separately for both the aerial (Figure 3) and satellite truncated datasets (Figure 4). Of the total locations available in the truncated datasets used to generate the usedensity surfaces, 7953 of the aerial locations and 13,241 of the satellite locations were in Colorado. Aerial and satellite use-density surfaces indicated similar high use-density areas. Satellite locations indicated broader spatial use by lynx because satellite collars provided more locations than flights. The use-density surface for lynx use in Colorado indicates two primary areas of use. The first is the Core Research Area (see Figure 1) and a secondary core centered in the Collegiate Peaks Wilderness (Figures 1, 3 and 4). High use is also documented for 1) the area east of Dillon, on both the north and south sides of I70 and 2) the area north of Hwy 50 centered around Gunnison and then north to Crested Butte. These last 2 high use areas are smaller in extent than the 2 core areas. Relative use-density surfaces were also generated for New Mexico, Wyoming and Utah and presented in detail in Shenk (2007). Home Range Reproductive females had the smallest 90% utilization distribution annual home ranges ( x = 75.2 km2, SE = 15.9 km2, n = 19), followed by attending males ( x = 102.5 km2, SE = 39.7 km2, n = 4). Nonreproductive females had the largest annual home ranges ( x = 703.9 km2, SE = 29.8 km2, n = 32) followed by non-reproductive males ( x = 387.0 km2, SE = 73.5 km2, n = 6). Combining all nonreproductive animals yielded a mean annual home range of 653.8 km2 (SE = 145.4 km2, n = 38). Survival Detailed analyses of lynx mortality was completed and described in Devineau et al. 2008 (in review). Monthly mortality rate was lower inside the study area than outside, and slightly higher for male than for female lynx, although 95% confidence intervals for sexes overlapped. Mortality was higher immediately after release (first month = 0.0368 [SE = 0.0140] inside the study area, and 0.1012 [SE = 0.0359] outside the study area), and then decreased according to a quadratic trend over time. As of August 27, 2008, CDOW was actively monitoring/tracking 45 of the 106 lynx still possibly alive (Table 2). There are 62 lynx that we have not heard signals on since at least August 27, 2007 and these animals are classified as ‗missing‘ (Table 2). One of these missing lynx is a mortality of unknown identity, thus only 61 are truly missing. Possible reasons for not locating these missing lynx include 1) 10 �long distance dispersal, beyond the areas currently being searched, 2) radio failure, or 3) destruction of the radio (e.g., run over by car). CDOW continues to search for all missing lynx during both aerial and ground searches. Two of the missing lynx released in 2000 are thought to have slipped their collars. Mortality Factors Of the total 218 adult lynx released, we have 112 known mortalities as of August 27, 2008 (Table 2). Starvation was a significant cause of mortality in the first year of releases only. The primary known causes of death included 30.4% human-induced deaths which were confirmed or probably caused by collisions with vehicles or gunshot (Table 3). Malnutrition and disease/illness accounted for 18.8% of the deaths. An additional 36.6% of known mortalities were from unknown causes. Mortalities occurred throughout the areas through which lynx moved, with 26.8% occurring outside of Colorado. The out of state mortalities included 14 in New Mexico, 4 in Wyoming, Utah and Nebraska, and 1 each in Arizona, Kansas, Iowa and Montana (Figure 2, Table 4). Reproduction Reproduction was first documented in 2003 when 6 dens and a total of 16 kittens were found in the lynx Core Release Area in southwestern Colorado. Reproduction was also documented in 2004, 2005 and 2006. No dens were found in 2007 or 2008 (Table 5). Field crews weighed, photographed, PIT-tagged the kittens and checked body condition. Beginning in 2005, we also collected blood samples from the kittens for genetic work in an attempt to confirm paternity Kittens were processed as quickly as possible (11-32 minutes) to minimize the time the kittens were without their mother. While working with the kittens the females remained nearby, often making themselves visible to the field crews. The females generally continued a low growling vocalization the entire time personnel were at the den. In all cases, the female returned to the den site once field crews left the area. At all dens the females appeared in excellent condition, as did the kittens. The kittens weighed from 270-500 grams. Lynx kittens weigh approximately 200 grams at birth and do not open their eyes until they are 10-17 days old. The percent of tracked females found with litters in 2006 was lower (0.095) than in the 3 previous years (0.413, SE = 0.032, Table 5). However, all demographic and habitat characteristics measured at the 4 dens that were found in 2006 were comparable to all other dens found. Mean number of kittens per litter from 2003-2006 was 2.78 (SE = 0.05) and sex ratio of females to males was equal ( x = 1.14, SE = 0.14). More details of reproduction in 2003-06 were presented in Shenk (2007). Den Sites.-- A total of 37 dens were found from 2003-2006. All of the dens except one have been scattered throughout the high elevation areas of Colorado, south of I-70. In 2004, 1 den was found in southeastern Wyoming, near the Colorado border. Dens were located on steep ( x slope = 30o , SE=2o), north-facing, high elevation ( x = 3354 m, SE = 31 m) slopes. The dens were typically in Engelmann spruce/subalpine fir forests in areas of extensive downfall of coarse woody debris (Shenk 2006). All dens were located within the winter use areas used by the females. No dens were found in either 2007 or 2008 even though up to 34 adult females were monitored intensively during the denning period (Table 5). Captures Two adult lynx were captured in 2001 for collar replacement. One lynx was captured in a tomahawk live-trap, the other was treed by hounds and then anesthetized using a jab pole. Five adult lynx were captured in 2002; 3 were treed by hounds and 2 were captured in padded leghold traps. In 2004, 1 lynx was captured with a Belisle snare and 6 adult lynx were captured in box-traps. Trapping effort was substantially increased in winter and spring 2005 and 12 adult lynx were captured and re-collared. Eight 11 �reintroduced lynx were captured in winter and spring 2006. In 2007, 11 reintroduced adult lynx were captured and re-collared and an additional 10 in 2008. All lynx captured in Colorado from 2005-2008 were caught in box-traps. In addition, as part of the collaring trapping effort, 14 Colorado-born kittens were captured and collared at approximately 10-months of age. Seven 2004-born kittens were collared in spring 2005, and 7, 2005-born kittens were collared in spring 2006. We were not successful at capturing and collaring any kittens born in 2006 in winter 2006-07. We did however, capture 2 adults (approximate age 2 years old) in winter 2006-07 that had no PIT-tags or radio collars. We assume these 2 lynx were from litters born in Colorado that were never found at dens (i.e., why there were no PIT-tags). All lynx captured for collaring or re-collaring were fitted with new Sirtrack TM dual VHF/satellite collars and re-released at their capture locations. Seven adult lynx were captured from March 1999-August 27, 2008 because they were in poor body condition (Table 6). Five of these lynx were successfully treated at the Frisco Creek Rehabilitation Center and re-released in the Core Release Area. One lynx, BC00F07, died from starvation and hypothermia within 1 day of capture at the rehabilitation center. Lynx QU04M07 died 3 days after capture at the rehabilitation center. Necropsy results documented starvation as the cause of death for this lynx that was precipitated by hydrocephalus and bronchopneumonia (unpublished data T. Spraker, CSUVTH). Seven lynx were captured (either by CDOW personnel or conservation personnel in other states) because they were in atypical habitat outside the state of Colorado (Table 6). They were held at Frisco Creek Rehabilitation Center for a minimum of 3 weeks, fitted with new Sirtrack TM dual VHF/satellite collars and re-released in the Core Release Area in Colorado. Five of these 7 lynx were still alive 6 months post-re-release but 3 had already dispersed out of Colorado and 1 stayed in Colorado through August 27, 2008. Two of these lynx died within 6 months of re-release: 1 died of starvation in Colorado and the other died of unknown causes in Nebraska. One lynx captured out of state and re-released currently remains in Colorado. HABITAT USE Landscape-scale daytime habitat use was documented from 9496 aerial locations of lynx collected from February 1999-June 30, 2007. Throughout the year Engelmann spruce - subalpine fir was the dominant cover used by lynx. A mix of Engelmann spruce, subalpine fir and aspen (Populus tremuloides) was the second most common cover type used throughout the year. Various riparian and riparian-mix areas were the third most common cover type where lynx were found during the daytime flights. Use of Engelmann spruce-subalpine fir forests and Engelmann spruce-subalpine fir-aspen forests was similar throughout the year. There was a trend in increased use of riparian areas beginning in July, peaking in November, and dropping off December through June. Site-scale habitat data collected from snow-tracking efforts indicate Engelmann spruce and subalpine fir were also the most common forest stands used by lynx for all activities during winter in southwestern Colorado. Comparisons were made among sites used for long beds, dens, travel and where they made kills. Little difference in aspect, mean slope and mean elevation were detected for 3 of the 4 site types including long beds, travel and kills where lynx typically use gentler slopes ( x = 15.7o ) at an mean elevation of 3173 m, and varying aspects with a slight preference for north-facing slopes. See Shenk (2006) for more detailed analyses of habitat use. 12 �DIET AND HUNTING BEHAVIOR Winter diet of lynx was documented through detection of kills found through snow-tracking. Prey species from failed and successful hunting attempts were identified by either tracks or remains. Scat analysis also provided information on foods consumed. A total of 548 kills were located from February 1999-April 2008. We collected over 950 scat samples from February 1999-April 2008 that will be analyzed for content. In each winter, the most common prey item was snowshoe hare, followed by red squirrel (Tamiusciurus hudsonicus; Table 7). The percent of snowshoe hare kills found however, varied annually from a low of 55.56% in 1999 to a high of 90.77% in winter 2002-2003. An annual mean of 73.29% (SE = 4.67) snowshoe hare kills in the diet has been documented. A comparison of percent overstory for successful and unsuccessful snowshoe hare chases indicated lynx were more successful at sites with slightly higher percent overstory, if the overstory species were Englemann spruce, subalpine fir or willow. Lynx were slightly less successful in areas of greater aspen overstory. This trend was repeated for percent understory at all 3 height categories except that higher aspen understory improved hunting success. Higher density of Engelmann spruce and subalpine fir increased hunting success while increased aspen density decreased hunting success. SNOWSHOE HARE ECOLOGY Two years of a 3-year study to evaluate snowshoe hare densities, demography and seasonal movement patterns among small and medium tree-sized lodgepole pine stands and mature spruce/fir stands have been completed and preliminary results presented (see Appendix I). DISCUSSION In an effort to establish a viable population of lynx in Colorado, CDOW initiated a reintroduction effort in 1997 with the first lynx released in winter 1999. From 1999 through spring 2007, 218 lynx were released in the Core Release Area. Locations of each lynx were collected through aerial- or satellite-tracking to document movement patterns and to detect mortalities. Most lynx remain in the high elevation, forested areas in southwestern Colorado. The use-density surfaces for lynx use in Colorado indicate two primary areas of use. The first is the Core Research Area (see Figure 1) and a secondary core centered in the Collegiate Peaks Wilderness (Figures 1, 3, 4). High use is also documented for 1) the area east of Dillon, on both the north and south sides of I70 and 2) the area north of Hwy 50 centered around Gunnison and then north to Crested Butte. These last 2 high use areas are smaller in extent than the 2 core areas. Dispersal movement patterns for lynx released in 2000 and subsequent years were similar to those of lynx released in 1999 (Shenk 2000). However, more animals released in 2000 and subsequent years remained within the Core Release Area than those released in 1999. This increased site fidelity may have been due to the presence of con-specifics in the area on release. Numerous travel corridors within Colorado have been used repeatedly by more than 1 lynx. These travel corridors include the Cochetopa Hills area for northerly movements, the Rio Grande Reservoir-Silverton-Lizardhead Pass for movements to the west, and southerly movements down the east side of Wolf Creek Pass to the southeast to the Conejos River Valley. Lynx appear to remain faithful to an area during winter months, and exhibit more extensive movements away from these areas in the summer. Reproductive females had the smallest 90% utilization distribution home ranges ( x = 75.2 km2, SE = 15.9 km2), followed by attending males ( x = 102.5 km2, SE = 39.7 km2) and non-reproductive animals ( x = 653.8 km2, SE = 145.4 km2). Most lynx currently being tracked are within the Core Release Area. During the summer months, lynx were documented to 13 �make extensive movements away from their winter use areas. Extensive summer movements away from areas used throughout the rest of the year have been documented in native lynx in Wyoming and Montana (Squires and Laurion 1999). Current data collection methods used for the Colorado lynx reintroduction program were not specifically designed to address the reintroduced lynx movements or use of areas in other states. In particular, the core research and release area were in Colorado. Therefore, the number of aerial locations obtained would be far fewer in other states than in Colorado which would bias low the number of lynx and intensity of lynx use documented outside the state. In contrast, obtaining satellite locations is not biased by the location of the lynx. Satellite locations are, however, biased by the shorter time the satellite transmitters function, approximately 18 months versus 60 months for the VHF transmitters used to obtain the aerial locations. However, data collected to meet objectives of the lynx reintroduction program were used to provide information to help address the question of lynx use outside of Colorado. Due to the rarity of flights conducted outside Colorado, only use-density surfaces generated from satellite locations were used to document relative lynx use of areas in New Mexico, Utah and Wyoming. New Mexico and Wyoming have been used continuously by lynx since the first year lynx were released in Colorado (1999) to the present. Lynx reintroduced in Colorado were first documented in Utah in 2000 and are still being documented there to date. In addition, all levels of lynx use-density documented throughout Colorado are also represented in New Mexico, Utah and Wyoming from none to the highest level of use (Shenk 2007). One den was found in Wyoming. Although no reproduction has been documented in New Mexico or Utah to date, documenting areas of the highest intensity of use and the continuous presence of lynx within these states for over six years does suggest the potential for yearround residency of lynx and reproduction in those states. From 1999-August 2008, there were 112 mortalities of released adult lynx. Human-caused mortality factors are currently the highest causes of death with approximately 30.4% attributed to collisions with vehicles or gunshot. Starvation and disease/illness accounted for 18.8% of the deaths while 36.6% of the deaths were from unknown causes. Lynx mortalities were documented throughout all areas lynx used, including 30 (26.8%) occurring in other states (Figure 2, Table 3). Nearly half (14 of 30) of the out-of-state mortalities were documented in New Mexico. Monthly mortality rate was lower inside the study area than outside, and slightly higher for male than for female lynx, although 95% confidence intervals for sexes overlapped. Mortality was higher immediately after release (first month = 0.0368 [SE = 0.0140] inside the study area, and 0.1012 [SE = 0.0359] outside the study area), and then decreased according to a quadratic trend over time. Reproduction is critical to achieving a self-sustaining viable population of lynx in Colorado. Reproduction was first documented from the 2003 reproduction season and again in 2004, 2005 and 2006. Lower reproduction occurred in 2006 (Table 5) but did include a Colorado-born female giving birth to 2 kittens, documenting the first recruitment of Colorado-born lynx into the Colorado breeding population. No reproduction was documented in 2007 or 2008. The cause of the decreased reproduction from 2006 08 is unknown. One possible explanation would be a decrease in prey abundance. Additional reproduction is likely to have occurred in all years from females we were no longer tracking, and from Colorado-born lynx that have not been collared. The dens we find are more representative of the minimum number of litters and kittens in a reproduction season. To achieve a viable population of lynx, enough kittens need to be recruited into the population to offset the mortality that occurs in that year and hopefully even exceed the mortality rate to achieve an increasing population. The use-density surfaces depict intensity of use by location. Why certain areas would be used more intensively than others should be explained by the quality of the habitat in those areas. 14 �Characteristics of areas used by lynx, as documented through aerial locations and snow-tracking of lynx in the Colorado core research area, include mature Engelmann spruce-subalpine fir forest stands with 4265% canopy cover and 15-20% conifer understory cover (Shenk 2006). Within these forest stand types, lynx appear to have a slight preference for north-facing, moderate slopes ( x = 15.7°) at high elevations ( x = 3173 m; Shenk 2006). Snow-tracking of released lynx also provided information on hunting behavior and diet through documentation of kills, food caches, chases, and diet composition estimated through prey remains. The primary winter prey species (n = 548) were snowshoe hare (Table 7) with an annual x = 73.3% (SE = 4.7, n = 10) and red squirrel (annual x = 18.2%, SE = 4.2, n = 10). Thus, areas of good habitat must also support populations of snowshoe hare and red squirrel. In winter, lynx reintroduced to Colorado appear to be feeding on their preferred prey species, snowshoe hare and red squirrel in similar proportions as those reported for northern lynx during lows in the snowshoe hare cycle (Aubry et al. 1999). Environmental conditions in the springs and summers of 2003 and 2006 resulted in high cone crops during their following winters based on field observations, resulting in increased red squirrel abundance. This may partially explain the higher percent of red squirrel kills, and thus a lower percent of snowshoe hare kills, found in winters 2003-04 and 2006-07 (Table 7). Caution must be used in interpreting the proportion of identified kills. Such a proportion ignores other food items that are consumed in their entirety and thus are biased towards larger prey and may not accurately represent the proportion of smaller prey items, such as microtines, in lynx winter diet. Through snow-tracking we have evidence that lynx are mousing and several of the fresh carcasses have yielded small mammals in the gut on necropsy. The summer diet of lynx has been documented to include less snowshoe hare and more alternative prey than in winter (Mowat et al., 1999). All evidence suggests reintroduced lynx are finding adequate food resources to survive. Mowat et al. (1999) suggest lynx and snowshoe hare select similar habitats except that hares select more dense stands than lynx. Very dense understory limits hunting success of the lynx and provides refugia for hares. Given the high proportion of snowshoe hare in the lynx diet in Colorado, we might then assume the habitats used by reintroduced lynx also depict areas where snowshoes hare are abundant and available for capture by lynx in Colorado. From both aerial locations taken throughout the year and from the site-scale habitat data collected in winter, the most common areas used by lynx are in stands of Engelmann spruce and subalpine fir. This is in contrast to adjacent areas of Ponderosa pine, pinyon juniper, aspen and oakbrush. The lack of lodgepole pine in the areas used by the lynx may be more reflective of the limited amount of lodgepole pine in southwestern Colorado, the Core Release Area, rather than avoidance of this tree species. Hodges (1999) summarized habitats used by snowshoe hare from 15 studies as areas of dense understory cover from shrubs, stands that are densely stocked, and stands at ages where branches have more lateral cover. Species composition and stand age appears to be less correlated with hare habitat use than is understory structure (Hodges 1999). The stands need to be old enough to provide dense cover and browse for the hares and cover for the lynx. In winter, the cover/browse needs to be tall enough to still provide browse and cover in average snow depths. Hares also use riparian areas and mature forests with understory. Site-scale habitat use documented for lynx in Colorado indicate lynx are most commonly using areas with Engelmann spruce understory present from the snow line to at least 1.5 m above the snow. The mean percent understory cover within the habitat plots is typically less than 15% regardless of understory species. However, if the understory species is willow, percent understory cover is typically double that, with mean number of shrubs per plot approximately 80, far greater than for any other understory species. 15 �In winter, hares browse on small diameter woody stems (<0.25"), bark and needles. In summer, hares shift their diet to include forbs, grasses, and other succulents as well as continuing to browse on woody stems. This shift in diet may express itself in seasonal shifts in habitat use, using more or denser coniferous cover in winter than in summer. The increased use of riparian areas by lynx in Colorado from July to November may reflect a seasonal shift in hare habitat use in Colorado. Major (1989) suggested lynx hunted the edge of dense riparian willow stands. The use of these edge habitats may allow lynx to hunt hares that live in habitats normally too dense to hunt effectively. The use of riparian areas and riparian-Engelmann spruce-subalpine fir and riparian-aspen mixes documented in Colorado may stem from a similar hunting strategy. However, too little is known about habitat use by hares in Colorado to test this hypothesis at this time. Lynx also require sufficient denning habitat. Denning habitat has been described by Koehler (1990) and Mowat et al. (1999) as areas having dense downed trees, roots, or dense live vegetation. We found this to be in true in Colorado as well (Shenk 2006). In addition, the dens used by reintroduced lynx were at high elevations and on steep north-facing slopes. All females that were documented with kittens denned in areas within their winter-use area. SUMMARY From results to date it can be concluded that CDOW developed release protocols that ensure high initial post-release survival of lynx, and on an individual level, lynx demonstrated they can survive longterm in areas of Colorado. We also documented that reintroduced lynx exhibited site fidelity, engaged in breeding behavior and produced kittens that were recruited into the Colorado breeding population. What is yet to be demonstrated is whether current conditions in Colorado can support the recruitment necessary to offset annual mortality in order to sustain the population. Monitoring of reintroduced lynx will continue in an effort to document such viability. ACKNOWLEDGMENTS The lynx reintroduction program involves the efforts of literally hundreds of people across North America, in Canada and USA. Any attempt to properly acknowledge all the people who played a role in this effort is at risk of missing many people. The following list should be considered to be incomplete. CDOW CLAWS Team (1998-2001): Bill Andree, Tom Beck, Gene Byrne, Bruce Gill, Mike Grode, Rick Kahn (Program Leader), Dave Kenvin, Todd Malmsbury, Jim Olterman, Dale Reed, John Seidel, Scott Wait, Margaret Wild. CDOW: John Mumma (Director 1996-2000), Russell George (Director 2001-2003), Bruce McCloskey (Director 2004-2007), Conrad Albert, Jerry Apker, Laurie Baeten, Cary Carron, Don Crane, Larry DeClaire, Phil Ehrlich, Lee Flores, Delana Friedrich, Dave Gallegos, Juanita Garcia, Drayton Harrison, Jon Kindler, Ann Mangusso, Jerrie McKee, Gary Miller, Melody Miller, Mike Miller, Kirk Navo, Robin Olterman, Jerry Pacheo, Mike Reid, Tom Remington, Ellen Salem, Eric Schaller, Mike Sherman, Jennie Slater, Steve Steinert, Kip Stransky, Suzanne Tracey, Anne Trainor, Scott Wait, Brad Weinmeister, Nancy Wild, Perry Will, Lisa Wolfe, Brent Woodward, Kelly Woods, Kevin Wright. Lynx Advisory Team (1998-2001): Steve Buskirk, Jeff Copeland, Dave Kenny, John Krebs, Brian Miller (Co-Leader), Mike Phillips, Kim Poole, Rich Reading (Co-Leader), Rob Ramey, John Weaver. U. S. Forest Service: Kit Buell, Joan Friedlander, Dale Gomez, Jerry Mastel, John Squires, Fred Wahl, Nancy Warren. U. S. Fish and Wildlife Service: Lee Carlson, Gary Patton (1998-2000), Kurt Broderdorp. 16 �State Agencies: Alaska: ADF&G: Cathie Harms, Mark Mcnay, Dan Reed (Regional Manager), Wayne Reglin (Director), Ken Taylor (Assist. Director), Ken Whitten, Randy Zarnke, Other:Ron Perkins (trapper), Dr. Cort Zachel (veterinarian). Washington: Gary Koehler. National Park Service: Steve King. Colorado State University: Alan Franklin, Gary White. Colorado Natural Heritage Program: Rob Schorr, Mike Wunder. Canada: British Columbia: Dr. Gary Armstrong (veterinarian), Mike Badry (government), Paul Blackwell (trapper coordinator), Trappers: Dennis Brown, Ken Graham, Tom Sbo, Terry Stocks, Ron Teppema, Matt Ounpuu. Yukon: Government: Arthur Hoole (Director), Harvey Jessup, Brian Pelchat, Helen Slama, Trappers: Roger Alfred, Ron Chamber, Raymond Craft, Lance Goodwin, Jerry Kruse, Elizabeth Hofer, Jurg Hofer, Guenther Mueller (YK Trapper‘s Association), Ken Reeder, Rene Rivard (Trapper coordinator), Russ Rose, Gilbert Tulk, Dave Young. Alberta: Al Cook. Northwest Territories: Albert Bourque, Robert Mulders (Furbearer Biologist), Doug Steward (Director NWT Renewable Res.), Fort Providence Native People. Quebec: Luc Farrell, Pierre Fornier. Colorado Holding Facility: Herman and Susan Dieterich, Kate Goshorn, Loree Harvey, Rachel Riling. Pilots: Dell Dhabolt, Larry Gepfert, Al Keith, Jim Olterman, Matt Secor, Brian Smith, Whitey Wannamaker, Steve Waters, Dave Younkin. Field Crews (1999-2007): Steve Abele, Brandon Barr, Bryce Bateman, Todd Bayless, Nathan Berg, Ryan Besser, Jessica Bolis, Mandi Brandt, Brad Buckley. Patrick Burke, Braden Burkholder, Paula Capece, Stacey Ciancone, Doug Clark, John DePue, Shana Dunkley, Tim Hanks, Carla Hanson, Dan Haskell, Nick Hatch, Matt Holmes, Andy Jennings, Susan Johnson, Paul Keenlance, Patrick Kolar, Tony Lavictoire, Jenny Lord, Clay Miller, Denny Morris, Kieran O‘Donovan, Gene Orth, Chris Parmater, Jake Powell, Jeremy Rockweit, Jenny Shrum, Josh Smith, Heather Stricker, Adam Strong, Dave Unger, David Waltz, Andy Wastell, Mike Watrobka, Lyle Willmarth, Leslie Witter, Kei Yasuda, Jennifer Zahratka. Research Associates: Bob Dickman, Grant Merrill. Data Analysts: Karin Eichhoff, Joanne Stewart, Anne Trainor. Data Entry: Charlie Blackburn, Patrick Burke, Rebecca Grote, Angela Hill, Mindy Paulek. Mary Schuette and Dave Theobald provided assistance with the GIS analysis and M. Schuette generated the maps used in this report Photographs: Tom Beck, Bruce Gill, Mary Lloyd, Rich Reading, Rick Thompson. Funding: CDOW, Great Outdoors Colorado (GOCO), Turner Foundation, U.S.D.A. Forest Service, Vail Associates, Colorado Wildlife Heritage Foundation. LITERATURE CITED Aubry, K. B., G. M. Koehler, J. R. Squires. 1999. Ecology of Canada lynx in southern boreal forests. Pages 373-396 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. Bartmann, R. M., and Byrne, G. (2001) Analysis and critique of the 1998 snowshoe hare pellet survey. Colorado Division of Wildlife Report No. 20. Fort Collins, Colorado. Byrne, G. 1998. Core area release site selection and considerations for a Canada lynx reintroduction in Colorado. Report for the Colorado Division of Wildlife. Devineau, O., T. M. Shenk, G. C. White, P. F. Doherty Jr., P. M. Lukacs, and R. H. Kahn. 2008. Estimating mortality for a widely dispersing reintroduced carnivore, the Canada lynx (Lynx canadensis). Ecology (in review). Hodges, K. E. 1999. Ecology of snowshoe hares in southern boreal and montane forests. Pages 163206 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires editors. Ecology and Conservation of Lynx in the United States. General 17 �Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. Koehler, G. M. 1990. Population and habitat characteristics of lynx and snowshoe hares in north central Washington. Canadian Journal of Zoology 68:845-851. Kolbe, J. A., J. R. Squires, T. W. Parker. 2003. An effective box trap for capturing lynx. Journal of Wildlife Management 31:980-985. Major, A. R. 1989. Lynx, Lynx canadensis canadensis (Kerr) predation patterns and habitat use in the Yukon Territory, Canada. M. S. Thesis, State University of New York, Syracuse. Mowat, G., K. G. Poole, and M. O‘Donoghue. 1999. Ecology of lynx in northern Canada and Alaska. Pages 265-306 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. Poole, K. G., G. Mowat, and B. G. Slough. 1993. Chemical immobilization of lynx. Wildlife Society Bulletin 21:136-140. Shenk, T. M. 1999. Program Narrative Study Plan: Post-release monitoring of reintroduced lynx (Lynx canadensis) to Colorado. Colorado Division of Wildlife, Fort Collins, Colorado. __________. 2001. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report, July: 7- 34. Colorado Division of Wildlife, Fort Collins, Colorado. __________. 2006. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report, July: 1-45. Colorado Division of Wildlife, Fort Collins, Colorado. __________. 2007. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report, July: 1-57. Colorado Division of Wildlife, Fort Collins, Colorado. Silverman, B.W. 1986. Density Estimation for Statistics and Data Analysis. Chapman and Hall. New York, New York, USA. Squires, J. R. and T. Laurion. 1999. Lynx home range and movements in Montana and Wyoming: preliminary results. Pages 337-349 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University Press of Colorado, Boulder, Colorado. U. S. Fish and Wildlife Service. 2000. Endangered and threatened wildlife and plants: final rule to list the contiguous United States distinct population segment of the Canada lynx as a threatened species. Federal Register 65, Number 58. Wild, M. A. 1999. Lynx veterinary services and diagnostics. Wildlife Research Report, Colorado Division of Wildlife, Fort Collins, Colorado. Zahratka, J. L. and T. M. Shenk. 2008. Population estimates of snowshoe hares in the southern Rocky Mountains. Journal of Wildlife Management 72: 906-912. Prepared by ___________________________________ Tanya M. Shenk, Wildlife Researcher 18 �Table 1. Number of wild-caught male (M) and female (F) Canada lynx (Lynx canadensis) from Alaska (AK) and Canada (BC = British Columbia, MB = Manitoba, QU = Quebec and YK = Yukon) released in southwestern Colorado per year from 1999–2006. State / Province of Origin Total Year %Released Sex AK BC MB QU YK 1999 19 2000 25 2003 15 2004 17 2005 17 2006 6 Total F 13 5 4 22 M 7 6 6 19 F 6 9 20 35 M 4 9 7 20 F 10 7 17 M 10 5 16 F 7 10 17 M 13 7 20 F 4 M 9 F M 30 1 3 8 3 18 8 3 20 4 3 7 5 2 7 48 218 91 4 45 Table 2. Status of adult Canada lynx (Lynx canadensis) reintroduced to Colorado as of August 27, 2008. Females Lynx Males Unknown TOTALS Released 115 103 218 Known Dead 62 49 1 112 Possible Alive 53 54 106 Missing 27 35 61a Monitoring/tracking 26 19 45 a 1 is unknown mortality Table 3. Causes of death for all Canada lynx (Lynx canadensis) released into southwestern Colorado 1999-2006 as of August 27, 2008. Mortalities Cause of Death Total (%) In Colorado (%) Outside Colorado (%) Unknown 41 (36.6) 27 (32.91) 14 (46.7) Gunshot 15 (13.4) 9 (11.0) 6 (20.0) Hit by Vehicle 14 (12.5) 9 (11.0) 5 (16.7) Starvation 11 (9.8) 10 (12.2) 1 (3.3) Other Trauma 8 (7.1) 7 (8.5) 1 (3.3) Plague 7 (6.3) 7 (8.5) 0 (0) Probable Gunshot 5 (4.5) 4 (4.9) 1 (3.3) Predation 5 (4.5) 5 (6.1) 0 (0) Probable Predation 3 (2.7) 2 (2.4) 1 (3.3) Illness 3 (2.7) 2 (2.4) 1 (3.3) Total Mortalities 112 82 (73.2) 30 (26.8) 19 �Table 4. Known lynx mortalities (n = 30) and causes of death documented by state outside of Colorado from February 1999 – August 27, 2008. Lynx ID State AK99F8 Unknown AK99M11 YK99M06 AK99F13 YK00F04 BC99M04 QU05M01 QU04F05 QU03F07 BC00M04 YK06F01 BC03M08 BC06F07 AK99M06 AK99M01 QU05M08 MB05F02 BC00F14 QU04F07 BC06M10 QU04F02 AK00M03 QU05M03 YK06M01 YK00F07 YK99F01 YK00M03 YK05M03 YK05M02 New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico Nebraska Nebraska Nebraska Nebraska Wyoming Wyoming Wyoming Wyoming Utah Utah Utah Utah Arizona Kansas Montana Iowa Date Mortality Recorded 7/30/1999 2000 1/27/2000 6/19/2000 6/22/2000 4/20/2001 6/7/2002 8/22/2005 8/26/2005 9/15/2005 7/19/2006 10/19/2006 10/19/2006 1/8/2007 11/16/1999 1/11/2005 10/1/2006 2/13/2007 7/28/2004 9/21/2004 8/15/2006 3/14/2007 7/2/2001 10/26/2005 12/4/2006 8/6/2007 9/15/2005 9/30/2005 11/8/2005 8/6/2007 Cause of Death Starvation Hit by Vehicle Unknown Probable Gunshot Unknown Gunshot Gunshot Unknown Hit by Vehicle Unknown Unknown Unknown Unknown Gunshot Gunshot Snared (Other Trauma) Unknown Gunshot Unknown Unknown Vehicle Collision Unknown Unknown Unknown Unknown Unknown Gunshot Vehicle Collision Unknown Vehicle Collision Table 5. Lynx reproduction summary statistics for 1999-2008. No reproduction was expected in 1999 because it was the first year of lynx releases and most animals were released after breeding season. Year Females Tracked 2000 2001 2002 2003 2004 2005 2006 2007 2008 TOTAL 9 25 21 17 26 40 42 34 28 Dens Found in May/June 0 0 0 6 11 17 4 0 0 Percent Tracked Females with Kittens 0.0 0.0 0.0 0.353 0.462 0.425 0.095 0.0 0.0 Additional Litters Found in Winter 0 0 0 0 2 1 0 0 0 20 Mean Kittens/Litte r (SE) 2.67 (0.33) 2.83 (0.24) 2.88 (0.18) 2.75 (0.47) Total Kittens Found Sex Ratio M/F (SE) 0 0 0 16 39 50 11 1.0 1.5 0.8 1.2 0 0 116 1.14 (0.14) �Table 6. Lynx captured because they were in poor body condition or were in atypical habitat and their fates 6 months post re-release and as of August 28, 2008. Lynx ID BC99F6 Date of Capture 3/25/1999 State Where Captured Colorado Reason For Capture Poor body condition Date of Re-release 5/28/1999 Status 6 Months Post Re-release Dead AK99M9 3/24/2000 Colorado 5/3/2000 Missing AK99F2 4/18/2000 Colorado 5/22/2000 BC00F7 2/11/2001 Colorado Alive in Colorado Dead BC00M13 3/21/2001 Wyoming BC03M08 9/5/2003 Colorado QU04M07 2/2/2006 Colorado Poor body condition Poor body condition Poor body condition Poor body condition Poor body condition Poor body condition BC04M01 11/5/2004 Utah QU04F02 4/10/2005 Nebraska QU05M08 11/25/2005 Wyoming QU04M04 12/5/2006 Utah YK00F07 12/12/2006 Utah YK05M02 1/1/2007 Kansas BC04M08 1/22/2007 Wyoming N/A 4/24/2001 1/1/2004 N/A Alive in Colorado Alive in Colorado Dead Atypical habitat Atypical habitat 12/5/2004 Atypical habitat Atypical habitat Atypical habitat Atypical habitat Atypical habitat 4/18/2006 Dead 1/20/2007 1/20/2007 Dead in Colorado Alive in Utah 2/2/2007 Alive in Iowa 2/15/2007 Alive in Colorado 5/7/2005 Alive in Colorado Alive in Wyoming Current Status Died 7/19/1999 in Colorado from vehicle collision Last located 5/3/2000, collar failure Last located 7/30/2003 in Colorado Died at Rehab Center on 2/12/2001 Last located 10/26/2004 in Colorado Died in New Mexico of unknown causes 10/19/06 Died at Rehab Center on 2/5/2006 from hydrocephalous and pneumonia In Colorado as of 8/27/2008 Died 3/14/2007 in Wyoming (good habitat) of unknown causes Died of unknown causes in Nebraska 10/1/2006 Died of starvation in Colorado, found 3/19/07 Died in Utah of unknown causes 8/6/2007 Died in Iowa from vehicle collision 8/6/2007 Died in Colorado from gunshot 1/4/2008 Table 7. Number of kills found each winter field season through snow-tracking of lynx and percent composition of kills of the three primary prey species. Field Season 1999 1999-2000 2000-2001 2001-2002 2002-2003 2003-2004 2004-2005 2005-2006 2006-2007 2007-2008 Total/Mean n 9 83 89 54 65 37 78 50 41 42 548 Snowshoe Hare 55.56 67.47 67.42 90.74 90.77 67.57 83.33 90.00 61.00 59.00 73.29 (SE=4.7) Prey (%) Red Squirrel Cottontail 22.22 0 19.28 1.20 19.10 8.99 5.56 0 6.15 0 27.03 2.70 10.26 0 0.08 0 39.0 0 33.3 0 18.2 (SE=4.2) 1.29 (SE=0.95) 21 Other 22.22 12.05 4.49 3.70 3.08 2.70 6.41 0.02 0 7.4 6.21 (SE=2.22) �Figure 1. Lynx are monitored throughout Colorado and by satellite throughout the western United States. The lynx core release area, where all lynx were released, is located in southwestern Colorado. A lynx-established core use area has developed in the Taylor Park and Collegiate Peak area in central Colorado. 22 �Figure 2. All documented lynx locations (non-truncated datasets) obtained from either aerial (red circles) or satellite (yellow circles) tracking from February 1999 through August 27, 2008. All known lynx mortality locations (n = 112) are displayed as black stars. 23 �Figure 3. Use-density surface for lynx aerial locations (truncated dataset) in Colorado from September 1999-March 2007. 24 �Figure 4. Use-density surface for lynx satellite locations (truncated dataset) in Colorado from September 1999-March 2007. 25 �Colorado Division of Wildlife July 2007 - June 2008 APPENDIX I WILDLIFE RESEARCH REPORT State of: Cost Center: Work Package: Task No.: Colorado 3430 0670 2 Federal Aid Project No. N/A : : : : : Division of Wildlife Mammals Research Lynx Conservation Density, Demography, and Seasonal Movements Of Snowshoe Hare in Colorado Period Covered: July 1, 2007- June 30, 2008 Author: J. S. Ivan, Ph.D. Candidate, Colorado State University Personnel: Dr. T. Shenk of CDOW and Dr. G. C. White of Colorado State University. All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT A program to reintroduce the threatened Canada lynx (Lynx canadensis) into Colorado was initiated in 1997. Analysis of scat collected from winter snow tracking indicates that snowshoe hares (Lepus americanus) comprise 65–90% of the winter diet of reintroduced lynx. Thus, existence of lynx in Colorado and success of the reintroduction hinge at least partly on maintaining adequate and widespread hare populations. Beginning in July 2006, I initiated a study to assess the relative value of 3 stand types for providing hare habitat in Colorado. These types include mature, uneven-aged spruce/fir forests, sapling lodgepole pine forests (―s mall lodgepole‖), and pole-sized lodgepole pine forests (―m edium lodgepole‖). Estimates and comparisons of survival, recruitment, finite population growth rate, and maximum (late summer) and minimum (late winter) snowshoe hare densities for each stand will provide the metrics for assessing these stands. Thus far, snowshoe hare densities on the study area are low compared to densities reported elsewhere. Within the study area, hare densities during summer were highest in small lodgepole stands, followed by mature spruce/fir and medium lodgepole, respectively. This pattern was consistent through the first 2 summers of this project, although absolute hare densities declined considerably in summer 2007. Hare density in small and medium lodgepole stands equalized during both winters of the project. However, as with summer, overall density was much lower during the second winter compared to the first. Hare survival from summer to winter has been relatively high. However the single winter to summer estimate I have to date is quite low. Extension of this time series will help determine whether low winter to summer survival is typical or somehow related to the decline in density. 26 �WILDIFE RESEARCH REPORT DENSITY, DEMOGRAPHY, AND SEASONAL MOVEMENTS OF SNOWSHOE HARES IN COLORADO JACOB S. IVAN P. N. OBJECTIVE Assess the relative value of 3 stand types (mature spruce/fir, sapling lodgepole, pole-sized lodgepole) that purportedly provide high quality hare habitat by estimating survival, recruitment, finite population growth rate, and maximum (late summer) and minimum (late winter) snowshoe hare densities for each type. SEGMENT OBJECTIVES 1. Complete mark-recapture work across all replicate stands during late summer (mid-July through midSeptember) and winter (mid-January through March). 2. Obtain daily telemetry locations on radio-tagged hares for 10 days immediately after capture periods, as well as monthly between primary trapping sessions. 3. Locate, retrieve, and refurbish radio tags as mortalities occur. 4. Summarize initial sampling efforts and provide initial density estimates for Progress Reports for Colorado Division of Wildlife (CDOW). INTRODUCTION NEED A program to reintroduce the threatened Canada lynx (Lynx canadensis) into Colorado was initiated in 1997. Since that time, 218 lynx have been released in the state, and an extensive effort to determine their movements, habitat use, reproductive success, and food habits has ensued (Shenk 2005, Shenk 2007). Analysis of scat collected from winter snow tracking indicates that snowshoe hares (Lepus americanus) comprise 65–90% of the winter diet of reintroduced lynx (T. Shenk, Colorado Division of Wildlife, unpublished data). Thus, as in the far north where the intimate relationship between lynx and snowshoe hares has captured the attention of ecologists for decades, it appears that the existence of lynx in Colorado and success of the reintroduction effort may hinge on maintaining adequate and widespread populations of hares. Colorado represents the extreme southern range limit for both lynx and snowshoe hares (Hodges 2000). At this latitude, habitat for each species is less widespread and more fragmented compared to the continuous expanse of boreal forest at the heart of lynx and hare ranges. Neither exhibits dramatic cycles as occur farther north, and typical lynx ( 2 3 lynx/100km2; Aubry et al. 2000) and hare ( 1 2 hares/ha; Hodges 2000) densities in the southern part of their range correspond to cyclic lows form northern populations (2-30 lynx/100 km2, 1 16 hares/ha; Aubry et al. 2000, Hodges 2000, Hodges et al. 2001). Whereas extensive research on lynx-hare ecology has occurred in the boreal forests of Canada, literature regarding the ecology of these species in the southern portion of their range is relatively sparse. This scientific uncertainty is acknowledged in the ― Canada Lynx Conservation Assessment and Strategy,‖ a formal agreement between federal agencies intended to provide a consistent approach to lynx conservation on public lands in the lower 48 states (Ruediger et al. 2000). In fact, one of the explicit guiding principles of this document is to ― retain future options…until more conclusive information concerning lynx management is developed.‖ Thus, management recommendations in this agreement are decidedly conservative, especially with respect to timber management, and are applied broadly to cover 27 �all habitats thought to be of possible value to lynx and hare. Accurate identification and detailed description of lynx-hare habitat in the southern Rocky Mountains would permit more informed and refined management recommendations. A commonality throughout the snowshoe hare literature, regardless of geographic location, is that hares are associated with dense understory vegetation that provides both browse and protection from elements and predators (Wolfe et al. 1982, Litvaitis et al. 1985, Hodges 2000, Homyack et al. 2003, Miller 2005). In western mountains, this understory can be provided by relatively young conifer stands regenerating after stand-replacing fires or timber harvest (Sullivan and Sullivan 1988, Koehler 1990a, Koehler 1990b, Bull et al. 2005) as well as mature, uneven-aged stands (Beauvais 1997, Griffin 2004). Hares may also take advantage of seasonally abundant browse and cover provided by deciduous, open habitats (e.g., riparian willow [Salix spp.], aspen [Populus tremuloides]; Wolff 1980, Miller 2005). In drier portions of hare range, such as Colorado, regenerating stands can be relatively sparse, and hares may be more associated with mesic, late-seral forest and/or riparian areas than with young stands (Ruggiero et al. 2000). Numerous investigators have sought to determine the relative importance of these distinctly different habitat types with regards to snowshoe hare ecology. Most previous evaluations were based on hare density or abundance (Bull et al. 2005), indices to hare density and abundance (Wolfe et al. 1982, Koehler 1990a, Beauvais 1997, Miller 2005), survival (Bull et al. 2005), and/or habitat use (Dolbeer and Clark 1975). Each of these approaches provides insight into hare ecology, but taken singly, none provide a complete picture and may even be misleading. For example, extensive use of a particular habitat type may not accurately reflect the fitness it imparts on individuals, and density can be high even in ― sink‖ habitats (Van Horne 1983). A more informative approach would be to measure density, survival, and habitat use simultaneously in addition to recruitment and population growth rate through time. Griffin (2004) employed such an approach and found that summer hare densities were consistently highest in young, dense stands. However, he also noted that only dense mature stands held as many hares in winter as in summer. Furthermore hare survival seemed to be higher in dense mature stands, and only dense mature stands were predicted (by matrix projection) to impart a mean positive population growth rate on hares. Griffin‘s (2004) study occurred in the relatively moist forests of Montana, which share many similarities but also many notable differences with Colorado forests including levels of fragmentation, species composition, elevation, and annual precipitation. Density estimation is a key component in assessing the value of a particular stand type and is the common currency by which hare populations are compared across time and space. However, it can be a difficult metric to estimate accurately. Abundance estimation based on capture-recapture methods is a well-developed field (Otis et al. 1978, White et al. 1982), but is often too costly and labor intensive to be implemented on scales necessary to effectively monitor density over a biologically meaningful area. Also, density can be difficult to assess from grid-trapping efforts because it is often unclear how much area was effectively sampled by the grid (Williams et al. 2002:314). Alternate approaches can produce density estimates that differ by an order of magnitude even when calculated from the same data (Zahratka 2004). Indices such as pellet plot counts and distance sampling of pellet groups can be used to estimate density, but each of these has limitations as well (Krebs et al. 1987, Eriksson 2006). The study outlined below is designed principally to evaluate the importance of young, regenerating lodgepole pine (Pinus contorta) and mature Engelmann spruce (Picea engelmannii)/ subalpine fir (Abies lasiocarpa) stands in Colorado by examining density and demography of snowshoe hares that reside in each. Secondarily, I intend to quantify movement between these stands and other seasonally available types (e.g., willow). My hope is that information gathered from this research will be drawn upon as managers make routine decisions, leading to landscapes that include stands capable of 28 �supporting abundant populations of hares. I assume that if management agencies focus on providing habitat, hares will persist. I will use mark-recapture techniques as data from such an approach can provide information on both density and demography. In the future, I will address the ― effective trapping area‖ issue using a new approach that augments mark-recapture data with telemetry locations of animals using the grid. However, for this report I used one of the more popular, traditional techniques. I determined that 2 classes of young, regenerating lodgepole stands could both provide adequate hare habitat. Thus, in addition to older spruce/fir forests, I am sampling ―s mall‖ (2.54-12.69 cm dbh) and ―m edium‖ (12.70-22.85 cm dbh) stands regenerating from clearcutting that took place 20 and 40 years ago, respectively (Figure 1). Additionally, medium lodgepole stands were pre-commercially thinned 20 years ago; small lodgepole stands have not yet been thinned. Hypotheses 1) In general, snowshoe hare density in Colorado will be relatively low ( 0.5 hares/ha) compared to densities reported in northern boreal forests, even immediately post-breeding when an influx of juveniles will bolster hare numbers. 2) Snowshoe hare density will be consistently highest in small lodgepole pine stands, followed by large spruce/fir and medium lodgepole pine, respectively. 3) Survival will generally be highest in mature (large) spruce/fir stands followed by small and medium lodgepole pine, respectively. 4) Finite population growth rate will be consistently at or above 1.0 in mature spruce/fir stands with survival contributing most significantly to the growth rate. Finite growth rates for the lodgepole pine stands will be more variable. 5) Snowshoe hares will significantly shift their home ranges to make use of abundant food and cover provided by riparian willow (and/or aspen) habitats in summer. 6) Snowshoe hare density, survival, and recruitment will be highly correlated with understory cover and stem density. STUDY AREA The study area stretches from Taylor Park to Pitkin in central Colorado (Figure 2). Elevation ranges from 2700 m to 4000 m. Sagebrush (Artemisia spp.) dominates broad, low-lying valleys. Most montane areas are covered by even-aged, large-diameter lodgepole pine forests with sparse understory. Moist, north-facing slopes and areas near tree line are dominated by large-diameter Engelmann spruce/subalpine fir. Interspersed along streams and rivers are corridors of willow. Patches of aspen occur sporadically on southern exposures. This area was chosen over other potential study areas in the state because 1) it contained numerous examples of the 3 stand types of interest (more southern regions lack naturally occurring stands of lodgepole pine), 2) it was not subject to confounding effects of largescale mountain pine beetle outbreak as were more northern stands, and 3) an adequate number of radio frequencies were available to support a large study with hundreds of radio-tagged individuals. Within the study area I selected sample stands based on the following: Potential replicate stands were required to be 1) close enough geographically to minimize differences due to climate, weather, and topography, but are far enough apart to be considered independent, 2) adjacent to one or more riparian willow corridors, 3) within 1 km of an access road for logistical purposes, 4) of suitable size and shape to admit a 16.5-ha trapping grid, and 5) consistent in their management history (i.e., replicate lodgepole pine stands were clear-cut and/or thinned within 1-2 years of each other). I queried the U.S. Forest Service R2VEG GIS database using the criteria listed above to initially develop a suite of potential sample stands. I further narrowed this suite after obtaining updated standlevel information from local USFS personnel (Art Haines, Silviculturalist, USFS Gunnison Ranger 29 �District, personal communication). Finally, I ground-truthed potential stands and qualitatively assessed their representativeness and similarity to other potential replicates. Given the numerous constraints imposed, very few stands met all criteria. Thus, I was unable to randomly select sample stands from a population of suitable stands. Rather, I subjectively chose the ― best‖ stands from among the handful that met my criteria. Small lodgepole stands rarely occur on the landscape in patches large enough to fit a full trapping grid. To accommodate this, I sampled 6 replicate small lodgepole stands (rather than 3) using half-sized trapping grids. METHODS Experimental Design/Procedures Variables.--The response variables of interest for this project include stand-specific snowshoe hare density (D), apparent survival ( ), recruitment (f), finite population growth rate (λ), and a metric of seasonal movement. Density is the number of hares per unit area and is estimated using a conventional techniques in this report. The stand-specific demographic parameters will be estimated primarily from capture-mark-recapture methods. As such, apparent survival is defined as the probability that a marked animal alive and in the population at time i survives and is in the population at time i + 1. Apparent survival encompasses losses due to both death and emigration. Estimates of recruitment, population growth, and seasonal movement are forthcoming and not provided in this report. Potential explanatory variables for snowshoe hare density, demographics, and movement include general species composition and structural stage of each stand in which response variables are measured. Additionally, stem density, horizontal cover, and canopy cover (to a lesser extent) are highly correlated with snowshoe hare abundance and habitat use (Wolfe et al. 1982, Litvaitis et al. 1985, Hodges 2000, Zahratka 2004, Miller 2005). Thus, I further characterized vegetation in each stand by measuring stem density by size class (1-7 cm, 7.1-10 cm, and >10 cm), percent canopy cover, percent horizontal cover of understory and basal area. Basal area is an easily obtainable metric that may be correlated with the other variables and is recorded routinely during timber cruises, whereas the others are not. Thus, it might prove a useful link for biologists designing management strategies for snowshoe hare. Additionally, I recorded physical covariates such as ambient temperature, precipitation, and snow depth at each stand during sampling. These metrics were not included in the current preliminary analyses, but will be used as covariates in future models. Sampling.--All trapping and handling procedures have been approved by the Colorado State University Animal Care and Use Committee and filed with the Colorado Division of Wildlife. Snowshoe hares breed synchronously and generally exhibit 2 birth pulses in Colorado (although in some years, some individuals may have 3 litters), with the first pulse terminating approximately June 5 20 and the second approximately July 15–25 (Dolbeer 1972). To obtain a maximum density estimate, I began data collection on the first suite of sites immediately following the second birth pulse in late July. Along with a crew of 5 technicians, I deployed one 7 12 trapping grid (50-m spacing between traps; grid covers 16.5 ha) in the large spruce/fir and medium lodgepole stands within the first suite, along with 2 6 7 grids in 2 small lodgepole stands. Grid set up and trap deployment followed Griffin (2004) and Zahratka (2004). Grid locations and orientation within each stand were chosen subjectively to accommodate logistical constraints and to ensure that hares using the grid had ample opportunity to use adjacent riparian willow zones. As traps were deployed, they were locked open and ― pre-baited‖ with apple slices, hay cubes, and commercial rabbit chow. Traps were pre-baited in this manner for a total of 3 nights to maximize capture rates when trapping began. This minimized the number of trap-nights needed to capture the desired number of animals which in turn minimized trap-related injuries and minimized problems with predators keying into trap lines. During pilot work in winter 2005, I observed low but increasing capture rates (<0.20) during the first 3 nights of trapping, with higher, more stable capture 30 �probabilities after 3 days (approximately 0.35–0.45). Thus 3 days of pre-baiting seemed reasonable. Traps were set on the afternoon of the 4th day and checked early each morning and again in the evening on days 5–9. By checking traps in both morning and evening I prevented hares from being entrapped >13 hours, which should minimize capture stress. A crew of 2 people worked together on each grid to check traps and process captures as quickly as possible. All captured hares were coaxed out of the trap and into a dark handling bag by blowing quick shots of air on them from behind. Hares remained in the handling bag, physically restrained with their eyes covered, for the entire handling process. Each individual was aged, sexed, marked with a passive integrated transponder (PIT) tag and temporary ear mark (to track PIT tag retention), then released. Aging consisted of assigning each individual as either juvenile (<1 year old, <1000 g) or adult ( 1 year old, 1000 g) based on weight. This criterion is accurate through the end of September at which point juveniles are difficult to distinguish from adults (K. Hodges, University of British Columbia; P. Griffin, University of Montana, personal communication). After the first day of trapping, all captured hares were scanned for a PIT tag prior to any handling and those already marked were recorded and immediately released. Traps and bait were completely removed from the grid on day 10. In addition to PIT tags and ear marks, I radio collared up to 10 hares captured on each grid with a 28-g mortality-sensing transmitter (BioTrack, LTD) to facilitate unbiased density estimation as well as assessment of seasonal movements. I expected heterogeneity in snowshoe hare movements and use of the grid area, with potential bias surfacing due to location at which a hare is captured (e.g., hares captured on the edge of a grid may use the grid area differently than those captured at the center), and differential behavioral responses to trapping (e.g., young individuals may have lower capture probabilities and thus may be more likely to be captured on later occasions). To guard against the first potential bias, I randomly selected a starting trap location each morning and ran the grid systematically from that point. Thus, the first several hares encountered (and collared) were as likely to be from the inner part of the grid as from the edge. To protect against the second potential source of bias, I refrained from deploying the final 3 collars until days 4 and 5 of the trapping session. Immediately following the removal of traps, the field crew began work locating each radiocollared hare 1–2 times per day for 10 days. Most locations were obtained by triangulation from relatively close proximity, but some were obtained by ―hom ing‖ on a signal (Samuel and Fuller 1996, Griffin 2004) taking care not to push hares while approaching them. Because hares are largely nocturnal (Keith 1964, Mech et al. 1966, Foresman and Pearson 1999), I made an effort to conduct telemetry work at various times of the night (safety and logistics permitting) and day to gather a representative sample of locations for each hare. Crews gathered telemetry locations for radio-collared hares on the initial suite of sites for 10 days. Then the 10 day trapping procedure and 8 to 10 day telemetry work were repeated on the grids comprising suites 2 and 3(Figure 3). The entire process was repeated during the winter when densities should have been at a minimum. Thus, during the period covered by this report, sampling occurred from July 16 – September 14 and from January 20 – March 24, 2008. Sampling occurred across similar dates during FY06/07 and will continue during FY08/09. During the interim between intensive trapping and telemetry work, monthly telemetry checks were conducted from the air to track mortalities and facilitate retrieval of collars from dead hares. Telemetry work also occurred during ―pr e-baiting‖ days after the initial summer sampling session to determine which hares were still alive and immediately available to be sampled by the grid during the ensuing trapping period. Vegetation sampling at each stand commenced in June 2008 and is nearly finished. I followed protocols established through previous snowshoe hare and lynx work in Colorado (Zahratka 2004, T. 31 �Shenk, Colorado Division of Wildlife, personal communication). Specifically, on each of the 12 livetrapping grids, I laid out 5 5 grids (3-m spacing) of vegetation sampling points centered on 15 of the 84 trap locations (Figure 4; 9 points were sampled on each of the ½-sized small lodgepole stands). At each of the 25 vegetation sampling points, I recorded canopy cover (present or absent) using a densitometer. I quantified downed coarse wood along the center transect of the 25-point grid following Brown (1974). From the centerpoint (i.e., trap location) I measured 1) distance to the nearest woody stem 1.0 7.0 cm, 7.1 10.0 cm, and >10.0 cm in diameter at heights of 0.1 m and 1.0 m above the ground (to capture both summer and winter stem density; Barbour et al. 1999), 2) horizontal cover in 0.5-m increments above the ground up to 2 m (Nudds 1977), 3) basal area, and 4) slope. Data Analysis Density, Survival, and Population Growth.--I analyzed mark-recapture data in a robust design framework (Williams et al. 2002:523-554) treating summer and winter sampling occasions as primary periods, and the 5-day trapping sessions within each as secondary periods. As such, I assumed hare populations were demographically and geographically closed during the short 5-day mark-recapture sampling periods, but were open to immigration, emigration, births, and deaths between these occasions. I specified the Pradel Robust Design data type in Program MARK (White and Burnham 1999) and chose the Huggins closed capture model (Huggins 1989, 1991) to obtain abundance estimates for each grid from the secondary periods. I obtained estimates of apparent survival ( ˆ i ) between each primary period. I employed a technique known as ½ Mean Maxmimum Distance Moved (MMDM; Wilson and Anderson 1985) to calculate the effective area trapped and obtain a density estimate for each grid from each secondary period. Future density analyses will employ a new estimator that employs telemetry data to correct for bias (Ivan 2005). I used Akaike‘s Information Criterion corrected for small sample size (AICc; Burnham and Anderson 1998) to select appropriate models from alternatives that included all 8 closed capture models (Otis et al. 1978) in combination with models that allowed survival to be constant, vary with time, and/or vary with stand type. RESULTS AND DISCUSSION I captured 30 hares 73 times during July-September 2007. I captured 48 hares 71 times during January-March 2008. During summer, density estimates have thus far followed hypotheses 1) and 2) above (Figure 5). Specifically, hare densities were clearly highest in small lodgepole stands and quite low in medium lodgepole stands. Spruce/fir was intermediate in density. This pattern remained consistent between summer 2006 to summer 2007, although the absolute density of hares dropped considerably during summer 2007. Why this decline occurred is unclear, although disease outbreak, natural population cycles, and response to increased predation due to lynx reintroduction are possibilities. Note that even the highest densities recorded here correspond to low estimates observed in other parts of hare range (Hodges 2000). Hare densities tend to equalize in lodgepole stands during winter (Figure 5). I submit that the interplay between food, cover, and snow depth provides a plausible explanation for this pattern. Medium lodgepole stands apparently provide very little forage/cover for hares during summer as the canopy in these stands is generally ≥1 meter off the ground. However, in winter, accumulated snow may make that canopy available again to hares. Conversely, small lodgepole stands provide abundant food and cover during summer, but accumulated snow during winter brings hares closer to the crowns of the young trees, which then provide less cover. Spruce/fir stands probably provide adequate access to both food and cover during both summer and winter due to their uneven-aged, multi-layered structure. Like the summer estimates, density during the second winter was much lower than during the first winter. 32 �Hare survival from the first sampling season into the first winter was relatively high (Figure 6). However, survival from the first winter to the second summer declined drastically. Survival from the second summer to the second winter was again quite high. Whether this pattern is typical is unclear. Survival from winter to summer is commonly lower than from summer to winter. However, the low survival from the first winter to second summer is coincident with the dramatic decline in hare density observed on spruce/fir and small lodgepole grids. Thus, low survival for this period is possibly reflective of, or maybe even a driver for, the decline in density. Extension of the time series and a breakdown of survival by stand type should provide more evidence for one or the other of these explanations. SUMMARY Snowshoe hare densities on my study sites appear to be relatively low compared to densities reported elsewhere. Densities during summer were highest in small lodgepole stands, followed by spruce/fir and medium lodgepole. During winter, densities equalize in lodgepole stands, possibly due to the interplay between snow depth and canopy height in small and medium lodgepole pine. Hare density declined considerably beginning in summer 2007. Summer to winter hare survival has been consistently high thus far in the study, but the lone winter to summer survival estimate is quite low. It is unclear whether winter to summer survival is typically this low or whether that estimate is related to coincident drop in density. ACKNOWLEDGMENTS Ken Wilson (CSU), Bill Romme (CSU), Paul Doherty (CSU), Dave Freddy (CDOW), Chad Bishop (CDOW), and Paul Lukacs (CDOW) provided helpful insight on the design of this study. We appreciate the invaluable logistical support provided by Mike Jackson (USFS), Art Haines (USFS), Jake Spritzer (USFS), Kerry Spetter (USFS), Margie Michaels (CDOW), Gabriele Engler (USGS), Dana Winkelman (USGS), Brandon Diamond (CDOW), Chris Parmeter (CDOW), Kathaleen Crane (CDOW), Lisa Wolfe (CDOW), and Laurie Baeten (CDOW). Jim Gammonley (CDOW), Dave Freddy (CDOW), Chad Bishop (CDOW), Jack Vayhinger (CDOW), Brandon Diamond (CDOW) assisted with trucks and equipment. The following hardy individuals collected the hard-won data presented in this report: Braden Burkholder, Matt Cuzzocreo, Brian Gerber, Belita Marine, Adam Behney, Pete Lundberg, Katie Yale, Britta Shielke, Cory VanStratt, Mike Watrobka, Meredith Goss, Sidra Blake, Keith Rutz, Rob Saltmarsh, Jennie Sinclair, Evan Wilson, Mat Levine, Matt Strauser, Greg Davidson, Leah Yandow, Renae Sattler, and Caylen Cummins. Funding was provided by the Colorado Division of Wildlife. LITERATURE CITED Aubry, K. B., G. M. Koehler, and J. R. Squires. 2000. Ecology of Canada lynx in southern boreal forests. Pages 373-396 in Ruggiero, L. F., K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation of lynx in the United States. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Collins, Colorado, USA. Barbour, M. G., J. H. Burk, W. E. Pitts, F. S. Gilliam, and M. W. Schwartz. 1999. Terrestrial plant ecology, Addison Wesley Longman, Inc., Menlo Park, California, USA. Beauvais, G. P. 1997. Mammals in fragmented forests in the Rocky Mountains: community structure, habitat selection, and individual fitness. Dissertation, University of Wyoming, Laramie, Wyoming, USA. Brown, J. K. 1974. Handbook for inventorying downed woody material. U.S. Department of Agriculture, Forest Service, Intermountain Forest and Range Experiment Station, General Technical Report 33 �INT-16. Burnham, K. P., and D. R. Anderson. 1998. Model selection and inference: a practical informationtheoretic approach. Springer-Verlag, New York, New York, USA. Bull, E. L., T. W. Heater, A. A. Clark, J. F. Shepherd, and A. K. Blumton. 2005. Influence of precommercial thinning on snowshoe hares. U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station, General Technical Report PNW-RP-562. Dolbeer, R. A. and W. R. Clark. 1975. Population ecology of snowshoe hares in the central Rocky Mountains. Journal of Wildlife Management 39:535-549. Dolbeer, R. A. 1972. Population dynamics of snowshoe hare in Colorado. Dissertation, Colorado State University, Fort Collins, Colorado, USA. Eriksson, H. M. 2006. Snowshoe hare densities in post-fire vegetation. Thesis, University of Alaska Fairbanks, Fairbanks, Alaska, USA. Foresman, K. R. and D. E. Pearson. 1999. Activity patterns of American martens, Martes americana, snowshoe hares, Lepus americanus, and red squirrels, Tamiasciurus hudsonicus, in westcentral Montana. The Canadian Field-Naturalist 113:386-389. Griffin, P. C. 2004. Landscape ecology of snowshoe hares in Montana. Dissertation, University of Montana, Missoula, Montana, USA. Hodges, K. E. 2000. Ecology of snowshoe hares in southern boreal and montane forests. Pages 163-206 in Ruggiero, L. F., K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation of lynx in the United States. University Press of Colorado, Boulder, Colorado, USA. Hodges, K. E., C. J. Krebs, D. S. Hik, C. I. Stefan, E. A. Gillis, and C. E. Doyle. 2001. Snowshoe hare demography. Pages 141-178 in Krebs, C. J., S. Boutin, and R. Boonstra, editors. Ecosystem dynamics of the boreal forest: the Kluane project. Oxford University Press, New York, New York, USA. Homyack, J. A., D. J. Harrison, and W. B. Krohn. 2003. Effects of precommercial thinning on select wildlife species in northern Maine, with special emphasis on snowshoe hare. Maine Cooperative Fish and Wildlife Research Unit, Orono, Maine, USA. Huggins, R. M. 1989. On the statistical analysis of capture-recapture experiments. Biometrika 76:133140. Huggins, R. M. 1991. Some practical aspects of a conditional likelihood approach to capture experiments. Biometrics 47:725-732. Ivan, J. S. 2006. Density, demography, and seasonal movements of snowshoe hares in Colorado. Program narrative study plan for mammals research. Pages 27-45 in T.M. Shenk. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report, July: 1-45. Colorado Division of Wildlife, Fort Collins, Colorado, USA. Keith, L. B. 1964. Daily activity pattern of snowshoe hares. Journal of Mammalogy 45:626-627. Koehler, G. M. 1990a. Snowshoe hare, Lepus americanus, use of forest successional stages and population changes during 1985-1989 in North-central Washington. Canadian Field-Naturalist 105:291-293. Koehler, G. M. 1990b. Population and habitat characteristics of lynx and snowshoe hares in north central Washington. Canadian Journal of Zoology 68:845-851. Krebs, C. J., B. Scott Gilbert, S. Boutin, and e. al. R. Boonstra. 1987. Estimation of snowshoe hare population density from turd transects. Canadian Journal of Zoology 65:565-567. Litvaitis, J. A., J. A. Sherburne, and J. A. Bissonette. 1985. Influence of understory characteristics on snowshoe hare habitat use and density. Journal of Wildlife Management 49:866-873. Mech, L. D., K. L. Heezen, and D. B. Siniff. 1966. Onset and cessation of activity in cottontail rabbit and snowshoe hares in relation to sunset and sunrise. Animal Behaviour 14:410-413. Miller, M. A. 2005. Snowshoe hare habitat relationships in northwest Colorado. Thesis, Colorado State University, Fort Collins, Colorado, USA. Nudds, T. D. 1977. Quantifying the vegetative structure of wildlife cover. Wildlife Society Bulletin 34 �5:113-117. Otis, D. L., K. P. Burnham, G. C. White, and D. R. Anderson. 1978. Statisical inference from capture data on closed animal populations. Wildlife Monographs 62: Ruediger, B., J. Claar, S. Gniadek, B. Holt, Lewis Lyle, S. Mighton, B. Naney, G. Patton , T. Rinaldi, J. Trick, A. Vendehey, F. Wahl, N. Warren, D. Wenger, and A. Williamson. 2000. Canada lynx conservation assessment and strategy. U.S. Department of Agriculture, Forest Service, U.S. Department of Interior, Fish and Wildlife Service, Bureau of Land Management, National Park Service R1-00-53 U.S. Department of Agriculture, Forest Service, U.S. Department of Interior, Fish and Wildlife Service, Bureau of Land Management, National Park Service, Missoula, Montana, USA. Ruggiero, L. F., K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires. 2000. The scientific basis for lynx conservation: qualified insights. Pages 443-454 in Ruggiero, L. F., K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation of lynx in the United States. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Collins, Colorado, USA. Samuel, M. D. and M. R. Fuller. 1996. Wildlife radiotelemetry. Pages 370-418 in Bookhout, T. A., editors. Research and Management Techniques for Wildlife and Habitats. Allen Press, Inc., Lawrence, Kansas, USA. Shenk, T. M. 2005. General locations of lynx (Lynx canadensis) reintroduced to southwestern Colorado from February 4, 1999 through February 1, 2005. Colorado Division of Wildlife Colorado Division of Wildlife, Fort Collins, Colorado, USA. Shenk, T. M. 2007. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report, July: 1-57. Colorado Division of Wildlife, Fort Collins, Colorado USA. Sullivan, T. P. and D. S. Sullivan. 1988. Influence of stand thinning on snowshoe hare population dynamics and feeding damage in lodgepole pine forest. Journal of Applied Ecology 25:791-805. Van Horne, B. 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47:893-901. White, G. C., D. R. Anderson, K. P. Burnham, and D. L. Otis. 1982. Capture-recapture and removal methods for sampling closed populations. Los Alamos National Laboratory, Los Alamos, New Mexico, USA. White, G. C., and K. P. Burnham. 1999. Program MARK: survival estimation from populations of marked animals. Bird Study 46 Supplement:120-138. Williams, B. K., J. D. Nichols, and M. J. Conroy. 2002. Analysis and management of animal populations, Academic Press, San Diego, California, USA. Wilson, K. R. and D. R. Anderson. 1985. Evaluation of two density estimators of small mammal population size. Journal of Mammalogy 66:13-21. Wolfe, M. L., N. V. Debyle, C. S. Winchell, and T. R. McCabe. 1982. Snowshoe hare cover relationships in northern Utah. Journal of Wildlife Management 46:662-670. Wolff, J. O. 1980. The role of habitat patchiness in the population dynamics of snowshoe hares. Ecological Monographs 50:111-130. Zahratka, J. L. 2004. The population and habitat ecology of snowshoe hares (Lepus americanus) in the southern Rocky Mountains. Thesis, The University of Wyoming, Laramie, Wyoming, USA. Prepared by _________________________________________________ Jacob S. Ivan, Graduate Student, Colorado State University 35 �Figure 1. Purported high quality snowshoe hare habitat in Colorado. From left to right: small lodgepole pine, medium lodgepole pine, and large Engelmann spruce/subalpine fir. Figure 2. Study area near Taylor Park and Pitkin, Colorado including medium lodgepole (squares), small lodgepole (circles), and spruce/fir (triangles) stands selected for mark-recapture sampling. 36 �Jul Aug Sep 1 Jan 1 2 3 4 2 3 4 5 6 7 8 5 6 7 8 9 10 11 9 10 11 12 13 14 15 12 13 14 15 16 17 18 16 17 18 19 20 21 22 19 20 21 22 23 24 25 23 24 25 26 27 28 29 26 27 28 29 30 31 1 30 31 1 2 3 4 5 2 3 4 5 6 7 8 6 7 8 9 10 11 12 9 10 11 12 13 14 15 13 14 15 16 17 18 19 16 17 18 19 20 21 22 20 21 22 23 24 25 26 23 24 25 26 27 28 1 27 28 29 30 31 1 2 2 3 4 5 6 7 8 Feb Mar 3 4 5 6 7 8 9 9 10 11 12 13 14 15 10 11 12 13 14 15 16 16 17 18 19 20 21 22 17 18 19 20 21 22 23 23 24 25 26 27 28 29 24 25 26 27 28 29 30 30 31 Figure 3. Approximate annual data collection schedule for trapping () and telemetry (). Dates and weeks changed depending on calendar year and pay schedule. During telemetry work, the 6-person crew was divided into 2 teams, only one of which worked at any given time. Monthly locations on radiocollared hares were also collected in the interim between the intensive sampling periods indicated here. Figure 4. 15 trap locations ( ) on 7 12 trapping grid where vegetation was sampled by measuring stem density, horizontal cover, downed woody material, and basal area. Additionally, the 25-point grid superimposed on each of the 15 trap locations (inset) was used to quantify canopy covert). 37 �Figure 5. Snowshoe hare density and 95% confidence intervals in 3 types of stands in central Colorado as determined by ½ mean maximum distance moved, summer 2006 through winter 2008. Figure 6. Snowshoe hare survival and 95% confidence intervals across summer (S) and winter (W) sampling seasons in central Colorado as determined by mark-recapture, 2006-2008. 38 �Colorado Division of Wildlife July 2008- Aug 2009 WILDLIFE RESEARCH REPORT State of: Cost Center: Work Package: Task No.: Colorado 3430 0670 1 Federal Aid Project No. N/A : : : : : Division of Wildlife Mammals Research Lynx Conservation Post-Release Monitoring of Lynx Reintroduced to Colorado Period Covered: July 1, 2008 – August 31, 2009 Author: T. M. Shenk Personnel: O. Devineau, R. Dickman, P. Doherty, D. Freddy, L. Gepfert, J. Ivan, R. Kahn, A. Keith, P. Lukacs, G. Merrill, B. Smith, T. Spraker, S. Waters, G. White, L. Wolfe All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT In an effort to establish a viable population of Canada lynx (Lynx canadensis) in Colorado, the Colorado Division of Wildlife (CDOW) initiated a reintroduction effort in 1997 with the first lynx released in February 1999. From 1999-2006, 218 wild-caught lynx from Canada and Alaska were released in Colorado. We documented survival, movement patterns, reproduction, and landscape habitatuse through aerial (n = 11,580) and satellite (n = 29,258) tracking. Most lynx remained near the core release area in southwestern Colorado. From 1999-August 2009, there were 118 mortalities of released adult lynx. Approximately 29.7% were either human-induced or likely human-induced through either collisions with vehicles or shot. Starvation and disease/illness accounted for 18.6% of the deaths while 37.3% of the deaths were from unknown causes. Of these mortalities, 26.3% occurred outside of Colorado. Monthly mortality rate was lower inside the study area than outside, and slightly higher for male than for female lynx, although 95% confidence intervals for sexes overlapped. Mortality was higher immediately after release (first month = 0.0368 [SE = 0.0140] inside the study area, and 0.1012 [SE = 0.0359] outside the study area), and then decreased according to a quadratic trend over time. Reproductive females had the smallest 90% utilization distribution home ranges ( x = 75.2 km2, SE = 15.9 km2), followed by attending males ( x = 102.5 km2, SE = 39.7 km2) and non-reproductive animals ( x = 653.8 km2, SE = 145.4 km2). Reproduction was first documented in 2003 with subsequent successful reproduction in 2004, 2005, 2006 and 2009. No dens were documented in 2007 or 2008. From snow-tracking, the primary winter prey species (n = 604 kills) were snowshoe hare (Lepus americanus, annual x = 69.4%, SE = 5.6, n = 11) and red squirrel (Tamiasciurus hudsonicus, annual x = 22.6%, SE = 5.7, n = 11); other mammals and birds formed a minor part of the winter diet. Lynx usedensity surfaces were generated to illustrate relative use of areas throughout Colorado. Within the areas of high use in southwestern Colorado, site-scale habitat use, documented through snow-tracking, supports 1 �mature Engelmann spruce (Picea engelmannii)-subalpine fir (Abies lasiocarpa) forest stands with 4265% canopy cover and 15-20% conifer understory cover as the most commonly used areas in southwestern Colorado. Little difference in aspect (slight preference for north-facing slopes), slope ( x = 15.7°) or elevation ( x = 3173 m) were detected for long beds, travel and kill sites (n = 1841). Den sites (n = 37) however, were located at higher elevations ( x = 3354 m, SE = 31 m) on steeper ( x = 30°, SE = 2°) and more commonly north-facing slopes with a dense understory of coarse woody debris. Three years of a study to evaluate snowshoe hare densities, demography and seasonal movement patterns among small and medium tree-sized lodgepole pine (Pinus contorta) stands and mature spruce/fir stands have been completed in 2006-2009 (see Appendix I of this report). A pilot study to evaluate the efficacy of using minimally-invasive monitoring techniques was developed to estimate the extent, stability and potential distribution of lynx throughout Colorado. Results to date have demonstrated that CDOW has developed lynx release protocols that ensure high initial post-release survival followed by high long-term survival, site fidelity, reproduction and recruitment of Colorado-born lynx into the Colorado breeding population. What is yet to be demonstrated is whether Colorado can support sufficient recruitment to offset annual mortality for a viable lynx population over time. Monitoring continues in an effort to document such viability. 2 �WILDLIFE RESEARCH REPORT POST RELEASE MONITORING OF LYNX (LYNX CANADENSIS) REINTRODUCED TO COLORADO TANYA M. SHENK P. N. OBJECTIVE The initial post-release monitoring of Canada lynx (Lynx canadensis) reintroduced into Colorado will emphasize 5 primary objectives: 1. Assess and modify release protocols to ensure the highest probability of survival for each lynx released. 2. Obtain regular locations of released lynx to describe general movement patterns and habitats used by lynx. 3. Determine causes of mortality in reintroduced lynx. 4. Estimate survival of lynx reintroduced to Colorado. 5. Estimate reproduction of lynx reintroduced to Colorado. Three additional objectives will be emphasized after lynx display site fidelity to an area: 6. Refine descriptions of habitats used by reintroduced lynx. 7. Refine descriptions of daily and overall movement patterns of reintroduced lynx. 8. Describe hunting habits and prey of reintroduced lynx. Information gained to achieve these objectives will form a basis for the development of lynx conservation strategies in the southern Rocky Mountains. SEGMENT OBJECTIVES 1. Complete winter 2008-09 field data collection on lynx habitat use at the landscape scale, hunting behavior, diet, mortalities, and movement patterns. 2. Complete winter 2008-09 lynx trapping field season to collar Colorado born lynx and re-collar adult lynx. 3. Complete spring 2009 field data on lynx reproduction. 4. Summarize and analyze data and publish information as Progress Reports, peer-reviewed manuscripts for appropriate scientific journals, or CDOW technical publications. 5. Complete the third and final year of field work to evaluate snowshoe hare (Lepus americanus) densities, demography and seasonal movement patterns among small and medium tree-sized lodgepole pine stands and mature spruce/fir stands (see Appendix I). 6. Complete a pilot study to evaluate the efficacy of using minimally-invasive monitoring techniques to estimate the extent, stability and potential distribution of lynx throughout Colorado (see Appendix II). INTRODUCTION The Canada lynx occurs throughout the boreal forests of northern North America. Colorado represents the southern-most historical distribution of lynx, where the species occupied the higher elevation, montane forests in the state. Little was known about the population dynamics or habitat use of this species in their southern distribution. Lynx were extirpated or reduced to a few animals in the state by the late 1970’s due, most likely, to predator control efforts such as poisoning and trapping. Given the isolation of Colorado to the nearest northern populations, the CDOW considered reintroduction as the only option to attempt to reestablish the species in the state. 3 �A reintroduction effort was begun in 1997, with the first lynx released in Colorado in 1999. To date, 218 wild-caught lynx from Alaska and Canada have been released in southwestern Colorado. The goal of the Colorado lynx reintroduction program is to establish a self-sustaining, viable population of lynx in this state. Evaluation of incremental achievements necessary for establishing viable populations is an interim method of assessing if the reintroduction effort is progressing towards success. There are 7 critical criteria for achieving a viable population: 1) development of release protocols that lead to a high initial post-release survival of reintroduced animals, 2) long-term survival of lynx in Colorado, 3) development of site fidelity by the lynx to areas supporting good habitat in densities sufficient to breed, 4) reintroduced lynx must breed, 5) breeding must lead to reproduction of surviving kittens 6) lynx born in Colorado must reach breeding age and reproduce successfully, and 7) recruitment must equal or be greater than mortality over an extended period of time. The post-release monitoring program for the reintroduced lynx has 2 primary goals. The first goal is to determine how many lynx remain in Colorado and their locations relative to each other. Given this information and knowing the sex of each individual, we can assess whether these lynx can form a breeding core from which a viable population might be established. From these data we can also describe general movement patterns and habitat use. The second primary goal of the monitoring program is to estimate survival of the reintroduced lynx and, where possible, determine causes of mortality for reintroduced lynx. Such information will help in assessing and modifying release protocols and management of lynx once they have been released to ensure their highest probability of survival. Documenting reproduction is critical to the success of the program and lynx are monitored intensively to document breeding, births, survival and recruitment of lynx born in Colorado. Site-scale habitat descriptions of den sites are also collected and compared to other sites used by lynx. Lynx populations in Canada and Alaska have long been known to cycle in response to the 10-year snowshoe hare (Lepus americana) cycle (Elton and Nicholson 1942). Northern populations of lynx respond to snowshoe hare lows first through a decline in reproduction followed by an increase in adult mortality; when snowshoe hare populations increase, lynx respond with increased survival and reproduction (O’Donoghue et al. 2001). Therefore, annual survival and reproduction are highly variable but must be sufficient, overall, to result in long-term persistence of the population. It is not known if snowshoe hare populations in Colorado cycle and if so, where in the approximate 10-year cycle we are currently. Given this uncertainty, documenting persistence of lynx in Colorado for a period of at least 1015 years would provide support that a viable population of lynx can be sustained in Colorado even in the event snowshoe hares do cycle in the state. Therefore, to document the continued viability of lynx in Colorado beyond the initial reintroduction period, some form of long-term monitoring must be used to determine whether recruitment exceeds mortality for a period of time long enough to encompass possible snowshoe hare cycles. In addition, a challenge facing CDOW is how efforts should be allocated between focusing on monitoring the persistence of those lynx that have established within the core release area (Shenk 2007, Shenk 2008) and those lynx that may be pioneering and expanding into other portions of the state. Reproduction and known recruitment have been observed to be sporadic in the core area. To continue to document lynx reproduction through den site visits and to document survival of those kittens through tracking the adult females in winter looking for accompanying kittens requires a continued trapping effort to capture and radio-collar adult females. Lynx trapping is typically a time consuming and expensive operation as the lynx are territorial with large home ranges that may be entirely located within or largely comprised of inaccessible areas (e.g., wilderness areas). Alternatively, occupancy modeling using minimally-invasive techniques could be a feasible alternative for ascertaining trends in population status. 4 �Additional goals of the post-release monitoring program for lynx reintroduced to the southern Rocky Mountains included refining descriptions of habitat use and movement patterns and describing successful hunting habitat once lynx established home ranges that encompassed their preferred habitat. Specific objectives for the site-scale habitat data collection include: 1) describe and quantify site-scale habitat use by lynx reintroduced to Colorado, 2) compare site-scale habitat use among types of sites (e.g., kills vs. long-duration beds), and 3) compare habitat features at successful and unsuccessful snowshoe hare chases. The program will also investigate the ecology of snowshoe hare in Colorado. A study comparing snowshoe hare densities among mature stands of Engelmann spruce (Picea engelmannii)/subalpine fir (Abies lasiocarpa), lodgepole pine (Pinus contorta) and Ponderosa pine (Pinus ponderosa) was completed in 2004 with highest hare densities found in Engelmann spruce/subalpine fir stands and no hares found in Ponderosa pine stands. A study to evaluate the importance of young, regenerating lodgepole pine and mature Engelmann spruce/subalpine fir stands in Colorado by examining density and demography of snowshoe hares that reside in each was initiated in 2005 and will continue through 2009 (see Appendix I). Lynx is listed as threatened under the Endangered Species Act (ESA) of 1973, as amended (16 U. S. C. 1531 et. seq.)(U. S. Fish and Wildlife Service 2000). Colorado is included in the federal listing as lynx habitat. Thus, an additional objective of the post-release monitoring program is to develop conservation strategies relevant to lynx in Colorado. To develop these conservation strategies, information specific to the ecology of the lynx in its southern Rocky Mountain range, such as habitat use, movement patterns, mortality factors, survival, and reproduction in Colorado is needed. STUDY AREA Byrne (1998) evaluated five areas within Colorado as potential lynx habitat based on (1) relative snowshoe hare densities (Bartmann and Byrne 2001), (2) road density, (3) size of area, (4) juxtaposition of habitats within the area, (5) historical records of lynx observations, and (6) public issues. Based on results from this analysis, the San Juan Mountains of southwestern Colorado were selected as the core reintroduction area, and where all lynx were reintroduced. Wild Canada lynx captured in Alaska, British Columbia, Manitoba, Quebec and Yukon were transported to Colorado and held at The Frisco Creek Wildlife Rehabilitation Center located within the reintroduction area prior to release. Post-release monitoring efforts were focused in a 20,684 km2 study area which included the core reintroduction area, release sites and surrounding high elevation sites (> 2,591 m). The area encompassed the southwest quadrant of Colorado and was bounded on the south by New Mexico, on the west by Utah, on the north by interstate highway 70, and on the east by the Sangre de Cristo Mountains (Figure 1). Southwestern Colorado is characterized by wide plateaus, river valleys, and rugged mountains that reach elevations over 4,200 m. Engelmann spruce/subalpine fir is the most widely distributed coniferous forest type within the study area. The lynx-established core area is roughly bounded by areas used by lynx in the Taylor Park/Collegiate Peak areas in central Colorado and includes areas of continuous use by lynx, including areas used during breeding and denning (Figure 1). METHODS REINTRODUCTION Effort Wild Canada lynx were captured in Alaska, British Columbia, Manitoba, Quebec and Yukon and transported to Colorado where they were held at the Frisco Creek Wildlife Rehabilitation Center prior to release. All lynx releases were conducted under the protocols found to maximize survival (see Shenk 5 �2001). Estimated age, sex and body condition were ascertained and recorded for each lynx prior to release (see Wild 1999). Lynx were transported from the rehabilitation facility to their release site in individual cages. Specific release site locations were recorded in Universal Transverse Mercator (UTM) coordinates and identification of all lynx released at the same location, on the same day, was recorded. Behavior of the lynx on release and movement away from the release site were documented. Movement, Distribution and Relative Use of Areas by Lynx To monitor lynx movements and thus determine distribution and relative use of areas all released lynx were fitted with radio collars. All lynx released in 1999 were fitted with TelonicsTM radio-collars. All lynx released since 1999, with the exception of 5 males released in spring 2000, were fitted with SirtrackTM dual satellite/VHF radio-collars. These collars have a mortality indicator switch that operated on both the satellite and VHF mode. The satellite component of each collar was programmed to be active for 12 hours per week. The 12-hour active periods for individual collars were staggered throughout the week. Signals from the collars allowed for locations of the animals to be made via Argos, NASA, and NOAA satellites. The location information was processed by ServiceArgos and distributed to the CDOW through e-mail messages. Datasets.-- To determine recent (post-reintroduction) movement and distribution of lynx reintroduced, born or initially trapped in Colorado and relative use of areas by these lynx, regular locations of lynx were collected through a combination of aerial and satellite tracking. Locations were recorded and general habitat descriptions for each aerial location was recorded. The first dataset of lynx locations included all locations obtained from daytime flights conducted with a Cessna 185 or similar aircraft to locate lynx by their VHF collar transmitters (hereafter aerial locations). VHF transmitters have been used on lynx since the first lynx were released in February 1999. The second type of lynx location data was collected via satellite from the satellite collar transmitters placed on the lynx (hereafter satellite locations). Satellite transmitter collars were first used for lynx in April 2000. These satellite collars also contained a VHF transmitter which also allowed locating lynx from the air or ground. All locations were recorded in Universal Transverse Mercator (UTM) coordinates using the CONUS NAD27 datum. Flights to obtain lynx aerial locations were typically conducted on a weekly basis throughout most summer and winter months and twice a week during the den search field season (May 15 – June 30), depending on weather and availability of planes and pilots. Flights were typically concentrated in the high elevation (> 2700 m) southwest quadrant of Colorado which encompasses the core lynx release and research area (Figure 1). Flights during the den seasons were conducted to obtain locations on all female lynx within the state wearing an active VHF transmitter. VHF transmitters were outfitted with sufficient batteries to last 60 months. The satellite transmitters were designed to provide locations on a weekly basis with sufficient batteries to last for 18 months. These data collections remain ongoing and all information will be used for future habitat use and survival analyses. Accuracy of both aerial and satellite locations varied with the environmental conditions at the time the location was obtained. Accuracy of aerial locations was influenced by weather with accuracy ranging from 50 - 500 meters. Satellite location accuracy was also influenced by atmospheric conditions and position of the satellites. Satellite location accuracy ranged from 150 meters -10 km. Movement and Distribution.-- To document all known lynx locations maps were generated with all aerial and satellite locations displayed. Due to lynx movements outside of Colorado, particularly into the states of New Mexico, Utah and Wyoming we further evaluated lynx use throughout those three states, as well as the data would allow. All individual lynx located at least once in these 3 states (nontruncated datasets) were identified and tallied for each year. To document consistency and known use of these states after the initial effect of being reintroduced was minimized (i.e., 180 days post-release), each individual lynx located at least once in these states from the truncated datasets were identified and tallied. 6 �Relative Use.-- To document relative use of areas by lynx, 90% kernel use-density surfaces were calculated for truncated satellite and aerial lynx locations using the ArcGIS Spatial Analyst Kernel Density Tool. Lynx may not be exhibiting typical behavior or habitat use within the first few months after their release in Colorado. Therefore, a subset of each of the aerial and satellite datasets was created that eliminated the first 180 days (approximately 6 months) of locations obtained for each lynx immediately after their initial release. As a result, the truncated aerial location dataset contained lynx locations from September 1999 through April 2009 while the truncated satellite location dataset began October 2000 and extended through April 2009. Due to differences in data collection frequency and accuracy between datasets, the truncated satellite and truncated aerial data were analyzed separately for generating the lynx use-density surfaces. These use-density surfaces fit a smoothly curved surface over each lynx location. The surface value was highest at the location of the point and diminished with increasing distance from the point. A fixed kernel was used with a smoothing parameter of 5 km, reaching 0 at the search radius distance from the point. Only a circular neighborhood was possible. The volume under the surface equaled the total value for the point. The use-density at each output GIS raster cell was calculated by adding the values of all the kernel surfaces from all the lynx point locations that overlaid each raster cell center. The kernel function was based on the quadratic kernel function described in Silverman (1986, p. 76, equation 4.5). The use-density surfaces were calculated at 100 m resolution. To enhance graphic displays of higher usedensity areas, density values representing single locations were not displayed. Home Range Preliminary estimates of annual home ranges were calculated as a 95% utilization distribution using a kernel home-range estimator for each lynx we had at least 30 locations for within a year. A year was defined as March 15 – March 14 of the following year. Locations used in the analyses were collected from September 1999 – January 2006 and all locations obtained for an individual during the first six months after its release were eliminated from any home range analyses as it was assumed movements of lynx initially post-release may not be representative of normal habitat use. Locations were obtained either through aerial VHF surveys or locations or the midpoint (ArcView Movement Extension) of all high quality (accuracy rating of 0-1km) satellite locations obtained within a single 24-hour period. All locations used within a single home range analysis were taken a minimum of 24 hours apart. Home range estimates were classified as being for a reproductive or non-reproductive animal. A reproductive female was defined as one that had kittens with her; a reproductive male was defined as a male whose movement patterns overlapped that of a reproductive female. If a litter was lost within the defined year a home range described for a reproductive animal were estimated using only locations obtained while the kittens were still with the female. Final estimates of annual home range size will completed with the addition of data collected through 2009 and in conjunction with current habitat use analyses and publications to be completed in 2009-2010. Survival Multi-state mark-recapture models were used to estimate monthly mortality rates and described in detail in Devineau et al. 2009a (in review) for the first year post-release and for 10 years post-release in Devineau et al. 2009b (in review). This approach accommodated missing data and allowed exploration of factors possibly affecting lynx survival such as sex, time spent in pre-release captivity, movement patterns, and origin. Mortality Factors When a mortality signal (75 beats per minute [bpm] vs. 50 bpm for the Telonics™ VHF transmitters, 20 bpm vs. 40 bpm for the Sirtrack™ VHF transmitters, 0 activity for Sirtrack™ PTT) was heard during either satellite, aerial or ground surveys, the location (UTM coordinates) was recorded. 7 �Ground crews then located and retrieved the carcass as soon as possible. The immediate area was searched for evidence of other predators and the carcass photographed in place before removal. Additionally, the mortality site was described and habitat associations and exact location were recorded. Any scat found near the dead lynx that appeared to be from the lynx was collected. All carcasses were transported to the Colorado State University Veterinary Teaching Hospital (CSUVTH) for a post mortem exam to 1) determine the cause of death and document with evidence, 2) collect samples for a variety of research projects, and 3) archive samples for future reference (research or forensic). The gross necropsy and histology were performed by, or under the lead and direct supervision of a board certified veterinary pathologist. At least one research personnel from the CDOW involved with the lynx program was also present. The protocol followed standard procedures used for thorough post-mortem examination and sample collection for histopathology and diagnostic testing (see Shenk 1999 for details). Some additional data/samples were routinely collected for research, forensics, and archiving. Other data/samples were collected based on the circumstances of the death (e.g., photographs, video, radiographs, bullet recovery, samples for toxicology or other diagnostic tests, etc.). From 1999–2004 the CDOW retained all samples and carcass remains with the exception of tissues in formalin for histopathology, brain for rabies exam, feces for parasitology, external parasites for ID, and other diagnostic samples. Since 2005 carcasses are disposed of at the CSUVTH with the exception of the lower canine, fecal samples, stomach content samples and tissue or bone marrow samples to be delivered by CDOW to the Center for Disease Control for plague testing. The lower canine, from all carcasses, is sent to Matson Labs (Missoula, Montana) for aging and the fecal and stomach content samples are evaluated for diet. Reproduction Females were monitored for proximity to males during each breeding season. We defined a possible mating pair as any male and female documented within at least 1 km of each other in breeding season through either flight data or snow-tracking data. Females were then monitored for site fidelity to a given area during each denning period of May and June. Each female that exhibited stationary movement patterns in May or June were closely monitored to locate possible dens. Dens were found when field crews walked in on females that exhibited virtually no movement for at least 10 days from both aerial and ground telemetry. Kittens found at den sites were weighed, sexed and photographed. Each kitten was uniquely marked by inserting a sterile passive integrated transponder (PIT, Biomark, Inc., Boise, Idaho, USA) tag subcutaneously between the shoulder blades. Time spent at the den was minimized to ensure the least amount of disturbance to the female and the kittens. Weight, PIT-tag number, sex and any distinguishing characteristics of each kitten was also recorded. Beginning in 2005, blood and saliva samples were collected and archived for genetic identification. During the den site visits, den site location was recorded as UTM coordinates. General vegetation characteristics, elevation, weather, field personnel, time at the den, and behavioral responses of the kittens and female were also recorded. Once the females moved the kittens from the natal den area, den sites were visited again and site-specific habitat data were collected (see Habitat Use section below). Captures Captures were attempted for either lynx that were in poor body condition or lynx that needed to have their radio-collars replaced due to failed or failing batteries or to radio-collar kittens born in Colorado once they reached at least 10-months of age when they were nearly adult size. Methods of recapture included 1) trapping using a Tomahawk™ live trap baited with a rabbit and visual and scent lures, 2) calling in and darting lynx using a Dan-Inject CO2 rifle, 3) custom box-traps modified from those 8 �designed by other lynx researchers (Kolbe et al. 2003) and 4) hounds trained to pursue felids were also used to tree lynx and then the lynx was darted while treed. Lynx were immobilized either with Telazol (3 mg/kg; modified from Poole et al. 1993 as recommended by M. Wild, DVM) or medetomidine (0.09mg/kg) and ketamine (3 mg/kg; as recommended by L. Wolfe, DVM)) administered intramuscularly (IM) with either an extendible pole-syringe or a pressurized syringe-dart fired from a Dan-Inject air rifle. Immobilized lynx were monitored continuously for decreased respiration or hypothermia. If a lynx exhibited decreased respiration 2mg/kg of Dopram was administered under the tongue; if respiration was severely decreased, the animal was ventilated with a resuscitation bag. If medetomidine/ketamine were the immobilization drugs, the antagonist Atipamezole hydrochloride (Antisedan) was administered. Hypothermic (body temperature < 95o F) animals were warmed with hand warmers and blankets. While immobilized, lynx were fitted with replacement SirtrackTM VHF/satellite collar and blood and hair samples were collected. Once an animal was processed, recovery was expedited by injecting the equivalent amount of the antagonist Antisedan IM as the amount of medetomidine given, if medetomodine/ketemine was used for immobilization. Lynx were then monitored while confined in the box-trap until they were sufficiently recovered to move safely on their own. No antagonist is available for Telezol so lynx anesthetized with this drug were monitored until the animal recovered on its own in the box-trap and then released. If captured and in poor body condition, lynx were anesthetized with either Telezol (2 mg/kg) or medetomodine/ketemine and returned to the Frisco Creek Wildlife Rehabilitation Center for treatment. HABITAT USE Gross habitat use was documented by recording canopy vegetation at aerial locations. More refined descriptions of habitat use by reintroduced lynx were obtained through following lynx tracks in the snow (i.e., snow-tracking) and site-scale habitat data collection conducted at sites found through this method to be used by lynx. See Shenk (2006) for detailed methodologies. DIET AND HUNTING BEHAVIOR Winter diet of reintroduced lynx was estimated by documenting successful kills through snowtracking. Prey species from failed and successful hunting attempts were identified by either tracks or remains. Scat analysis also provided information on foods consumed. Scat samples were collected wherever found and labeled with location and individual lynx identification. Only part of the scat was collected (approximately 75%); the remainder was left in place in the event that the scat was being used by the animal as a territory mark. Site-scale habitat data collected for successful and unsuccessful snowshoe hare kills were compared. SNOWSHOE HARE ECOLOGY To further our understanding of snowshoe hare ecology in Colorado, a study was conducted comparing snowshoe hare densities among mature stands of Engelmann spruce/subalpine fir, lodgepole pine (Pinus contorta) and Ponderosa pine (Pinus ponderosa). The highest hare densities were found in Engelmann spruce/subalpine fir stands and no hares found in Ponderosa pine stands (Zahratka and Shenk 2008). A second study was initiated in 2005 to evaluate the importance of young, regenerating lodgepole pine and mature Engelmann spruce / subalpine fir stands in Colorado by examining density and demography of snowshoe hares that reside in each (Ivan 2005). Specifically, this study was designed to evaluate small and medium lodgepole pine stands and large spruce/fir stands where the classes “small”, “medium”, and “large” refer to the diameter at breast height (dbh) of overstory trees as defined in the United States Forest Service R2VEG Database (small = 2.54−12.69 cm dbh, medium = 12.70−22.85 cm, and large = 22.86−40.64 cm dbh; J. Varner, United States Forest Service, personal communication). The study design was also developed to identify which 9 �of the numerous hare density-estimation procedures available perform accurately and consistently using an innovative, telemetry augmentation approach as a baseline. In addition, movement patterns and seasonal use of deciduous cover types such as riparian willow were assessed. Finally, the study was designed to further expound on the relationship between density, demography, and stand-type by examining how snowshoe hare density and demographic rates vary with specific vegetation, physical, and landscape characteristics of a stand. RESULTS REINTRODUCTION Effort From 1999 through 2006, 218 wild-caught lynx were reintroduced into southwestern Colorado (Table 1). No lynx were released in 2007, 2008 or 2009. All lynx were released with either VHF or dual VHF/satellite radio collars so they could be monitored for movement, reproduction and survival. The CDOW does not plan to release any additional lynx in 2010. Movement Patterns and Distribution Numerous travel corridors were used repeatedly by more than one lynx. These travel corridors include the Cochetopa Hills area for northerly movements, the Rio Grande Reservoir-SilvertonLizardhead Pass for movements to the west, and southerly movements down the east side of Wolf Creek Pass to the southeast through the Conejos River Valley. Lynx appear to remain faithful to an area during winter months, and exhibit more extensive movements away from these areas in the summer. A total of 11,580 aerial and 29,258 satellite locations were obtained from the 218 reintroduced lynx, radio-collared Colorado kittens (n = 16) and unmarked lynx captured in Colorado (n = 3) as of August 31, 2009. The majority of these locations were in Colorado (Figure 2). Some reintroduced lynx dispersed outside of Colorado into Arizona, Idaho, Iowa, Kansas, Montana, Nebraska, Nevada, New Mexico, South Dakota, Utah and Wyoming (Figure 2). The majority of surviving lynx from the reintroduction effort currently continue to use high elevation (> 2900 m), forested terrain in an area bounded on the south by New Mexico north to Independence Pass, west as far as Taylor Mesa and east to Monarch Pass. Most movements away from the Core Release Area were to the north. Relative Use The lynx use-density surfaces resulting from the fixed kernel analyses provided relative probabilities of finding lynx in areas throughout their distribution. All 218 lynx released in Colorado, all radio-collared kittens and 3 captured unmarked adults were located at least once in Colorado. The majority of these lynx remained in Colorado. Single use density surfaces were calculated for both truncated aerial and truncated satellite datasets in Colorado up to March 2007 and presented in Shenk (2008). Relative use-density surfaces were also generated for New Mexico, Wyoming and Utah and presented in detail in Shenk (2007). Aerial and satellite use-density surfaces indicated similar high usedensity areas. Satellite locations indicated broader spatial use by lynx because satellite collars provided more locations than flights. A single use-density surface was calculated for the satellite non-truncated dataset from April 2000-April 2009 (n = 18,240). The use-density surface was displayed for the satellite non-truncated dataset in Colorado (Figure 3) and for all documented use (Figure 4). The use-density surface for lynx use in Colorado indicates two primary areas of use. The first is the Core Research Area (see Figure 1) and a secondary core centered in the Collegiate Peaks Wilderness (Figures 1, 3 and 4). High use is also documented for 1) the area east of Dillon, on both the north and south sides of I70 and 2) the area north of Hwy 50 centered around Gunnison and then north to Crested Butte. These last 2 high use areas are smaller in extent than the 2 core areas. 10 �Home Range Reproductive females had the smallest 90% utilization distribution annual home ranges ( x = 75.2 km2, SE = 15.9 km2, n = 19), followed by attending males ( x = 102.5 km2, SE = 39.7 km2, n = 4). Nonreproductive females had the largest annual home ranges ( x = 703.9 km2, SE = 29.8 km2, n = 32) followed by non-reproductive males ( x = 387.0 km2, SE = 73.5 km2, n = 6). Combining all nonreproductive animals yielded a mean annual home range of 653.8 km2 (SE = 145.4 km2, n = 38). Survival Detailed analysis of lynx mortality was completed and described in Devineau et al. 2009a (in review) to evaluate how the different release protocols used to reintroduce lynx in Colorado (Shenk 2001) affected mortality within the first year post-release. Average monthly mortality in the study area during the first year decreased with time in captivity from 0.205 [95% CI 0.069, 0.475] for lynx having spent up to 7 days in captivity to 0.028 [95% CI 0.012, 0.064] for lynx spending > 45 days in captivity before release (Devineau et al. 2009). The results also suggest that keeping lynx in captivity beyond 5 or 6 weeks accrued little benefit in terms of monthly survival. On a monthly average basis, lynx were as likely to move out (probability = 0.196, SE=0.032) as well as back on (probability = 0.143, SE=0.034) the reintroduction area (i.e., study area) during the first year after release. Mortality was 1.6x greater outside of the reintroduction area. Detailed analysis of lynx mortality over the first 10 years post-reintroduction was completed and described in Devineau et al. 2009b (in review). In summary, monthly mortality rate was lower inside the study area than outside, and slightly higher for male than for female lynx, although 95% confidence intervals for sexes overlapped. Mortality was higher immediately after release (first month = 0.0368 [SE = 0.0140]; inside the study area, and 0.1012 [SE = 0.0359] outside the study area), and then decreased according to a quadratic trend over time. As of August 31, 2009, CDOW was actively monitoring/tracking 37 of the 100 lynx still possibly alive (Table 2). There are 61 lynx that we have not heard signals on since at least August 31, 2008 and these animals are classified as ‘missing’ (Table 2). One of these missing lynx is a mortality of unknown identity, thus only 60 are truly missing. Possible reasons for not locating these missing lynx include 1) long distance dispersal, beyond the areas currently being searched, 2) radio failure, or 3) destruction of the radio (e.g., run over by car). CDOW continues to search for all missing lynx during both aerial and ground searches. Two of the missing lynx released in 2000 are thought to have slipped their collars. Mortality Factors Of the total 218 adult lynx released, we have 118 known mortalities as of August 31, 2009 (Table 2). Starvation was a significant cause of mortality in the first year of releases only. The primary known causes of death included 29.7% human-induced deaths which were confirmed or probably caused by collisions with vehicles or gunshot (Table 3). Malnutrition and disease/illness accounted for 18.6% of the deaths. An additional 37.3% of known mortalities were from unknown causes. Mortalities occurred throughout the areas through which lynx moved, with 26.3% (n=31) occurring outside of Colorado. The out of state mortalities included 14 in New Mexico, 5 in Utah, 4 in Wyoming and Nebraska, and 1 each in Arizona, Kansas, Iowa and Montana (Figure 2, Table 4). 11 �Reproduction Reproduction was first documented in 2003 when 6 dens and a total of 16 kittens were found in the lynx Core Release Area in southwestern Colorado. Reproduction was also documented in 2004, 2005, 2006, and 2009. No dens were found in 2007 or 2008 (Table 5). Field crews weighed, photographed, PIT-tagged the kittens and checked body condition. Beginning in 2005, we also collected blood samples from the kittens for genetic work in an attempt to confirm paternity. Kittens were processed as quickly as possible (11-32 minutes) to minimize the time the kittens were without their mother. While working with the kittens the females remained nearby, often making themselves visible to the field crews. The females generally continued a low growling vocalization the entire time personnel were at the den. In all cases, the female returned to the den site once field crews left the area. At all dens the females appeared in excellent condition, as did the kittens. The kittens weighed from 270-500 grams. Lynx kittens weigh approximately 200 grams at birth and do not open their eyes until they are 10-17 days old. The proportion of tracked females found with litters in 2006 was lower (0.095) than in the 3 previous years (0.413, SE = 0.032, Table 5). However, all demographic and habitat characteristics measured at the 4 dens that were found in 2006 were comparable to all other dens found. Mean number of kittens per litter from 2003-2006 was 2.78 (SE = 0.05) and sex ratio of females to males was equal ( x = 1.14, SE = 0.14). More details of reproduction in 2003-06 were presented in Shenk (2007). No dens were found in either 2007 or 2008, even though up to 34 adult females were monitored intensively during the denning period (Table 5). In 2009, 22.7% of females being monitored (n = 22) had dens. Two kittens were found at each of these 5 dens, a decrease in the mean of 2.78 (SE= 0.05) kittens per litter found in other years. Sex ratio was also more biased towards female kittens in 2009 (0.4 males/females) than found in previous years. Den Sites.-- A total of 42 dens were found from 2003-2009. All of the dens except one have been scattered throughout the high elevation areas of Colorado, south of I-70. In 2004, 1 den was found in southeastern Wyoming, near the Colorado border. Habitat measurements conducted through 2006 (n=37) document that dens were located on steep ( x slope = 30o , SE=2o), north-facing, high elevation ( x = 3354 m, SE = 31 m) slopes. The dens were typically in Engelmann spruce/subalpine fir forests in areas of extensive downfall of coarse woody debris (Shenk 2006). All dens (n = 42) were located within the winter use areas used by the females. Captures Two adult lynx were captured in 2001 for collar replacement. One lynx was captured in a tomahawk live-trap, the other was treed by hounds and then anesthetized using a jab pole. Five adult lynx were captured in 2002; 3 were treed by hounds and 2 were captured in padded leghold traps. In 2004, 1 lynx was captured with a Belisle snare and 6 adult lynx were captured in box-traps. Trapping effort was substantially increased in winter and spring 2005 and 12 adult lynx were captured and re-collared. Eight reintroduced lynx were captured in winter and spring 2006. In 2007, 11 reintroduced adult lynx were captured and re-collared; 10 in 2008 and 11 in 2009. All lynx captured in Colorado from 2005-2009were caught in box-traps. In addition, as part of the collaring trapping effort, 16 Colorado-born kittens were captured and collared at approximately 10-months of age. Seven 2004-born kittens were collared in spring 2005; 7 2005-born kittens were collared in spring 2006; and 1 2004- and 1 2005 born kitten were first captured and collared in 2009. We also captured 3 adults (approximate age 2 years old) in winters 2006-09 that had no PIT-tags or radio collars. We assume these 3 lynx were from litters born in Colorado that were 12 �never found at dens (i.e., why there were no PIT-tags). All lynx captured for collaring or re-collaring were fitted with new Sirtrack TM dual VHF/satellite collars and re-released at their capture locations. Seven adult lynx were captured from March 1999-August 31, 2009 because they were in poor body condition (Table 6). Five of these lynx were successfully treated at the Frisco Creek Rehabilitation Center and re-released in the Core Release Area. One lynx, BC00F07, died from starvation and hypothermia within 1 day of capture at the rehabilitation center. Lynx QU04M07 died 3 days after capture at the rehabilitation center. Necropsy results documented starvation as the cause of death for this lynx that was precipitated by hydrocephalus and bronchopneumonia (unpublished data T. Spraker, CSUVTH). There were no apparent commonalities among these animals. Seven lynx were captured (either by CDOW personnel or conservation personnel in other states) because they were in atypical habitat outside the state of Colorado (Table 6). They were held at Frisco Creek Rehabilitation Center for a minimum of 3 weeks, fitted with new Sirtrack TM dual VHF/satellite collars and re-released in the Core Release Area in Colorado. Five of these 7 lynx were still alive 6 months post-re-release but 3 had already dispersed out of Colorado and 1 stayed in Colorado through August 31, 2009. Two of these lynx died within 6 months of re-release: 1 died of starvation in Colorado and the other died of unknown causes in Nebraska. One lynx captured out of state and re-released currently remains in Colorado. HABITAT USE Landscape-scale daytime habitat use was documented from 9496 aerial locations of lynx collected from February 1999-June 30, 2007. Throughout the year Engelmann spruce - subalpine fir was the dominant cover used by lynx. A mix of Engelmann spruce, subalpine fir and aspen (Populus tremuloides) was the second most common cover type used throughout the year. Various riparian and riparian-mix areas were the third most common cover type where lynx were found during the daytime flights. Use of Engelmann spruce-subalpine fir forests and Engelmann spruce-subalpine fir-aspen forests was similar throughout the year. There was a trend in increased use of riparian areas beginning in July, peaking in November, and dropping off December through June. Site-scale habitat data collected from snow-tracking efforts indicate Engelmann spruce and subalpine fir were also the most common forest stands used by lynx for all activities during winter in southwestern Colorado. Comparisons were made among sites used for long beds, dens, travel and where they made kills. Little difference in aspect, mean slope and mean elevation were detected for 3 of the 4 site types including long beds, travel and kills where lynx typically use gentler slopes ( x = 15.7o ) at a mean elevation of 3173 m, and varying aspects with a slight preference for north-facing slopes. See Shenk (2006) for more detailed analyses of habitat use. DIET AND HUNTING BEHAVIOR Winter diet of lynx was documented through detection of kills found through snow-tracking. Prey species from failed and successful hunting attempts were identified by either tracks or remains. Scat analysis also provided information on foods consumed. A total of 604 kills were located from February 1999-April 2009. We collected over 990 scat samples from February 1999-April 2009 that will be analyzed for content. In each winter, the most common prey item was snowshoe hare, followed by red squirrel (Tamiusciurus hudsonicus; Table 7). The percent of snowshoe hare kills found however, varied annually from a low of 30.4% in 2009 to a high of 90.77% in winter 2002-2003. An annual mean of 69.39% (SE = 5.6) snowshoe hare kills in the diet has been documented. A comparison of percent overstory for successful and unsuccessful snowshoe hare chases indicated lynx were more successful at sites with slightly higher percent overstory, if the overstory 13 �species were Englemann spruce, subalpine fir or willow. Lynx were slightly less successful in areas of greater aspen overstory. This trend was repeated for percent understory at all 3 height categories except that higher aspen understory improved hunting success. Higher density of Engelmann spruce and subalpine fir increased hunting success while increased aspen density decreased hunting success. SNOWSHOE HARE ECOLOGY Three years of a 3-year study to evaluate snowshoe hare densities, demography and seasonal movement patterns among small and medium tree-sized lodgepole pine stands and mature spruce/fir stands have been completed and preliminary results presented (see Appendix I). DISCUSSION In an effort to establish a viable population of lynx in Colorado, CDOW initiated a reintroduction effort in 1997 with the first lynx released in winter 1999. From 1999 through spring 2006, 218 lynx were released in the Core Release Area. Locations of each lynx were collected through aerial- or satellite-tracking to document movement patterns and to detect mortalities. Most lynx remain in the high elevation, forested areas in southwestern Colorado. The use-density surfaces for lynx use in Colorado indicate two primary areas of use. The first is the Core Research Area (see Figure 1) and a secondary core centered in the Collegiate Peaks Wilderness (Figures 1, 3, 4). High use is also documented for 1) the area east of Dillon, on both the north and south sides of I70 and 2) the area north of Hwy 50 centered around Gunnison and then north to Crested Butte. These last 2 high use areas are smaller in extent than the 2 core areas. Dispersal movement patterns for lynx released in 2000 and subsequent years were similar to those of lynx released in 1999 (Shenk 2000). However, more animals released in 2000 and subsequent years remained within the Core Release Area than those released in 1999. This increased site fidelity may have been due to the presence of con-specifics in the area on release. Numerous travel corridors within Colorado have been used repeatedly by more than 1 lynx. These travel corridors include the Cochetopa Hills area for northerly movements, the Rio Grande Reservoir-Silverton-Lizardhead Pass for movements to the west, and southerly movements down the east side of Wolf Creek Pass to the southeast to the Conejos River Valley. Lynx appear to remain faithful to an area during winter months, and exhibit more extensive movements away from these areas in the summer. Reproductive females had the smallest 90% utilization distribution home ranges ( x = 75.2 km2, SE = 15.9 km2), followed by attending males ( x = 102.5 km2, SE = 39.7 km2) and non-reproductive animals ( x = 653.8 km2, SE = 145.4 km2). Most lynx currently being tracked are within the Core Release Area. During the summer months, lynx were documented to make extensive movements away from their winter use areas. Extensive summer movements away from areas used throughout the rest of the year have been documented in native lynx in Wyoming and Montana (Squires and Laurion 1999). Current data collection methods used for the Colorado lynx reintroduction program were not specifically designed to address the reintroduced lynx movements or use of areas in other states. In particular, the core research and release area were in Colorado. Therefore, the number of aerial locations obtained would be far fewer in other states than in Colorado which would bias low the number of lynx and intensity of lynx use documented outside the state. In contrast, obtaining satellite locations is not biased by the location of the lynx. Satellite locations are, however, biased by the shorter time the satellite transmitters function, approximately 18 months versus 60 months for the VHF transmitters used to obtain the aerial locations. However, data collected to meet objectives of the lynx reintroduction program were 14 �used to provide information to help address the question of lynx use outside of Colorado. Due to the rarity of flights conducted outside Colorado, only use-density surfaces generated from satellite locations were used to document relative lynx use of areas in New Mexico, Utah and Wyoming. New Mexico and Wyoming have been used continuously by lynx since the first year lynx were released in Colorado (1999) to the present. Lynx reintroduced in Colorado were first documented in Utah in 2000 and are still being documented there to date. In addition, all levels of lynx use-density documented throughout Colorado are also represented in New Mexico, Utah and Wyoming from none to the highest level of use (Shenk 2007). One den was found in Wyoming. Although no reproduction has been documented in New Mexico or Utah to date, documenting areas of the highest intensity of use and the continuous presence of lynx within these states for over six years does suggest the potential for yearround residency of lynx and reproduction in those states. From 1999-August 2009, there were 118 mortalities of released adult lynx. Human-caused mortality factors are currently the highest causes of death with approximately 29.7% attributed to collisions with vehicles or gunshot. Starvation and disease/illness accounted for 18.6% of the deaths while 37.3% of the deaths were from unknown causes. Lynx mortalities were documented throughout all areas lynx used, including 31 (26.3%) occurring in other states (Figure 2, Table 3). Nearly half (14 of 30) of the out-of-state mortalities were documented in New Mexico. Detailed analysis of lynx mortality was completed and described in Devineau et al. 2009a to evaluate how the different release protocols used to reintroduce lynx in Colorado (Shenk 2002) affected mortality within the first year post-release. Average monthly mortality in the study area during the first year decreased with time in captivity from 0.205 [95% CI 0.069, 0.475] for lynx having spent up to 7 days in captivity to 0.028 [95% CI 0.012, 0.064] for lynx spending > 45 days in captivity before release (Devineau et al. 2009a). The results also suggest that keeping lynx in captivity beyond 5 or 6 weeks accrued little benefit in terms of monthly survival. On a monthly average basis, lynx were as likely to move out (probability = 0.196, SE=0.032) as well as back on (probability = 0.143, SE=0.034) the reintroduction area during the first year after release. Mortality was 1.6x greater outside of the study area suggesting that permanent emigration and differential mortality rates on and off reintroduction areas should be factored into sample size calculations for an effective reintroduction effort. A post-release monitoring plan is critical to providing information to assess aspects of release protocols in order to improve the survival of individuals. Future lynx, as well as other carnivore, reintroductions may use our results to help design reintroduction programs including both their release and post-release monitoring protocols. Over the 10 years of the reintroduction effort, monthly mortality rate was lower inside the study area than outside, and slightly higher for male than for female lynx, although 95% confidence intervals for sexes overlapped (Devineau et al. 2009b). Mortality was higher immediately after release (first month = 0.0368 [SE = 0.0140] inside the study area, and 0.1012 [SE = 0.0359] outside the study area), and then decreased according to a quadratic trend over time (Devineau et al. 2009, in review). Reproduction is critical to achieving a self-sustaining viable population of lynx in Colorado. Reproduction was first documented from the 2003 reproduction season and again in 2004, 2005 and 2006. Lower reproduction occurred in 2006 (Table 5) but did include a Colorado-born female giving birth to 2 kittens, documenting the first recruitment of Colorado-born lynx into the Colorado breeding population. No reproduction was documented in 2007 or 2008. The cause of the decreased reproduction from 2006 08 is unknown. One possible explanation would be a decrease in prey abundance. Reproduction was again observed in 2009 with 5 dens and 10 kittens found in Colorado. Litter size was smaller than previously documented with only2 kittens found in each litter in comparison to a mean of 2.78 found in previous years. In addition, a sex bias towards female kittens was evident in 2009 which was not evident 15 �in prior years. Two litters found in 2009 had both parents born in Colorado, resulting in the first documented third generation Colorado lynx from the reintroduction. Additional reproduction is likely to have occurred in all years from females we were no longer tracking, and from Colorado-born lynx that have not been collared. The dens we find are more representative of the minimum number of litters and kittens in a reproduction season. To achieve a viable population of lynx, enough kittens need to be recruited into the population to offset the mortality that occurs in that year and hopefully even exceed the mortality rate to achieve an increasing population. The use-density surfaces depict intensity of use by location. Why certain areas would be used more intensively than others should be explained by the quality of the habitat in those areas. Characteristics of areas used by lynx, as documented through aerial locations and snow-tracking of lynx in the Colorado core research area, include mature Engelmann spruce-subalpine fir forest stands with 4265% canopy cover and 15-20% conifer understory cover (Shenk 2006). Within these forest stand types, lynx appear to have a slight preference for north-facing, moderate slopes ( x = 15.7°) at high elevations ( x = 3173 m; Shenk 2006). Snow-tracking of released lynx also provided information on hunting behavior and diet through documentation of kills, food caches, chases, and diet composition estimated through prey remains. Primary winter prey species (n = 604) were snowshoe hare and red squirrel (Table 7), which comprised 69.4% (SE = 5.6, n = 11) and 22.6.2% (SE = 5.7, n = 11) of the annual diet, respectively. Thus, areas of good habitat must also support populations of snowshoe hare and red squirrel. In winter, lynx reintroduced to Colorado appear to be feeding on their preferred prey species, snowshoe hare and red squirrel in similar proportions as those reported for northern lynx during lows in the snowshoe hare cycle (Aubry et al. 1999). Environmental conditions in the springs and summers of 2003, 2006 and 2008 resulted in high cone crops during their following winters based on field observations, resulting in increased red squirrel abundance. This may partially explain the higher percent of red squirrel kills, and thus a lower percent of snowshoe hare kills, found in winters 2003-04, 2006-07 and 2008-09 (Table 7). Caution must be used in interpreting the proportion of identified kills. Such a proportion ignores other food items that are consumed in their entirety and thus are biased towards larger prey and may not accurately represent the proportion of smaller prey items, such as microtines, in lynx winter diet. Through snow-tracking we have evidence that lynx are mousing and several of the fresh carcasses have yielded small mammals in the gut on necropsy. The summer diet of lynx has been documented to include less snowshoe hare and more alternative prey than in winter (Mowat et al., 1999). All evidence suggests that most reintroduced lynx are finding adequate food resources to survive. Mowat et al. (1999) suggest lynx and snowshoe hare select similar habitats except that hares select more dense stands than lynx. Very dense understory limits hunting success of the lynx and provides refugia for hares. Given the high proportion of snowshoe hare in the lynx diet in Colorado, we might then assume the habitats used by reintroduced lynx also depict areas where snowshoes hare are abundant and available for capture by lynx in Colorado. From both aerial locations taken throughout the year and from the site-scale habitat data collected in winter, the most common areas used by lynx are in stands of Engelmann spruce and subalpine fir. This is in contrast to adjacent areas of Ponderosa pine, pinyon juniper, aspen and oakbrush. The lack of lodgepole pine in the areas used by the lynx may be more reflective of the limited amount of lodgepole pine in southwestern Colorado, the Core Release Area, rather than avoidance of this tree species. Hodges (1999) summarized habitats used by snowshoe hare from 15 studies as areas of dense understory cover from shrubs, stands that are densely stocked, and stands at ages where branches have 16 �more lateral cover. Species composition and stand age appears to be less correlated with hare habitat use than is understory structure (Hodges 1999). The stands need to be old enough to provide dense cover and browse for the hares and cover for the lynx. In winter, the cover/browse needs to be tall enough to still provide browse and cover in average snow depths. Hares also use riparian areas and mature forests with understory. Site-scale habitat use documented for lynx in Colorado indicate lynx are most commonly using areas with Engelmann spruce understory present from the snow line to at least 1.5 m above the snow. The mean percent understory cover within the habitat plots is typically less than 15% regardless of understory species. However, if the understory species is willow, percent understory cover is typically double that, with mean number of shrubs per plot approximately 80, far greater than for any other understory species. In winter, hares browse on small diameter woody stems (<0.25"), bark and needles. In summer, hares shift their diet to include forbs, grasses, and other succulents as well as continuing to browse on woody stems. This shift in diet may express itself in seasonal shifts in habitat use, using more or denser coniferous cover in winter than in summer. The increased use of riparian areas by lynx in Colorado from July to November may reflect a seasonal shift in hare habitat use in Colorado. Major (1989) suggested lynx hunted the edge of dense riparian willow stands. The use of these edge habitats may allow lynx to hunt hares that live in habitats normally too dense to hunt effectively. The use of riparian areas and riparian-Engelmann spruce-subalpine fir and riparian-aspen mixes documented in Colorado may stem from a similar hunting strategy. However, too little is known about habitat use by hares in Colorado to test this hypothesis at this time. Lynx also require sufficient denning habitat. Denning habitat has been described by Koehler (1990) and Mowat et al. (1999) as areas having dense downed trees, roots, or dense live vegetation. We found this to be in true in Colorado as well (Shenk 2006). In addition, the dens used by reintroduced lynx were at high elevations and on steep north-facing slopes. All females that were documented with kittens denned in areas within their winter-use area. FUTURE STUDIES Monitoring of individuals through telemetry continues in an effort to document the viability of the reintroduced lynx population. However, as time since release increases, battery failure of telemetry collars also increases resulting in fewer released animals having working collars. In addition, few Colorado-born lynx have been captured and fitted with telemetry collars. Although trapping efforts have been conducted in earnest since 2003 to capture and fit animals with working telemetry collars, we have not been able to collar a sufficient number of animals throughout the state to document the status and trends of lynx distribution and demography throughout Colorado from these collared animals. The extent of lynx dispersal and current distribution beyond the Core Research Area and the difficulty of trapping lynx in all areas they inhabit, particularly large tracts of wilderness, requires redesigning our sampling and monitoring efforts to provide valid estimates of lynx distribution. Exploring occupancy modeling using non-invasive techniques may be a feasible alternative for ascertaining trends in population status and forming a basis for a large scale area monitoring program Therefore, we propose that monitoring lynx distribution would consist of 3 potential primary objectives to document the extent, stability and potential distribution of lynx (at the species and individual level) in Colorado. To estimate patterns in lynx distribution in Colorado a monitoring program could be developed that will: 1) annually estimate the spatial distribution of lynx in the core area and assess changes in lynx distribution over time; 2) detect colonization or expansion of lynx into other portions of the state, and 3) determine whether distribution or persistence are associated with habitat features, measured at the landscape-scale (stand age or composition). 17 �In order to design the most efficient statewide monitoring program, however, we will first evaluate the detection probabilities and efficacy of 3 methods of detection. These include snow-tracking, hair snares and camera surveillance. All of these methods can be conducted with minimal (camera surveillance or collection of hair) or non-invasive approaches (collection of scat samples) to individual animals. A pilot study will be conducted first to establish the most valid, efficient method to estimate the distribution and persistence of lynx. (see Appendix II for the detailed study plan). Information from the pilot study will then be used to design the most efficient strategy to meet the objectives of larger-scale monitoring programs to detect changes in lynx persistence and distribution as a foundation for assessing whether lynx have become established and will persist in Colorado. First, a minimally invasive monitoring program will be designed and implemented within the Core Research Area to describe lynx distribution and distribution trends in this area. A statewide plan could then be implemented to describe lynx distribution and distribution trends throughout Colorado. This monitoring protocol could result in the development of a standardized methodology that might be used by multiple entities to monitor the status of lynx throughout their range in North America. SUMMARY From results to date it can be concluded that CDOW developed release protocols that ensure high initial post-release survival of lynx, and on an individual level, lynx demonstrated they can survive longterm in areas of Colorado. We also documented that reintroduced lynx exhibited site fidelity, engaged in breeding behavior and produced kittens that were recruited into the Colorado breeding population. What is yet to be demonstrated is whether current conditions in Colorado can support the recruitment necessary to offset annual mortality in order to sustain the population. Monitoring of reintroduced lynx will continue in an effort to document such viability. ACKNOWLEDGMENTS The lynx reintroduction program involves the efforts of literally hundreds of people across North America, in Canada and USA. Any attempt to properly acknowledge all the people who played a role in this effort is at risk of missing many people. The following list should be considered to be incomplete. CDOW CLAWS Team (1998-2001): Bill Andree, Tom Beck, Gene Byrne, Bruce Gill, Mike Grode, Rick Kahn (Program Leader), Dave Kenvin, Todd Malmsbury, Jim Olterman, Dale Reed, John Seidel, Scott Wait, Margaret Wild. CDOW: John Mumma (Director 1996-2000), Russell George (Director 2001-2003), Bruce McCloskey (Director 2004-2007), Conrad Albert, Jerry Apker, Laurie Baeten, Cary Carron, Don Crane, Larry DeClaire, Phil Ehrlich, Lee Flores, Delana Friedrich, Dave Gallegos, Juanita Garcia, Drayton Harrison, Jon Kindler, Ann Mangusso, Jerrie McKee, Gary Miller, Melody Miller, Mike Miller, Kirk Navo, Robin Olterman, Jerry Pacheo, Mike Reid, Tom Remington, Ellen Salem, Eric Schaller, Mike Sherman, Jennie Slater, Steve Steinert, Kip Stransky, Suzanne Tracey, Anne Trainor, Scott Wait, Brad Weinmeister, Nancy Wild, Perry Will, Lisa Wolfe, Brent Woodward, Kelly Woods, Kevin Wright. Lynx Advisory Team (1998-2001): Steve Buskirk, Jeff Copeland, Dave Kenny, John Krebs, Brian Miller (Co-Leader), Mike Phillips, Kim Poole, Rich Reading (Co-Leader), Rob Ramey, John Weaver. U. S. Forest Service: Kit Buell, Joan Friedlander, Dale Gomez, Jerry Mastel, John Squires, Fred Wahl, Nancy Warren. U. S. Fish and Wildlife Service: Lee Carlson, Gary Patton (1998-2000), Kurt Broderdorp. State Agencies: Alaska: ADF&G: Cathie Harms, Mark Mcnay, Dan Reed (Regional Manager), Wayne Reglin (Director), Ken Taylor (Assist. Director), Ken Whitten, Randy Zarnke, Other:Ron Perkins (trapper), Dr. Cort Zachel (veterinarian). Washington: Gary Koehler. 18 �National Park Service: Steve King. Colorado State University: Alan Franklin, Gary White. Colorado Natural Heritage Program: Rob Schorr, Mike Wunder. Canada: British Columbia: Dr. Gary Armstrong (veterinarian), Mike Badry (government), Paul Blackwell (trapper coordinator), Trappers: Dennis Brown, Ken Graham, Tom Sbo, Terry Stocks, Ron Teppema, Matt Ounpuu. Yukon: Government: Arthur Hoole (Director), Harvey Jessup, Brian Pelchat, Helen Slama, Trappers: Roger Alfred, Ron Chamber, Raymond Craft, Lance Goodwin, Jerry Kruse, Elizabeth Hofer, Jurg Hofer, Guenther Mueller (YK Trapper’s Association), Ken Reeder, Rene Rivard (Trapper coordinator), Russ Rose, Gilbert Tulk, Dave Young. Alberta: Al Cook. Northwest Territories: Albert Bourque, Robert Mulders (Furbearer Biologist), Doug Steward (Director NWT Renewable Res.), Fort Providence Native People. Quebec: Luc Farrell, Pierre Fornier. Colorado Holding Facility: Herman and Susan Dieterich, Kate Goshorn, Loree Harvey, Rachel Riling. Pilots: Dell Dhabolt, Larry Gepfert, Al Keith, Jim Olterman, Matt Secor, Brian Smith, Whitey Wannamaker, Steve Waters, Dave Younkin. Field Crews (1999-2009): Steve Abele, Brandon Barr, Bryce Bateman, Todd Bayless, Nathan Berg, Ryan Besser, Jessica Bolis, Mandi Brandt, Keith Bruno, Brad Buckley. Patrick Burke, Braden Burkholder, Paula Capece, Matthew Chappell, Stacey Ciancone, Doug Clark, John DePue, Shana Dunkley, Brady Dunne, Tim Hanks, Carla Hanson, Dan Haskell, Nick Hatch, Matt Holmes, Allie Hunter, Andy Jennings, Susan Johnson, Paul Keenlance, Darrin Kite, Patrick Kolar, Tony Lavictoire, Jenny Lord, Clay Miller, Denny Morris, Kieran O’Donovan, Gene Orth, Chris Parmater, Bob Peterson, Jake Powell, Jeremy Rockweit, Britta Schielke, Jenny Shrum, Josh Smith, Heather Stricker, Adam Strong, Dave Unger, James Waddell, David Waltz, Andy Wastell, Mike Watrobka, Lyle Willmarth, Leslie Witter, Kei Yasuda, Jennifer Zahratka. Research Associates: Bob Dickman, Grant Merrill. Data Analysts: Karin Eichhoff, Joanne Stewart, Anne Trainor. Data Entry: Charlie Blackburn, Patrick Burke, Rebecca Grote, Angela Hill, Mindy Paulek. Mary Schuette , Dave Theobald and Chris Woodward provided assistance with the GIS analysis. . Funding: CDOW, Great Outdoors Colorado (GOCO), Turner Foundation, U.S.D.A. Forest Service, Vail Associates, Colorado Wildlife Heritage Foundation. LITERATURE CITED Aubry, K. B., G. M. Koehler, J. R. Squires. 1999. Ecology of Canada lynx in southern boreal forests. Pages 373-396 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. Bartmann, R. M., and Byrne, G. (2001) Analysis and critique of the 1998 snowshoe hare pellet survey. Colorado Division of Wildlife Report No. 20. Fort Collins, Colorado. Byrne, G. 1998. Core area release site selection and considerations for a Canada lynx reintroduction in Colorado. Report for the Colorado Division of Wildlife. Devineau, O., T. M. Shenk, P. F. Doherty Jr., G. C. White, and R. H. Kahn. 2009. Assessing release protocols for the Colorado Canada lynx (Lynx canadensis) reintroduction. Journal of Wildlife Management (in review). Devineau, O., T. M. Shenk, G. C. White, P. F. Doherty Jr., P. M. Lukacs, and R. H. Kahn. 2009. Estimating mortality for a widely dispersing carnivore, the Canada lynx (Lynx canadensis) reintroduced to Colorado. Journal of Applied Ecology (in review). Elton, C. and M. Nicholson 1942. The ten-year cycle in numbers of lynx in Canada. Journal of Animal Ecology 11: 215-244. Hodges, K. E. 1999. Ecology of snowshoe hares in southern boreal and montane forests. Pages 163206 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, 19 �and J. R. Squires editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. Koehler, G. M. 1990. Population and habitat characteristics of lynx and snowshoe hares in north central Washington. Canadian Journal of Zoology 68:845-851. Kolbe, J. A., J. R. Squires, T. W. Parker. 2003. An effective box trap for capturing lynx. Journal of Wildlife Management 31:980-985. Major, A. R. 1989. Lynx, Lynx canadensis canadensis (Kerr) predation patterns and habitat use in the Yukon Territory, Canada. M. S. Thesis, State University of New York, Syracuse. Mowat, G., K. G. Poole, and M. O’Donoghue. 1999. Ecology of lynx in northern Canada and Alaska. Pages 265-306 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. O’Donoghue, M, S. Boutin, D. L. Murray, C. J. Krebs, E. J. Hofer, U. Breitenmoser, C. BreitenmoserWuersten, G. Zuleta, C. , C. Doyle, and V. O. Nams. 2001. Mammalian predators: Coyotes and lynx. in Ecosystem Dynamics of the Boreal Forest: The Kluane Project. eds. C. J. Krebs, S. Boutin and R. Boonstra. Oxford University Press, Inc. New York, New York Poole, K. G., G. Mowat, and B. G. Slough. 1993. Chemical immobilization of lynx. Wildlife Society Bulletin 21:136-140. Shenk, T. M. 1999. Program Narrative Study Plan: Post-release monitoring of reintroduced lynx (Lynx canadensis) to Colorado. Colorado Division of Wildlife, Fort Collins, Colorado. __________. 2001. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report, July: 7- 34. Colorado Division of Wildlife, Fort Collins, Colorado. __________. 2006. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report, July: 1-45. Colorado Division of Wildlife, Fort Collins, Colorado. __________. 2007. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report, July: 1-57. Colorado Division of Wildlife, Fort Collins, Colorado. __________. 2008. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report, July: 1-57. Colorado Division of Wildlife, Fort Collins, Colorado Silverman, B.W. 1986. Density Estimation for Statistics and Data Analysis. Chapman and Hall. New York, New York, USA. Squires, J. R. and T. Laurion. 1999. Lynx home range and movements in Montana and Wyoming: preliminary results. Pages 337-349 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University Press of Colorado, Boulder, Colorado. U. S. Fish and Wildlife Service. 2000. Endangered and threatened wildlife and plants: final rule to list the contiguous United States distinct population segment of the Canada lynx as a threatened species. Federal Register 65, Number 58. Wild, M. A. 1999. Lynx veterinary services and diagnostics. Wildlife Research Report, Colorado Division of Wildlife, Fort Collins, Colorado. Zahratka, J. L. and T. M. Shenk. 2008. Population estimates of snowshoe hares in the southern Rocky Mountains. Journal of Wildlife Management 72: 906-912. Prepared by ___________________________________ Tanya M. Shenk, Wildlife Researcher 20 �Table 1. Number of wild-caught male (M) and female (F) Canada lynx (Lynx canadensis) from Alaska (AK) and Canada (BC = British Columbia, MB = Manitoba, QU = Quebec and YK = Yukon) released in southwestern Colorado per year from 1999–2006. State / Province of Origin Total Year %Released Sex AK BC MB QU YK 1999 19 2000 25 2003 15 2004 17 2005 17 2006 6 Total F 13 5 4 22 M 7 6 6 19 F 6 9 20 35 M 4 9 7 20 F 10 7 17 M 10 5 16 F 7 10 17 M 13 7 20 F 4 M 9 F M 30 1 3 8 3 18 8 3 20 4 3 7 5 2 7 48 218 91 4 45 Table 2. Status of adult Canada lynx (Lynx canadensis) reintroduced to Colorado as of August 31, 2009. Females Lynx Males Unknown TOTALS Released 115 103 218 Known Dead 65 52 1 118 Possible Alive 50 51 100 Missing 27 35 61a Monitoring/tracking 20 17 37 a 1 is unknown mortality Table 3. Causes of death for all Canada lynx (Lynx canadensis) released into southwestern Colorado 1999-2006 as of August 31, 2009. Mortalities Cause of Death Total (%) In Colorado (%) Outside Colorado (%) Unknown 44 (37.3) 29 (24.6) 15 (12.7) Gunshot 16 (13.6) 10 (8.5) 6 (5.1) Hit by Vehicle 14 (11.9) 9 (7.6) 5 (4.2) Starvation 12 (10.2) 11 (9.3) 1 (0.8) Other Trauma 8 (6.8) 7 (5.9) 1 (0.8) Plague 7 (5.9) 7 (5.9) 0 (0) Predation 6 (5.1) 6 (5.1) 0 (0) Probable Gunshot 5 (4.2) 4 (3.4) 1 (0.8) Probable Predation 3 (2.5) 2 (1.7) 1 (0.8) Illness 3 (2.5) 2 (1.7) 1 (0.8) Total Mortalities 118 87 (73.7) 31 (26.3) 21 �Table 4. Known lynx mortalities (n = 31) and causes of death documented by state outside of Colorado from February 1999 – August 31, 2009. Lynx ID AK99F8 Unknown AK99M11 YK99M06 AK99F13 YK00F04 BC99M04 QU05M01 QU04F05 QU03F07 BC00M04 YK06F01 BC03M08 BC06F07 AK99M06 AK99M01 QU05M08 MB05F02 BC00F14 QU04F07 BC06M10 QU04F02 AK00M03 QU05M03 YK06M01 YK00F07 BC06M13 YK99F01 YK00M03 YK05M03 YK05M02 State Date Mortality Recorded Cause of Death New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico New Mexico Nebraska Nebraska Nebraska Nebraska Wyoming Wyoming Wyoming Wyoming Utah Utah Utah Utah Utah Arizona Kansas Montana Iowa 7/30/1999 2000 1/27/2000 6/19/2000 6/22/2000 4/20/2001 6/7/2002 8/22/2005 8/26/2005 9/15/2005 7/19/2006 10/19/2006 10/19/2006 1/8/2007 11/16/1999 1/11/2005 10/1/2006 2/13/2007 7/28/2004 9/21/2004 8/15/2006 3/14/2007 7/2/2001 10/26/2005 12/4/2006 8/6/2007 12/11/08 9/15/2005 9/30/2005 11/8/2005 8/6/2007 Starvation Hit by Vehicle Unknown Probable Gunshot Unknown Gunshot Gunshot Unknown Hit by Vehicle Unknown Unknown Unknown Unknown Gunshot Gunshot Snared (Other Trauma) Unknown Gunshot Unknown Unknown Vehicle Collision Unknown Unknown Unknown Unknown Unknown Unknown Gunshot Vehicle Collision Unknown Vehicle Collision Table 5. Lynx reproduction summary statistics for 1999-2009. No reproduction was expected in 1999 because it was the first year of lynx releases and most animals were released after breeding season. Year Females Tracked 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 TOTAL /MEAN 9 25 21 17 26 40 42 34 28 22 Dens Found in May/June 0 0 0 6 11 17 4 0 0 5 Percent Tracked Females with Kittens 0.0 0.0 0.0 35.3 46.2 42.5 9.5 0.0 0.0 22.7 Additional Litters Found in Winter 0 0 0 0 2 1 0 0 0 - 22 Total Kittens Found Sex Ratio M/F (SE) 2.67 (0.33) 2.83 (0.24) 2.88 (0.18) 2.75 (0.47) 0 0 0 16 39 50 11 1.0 1.5 0.8 1.2 2.00 (0.00) 0 0 10 0.4 2.63(0.16) 126 0.98 (0.18) Mean Kittens/Litter (SE) �Table 6. Lynx captured because they were in poor body condition or were in atypical habitat and their fates 6 months post re-release as of August 31, 2009. Lynx ID Date of Capture State Where Captured Reason For Capture BC99F6 AK99M9 AK99F2 BC00F7 BC00M13 BC03M08 QU04M07 BC04M01 QU04F02 QU05M08 QU04M04 YK00F07 YK05M02 BC04M08 3/25/1999 3/24/2000 4/18/2000 2/11/2001 3/21/2001 9/5/2003 2/2/2006 11/5/2004 4/10/2005 11/25/2005 12/5/2006 12/12/2006 1/1/2007 1/22/2007 Colorado Colorado Colorado Colorado Wyoming Colorado Colorado Utah Nebraska Wyoming Utah Utah Kansas Wyoming Poor body condition Poor body condition Poor body condition Poor body condition Poor body condition Poor body condition Poor body condition Atypical habitat Atypical habitat Atypical habitat Atypical habitat Atypical habitat Atypical habitat Atypical habitat Date of Re-release 5/28/1999 5/3/2000 5/22/2000 N/A 4/24/2001 1/1/2004 N/A 12/5/2004 5/7/2005 4/18/2006 1/20/2007 1/20/2007 2/2/2007 2/15/2007 Status 6 Months Post Re-release Dead Missing Alive in Colorado Dead Alive in Colorado Alive in Colorado Dead Alive in Colorado Alive in Wyoming Dead Dead in Colorado Alive in Utah Alive in Iowa Alive in Colorado Table 7. Number of kills found each winter field season through snow-tracking of lynx and percent composition of kills of the three primary prey species. Field Season 1999 1999-2000 2000-2001 2001-2002 2002-2003 2003-2004 2004-2005 2005-2006 2006-2007 2007-2008 2008-2009 Total/Mean n 9 83 89 54 65 37 78 50 41 42 56 604 Snowshoe Hare 55.56 67.47 67.42 90.74 90.77 67.57 83.33 90.00 61.00 59.00 30.4 69.39 (SE=5.6) Prey (%) Red Squirrel Cottontail 22.22 0 19.28 1.20 19.10 8.99 5.56 0 6.15 0 27.03 2.70 10.26 0 0.08 0 39.0 0 33.3 0 66.1 0 22.55 (SE=5.7) 1.17 (SE=0.82) 23 Other 22.22 12.05 4.49 3.70 3.08 2.70 6.41 0.02 0 7.4 3.5 5.96 (SE=1.92) �Figure 1. Lynx are monitored throughout Colorado and by satellite throughout the western United States. The lynx core release area, where all lynx were released, is located in southwestern Colorado (outlines in white). A lynx-established core use area has developed in the Taylor Park and Collegiate Peak area in central Colorado. 24 �Figure 2. All documented lynx locations (non-truncated datasets) obtained from either aerial (red circles) or satellite (yellow circles) tracking from February 1999 through August 31, 2009 All known lynx mortality locations (n = 112) are displayed as black stars. 25 �Figure 3. Use-density surface for lynx satellite locations (non-truncated dataset) in Colorado from April 2000-April 2009. 26 �Figure 4. Use-density surface for lynx satellite locations (non- truncated dataset) in Colorado from April 2000-April 2009 27 �APPENDIX I Colorado Division of Wildlife August 2009 WILDLIFE RESEARCH REPORT State of Colorado Cost Center_______3430 Work Package_____0670 Task No.___________2 : : : : Federal Aid Project: N/A : Division of Wildlife Mammals Research Lynx Reintroduction Density, Demography, and Seasonal Movements of Snowshoe Hare in Colorado Period Covered: July 1, 2008- June 30, 2009 Author: J. S. Ivan, Ph.D. Candidate, Colorado State University Personnel: Dr. T. Shenk of CDOW and Dr. G. C. White of Colorado State University. All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT A program to reintroduce the threatened Canada lynx (Lynx canadensis) into Colorado was initiated in 1997. Analysis of scat collected from winter snow tracking indicates that snowshoe hares (Lepus americanus) comprise 65–90% of the winter diet of reintroduced lynx in most winters. Thus, existence of lynx in Colorado and success of the reintroduction hinge at least partly on maintaining adequate and widespread hare populations. Beginning in July 2006, I initiated a study to assess the relative value of 3 stand types for providing hare habitat in Colorado. These types include mature, uneven-aged Engelmann spruce (Picea engelmannii)-subalpine fir (Abies lasiocarpa) forests, sapling lodgepole pine (Pinus contorta) forests (“small lodgepole”), and pole-sized lodgepole pine forests (“medium lodgepole”). Estimates and comparisons of survival, recruitment, finite population growth rate, and maximum (late summer) and minimum (late winter) snowshoe hare densities for each stand will provide the metrics for assessing these stands. Snowshoe hare densities on the study area are low compared to densities reported elsewhere. Within the study area, hare densities during summer were generally highest in small lodgepole stands, followed by mature spruce/fir and medium lodgepole, respectively. Absolute hare densities declined considerably in summer 2007 and rebounded only slightly during summer 2008. Hare density in small and medium lodgepole stands equalized during winters. However, as with summer, overall density was much lower during the second winter compared to the first and rebounded somewhat during the last winter. Hare survival from summer to winter was relatively high whereas winter to summer survival is quite low. Survival does not appear to differ between stand types or years, although a much more thorough analysis that will include known-fate telemetry data is forthcoming. This combined analysis will provide a final winter-summer estimate, will bring much more information to bear on the estimation process, and should increase precision of all estimates by a fair amount. 28 �WILDIFE RESEARCH REPORT DENSITY AND SURVIVAL OF SNOWSHOE HARES IN TAYLOR PARK AND PITKIN JACOB S. IVAN P. N. OBJECTIVE Assess the relative value of 3 stand types (mature spruce/fir, sapling lodgepole, pole-sized lodgepole) that purportedly provide high quality hare habitat by estimating survival, recruitment, finite population growth rate, and maximum (late summer) and minimum (late winter) snowshoe hare densities for each type. SEGMENT OBJECTIVES 1. Complete mark-recapture work across all replicate stands during late summer (mid-July through midSeptember) and winter (mid-January through March). 2. Obtain daily telemetry locations on radio-tagged hares for 10 days immediately after capture periods, as well as monthly between primary trapping sessions. 3. Locate, retrieve, and refurbish radio tags as mortalities occur. INTRODUCTION A program to reintroduce the threatened Canada lynx (Lynx canadensis) into Colorado was initiated in 1997. Since that time, 218 lynx have been released in the state, and an extensive effort to determine their movements, habitat use, reproductive success, and food habits has ensued (Shenk 2005). Analysis of scat collected from winter snow tracking indicates that snowshoe hares (Lepus americanus) comprise 65–90% of the winter diet of reintroduced lynx during most winters (T. Shenk, Colorado Division of Wildlife, unpublished data). Thus, as in the far north where the relationship between lynx and snowshoe hares has captured the attention of ecologists for decades, it appears that the existence of lynx in Colorado and success of the reintroduction effort may hinge on maintaining adequate and widespread populations of hares. Colorado represents the extreme southern range limit for both lynx and snowshoe hares (Hodges 2000). At this latitude, habitat for each species is less widespread and more fragmented compared to the continuous expanse of boreal forest at the heart of lynx and hare ranges. Neither exhibits dramatic cycles as occur farther north, and typical lynx (≤2−3 lynx/100km2; Aubry et al. 2000) and hare (≤1−2 hares/ha; Hodges 2000) densities in the southern part of their range correspond to cyclic lows form northern populations (2-30 lynx/100 km2, 1−16 hares/ha; Aubry et al. 2000, Hodges 2000, Hodges et al. 2001). Whereas extensive research on lynx-hare ecology has occurred in the boreal forests of Canada, literature regarding the ecology of these species in the southern portion of their range is relatively sparse. This scientific uncertainty is acknowledged in the “Canada Lynx Conservation Assessment and Strategy,” a formal agreement between federal agencies intended to provide a consistent approach to lynx conservation on public lands in the lower 48 states (Ruediger et al. 2000). In fact, one of the explicit guiding principles of this document is to “retain future options…until more conclusive information concerning lynx management is developed.” Thus, management recommendations in this agreement are decidedly conservative, especially with respect to timber management, and are applied broadly to cover all habitats thought to be of possible value to lynx and hare. Accurate identification and detailed 29 �description of lynx-hare habitat in the southern Rocky Mountains would permit more informed and refined management recommendations. A commonality throughout the snowshoe hare literature, regardless of geographic location, is that hares are associated with dense understory vegetation that provides both browse and cover (Wolfe et al. 1982, Litvaitis et al. 1985, Hodges 2000, Homyack et al. 2003, Miller 2005). In western mountains, this understory can be provided by relatively young conifer stands regenerating after stand-replacing fires or timber harvest (Sullivan and Sullivan 1988, Koehler 1990a, Koehler 1990b, Bull et al. 2005) as well as mature, uneven-aged stands (Beauvais 1997, Griffin 2004). Hares may also take advantage of seasonally abundant browse and cover provided by deciduous shrubs (e.g., riparian willow [Salix spp.], aspen [Populus tremuloides]; Wolff 1980, Miller 2005). In drier portions of hare range, such as Colorado, regenerating stands can be relatively sparse, and hares may be more associated with mesic, late-seral forest and/or riparian areas than with young stands (Ruggiero et al. 2000). Numerous investigators have sought to determine the relative importance of these distinctly different habitat types with regards to snowshoe hare ecology. Most previous evaluations were based on hare density or abundance (Bull et al. 2005), indices to hare density and abundance (Wolfe et al. 1982, Koehler 1990a, Beauvais 1997, Miller 2005), survival (Bull et al. 2005), and/or habitat use (Dolbeer and Clark 1975). Each of these approaches provides insight into hare ecology, but taken singly, none provide a complete picture and may even be misleading. For example, extensive use of a particular habitat type may not accurately reflect the fitness it imparts on individuals, and density can be high even in “sink” habitats (Van Horne 1983). A more informative approach would be to measure density, survival, and habitat use simultaneously in addition to recruitment and population growth rate through time. Griffin (2004) employed such an approach and found that summer hare densities were consistently highest in young, dense stands. However, he also noted that only dense mature stands held as many hares in winter as in summer. Furthermore hare survival seemed to be higher in dense mature stands, and only dense mature stands were predicted (by matrix projection) to impart a mean positive population growth rate on hares. Griffin’s (2004) study occurred in the relatively moist forests of Montana, which share many similarities but also many notable differences with Colorado forests including levels of fragmentation, species composition, elevation, and annual precipitation. The study outlined below is designed principally to evaluate the importance of young, regenerating lodgepole pine (Pinus contorta) and mature Engelmann spruce (Picea engelmannii)/ subalpine fir (Abies lasiocarpa) stands in Colorado by examining density and demography of snowshoe hares that reside in each. I determined that 2 classes of regenerating lodgepole could provide adequate hare habitat. Thus, I sampled both “small” (2.54-12.69 cm dbh) and “medium” (12.70-22.85 cm dbh) stands regenerating from clearcutting 20 and 40 years ago, respectively (Figure 1). Medium lodgepole stands were pre-commercially thinned 20 years ago; small lodgepole stands have not yet been thinned. Density and demography will be estimated primarily from mark-recapture techniques as data from such approaches can simultaneously provide information on both aspects of hare ecology. However, I will augment both density and demographic analyses with telemetry data to improve the accuracy and precision of estimates. The estimates reported here do not yet reflect addition of telemetry information. My hope is that information gathered from this research will be drawn upon as managers make routine decisions, leading to landscapes that include stands capable of supporting abundant populations of hares. I assume that if management agencies focus on providing habitat, hares will persist. 30 �Hypotheses 1) In general, snowshoe hare density in Colorado will be relatively low (≤0.5 hares/ha) compared to densities reported in northern boreal forests, even immediately post-breeding when an influx of juveniles will bolster hare numbers. 2) Snowshoe hare density will be consistently highest in small lodgepole pine stands, followed by large spruce/fir and medium lodgepole pine, respectively. 3) Survival will generally be highest in mature (large) spruce/fir stands followed by small and medium lodgepole pine, respectively. 4) Finite population growth rate will be consistently at or above 1.0 in mature spruce/fir stands with survival contributing most significantly to the growth rate. Finite growth rates for the lodgepole pine stands will be more variable. 5) Snowshoe hares will significantly shift their home ranges to make use of abundant food and cover provided by riparian willow (and/or aspen) habitats in summer. 6) Snowshoe hare density, survival, and recruitment will be highly correlated with understory cover and stem density. STUDY AREA The study area stretches from Taylor Park to Pitkin in central Colorado (Figure 2). Elevation ranges from 2700 m to 4000 m. Sagebrush (Artemisia spp.) dominates broad, low-lying valleys. Most montane areas are covered by even-aged, large-diameter lodgepole pine forests with sparse understory. Moist, north-facing slopes and areas near tree line are dominated by large-diameter Engelmann spruce/subalpine fir. Interspersed along streams and rivers are corridors of willow. Patches of aspen occur sporadically on southern exposures. This area was chosen over other potential study areas in the state because 1) it contained numerous examples of the 3 stand types of interest (more southern regions lack naturally occurring stands of lodgepole pine), 2) it was not subject to confounding effects of largescale mountain pine beetle outbreak as were more northern stands, and 3) an adequate number of radio frequencies were available to support a large study with hundreds of radio-tagged individuals. Within the study area I selected sample stands based on the following: Potential replicate stands were required to be 1) close enough geographically to minimize differences due to climate, weather, and topography, but are far enough apart to be considered independent, 2) adjacent to one or more riparian willow corridors, 3) within 1 km of an access road for logistical purposes, 4) of suitable size and shape to admit a 16.5-ha trapping grid, and 5) consistent in their management history (i.e., replicate lodgepole pine stands were clear-cut and/or thinned within 1-2 years of each other). I queried the U.S. Forest Service R2VEG GIS database using the criteria listed above to initially develop a suite of potential sample stands. I further narrowed this suite after obtaining updated standlevel information from local USFS personnel (Art Haines, Silviculturalist, USFS Gunnison Ranger District, personal communication). Finally, I ground-truthed potential stands and qualitatively assessed their representativeness and similarity to other potential replicates. Given the numerous constraints imposed, very few stands met all criteria. Thus, I was unable to randomly select sample stands from a population of suitable stands. Rather, I subjectively chose the “best” stands from among the handful that met my criteria. Small lodgepole stands rarely occur on the landscape in patches large enough to fit a full trapping grid. To accommodate this, I sampled 6 replicate small lodgepole stands (rather than 3) using half-sized trapping grids. 31 �METHODS Experimental Design/Procedures Variables.--The response variables of interest for this project include stand-specific snowshoe hare density (D), apparent survival (φ), recruitment (f), finite population growth rate (λ), and a metric of seasonal movement. Density is the number of hares per unit area and is estimated using conventional “boundary strip” techniques (Wilson and Anderson 1985) in this report. Stand-specific demographic parameters were estimated primarily from capture-mark-recapture methods. As such, apparent survival was defined as the probability that a marked animal alive and in the population at time i survived and was in the population at time i + 1. Apparent survival encompassed losses due to both death and emigration. Estimates of recruitment, population growth, and seasonal movement are forthcoming and not provided in this report. Potential explanatory variables for snowshoe hare density, demographics, and movement include general species composition and structural stage of each stand in which response variables are measured. Additionally, stem density, horizontal cover, and canopy cover (to a lesser extent) are highly correlated with snowshoe hare abundance and habitat use (Wolfe et al. 1982, Litvaitis et al. 1985, Hodges 2000, Zahratka 2004, Miller 2005). Thus, I further characterized vegetation in each stand by measuring stem density by size class (1-7 cm, 7.1-10 cm, and >10 cm), percent canopy cover, percent horizontal cover of understory and basal area. Basal area is an easily obtainable metric that may be correlated with the other variables and is recorded routinely during timber cruises, whereas the others are not. Thus, it might prove a useful link for biologists designing management strategies for snowshoe hare. Additionally, I recorded physical covariates such as ambient temperature, precipitation, and snow depth at each stand during sampling. These metrics were not included in the current preliminary analyses, but will be used as covariates in future models. Sampling.--All trapping and handling procedures have been approved by the Colorado State University Animal Care and Use Committee and filed with the Colorado Division of Wildlife. Snowshoe hares breed synchronously and generally exhibit 2 birth pulses in Colorado (although in some years, some individuals may have 3 litters), with the first pulse terminating approximately June 5−20 and the second approximately July 15–25 (Dolbeer 1972). To obtain a maximum density estimate, I began data collection on the first suite of sites immediately following the second birth pulse in late July. Along with a crew of 5 technicians, I deployed one 7 × 12 trapping grid (50-m spacing between traps; grid covers 16.5 ha) in the large spruce/fir and medium lodgepole stands within the first suite, along with 2 6 × 7 grids in 2 small lodgepole stands. Grid set up and trap deployment followed Griffin (2004) and Zahratka (2004). Grid locations and orientation within each stand were chosen subjectively to accommodate logistical constraints and to ensure that hares using the grid had ample opportunity to use adjacent riparian willow zones. As traps were deployed, they were locked open and “pre-baited” with apple slices, hay cubes, and commercial rabbit chow. Traps were pre-baited in this manner for a total of 3 nights to maximize capture rates when trapping began. This minimized the number of trap-nights needed to capture the desired number of animals which in turn minimized trap-related injuries and minimized problems with predators keying into trap lines. During pilot work in winter 2005, I observed low but increasing capture rates (<0.20) during the first 3 nights of trapping, with higher, more stable capture probabilities after 3 days (approximately 0.35–0.45). Thus 3 days of pre-baiting seemed reasonable. Traps were set on the afternoon of the 4th day and checked early each morning and re-set again in the evening on days 5–9. By checking traps in both morning and evening I prevented hares from being entrapped >13 hours, which minimized capture stress. A crew of 2 people worked together on each grid to check traps and process captures as quickly as possible. All captured hares were coaxed out of the trap and into a dark handling bag by blowing quick shots of air on them from behind. Hares remained in the 32 �handling bag, physically restrained with their eyes covered, for the entire handling process. Each individual was aged, sexed, marked with a passive integrated transponder (PIT) tag and temporary ear mark (to track PIT tag retention), then released. Aging consisted of assigning each individual as either juvenile (<1 year old, <1000 g) or adult (≥1 year old, ≥1000 g) based on weight and development of genitalia. This criterion is accurate through the end of September at which point juveniles are difficult to distinguish from adults (K. Hodges, University of British Columbia; P. Griffin, University of Montana, personal communication). After the first day of trapping, all captured hares were scanned for a PIT tag prior to any handling and those already marked were recorded and immediately released. Traps and bait were completely removed from the grid on day 10. In addition to PIT tags and ear marks, I radio collared up to 10 hares captured on each grid with a 28-g mortality-sensing transmitter (BioTrack, LTD) to facilitate unbiased density estimation as well as assessment of seasonal movements. I expected heterogeneity in snowshoe hare movements and use of the grid area, with potential bias surfacing due to location at which a hare is captured (e.g., hares captured on the edge of a grid may use the grid area differently than those captured at the center), and differential behavioral responses to trapping (e.g., young individuals may have lower capture probabilities and thus may be more likely to be captured on later occasions). To guard against the first potential bias, I randomly selected a starting trap location each morning and ran the grid systematically from that point. Thus, the first several hares encountered (and collared) were as likely to be from the inner part of the grid as from the edge. To protect against the second potential source of bias, I refrained from deploying the final 3 collars until days 4 and 5 of the trapping session. Immediately following the removal of traps, the field crew began work locating each radiocollared hare 1–2 times per day for 10 days. Most locations were obtained by triangulation from relatively close proximity, but some were obtained by “homing” on a signal (Samuel and Fuller 1996, Griffin 2004) taking care not to push hares while approaching them. Because hares are largely nocturnal (Keith 1964, Mech et al. 1966, Foresman and Pearson 1999), I made an effort to conduct telemetry work at various times of the night (safety and logistics permitting) and day to gather a representative sample of locations for each hare. Crews gathered telemetry locations for radio-collared hares on the initial suite of sites for 10 days. Then the 10−day trapping procedure and 8 to 10−day telemetry work were repeated on the grids comprising suites 2 and 3(Figure 3). The entire process was repeated during the winter when densities should have been at a minimum. Thus, during the period covered by this report, sampling occurred between July 16 – September 22 and between January 20−March 26. Telemetry work also occurred during “pre-baiting” days after the initial summer sampling session to determine which hares were still alive and immediately available to be sampled by the grid during the ensuing trapping period. Vegetation sampling was conducted in June and July 2008. I followed protocols established through previous snowshoe hare and lynx work in Colorado (Zahratka 2004, T. Shenk, Colorado Division of Wildlife, personal communication). Specifically, on each of the 12 live-trapping grids, I laid out 5 × 5 grids (3-m spacing) of vegetation sampling points centered on 15 of the 84 trap locations (Figure 4; 9 points were sampled on each of the ½-sized small lodgepole stands). At each of the 25 vegetation sampling points, I recorded canopy cover (present or absent) using a densitometer. I quantified downed coarse wood along the center transect of the 25-point grid following Brown (1974). From the center point (i.e., trap location) I measured 1) distance to the nearest woody stem 1.0−7.0 cm, 7.1−10.0 cm, and >10.0 cm in diameter at heights of 0.1 m and 1.0 m above the ground (to capture both summer and winter stem density; Barbour et al. 1999), 2) horizontal cover in 0.5-m increments above the ground up to 2 m (Nudds 1977), 3) basal area, and 4) slope. 33 �Data Analysis Density, Survival, and Population Growth.--I analyzed mark-recapture data in a robust design framework (Williams et al. 2002:523-554) treating summer and winter sampling occasions as primary periods, and the 5-day trapping sessions within each as secondary periods. As such, I assumed hare populations were demographically and geographically closed during the short 5-day mark-recapture sampling periods, but were open to immigration, emigration, births, and deaths between these occasions. I specified the Robust Design data type in Program MARK (White and Burnham 1999) and used the Huggins closed capture model (Huggins 1989, 1991) for secondary periods. I obtained estimates of apparent survival ( φˆ i )between each primary period. I followed Wilson and Anderson (1985) to calculate the effective area trapped and obtain a density estimate for each grid from each secondary period. Future density analyses will employ a new estimator that employs telemetry data to correct for bias (Ivan 2005). For this report, I used a relatively simple model where capture probability varied by stand type and season (i.e., winter and summer), while survival was allowed to vary by stand type, season, and time. RESULTS AND DISCUSSION During summer, density estimates followed hypotheses 1) and 2) above (Figure 5). Specifically, hare densities were clearly highest in small lodgepole stands and quite low in medium lodgepole stands. Spruce/fir was generally intermediate in density with the exception of the final summer. Telemetry data collected during this last sampling period suggests that many hares were present on spruce/fir sites, but were never caught. Therefore, I believe spruce/fir densities were much higher than actually measured during the final summer. While the relationship in density between stand types remained fairly constant throughout the study, the absolute density of hares dropped considerably from summer 2006 to summer 2007 and rebounded only slightly during summer 2008. It is unclear why this sharp decline occurred, although disease outbreak, natural population cycles, and response to increased predation due to lynx reintroduction are possibilities. Note that even the highest densities recorded here correspond to low estimates observed in other parts of hare range (Hodges 2000). Hare densities tend to equalize in lodgepole stands during winter (Figure 5). I submit that the interplay between food, cover, and snow depth provides a plausible explanation for this pattern. Medium lodgepole stands apparently provide very little forage/cover for hares during summer as the canopy in these stands is generally ≥1 meter off the ground. However, in winter, accumulated snow may make that canopy available again to hares. Conversely, small lodgepole stands provide abundant food and cover during summer, but accumulated snow during winter brings hares closer to the crowns of the young trees, which then provide less cover. Spruce/fir stands probably provide adequate access to both food and cover during both summer and winter due to their uneven-aged, multi-layered structure. Like the summer estimates, density during the second winter was much lower than during the first winter. Hare survival is quite high from summer to winter but very low from winter to summer (Figure 6). However, survival did not appear to differ between stand types or among years of this study. A deeper analysis of these data will occur over the next several months in which known-fate telemetry data will be combined with the current mark-recapture dataset. This combined analysis will bring significantly more information to bear on the process which should improve precision of estimates and may elucidate differences between stands or years that are not yet apparent. A much larger suite of models will be considered in that analysis. Model selection and model averaging (Burnham and Anderson 2002) will be used to more thoroughly assess survival of hares. Additionally, combining telemetry data with the current dataset will allow for another estimate of survival from winter 2009 to summer 2009. Hare recruitment and finite population growth rate will be estimated as derived parameters following the combined survival analysis. 34 �SUMMARY • • • • Snowshoe hare densities on my study sites appear to be relatively low compared to densities reported elsewhere. Densities during summer were highest in small lodgepole stands, followed by spruce/fir and medium lodgepole. During winter, densities equalize in lodgepole stands, possibly due to the interplay between snow depth and canopy height in small and medium lodgepole pine. Hare density declined considerably from winter to summer 2007 but has recovered somewhat since then. Summer to winter hare survival was consistently high but winter to summer survival is quite low. A more thorough analysis including known-fate survival data is forthcoming. This new analysis should improve precision of estimates and will add a sixth survival estimate to the current time series. ACKNOWLEDGMENTS Ken Wilson (CSU), Bill Romme (CSU), Paul Doherty (CSU), Dave Freddy (CDOW), Chad Bishop (CDOW), and Paul Lukacs (CDOW) provided helpful insight on the design of this study. We appreciate the invaluable logistical support provided by Mike Jackson (USFS), Art Haines (USFS), Jake Spritzer (USFS), Kerry Spetter (USFS), Margie Michaels (CDOW), Gabriele Engler (USGS), Dana Winkelman (USGS), Brandon Diamond (CDOW), Chris Parmeter (CDOW), Kathaleen Crane (CDOW), Lisa Wolfe (CDOW), and Laurie Baeten (CDOW). Jim Gammonley (CDOW), Dave Freddy (CDOW), Chad Bishop (CDOW), Jack Vayhinger (CDOW), Brandon Diamond (CDOW), and Brent Bibles (CDOW) assisted with trucks and equipment. The following hardy individuals collected the hard-won data presented in this report: Braden Burkholder, Matt Cuzzocreo, Brian Gerber, Belita Marine, Adam Behney, Pete Lundberg, Katie Yale, Britta Shielke, Cory VanStratt, Mike Watrobka, Meredith Goss, Sidra Blake, Keith Rutz, Rob Saltmarsh, Jennie Sinclair, Evan Wilson, Mat Levine, Matt Strauser, Greg Davidson, Leah Yandow, Renae Sattler, Caylen Cummins, DeVaughn Fraser, Mark Ratchford, Mike Petriello, Cynthia Soria, Roblyn Stitt, Sarah Ryan, Eric Newkirk, Kyle Heinrick, Matt Strauser, Doug Miles, and Cate Brown. Funding was provided by the Colorado Division of Wildlife. LITERATURE CITED Aubry, K. B., G. M. Koehler, and J. R. Squires. 2000. Ecology of Canada lynx in southern boreal forests. Pages 373-396 in Ruggiero, L. F., K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation of lynx in the United States. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Collins, Colorado, USA. Barbour, M. G., J. H. Burk, W. E. Pitts, F. S. Gilliam, and M. W. Schwartz. 1999. Terrestrial plant ecology, Addison Wesley Longman, Inc., Menlo Park, California, USA. Beauvais, G. P. 1997. Mammals in fragmented forests in the Rocky Mountains: community structure, habitat selection, and individual fitness. Dissertation, University of Wyoming, Laramie, Wyoming, USA. BROWN, J. K. 1974. Handbook for inventorying downed woody material. U.S. Department of Agriculture, Forest Service, Intermountain Forest and Range Experiment Station, General Technical Report INT-16. Burnham, K. P., and D. R. Anderson. 1998. Model selection and inference: a practical informationtheoretic approach. Springer-Verlag, New York, New York, USA. 35 �Bull, E. L., T. W. Heater, A. A. Clark, J. F. Shepherd, and A. K. Blumton. 2005. Influence of precommercial thinning on snowshoe hares. U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station, General Technical Report PNW-RP-562. Dolbeer, R. A. and W. R. Clark. 1975. Population ecology of snowshoe hares in the central Rocky Mountains. Journal of Wildlife Management 39:535-549. Dolbeer, R. A. 1972. Population dynamics of snowshoe hare in Colorado. Dissertation, Colorado State University, Fort Collins, Colorado, USA. Foresman, K. R. and D. E. Pearson. 1999. Activity patterns of American martens, Martes americana, snowshoe hares, Lepus americanus, and red squirrels, Tamiasciurus hudsonicus, in westcentral Montana. The Canadian Field-Naturalist 113:386-389. Griffin, P. C. 2004. Landscape ecology of snowshoe hares in Montana. Dissertation, University of Montana, Missoula, Montana, USA. Hodges, K. E. 2000. Ecology of snowshoe hares in southern boreal and montane forests. Pages 163-206 in Ruggiero, L. F., K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation of lynx in the United States. University Press of Colorado, Boulder, Colorado, USA. Hodges, K. E., C. J. Krebs, D. S. Hik, C. I. Stefan, E. A. Gillis, and C. E. Doyle. 2001. Snowshoe hare demography. Pages 141-178 in Krebs, C. J., S. Boutin, and R. Boonstra, editors. Ecosystem dynamics of the boreal forest: the Kluane project. Oxford University Press, New York, New York, USA. Homyack, J. A., D. J. Harrison, and W. B. Krohn. 2003. Effects of precommercial thinning on select wildlife species in northern Maine, with special emphasis on snowshoe hare. Maine Cooperative Fish and Wildlife Research Unit, Orono, Maine, USA. Huggins, R. M. 1989. On the statistical analysis of capture-recapture experiments. Biometrika 76:133140. Huggins, R. M. 1991. Some practical aspects of a conditional likelihood approach to capture experiments. Biometrics 47:725-732. Ivan, J. S. 2005. Density, demography, and seasonal movements of snowshoe hares in Colorado. Program narrative study plan for mammals research, Colorado Division of Wildlife, Fort Collins, Colorado, USA. Keith, L. B. 1964. Daily activity pattern of snowshoe hares. Journal of Mammalogy 45:626-627. Koehler, G. M. 1990a. Snowshoe hare, Lepus americanus, use of forest successional stages and population changes during 1985-1989 in North-central Washington. Canadian Field-Naturalist 105:291-293. Koehler, G. M. 1990b. Population and habitat characteristics of lynx and snowshoe hares in north central Washington. Canadian Journal of Zoology 68:845-851. Litvaitis, J. A., J. A. Sherburne, and J. A. Bissonette. 1985. Influence of understory characteristics on snowshoe hare habitat use and density. Journal of Wildlife Management 49:866-873. Mech, L. D., K. L. Heezen, and D. B. Siniff. 1966. Onset and cessation of activity in cottontail rabbit and snowshoe hares in relation to sunset and sunrise. Animal Behaviour 14:410-413. Miller, M. A. 2005. Snowshoe hare habitat relationships in northwest Colorado. Thesis, Colorado State University, Fort Collins, Colorado, USA. Nudds, T. D. 1977. Quantifying the vegetative structure of wildlife cover. Wildlife Society Bulletin 5:113-117. Ruediger, B., J. Claar, S. Gniadek, B. Holt, Lewis Lyle, S. Mighton, B. Naney, G. Patton , T. Rinaldi, J. Trick, A. Vendehey, F. Wahl, N. Warren, D. Wenger, and A. Williamson. 2000. Canada lynx conservation assessment and strategy. U.S. Department of Agriculture, Forest Service, U.S. Department of Interior, Fish and Wildlife Service, Bureau of Land Management, National Park Service R1-00-53 U.S. Department of Agriculture, Forest Service, U.S. Department of Interior, Fish and Wildlife Service, Bureau of Land Management, National Park Service, Missoula, Montana, USA. 36 �Ruggiero, L. F., K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires. 2000. The scientific basis for lynx conservation: qualified insights. Pages 443-454 in Ruggiero, L. F., K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation of lynx in the United States. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Collins, Colorado, USA. Samuel, M. D. and M. R. Fuller. 1996. Wildlife radiotelemetry. Pages 370-418 in Bookhout, T. A., editors. Research and Management Techniques for Wildlife and Habitats. Allen Press, Inc., Lawrence, Kansas, USA. Shenk, T. M. 2005. General locations of lynx (Lynx canadensis) reintroduced to southwestern Colorado from February 4, 1999 through February 1, 2005. Colorado Division of Wildlife Colorado Division of Wildlife, Fort Collins, Colorado, USA. Sullivan, T. P. and D. S. Sullivan. 1988. Influence of stand thinning on snowshoe hare population dynamics and feeding damage in lodgepole pine forest. Journal of Applied Ecology 25:791-805. Van Horne, B. 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47:893-901. White, G. C., and K. P. Burnham. 1999. Program MARK: survival estimation from populations of marked animals. Bird Study 46 Supplement:120-138. Williams, B. K., J. D. Nichols, and M. J. Conroy. 2002. Analysis and management of animal populations, Academic Press, San Diego, California, USA. Wilson, K. R. and D. R. Anderson. 1985. Evaluation of two density estimators of small mammal population size. Journal of Mammalogy 66:13-21. Wolfe, M. L., N. V. Debyle, C. S. Winchell, and T. R. McCabe. 1982. Snowshoe hare cover relationships in northern Utah. Journal of Wildlife Management 46:662-670. Wolff, J. O. 1980. The role of habitat patchiness in the population dynamics of snowshoe hares. Ecological Monographs 50:111-130. Zahratka, J. L. 2004. The population and habitat ecology of snowshoe hares (Lepus americanus) in the southern Rocky Mountains. Thesis, The University of Wyoming, Laramie, Wyoming, USA. Prepared by _________________________________________________ Jacob S. Ivan, Graduate Student, Colorado State University 37 �Figure 1. Purported high quality snowshoe hare habitat in Colorado. From left to right: small lodgepole pine, medium lodgepole pine, and large Engelmann spruce/subalpine fir. Figure 2. Study area near Taylor Park and Pitkin, Colorado including medium lodgepole (squares), small lodgepole (circles), and spruce/fir (triangles) stands selected for mark-recapture sampling. 38 �Jul Aug Sep 1 2 3 4 2 3 4 5 6 7 8 1 Jan 5 6 7 8 9 10 11 9 10 11 12 13 14 15 12 13 14 15 16 17 18 16 17 18 19 20 21 22 19 20 21 22 23 24 25 23 24 25 26 27 28 29 26 27 28 29 30 31 1 30 31 1 2 3 4 5 2 3 4 5 6 7 8 6 7 8 9 10 11 12 9 10 11 12 13 14 15 13 14 15 16 17 18 19 16 17 18 19 20 21 22 20 21 22 23 24 25 26 23 24 25 26 27 28 1 27 28 29 30 31 1 2 2 3 4 5 6 7 8 3 4 5 6 7 8 9 9 10 11 12 13 14 15 10 11 12 13 14 15 16 16 17 18 19 20 21 22 17 18 19 20 21 22 23 23 24 25 26 27 28 29 24 25 26 27 28 29 30 30 31 Feb Mar Figure 3. Approximate annual data collection schedule for trapping () and telemetry (). Dates and weeks changed depending on calendar year and pay schedule. During telemetry work, the 6-person crew was divided into 2 teams, only one of which worked at any given time. Monthly locations on radio-collared hares were also collected in the interim between the intensive sampling periods indicated here. Figure 4. 15 trap locations (•) on 7 × 12 trapping grid where vegetation was sampled by measuring stem density, horizontal cover, downed woody material, and basal area. Additionally, the 25-point grid superimposed on each of the 15 trap locations (inset) was used to quantify canopy cover). 39 �Figure 5. Snowshoe hare density and 95% confidence intervals in 3 types of stands in central Colorado as determined by ½ mean maximum distance moved, summer 2006 through winter 2009. Figure 6. Snowshoe hare survival and 95% confidence intervals between summer and winter sampling seasons in 3 types of stands in central Colorado as determined by mark-recapture, 2006-2009. 40 �APPENDIX II PROGRAM NARRATIVE STUDY PLAN FOR MAMMALS RESEARCH FY 2009 – 10 State of: Cost Center: Work Package: Task No.: Colorado 3430 0670 3 Federal Aid Project No. N/A : : : : : Division of Wildlife Mammals Research Lynx Conservation Estimating Potential Changes in Distribution of Canada Lynx in Colorado: A Pilot Study Plan to Estimate Lynx Detection Probabilities ESTIMATING POTENTIAL CHANGES IN DISTRIBUTION OF CANADA LYNX IN COLORADO; A PILOT STUDY PLAN TO ESTIMATE LYNX DETECTION PROBABILITIES Principal Investigator Tanya M. Shenk, Wildlife Researcher, Mammals Research Cooperators Rick H. Kahn, Terrestrial Management Coordinator, CDOW Paul M. Lukacs, Biometrician, CDOW Grant J. Merrill, Research Associate, CSU Cooperative Research Unit Robert D. Dickman, CDOW Mike Miller, Acting Mammals Research Leader, CDOW STUDY PLAN APPROVAL Prepared by: Date: Submitted by; Date: Reviewed by: Date: Date: Date: Biometrician Review Date: Approved by: Date: Mammals Research Leader 41 �PROGRAM NARRATIVE STUDY PLAN FOR MAMMALS RESEARCH FY 2009-10 ESTIMATING THE EXTENT, STABILITY AND POTENTIAL DISTRIBUTION OF CANADA LYNX (LYNX CANADENSIS) IN COLORADO: A PILOT STUDY TO ESTIMATE LYNX DETECTION PROBABILITIES A Research Proposal Submitted By Tanya M. Shenk, Wildlife Researcher, Mammals Research A. Background: The Canada lynx (Lynx canadensis) occurs throughout the boreal forests of northern North America. While Canada and Alaska support healthy populations of the species, the lynx is currently listed as threatened under the Endangered Species Act (ESA) of 1973, as amended (16 U. S. C. 1531 et. seq.; U. S. Fish and Wildlife Service 2000) in the coterminous United States. Colorado represents the southern-most historical distribution of naturally occurring lynx, where the species occupied the higher elevation, montane forests in the state (U. S. Fish and Wildlife Service 2000). Thus, Colorado is included in the federal listing as lynx habitat. Lynx were extirpated or reduced to a few animals in Colorado, however, by the late 1970’s (U. S. Fish and Wildlife Service 2000), most likely due to multiple humanassociated factors, including predator control efforts such as poisoning and trapping (Meaney 2002). Given the isolation of and distance from Colorado to the nearest northern populations of lynx, the Colorado Division of Wildlife (CDOW) considered reintroduction as the only option to attempt to reestablish the species in the state. Therefore, a reintroduction effort was begun in 1997, with the first lynx released in Colorado in 1999. To date, 218 wild lynx were captured in Alaska or Canada and released in southwestern Colorado. The goal of the Colorado lynx reintroduction program is to establish a self-sustaining, viable population of lynx in this state. Evaluation of incremental achievements necessary for establishing viable populations is an interim method of assessing the success of the reintroduction effort. Seven critical criteria were identified that must be met before concluding a viable population had been established: 1) development of release protocols that lead to a high initial post-release survival of reintroduced animals, 2) long-term survival of lynx in Colorado, 3) site fidelity by lynx to areas supporting good habitat and in densities sufficient to breed, 4) reintroduced lynx must breed, 5) breeding must lead to production of surviving kittens, 6) lynx born in Colorado must reach breeding age and reproduce successfully, and 7) recruitment must equal to or be greater than mortality over an extended (~10 year) period of time (Shenk 2006). The fundamental approach taken to evaluate the status of each of these criteria was to PIT-tag and place telemetry collars on every lynx released and as many Colorado-born kittens surviving to adulthood as possible, followed by intensive monitoring of these animals through satellite, aerial and groundtracking. All establishment criteria, except (7) have been achieved. Lynx populations in Canada and Alaska have long been known to cycle in response to the 10-year snowshoe hare (Lepus americana) cycle (Elton and Nicholson 1942). Northern populations of lynx respond to snowshoe hare lows first through a decline in reproduction followed by an increase in adult mortality; when snowshoe hare populations increase, lynx respond with increased survival and reproduction (O’Donoghue et al. 2001). Therefore, annual survival and reproduction are highly variable but must be sufficient, overall, to result in long-term persistence of the population. It is not known if snowshoe hare populations in Colorado cycle and if so, where in the approximate 10-year cycle we are currently. Given this uncertainty, documenting persistence of lynx in Colorado for a period of at least 10- 42 �15 years would provide support that a viable population of lynx can be sustained in Colorado even in the event snowshoe hares do cycle in the state. Therefore, to document viability of the lynx population in Colorado, some form of long-term monitoring must be used to determine whether recruitment exceeds mortality for a period of time long enough to encompass a possible snowshoe hare cycle, and thus, determine the reintroduction a success. A challenge facing CDOW is how efforts should be allocated between focusing on monitoring the persistence of those lynx that have established within the core release area (Shenk 2007, Shenk 2008) and those lynx that may be pioneering and expanding into other portions of the state. Reproduction and known recruitment have been observed to be sporadic in the core area. To continue to document lynx reproduction through den site visits and to document survival of those kittens through tracking the adult females in winter looking for accompanying kittens requires a continued trapping effort to capture and radio-collar adult females. Lynx trapping is typically a time consuming and expensive operation as the lynx are territorial with large home ranges that may be entirely located within or largely comprised of inaccessible areas (e.g., wilderness areas). Alternatively, exploring occupancy modeling using noninvasive techniques may be a feasible alternative for ascertaining trends in population status and forming a basis for a large scale area monitoring program. Monitoring of individuals through telemetry continues in an effort to document the viability of the reintroduced lynx population. However, as time since release increases, battery failure of telemetry collars also increases resulting in fewer released animals having working collars. In addition, few Colorado-born lynx have been captured and fitted with telemetry collars. Although trapping efforts have been conducted in earnest since 2003 to capture and fit animals with working telemetry collars, we have not been able to collar a sufficient number of animals throughout the state to document the status and trends of lynx distribution and demography throughout Colorado from these collared animals. The extent of lynx dispersal and current distribution beyond the Core Research Area and the difficulty of trapping lynx in all areas they inhabit, particularly large tracts of wilderness, requires redesigning our sampling and monitoring efforts to provide valid estimates of lynx distribution. We propose that monitoring lynx distribution would consist of 3 potential primary objectives to document the extent, stability and potential distribution of lynx (at the species and individual level) in Colorado. To estimate patterns in lynx distribution in Colorado a monitoring program could be developed that will: 1) annually estimate the spatial distribution of lynx in the core area and assess changes in lynx distribution over time; 2) detect colonization or expansion of lynx into other portions of the state, and 3) determine whether distribution or persistence are associated with habitat features, measured at the landscape-scale (stand age or composition). A pilot study will be conducted first to establish the most valid, efficient method to estimate the distribution and persistence of lynx. B. Need The primary goal of the Colorado lynx reintroduction program is to establish a self-sustaining, viable population of Canada lynx in Colorado. The approach taken to reach this goal was to initially establish a lynx population within a core reintroduction area in southwestern Colorado. From this core reintroduction area, lynx could disperse on their own throughout the suitable habitat in the state, or additional reintroductions north of the core area could be conducted. The current lynx population in Colorado is comprised of surviving reintroduced adults, lynx born in Colorado from the reintroduced animals and possibly some naturally occurring lynx. Research and monitoring efforts over the last 9 years, since the first lynx were released, have focused primarily on monitoring reintroduced animals through VHF and satellite telemetry and estimating demographic parameters of these animals (e.g., Devineau et al. 2009). However, as more of these animals become unavailable for monitoring due to failed telemetry collars, death or movement out of the Core 43 �Research Area, it becomes more difficult to accurately evaluate the status of the entire lynx population in Colorado, including the Core Research Area. A dual monitoring approach will provide a comprehensive, feasible and valid estimation of the demography of the lynx population throughout the state. The first approach would continue to estimate reproduction within the Core Research Area through the use of telemetry. The second approach would obtain information on the status and trend of the distribution of lynx throughout the high elevation, montane areas of Colorado. Below we first outline the objectives and approach for the statewide distribution study and then propose a pilot study to establish the most valid, efficient methods to estimate the statewide distribution and persistence of lynx. A minimally-invasive monitoring program can be developed to estimate the extent, stability and potential distribution of lynx throughout Colorado. The primary objectives of the monitoring program will be to document the current distribution of lynx throughout Colorado, the stability, growth or shrinkage of this distribution over time, and to identify potential areas lynx may occupy in the future. The proposed goal would be to annually monitor lynx into the long-term future, with regular analyses of change (e.g., every 5 years). The fundamental structure of such a monitoring program will consist of: 1. 2. 3. 4. Creating a sampling frame of all potential lynx home range sized primary sampling units within Colorado. Annually estimating winter site occupancy and persistence within this sampling frame. Measuring key habitat features that have been documented to be important for both snowshoe hare and lynx at the landscape-scale within annually sampled sites. Predicting potential distribution of lynx throughout Colorado based on these habitat relationships. In the past, biologists referred to presence/absence as present/not detected, because absence cannot be absolutely determined. This term, however, confuses the status of being present or not present with the activity of either detecting or not detecting an animal. This monitoring program adopts the term presence/absence with the argument that although absence cannot be determined, it can be estimated statistically using a known or estimated detection probability. The indicator used to determine the distribution of occurrence of lynx is P, the proportion of primary sampling units (PSU’s) (Levy and Lemeshow 1999) with lynx presence. A PSU is a square sampling unit of 75km2, the approximate mean size of a lynx winter home range as estimated by a 90% kernel utilization distribution (Shenk 2007). For the statewide monitoring program, the sampling frame would consist of a grid of PSU’s laid over all areas of Colorado above 2591 meters (8500 feet). We would then estimate P from a random sample of PSU’s, using a sample size that is sufficient for attaining an estimate that is within 10% of the actual frequency 90% of the time (see Table 6.1, pg. 168 in MacKenzie et al. 2006). In order to design the most efficient statewide monitoring program, however, we will first evaluate the detection probabilities and efficacy of 3 methods of detection. These include snow-tracking, hair snares and camera surveillance. All of these methods can be conducted with minimal (camera surveillance or collection of hair) or non-invasive approaches (collection of scat samples) to individual animals. Identification of species will allow us to determine the presence of lynx in a PSU; identifying individual lynx within PSU’s will allow for monitoring individual movement patterns across PSU’s, reproduction, social structure and possibly apparent survival rates. Such non-invasive techniques are widely desirable because they are considered to have a minimal impact on animals and are inexpensive relative to other methods. Methodologies for identifying the species and individual lynx from blood and scat samples has been completed by the USFS Conservation Genetics Laboratory in Missoula, Montana. Thus, development costs have already been expended (by other agencies) and we need only cover the 44 �costs of genetic sample processing and interpretation of results. In order to begin genetic tracking of individual lynx a genetic library should be created from all lynx released in Colorado as part of the Colorado lynx reintroduction program, all documented kittens and lynx of unknown origin captured in Colorado. These samples have already been collected and are currently archived at the CDOW. This genetic library would be used to help determine paternity of Colorado-born kittens for future, detailed reproduction studies, document the dispersal of individuals throughout Colorado and also be available for research conducted on continent-wide studies of Canada lynx (e.g., Schwartz et al. 2002, Schwartz et al. 2003). Collecting scat samples during the pilot study will allow a test of these methodologies for the larger study as well as providing an opportunity to establish the protocols with the conservation genetics lab for collection, transport and analysis of the samples. This pilot study will provide necessary information to (1) identify the most efficient method of detecting lynx in a PSU and (2) provide an estimate of detection probability within a PSU. This detection probability will then be used to design the most efficient strategy to meet the objectives of larger-scale monitoring programs to detect changes in lynx persistence and distribution as a foundation for assessing whether lynx have become established and will persist in Colorado. First, a minimally invasive monitoring program will be designed and implemented within the Core Research Area to describe lynx distribution and distribution trends in this area. A statewide plan could then be implemented to describe lynx distribution and distribution trends throughout Colorado. This monitoring protocol could result in the development of a standardized methodology that might be used by multiple entities to monitor the status of lynx throughout their range in North America. This monitoring design will not provide a means of estimating total population size in the state because detection of a lynx may represent a single territorial animal, a breeding pair or a family unit. To obtain a statewide lynx abundance estimate, further efforts beyond this sampling design would be needed to establish the actual or estimated number of lynx in a PSU. Furthermore, this monitoring program is not designed to provide information on reproductive success or estimate survival. C. Objectives: The primary objectives of this pilot study are to: 1. Provide information needed to estimate the detection probability (p) of 3 different, minimally-invasive methods to detect lynx in a PSU in winter, where lynx are known to occur but in extremely low densities (approximately 1 per 75 km2). 2. Evaluate and compare the efficacy of the 3 methods of lynx detection in winter within a PSU. 3. Develop a standardized, valid methodology for describing various landscape-scale habitat features, including those important to snowshoe hare, within a PSU. D. Expected Results or Benefits: The methodologies developed during this pilot study will be used to develop a valid, non-invasive or minimally invasive inventory and monitoring program to estimate the distribution of Canada lynx in Colorado. The monitoring program will provide information on the annual winter distribution, extent and habitat relationships of these parameters as well as their long-term trend which will be evaluated every 5 years. The protocols developed will be made available to any other agencies or entities that want to monitor lynx. The proposed methodology to estimate and monitor trends in lynx distribution throughout Colorado is designed to make use of technologies (e.g., genetic identification) reliant only on noninvasive or minimally invasive techniques. Such non-invasive techniques are widely desirable because they require minimal impact to the animals and because of their cost efficiencies. 45 �E. Approach The primary objective of the pilot study is to evaluate the efficacy of the proposed sampling techniques for detecting lynx presence. However, the pilot study will also include qualitative evaluation of all design methods that will be employed in a future, larger research area and statewide monitoring efforts, (i.e., the complete sampling frame). Sampling Frame and Primary Sampling Unit Selection The sampling frame will consist of all forested areas in Colorado >2591 m (8500 ft) in elevation. The sampling frame will be randomly overlayed with a contiguous grid of 75 km2 squares. The size of the square reflects a mean annual home range size of a reproducing lynx in Colorado (Shenk 2007) and similar to home range estimates obtained for lynx in Montana (Squires and Laurion 1999). If a grid square is >50% forested it will be identified as a PSU. We will assume the lowest detection probabilities for lynx would occur in a PSU occupied by only 1 lynx. Given that we want to estimate lynx detection probabilities under the worst case scenario, we will eliminate all PSU’s where we know, through VHF or satellite-tracking, there is more than one lynx occupying the area. We will then select 6 PSU’s where we know at least 1 but not likely more than 1 lynx occupies the area. The assumptions that must be met in estimating occupancy are 1) surveyed sites can be occupied by the species of interest throughout the duration of the study, with no sites becoming occupied or unoccupied during the survey period (i.e., the system is closed), 2) species are not falsely detected, but can remain undetected if present, and 3) species detection at a site is assumed to be independent of species detection at other sites (MacKenzie et al. 2006). For this pilot study, there will be 3 different methods of detection (snow-tracking, hair snares and camera surveillance). Snow-tracking and camera surveillance will be evaluated at 2 different levels of effort; hair snares will be evaluated at 3 levels of effort resulting in 7 total detection approaches. In order to meet the assumptions for estimating occupancy and assuming the different detection approaches don’t influence each other, each of the 6 PSU’s will be assigned all detection approaches (except for the higher level of hair-snaring) for 3 weeks, allowing for completing surveys of 2 PSU’s per month. The increased hair snare effort will be conducted on a PSU the month following the initial survey effort (see below). Thus, by the end of four months each PSU will have had each detection approach applied to it. This will result in 6 spatial replications of each of 3 detection approaches applied to a PSU for 3 weeks. Maximum levels of effort will be applied to each PSU and then the data sub-sampled to evaluate lower levels of effort. Field Methods Temporal aspects of the sampling design In order to verify the detection methods being evaluated in this pilot study are effective at detecting lynx when they are present, we need to conduct the study while we have active radio collars on lynx. Currently, we are continuing to monitor and re-collar lynx within the Core Research Area for data on the demography and movement patterns of the reintroduced lynx. Thus, completing this pilot study at the same time that active monitoring is being conducted in the research area eliminates the need for future radio-collaring efforts to conduct this pilot study. All data collection will be conducted from January 1-March 31 (Table 1). This is within the time period (October–April) when lynx typically maintain fidelity to a winter home range and when breeding occurs, the period of interest for document long-term persistence of lynx. 46 �Table 1. Data collection and crew work schedule for the six PSU’s to be sampled. PSU Month Week Crew Activity 1 January 1 I Set-up detection routes and 5 detection stations with hair snares and cameras; Snow-track (2 10-hour days) 2 I Snow-track (4 10-hour days) 3 I Snow-track (4 10-hour days)4 I Snow-track (2 10-hour days); Retrieve cameras and hair snares at the 5 detection stations, place 20 hair snares along the detection route; Travel to next PSU 2 January 1 II Set-up detection routes and stations with hair snares and cameras; Snow-track (2 10-hour days) 2 II Snow-track (4 10-hour days) 3 II Snow-track (4 10-hour days)4 II Snow-track (2 10-hour days); Retrieve cameras and hair snares at the 5 detection stations, place 20 hair snares along the detection route; Travel to next PSU 3 February 1 I Set-up detection routes and stations with hair snares and cameras; Snow-track (2 10-hour days) 2 I Snow-track (4 10-hour days) 3 I Snow-track (4 10-hour days)4 I Snow-track (2 10-hour days); Retrieve cameras and hair snares at the 5 detection stations, place 20 hair snares along the detection route; Travel to next PSU 4 February 1 II Set-up detection routes and stations with hair snares and cameras; Snow-track (2 10-hour days) 2 II Snow-track (4 10-hour days) 3 II Snow-track (4 10-hour days)4 II Snow-track (2 10-hour days); Retrieve cameras and hair snares at the 5 detection stations, place 20 hair snares along the detection route; Travel to next PSU 5 March 1 I Set-up detection routes and stations with hair snares and cameras; Snow-track (2 10-hour days) 2 I Snow-track (4 10-hour days) 3 I Snow-track (4 10-hour days)4 I Snow-track (2 10-hour days); Retrieve cameras and hair snares at the 5 detection stations, place 20 hair snares along the detection route; Travel to next PSU 6 March 1 II Set-up detection routes and stations with hair snares and cameras; Snow-track (2 10-hour days) 2 II Snow-track (4 10-hour days) 3 II Snow-track (4 10-hour days)4 II Snow-track (2 10-hour days); Retrieve cameras and hair snares at the 5 detection stations, place 20 hair snares along the detection route; Travel to next PSU Lynx Detection Data Collection Three methods will be evaluated to determine which is most efficient in detecting the presence of lynx. These methods include 1) documenting the presence of lynx tracks in the snow coupled with a DNA sample collection (hair or scat found through snow-tracking), 2) a photograph of a lynx captured by 47 �a surveillance camera, or 3) documenting the presence of lynx from a hair DNA sample collected on a hair snag at a scent and visual lure station. All methods will be applied to the same stations within a PSU at the same time. Each method will be implemented in the areas within the selected PSU that a lynx would most likely use. Based on lynx habitat use in Colorado (Shenk 2005), this will include areas of mature Engelmann spruce-subalpine fir forest stands with 42-65% canopy cover and 15-20% conifer understory cover, mean slopes of 16° and elevations above 2591 m. In addition, selection of specific detection stations will be based on natural travel routes or the presence of lynx sign (i.e., tracks or scat). Chances of detecting lynx at these locations will be further enhanced by placing scent and visual lures at these sites. Other feline species may be attracted to these same lures, however, the probability will be low as the study will be conducted in winter and the deep snows at these elevations should preclude species such as mountain lion (Puma concolor) and bobcat (Lynx rufus) from using these areas. Different levels of sampling intensity will be evaluated for each method to determine the most efficient sampling design. Establishing Detection Stations & Routes. – To eliminate bias in site selection of detection stations and routes, any known lynx locations in the selected PSU’s will not be made available to the field technicians who will be establishing the detection routes, detection stations and collecting the detection data. Field personnel will be provided information to select routes that are both the most feasible and likely areas to detect lynx within a PSU (see above). Detection stations will be set up in areas along those selected routes in areas of good lynx habitat. Commercial scent lures and visual lures (e.g., CD’s, waterfowl wings) will be used at each detection station to enhance the probability of drawing a lynx into the station. To increase the probability of lynx using the hair snares, the hair snares will be placed on landscape features at the detection station known to be used as scent posts by lynx such as tree stumps, small trees and broken logs protruding from the snow at approximate head height of a lynx (Schmidt and Kowalczyk 2006). Snow-Tracking. –Searches for tracks will be attempted by hiking, driving or snowmobiling detection station routes in the PSU once enough snow has accumulated. Due to the inaccessibility of wilderness and roadless areas after significant snowfall, surveys will be conducted in these areas first, while snow accumulations are great enough to detect tracks but not so great as to preclude human access to the area. Once tracks are observed, personnel will follow the tracks until either lynx hair or scat are found and collected or the distance tracks are followed exceeds 1 km. All hair found in day beds or a single scat will constitute a sample. Because lynx are a federally listed species, which can result in regulatory protection, we will eliminate doubt about the presence of lynx by submitting hair or scat sampled to a conservation genetics lab to confirm species identification (see McKelvey et al. 2006). All hair and fecal samples will be submitted to a conservation genetics lab for identification to species and individual, if possible. The distance a track is followed will be limited to 1 km to increase efficiency in lynx detection within the PSU (i.e., it will be assumed it is quicker to find a new lynx track to follow to locate hair or scat than to pursue a single track for more than 1 km; see McKelvey et al. 2006). Two levels of search effort for lynx tracks will be implemented within a PSU. The first tracking intensity will be 4 consecutive tracking days (although there may be days of no tracking within this period – e.g., days off, cancellation of tracking effort due to weather etc.), the second will be 8 consecutive days of tracking. All PSU’s will be snow-tracked for 12 days (3 week field effort, see Table 1). This will provide 3 replicates of a 4-day tracking session and 2 replicates of an 8-day tracking session (replicating one of the 4-day tracking sessions). Camera Traps. – Digital infrared surveillance cameras (RECONYX RapidFireTM Professional PC85) will be placed at 5 randomly selected detection stations among those that appear the most likely places where lynx would encounter them within the PSU, as defined above. Cameras will be encased in heavy duty 16 gauge steel security enclosure, attached to a tree with a Master Lock TM PythonTM cable lock and powered by 3-volt C-cell lithium batteries. 48 �We will evaluate detection probabilities for 2 levels of camera surveillance, placing either 2 cameras within the grid or 5 cameras. Five cameras will be placed in all PSU’s, a random subset of 2 cameras from these 5 will be selected to evaluate the efficacy of the lesser effort. Cameras will run continuously for the 3.5 week period. We can evaluate the most efficient number of days required to detect a lynx and the interaction between number of cameras and length of time cameras are active. Hair-Snares. - Barbed wire and carpet hair traps, scented with commercial lynx lures as described by McDaniel et al. (2000) will be placed at each of the detection stations within the PSU in areas where lynx would most likely encounter them (see above). A sample will be defined as all hairs from a single hair snare. Each hair sample will be placed in a uniquely numbered paper envelop, and a flame passed under the barbs to remove any genetic material so that the hair snare can be used again without contaminating future samples. All hair samples will be submitted to a conservation genetics lab for identification to species. Hair snares have been shown to be highly reliable for lynx identification to species (Schwartz et al. 2002) but not for individual lynx identification (Lukacs 2005). We will evaluate detection probabilities of lynx for 3 sample intensity levels of hair snares. First, hair snares will be set up within the PSU at each of the 5 detection stations. A the end of the 3.5 week monitoring session of a PSU, 20 hair snares, at least 100 meters apart (McDaniel et al. 2000) will be placed along the detection route (assuming detection routes will be approximately 25 km long) and collected approximately 1 month later (by the crew leader). Both the detection probability for the 20 hair snares and a random subset of 10 hair snares from these 20 will be selected to evaluate the efficacy of the lesser effort. This larger effort of 20 hair snares will be completed in a PSU after the monitoring conducted by snow-tracking and camera traps as the presence of additional scent stations may affect the use of the 5 camera detection stations. Data Analysis We will estimate the probability of detecting a lynx (p) on each of the PSU’s for each of the detection methods and level of effort for each of those methods. Aerial or satellite telemetry will be used to confirm the presence of at least one lynx in each of the six sampled PSU’s. An evaluation of each of the detection methods will be completed to determine the most reliable, efficient (e.g., cost of equipment, labor) and feasible method of detecting a lynx on a PSU when at least one lynx is present. Project Schedule Completed by Dec. 2009 1. Complete sampling frame and selection of primary sampling units. 2. Purchase and test equipment. Jan.–Mar. 2010 1. Set up detection stations. 2. Conduct lynx snow-tracking surveys. 3. Conduct lynx hair snare sampling. 4. Conduct camera surveillance surveys. 5. Process and submit all genetic samples collected during surveys to a genetic conservation lab (e.g., USDAFS Conservation Genetics Lab in Missoula, Montana, USGS Conservation Genetics Lab in Denver, Colorado). Apr.–May 2010 1. Data entry, analyses and complete report. 49 �Personnel: Project Leader: Tanya Shenk, Wildlife Researcher, CDOW Responsibilities: Design study, work with research associate to implement and complete field work and data entry, complete analysis, write report. Crew Leader: Responsibilities: Assist is study design and selection of PSU’s, supervise field technician, complete all data entry, and perform other duties as needed associated with the post-release monitoring program and the reproduction study. Field Technicians Responsibilities. To establish detection routes, detection stations, place hair snags, cameras and conduct all snow-tracking. Data Analysis: Tanya Shenk, Wildlife Researcher, CDOW Paul Lukacs, Biometrician CDOW Gary White, Professor Emeritus, CSU Paul Doherty, Associate Professor, CSU Estimated Annual Budget: January 2009 – April 2010 Salary (Tech III, Jan 2009 –Apr 2010) Salary (4 Field Technicians, Tech II Jan 2010 – Mar 2010) Travel, housing Misc. Supplies/Operating Equipment Repair, maintenance (snowmobiles) Detection cameras (11 @$1,000 each) Processing of genetic samples collected during monitoring Vehicles (3) $ 15,000 $ 36,100 $ 5,000 $ 4,000 $ 5000 $ 11,000 $ 4,000 $ 6,000 TOTAL $86,100 G. Location: Southwestern and central Colorado is characterized by wide plateaus, river valleys, and rugged mountains that reach elevations over 4200 m. Engelmann spruce-subalpine fir is the most widely distributed coniferous forest type at elevations most typically used by lynx (2591-3353 m). The Core Reintroduction Research Area is defined as areas >2591 m in elevation within the area bounded by the New Mexico state line to the south, Taylor Mesa to the west and Monarch Pass on the north and east (Figure 1). Project headquarters will at the Fort Collins CDOW Research Center. H. Literature Cited: Aubry, K. B., G. M. Koehler, J. R. Squires. 1999. Ecology of Canada lynx in southern boreal forests. Pages 373-396 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. 50 �Byrne, G. 1998. Core area release site selection and considerations for a Canada lynx reintroduction in Colorado. Report for the Colorado Division of Wildlife. Curtis, J. T. 1959. The vegetation of Wisconsin. University of Wisconsin Pres, Madison. Devineau, O., T. M. Shenk, G. C. White, P. F. Doherty Jr., P. M. Lukacs, and R. H. Kahn. 2008. Estimating mortality for a widely dispersing reintroduced carnivore, the Canada lynx (Lynx canadensis). Ecology (in review). Elton, C. and M. Nicholson 1942. The ten-year cycle in numbers of lynx in Canada. Journal of Animal Ecology 11: 215-244. Hodges, K. E. 1999. Ecology of snowshoe hares in southern boreal and montane forests. Pages 163206 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. Koehler, G. M. 1990. Population and habitat characteristics of lynx and snowshoe hares in north central Washington. Canadian Journal of Zoology 68:845-851. Kolbe, J. A., J. R. Squires, T. W. Parker. 2003. An effective box trap for capturing lynx. Journal of Wildlife Management 31:980-985. Laymon, S. A. 1988. The ecology of the spotted owl in the central Sierra Nevada, California. PhD Dissertation, University of California, Berkeley, California. Lukacs, P. M. 2005. Statistical aspects of using genetic markers for individual identification in capturerecapture studies. PhD Dissertation, Colorado State University, Fort Collins, Colorado. MacKenzie, D. I., J. D. Nichols, J. A. Royle, K. H. Pollock, L. L. Bailey, and J. E. Hines. 2006. Occupancy Estimation and Modeling: Inferring Patterns and Dynamics of Species Occurrence. Elsevier Academic Press. Oxford, UK. Major, A. R. 1989. Lynx, Lynx canadensis canadensis (Kerr) predation patterns and habitat use in the Yukon Territory, Canada. M. S. Thesis, State University of New York, Syracuse. McDaniel, G. W., K. S. McKelvey, J. R. Squires. and L. F. Ruggiero. 2000. Efficacy of lures and hair snares to detect lynx. Wildlife Society Bulletin 28:119-123. McKelvey, K. S., J. von Kienast; K.B. Aubry; G. M. Koehler; B. T. Maletzke; J. R. Squires; E. L. Lindquist; S. Loch; M. K. Schwartz. 2006. DNA analysis of hair and scat collected along snow tracks to document the presence of Canada lynx. Wildlife Society Bulletin 34: 451-455. Meaney C. 2002. A review of Canada lynx (Lynx canadensis) abundance records from Colorado in the first quarter of the 20Th century. Colorado Department of Transportation Report. Mowat, G., K. G. Poole, and M. O’Donoghue. 1999. Ecology of lynx in northern Canada and Alaska. Pages 265-306 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University of Colorado Press, Boulder, Colorado. O’Donoghue, M, S. Boutin, D. L. Murray, C. J. Krebs, E. J. Hofer, U. Breitenmoser, C. BreitenmoserWuersten, G. Zuleta, C. , C. Doyle, and V. O. Nams. 2001. Mammalian predators: Coyotes and lynx. in Ecosystem Dynamics of the Boreal Forest: The Kluane Project. eds. C. J. Krebs, S. Boutin and R. Boonstra. Oxford University Press, Inc. New York, New York. Poole, K. G., G. Mowat, and B. G. Slough. 1993. Chemical immobilization of lynx. Wildlife Society Bulletin 21:136-140. Schmidt, K. and R. Kowalcyzk. 2006. Using scent-marking stations to collect hair samples to monitor Eurasian lynx populations. Wildlife Society Bulletin 34: 462-466. Schwartz, M. K. , L. S. Mills, K. S. McKelvey, L. F. Ruggiero, AND F. W. Allendorf. 2002. DNA reveals high dispersal synchronizing the population dynamics of Canada lynx. Nature 415:520522. Schwartz, M. K., L. S. Mills, Y. Ortega, L. F. Ruggiero, and F. W. Allendorf. 2003. Landscape location affects genetic variation of Canada lynx (Lynx canadensis). Molecular Ecology 12:1807-1816. 51 �Shenk, T. M. 1999. Program narrative Study Plan: Post-release monitoring of reintroduced lynx (Lynx canadensis) to Colorado. Colorado Division of Wildlife, Fort Collins, Colorado. __________. 2001. Post-release monitoring of lynx reintroduced to Colorado. Job Progress Report, Colorado Division of Wildlife, Fort Collins, Colorado. __________. 2006. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report, Colorado Division of Wildlife, Fort Collins, Colorado __________. 2007. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report, Colorado Division of Wildlife, Fort Collins, Colorado Silverman, B.W. 1986. Density Estimation for Statistics and Data Analysis. Chapman and Hall. New York, New York, USA. Squires, J. R. and T. Laurion. 1999. Lynx home range and movements in Montana and Wyoming: preliminary results. Pages 337-349 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S McKelvey, and J. R. Squires, editors. Ecology and Conservation of Lynx in the United States. General Technical Report for U. S. D. A. Rocky Mountain Research Station. University Press of Colorado, Boulder, Colorado. U. S. Fish and Wildlife Service. 2000. Endangered and threatened wildlife and plants: final rule to list the contiguous United States distinct population segment of the Canada lynx as a threatened species. Federal Register 65, Number 58. White, G.C. and K. P. Burnham. 1999. Program MARK: Survival estimation from populations of marked animals. Bird Study 46 Supplement, 120-138. Wild, M. A. 1999. Lynx veterinary services and diagnostics. Job Progress Report for the Colorado Division of Wildlife. Fort Collins, Colorado. 52 �Figure 3. Study area depicting the Core Research Area, Lynx-established Core Area and relative lynx use (red is high intensity use, yellow is low intensity use). 53 �Colorado Division of Wildlife July 2009–June 2010 WILDLIFE RESEARCH REPORT State of: Cost Center: Work Package: Task No.: Colorado 3430 0670 1 Federal Aid Project No. N/A : : : : : Division of Wildlife Mammals Research Lynx Conservation Post-Release Monitoring of Lynx Reintroduced to Colorado Period Covered: July 1, 2009 – June 30, 2010 Author: T. M. Shenk Personnel: O. Devineau, R. Dickman, P. Doherty, L. Gepfert, J. Ivan, R. Kahn, A. Keith, P. Lukacs, G. Merrill, B. Smith, T. Spraker, S. Waters, G. White, L. Wolfe All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT In an effort to establish a viable population of Canada lynx (Lynx canadensis) in Colorado, the Colorado Division of Wildlife (CDOW) initiated a reintroduction effort in 1997 with the first lynx released in February 1999. From 1999-2006, 218 wild-caught lynx from Canada and Alaska were released in Colorado. Post-release monitoring was critical to assess and modify the release protocols as they were implemented to improve the survival of released individuals. Average monthly mortality rate in the reintroduction area during the first year post-release decreased with time in captivity from 0.205 [95% CI 0.069, 0.475] for lynx spending up to 7 days in captivity to 0.028 [95% CI 0.012, 0.064] for lynx spending > 45 days in captivity before release. Under the final release protocol, lynx were held in captivity and fed a high quality diet for a minimum of three weeks before release. Results suggested that keeping lynx in captivity beyond 5 or 6 weeks accrued little benefit in terms of monthly survival. We documented survival, movement patterns, reproduction, and landscape habitat-use through aerial (n = 11,580) and satellite (n = 29,258) tracking. Monthly mortality rate was estimated as lower inside the reintroduction area than outside the reintroduction area, and slightly higher for male than for female lynx, although 95% confidence intervals for sexes overlapped. Mortality was higher immediately after release (first month = 0.0368 [SE = 0.0140] inside the study area, and 0.1012 [SE = 0.0359] outside the study area), and then decreased according to a quadratic trend over time. Given the importance of adult survival in the dynamics of long-lived species, the long-term, high survival rates estimated for the reintroduced lynx both inside (0.9315, SE = 0.0325) and outside (0.8219, SE = 0.0744) the reintroduction area are promising for the establishment of a viable population of lynx in Colorado. From 1999-June 2010, there were 122 known mortalities of released adult lynx. Human-caused mortality factors were the highest causes of death with approximately 29.7% attributed to collisions with vehicles or gunshot. Starvation and disease/illness accounted for 18.6% of the deaths while 37.3% of the deaths were from unknown causes. Reproduction was first documented in 2003 with subsequent successful reproduction 1 �in 2004, 2005, 2006, 2009, and 2010. No dens were documented in 2007 or 2008. Reproduction followed a pattern of good and bad years followed by a return to good years in both the reintroduction area and outside the reintroduction area suggesting there may be a cyclic pattern to reproductive output of lynx in Colorado. If the pattern of annual reproductive and survival parameters estimated to date for lynx within the core reintroduction area would repeat over the next 20 years, the population currently in the core reintroduction area would sustain itself at existing densities. To document the continued viability of lynx in Colorado beyond the reintroduction period, some form of long-term monitoring will be needed. A site-occupancy monitoring program using cost-effective, minimally invasive techniques is currently being developed to estimate the extent, stability and potential distribution of lynx throughout Colorado. 2 �WILDLIFE RESEARCH REPORT POST RELEASE MONITORING OF LYNX (LYNX CANADENSIS) REINTRODUCED TO COLORADO TANYA M. SHENK P. N. OBJECTIVE The post-release monitoring of Canada lynx (Lynx canadensis) reintroduced into Colorado emphasized 5 primary objectives: 1. Assess and modify release protocols to ensure the highest probability of survival for each lynx released. 2. Obtain regular locations of released lynx to describe general movement patterns and habitats used by lynx. 3. Determine causes of mortality in reintroduced lynx. 4. Estimate survival of lynx reintroduced to Colorado. 5. Estimate reproduction of lynx reintroduced to Colorado. Three additional objectives were emphasized after lynx displayed site fidelity to an area: 6. Refine descriptions of habitats used by reintroduced lynx. 7. Refine descriptions of daily and overall movement patterns of reintroduced lynx. 8. Describe hunting habits and prey of reintroduced lynx. Information gained to achieve these objectives will form a basis for the development of lynx conservation strategies in the southern Rocky Mountains. SEGMENT OBJECTIVES 1. Complete winter 2009-10 field data collection on lynx habitat use at the landscape scale, hunting behavior, diet, mortalities, and movement patterns. 2. Complete data collection for the pilot study designed to estimate lynx detection probabilities using non-invasive techniques. 3. Complete spring 2010 field data on lynx reproduction. 4. Summarize and analyze data and publish information as Progress Reports, peer-reviewed manuscripts for appropriate scientific journals, or CDOW technical publications (see Appendix I). 5. Complete field research on the post-release monitoring of lynx reintroduced to Colorado and prepare a final report describing status of the lynx reintroduction. INTRODUCTION The Colorado Division of Wildlife implemented the largest Canada lynx (Lynx canadensis), and one of the largest carnivore, reintroductions programs undertaken to date. Thus, evaluating success of this program is critical, and assessing the methods used may prove useful for other ongoing or future carnivore reintroductions. The reintroduction effort was begun in Colorado in 1997, with the first lynx released in the state in 1999. The goal of the Colorado lynx reintroduction program was to establish a self-sustaining, viable population of lynx in this state. The approach taken to reach this goal was to first establish a viable lynx population within a core reintroduction area in southwestern Colorado. From this core reintroduction area, it was hoped that lynx would remain in this area and disperse on their own into 3 �suitable habitat throughout the state. Thus, 218 wild-caught lynx from Canada and Alaska were reintroduced in the core reintroduction area from 1999-2006. There were 7 critical criteria established for achieving a viable lynx population in Colorado: 1) development of release protocols that lead to a high initial post-release survival of reintroduced animals, 2) long-term survival of lynx in Colorado, 3) development of site fidelity by the lynx to areas supporting good habitat in densities sufficient to breed, 4) reintroduced lynx must breed, 5) breeding must lead to reproduction of surviving kittens 6) lynx born in Colorado must reach breeding age and reproduce successfully, and 7) recruitment must equal or be greater than mortality over an extended period of time. These criteria were evaluated incrementally over time to gauge whether the reintroduction effort was progressing toward success (Shenk and Kahn 2002). All seven criteria have now been met. STUDY AREA Byrne (1998) evaluated five areas within Colorado as potential lynx habitat based on (1) relative snowshoe hare densities (Bartmann and Byrne 2001), (2) road density, (3) size of area, (4) juxtaposition of habitats within the area, (5) historical records of lynx observations, and (6) public issues. Based on results from this analysis, the San Juan Mountains of southwestern Colorado were selected as the core reintroduction area, and where all lynx were reintroduced. Wild Canada lynx captured in Alaska, British Columbia, Manitoba, Quebec and Yukon were transported to Colorado and held at The Frisco Creek Wildlife Rehabilitation Center located within the reintroduction area prior to release. Post-release monitoring efforts were focused in a 20,684 km2 study area which included the core reintroduction area, release sites and surrounding high elevation sites (> 2,591 m). The area encompassed the southwest quadrant of Colorado and was bounded on the south by New Mexico, on the west by Utah, on the north by interstate highway 70, and on the east by the Sangre de Cristo Mountains (Figure 1). Southwestern Colorado is characterized by wide plateaus, river valleys, and rugged mountains that reach elevations over 4,200 m. Engelmann spruce/subalpine fir is the most widely distributed coniferous forest type within the study area. The lynx-established core area is roughly bounded by areas used by lynx in the Taylor Park/Collegiate Peak areas in central Colorado and includes areas of continuous use by lynx, including areas used during breeding and denning (Figure 1). METHODS, RESULTS AND DISCUSSION Development of Release Protocols Post-release monitoring was critical to assess and modify the release protocols as they were implemented to improve the survival of released individuals (Shenk 1999). Under the final release protocol, lynx were held in captivity and fed a high quality diet for a minimum of three weeks before release. Thus, they were released in good body condition and one could expect that the longer the captivity, the lower the post-release mortality. This final protocol resulted in high initial post-release survival. Later, detailed analysis of lynx mortality was completed to evaluate how the different release protocols affected mortality within the first year post-release. From this analysis, it was documented that the average monthly mortality rate in the reintroduction area during the first year post-release decreased with time in captivity from 0.205 [95% CI 0.069, 0.475] for lynx spending up to 7 days in captivity to 0.028 [95% CI 0.012, 0.064] for lynx spending > 45 days in captivity before release (Devineau et al. 2010a). The results also suggested that keeping lynx in captivity beyond 5 or 6 weeks accrued little benefit in terms of monthly survival. On a monthly average basis, lynx were as likely to move out (probability = 0.196, SE=0.032) as to move back on (probability = 0.143, SE=0.034) the reintroduction area during the first year after release. Mortality was 1.6x greater outside of the reintroduction area 4 �suggesting that permanent emigration and differential mortality rates on and off reintroduction areas should be factored into sample size calculations for an effective reintroduction effort. Our results will be useful in the development of release and post-release monitoring protocols for future lynx, as well as other carnivore, reintroductions. Long-Term Survival Viability of a reintroduced population requires long-term survival and site fidelity of individuals to the reintroduction area. Over a 10-year period of the reintroduction effort (1999-2009), monthly mortality rate was estimated as lower inside the reintroduction area than outside the reintroduction area, and slightly higher for male than for female lynx, although 95% confidence intervals for sexes overlapped (Devineau et al. 2010). Mortality was higher immediately after release (first month = 0.0368 [SE = 0.0140] inside the study area, and 0.1012 [SE = 0.0359] outside the study area), and then decreased according to a quadratic trend over time. Given the importance of adult survival in the dynamics of longlived species, the long-term, high survival rates estimated for the reintroduced lynx both inside (0.9315, SE = 0.0325) and outside (0.8219, SE = 0.0744) the reintroduction area are promising for the establishment of a viable population of lynx in Colorado (Figure 2, Devineau et al. 2010b). The higher mortality outside the reintroduction area may have been influenced by habitat fragmentation, increased road density and more opportunities for human interactions. From 1999-June 2010, there were 122 known mortalities of released adult lynx. Human-caused mortality factors are currently the highest causes of death with approximately 29.7% attributed to collisions with vehicles or gunshot. Starvation and disease/illness accounted for 18.6% of the deaths while 37.3% of the deaths were from unknown causes. Lynx mortalities were documented throughout all areas lynx used, including 31 (26.3%) occurring in other states. Reproduction Reproduction is necessary to achieve a self-sustaining viable population of lynx in Colorado. Reproduction was first documented from the 2003 reproduction season and again in 2004, 2005 and 2006. Lower reproduction occurred in 2006, although a Colorado-born female gave birth to 2 kittens, documenting the first recruitment of Colorado-born lynx into the Colorado breeding population. No reproduction was documented in 2007 or 2008. The cause of the decreased reproduction from 2006 -08 is unknown. One possible explanation would be a decrease in prey abundance. Reproduction was again observed in 2009 with 5 dens and 10 kittens found in Colorado. Litter size was smaller than previously documented with only 2 kittens found in each litter in comparison to a mean of 2.8 found in previous years. In addition, a sex bias towards female kittens was evident in 2009 which was not evident in prior years. Two litters found in 2009 had both parents born in Colorado, resulting in the first documented third generation Colorado lynx from the reintroduction. The percent of females having dens increased in 2010 to 33%, similar to the highest years documented in 2004-2005. The average number of kittens per litter also returned to the previously observed mean of 2.8. Breeding males and females in 2010 included Colorado-born lynx that have established territories and are now contributing to the breeding population. Reproduction has followed a pattern of good and bad years followed by a return to good years in both the reintroduction area (Figure 3) and outside the reintroduction area suggesting there may be a cyclic pattern to reproductive output of lynx in Colorado. Such a pattern matches the classic Canada lynx-snowshoe hare (Lepus americanus) cycle (Elton 1942). Long-term studies spanning an additional10-20 years would be required to document such a cycle in Colorado. Viability The current lynx population in Colorado is comprised of surviving reintroduced adults, lynx born in Colorado from the reintroduced animals and their offspring and possibly some naturally occurring lynx. To achieve a self-sustaining, viable population of lynx, enough kittens need to be born and 5 �recruited into this population to offset the mortality that occurs and hopefully even exceed the mortality rate to achieve an increasing population. If the pattern of annual reproductive and survival parameters estimated to date for lynx within the core reintroduction area would repeat over the next 20 years, the population currently in the core reintroduction area would sustain itself at existing densities (Figure 4). FUTURE DIRECTIONS Research and monitoring efforts over the last 11 years, since the first lynx were released, have focused primarily on monitoring reintroduced animals through VHF and satellite telemetry and estimating demographic parameters of these animals. However, as more of these animals become unavailable for monitoring due to failed telemetry collars, death or movement out of the core reintroduction area, it becomes more difficult to accurately evaluate the status of the entire lynx population in Colorado, including the core reintroduction area. To document the continued viability of lynx in Colorado beyond the reintroduction period, some form of long-term monitoring will be needed to determine viability for a period of time long enough to encompass possible snowshoe hare cycles. In addition, a challenge facing Colorado Division of Wildlife is how efforts should be allocated between monitoring persistence of lynx that have established within the core reintroduction area and lynx that may be pioneering and expanding into other portions of the state. A site-occupancy monitoring program using cost-effective, minimally invasive techniques is currently being developed to estimate the extent, stability and potential distribution of lynx throughout Colorado (Shenk 2009, Appendix 2). The primary objectives of this monitoring program would be to document the distribution of lynx throughout Colorado and the stability, growth or shrinkage of this distribution over time, and to identify potential areas lynx may occupy in the future. Minimally invasive techniques (e.g., genetic identification, cameras) would be used to detect changes in lynx persistence and distribution as a foundation for assessing whether lynx continue to persist in Colorado. Such noninvasive techniques are widely desirable because they require minimal impact to the animals and are costeffective. The protocols developed will also be made available to any other agencies or entities that want to monitor lynx. Methods to extend this monitoring effort to estimate lynx density are currently being pursued. ADDITIONAL EFFORTS Additional goals of the post-release monitoring program for lynx reintroduced to the southern Rocky Mountains included refining descriptions of habitat use and movement patterns of lynx once lynx established home ranges that encompassed their preferred habitat. This work is ongoing. The program also investigated the ecology of snowshoe hare in Colorado. A study comparing snowshoe hare densities among mature stands of Engelmann spruce (Picea engelmannii)/subalpine fir (Abies lasiocarpa), lodgepole pine (Pinus contorta) and Ponderosa pine (Pinus ponderosa) was completed in 2004 with highest hare densities found in Engelmann spruce/subalpine fir stands and no hares found in Ponderosa pine stands (Zahratka and Shenk 2008). A study to evaluate the importance of young, regenerating lodgepole pine and mature Engelmann spruce/subalpine fir stands in Colorado by examining density and demography of snowshoe hares that reside in each was completed in 2010. Small lodgepole stands supported the highest densities of hares as well as the highest and most consistent recruitment rates. Hares survived best in spruce/fir stands while density and recruitment in these stands were intermediate. Thus, small lodgepole and mature spruce/fir likely provide the most important hare habitat in Colorado; while thinned, medium lodgepole stands appear to be relatively unimportant based on the density and demography measures in this study (J. Ivan, Colorado State University, unpublished data, Appendix 3). However, within the study area, small lodgepole stands occupied only 10% of the area 6 �covered by mature spruce/fir, and we suspect a similar pattern statewide. Additionally, the structure provided by mature spruce/fir stands is less transient than that provided by regenerating lodgepole. Thus, while density and recruitment estimates in spruce/fir stands were somewhat inferior to those collected in small lodgepole, the areal coverage and longevity of spruce/fir likely renders it as important, if not more important, to snowshoe hare and lynx management in Colorado as regenerating lodgepole (J. Ivan, Colorado State University, unpublished data, Appendix 3). Lynx is listed as threatened under the Endangered Species Act (ESA) of 1973, as amended (16 U. S. C. 1531 et. seq.)(U. S. Fish and Wildlife Service 2000). Colorado is included in the federal listing as lynx habitat. Thus, an additional objective of the post-release monitoring program is to develop conservation strategies relevant to lynx in Colorado. To develop these conservation strategies, information specific to the ecology of the lynx in its southern Rocky Mountain range, such as habitat use, movement patterns, mortality factors, survival, and reproduction in Colorado have been and will continue to be provided to regulatory agencies. SUMMARY From results to date it can be concluded that the Colorado Division of Wildlife developed release protocols that ensured high initial post-release survival of lynx, and on an individual level, lynx demonstrated they can survive long-term in areas of Colorado. We also documented that reintroduced lynx exhibited site fidelity, engaged in breeding behavior and produced kittens that were recruited into the Colorado breeding population. Following the successful reproduction in 2010, we have now documented that if the population would repeat the reproduction and mortality patterns documented over the last 10 years the lynx population would continue into the future at sustainable numbers. Thus, the final criterion of a successful reintroduction, documenting recruitment necessary to offset annual mortality, is now supported. To build upon the success of this reintroduction effort, effective conservation and management strategies will need to be developed and implemented to ensure the long-term viability of Canada lynx in Colorado. ACKNOWLEDGEMENTS The Colorado Lynx Reintroduction Program required the continued efforts of numerous personnel in the Colorado Division of Wildlife, other agencies and the general public. Such sustained dedication has resulted in the successful reintroduction of this species to our ecosystems. Funding for the reintroduction program was provided by Colorado Division of Wildlife, Great Outdoors Colorado (GOCO), Vail Associates, Colorado Wildlife Heritage Foundation, Turner Endangered Species Foundation and the U.S.D.A. Forest Service. LITERATURE CITED Bartmann, R. M., and G. Byrne. 2001. Analysis and critique of the 1998 snowshoe hare pellet survey. Colorado Division of Wildlife Report No. 20. Fort Collins, Colorado. Byrne, G. 1998. Core area release site selection and considerations for a Canada lynx reintroduction in Colorado. Report for the Colorado Division of Wildlife. Devineau, O., T. M. Shenk, P. F. Doherty Jr., G. C. White, and R. H. Kahn. 2010. Assessing release protocols for the Colorado Canada lynx (Lynx canadensis) reintroduction. Journal of Wildlife Management (in review). Devineau, O., T. M. Shenk, G. C. White, P. F. Doherty Jr., P. M. Lukacs, and R. H. Kahn. 2010. Evaluating the Canada lynx reintroduction programme in Colorado: patterns in mortality. Journal of Applied Ecology 47:524-531. 7 �Elton, C. and M. Nicholson 1942. The ten-year cycle in numbers of lynx in Canada. Journal of Animal Ecology 11: 215-244. Shenk, T. M. 2002. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report, July: 7- 34. Colorado Division of Wildlife, Fort Collins, Colorado. __________. 2009. Post-release monitoring of lynx reintroduced to Colorado. Wildlife Research Report, July: 1-57. Colorado Division of Wildlife, Fort Collins, Colorado Shenk, T. M. and R. H. Kahn. Lynx reintroduction: report to wildlife commission. Colorado Division of Wildlife. U. S. Fish and Wildlife Service. 2000. Endangered and threatened wildlife and plants: final rule to list the contiguous United States distinct population segment of the Canada lynx as a threatened species. Federal Register 65, Number 58. Zahratka, J. L. and T. M. Shenk. 2008. Population estimates of snowshoe hares in the southern Rocky Mountains. Journal of Wildlife Management 72:906-912. Prepared by __________________________________________________________ Tanya Shenk, Wildlife Researcher & Jake Ivan, Wildlife Researcher 8 �Figure 1. Lynx are monitored throughout Colorado and by satellite throughout the western United States. The lynx core release area, where all lynx were released, is located in southwestern Colorado (outlines in white). A lynx-established core use area has developed in the Taylor Park and Collegiate Peak area in central Colorado. 9 �Figure 2. Variation of monthly mortality rate with time since release for Canada lynx reintroduced to Colorado, inside and outside of the study area, according to the best-AICc model (from Devineau et al. 2010). Only the first 50 months following release are shown. Figure 3. Percent of tracked Canada lynx females in the reintroduction area found with kittens in May or June from 2003 through 2010. 10 �Figure 4. Projected Canada lynx population trend in the core reintroduction area over 20 years if the pattern of reproductive and survival parameters observed over the last 8 years would repeat. The initial population sizes of 50 males and 50 females for this projection was not based on a current population estimate, however, they are not unreasonable assumptions for the study area. Using alternative initial population sizes would not change the projected pattern. 11 �APPENDIX I STATUS OF PUBLICATIONS ASSOCIATED WITH THE COLORADO LYNX REINTRODUCTION PROGRAM Five papers have been published: Devineau, O., T. M. Shenk, G. C. White, P. F. Doherty, Jr., P. M. Lukacs, and R. H. Kahn. 2010. Evaluating the Canada lynx reintroduction programme in Colorado: patterns in mortality. Journal of Applied Ecology 47:524–531. Shenk, T. M., R. H. Kahn, G. Byrne, D. Kenvin, S. Wait, J. Seidel, and J. Mumma. 2009. Canada lynx (Lynx canadensis) reintroduction in Colorado. Pages 410-421 in A. Vargas, C. Breitenmoser, and U. Breitenmoser, editors. Iberian Lynx Ex situ Conservation: An Interdisciplinary Approach. Fundacion Biodiversidad, Madrid, Spain. Shenk, T. M and, R. H. Kahn. 2009. Reintroduction of the Canada lynx (Lynx canadensis) to Colorado. in Proceedings of the Third Iberian Lynx Symposium. eds. A. Vargas, C. Breitenmoser, U. Breitenmoser, Fundacion Biodiversidad and IUCN Cat Specialist Group. Fundacion Biodiversidad, Spain. Wild, M. A., T. M. Shenk, and T. R. Spraker. 2006. Plague as a mortality factor in Canada lynx (Lynx canadensis) reintroduced to Colorado. Journal of Wildlife Diseases 42:646–650. Zahratka, J. L., and T. M. Shenk. 2008. Population estimates of snowshoe hares in the southern Rocky Mountains. Journal of Wildlife Management 72:906–912. Five additional papers are currently in review: Devineau, O., T. M. Shenk, P. F. Doherty, Jr., G. C. White, and R. H. Kahn. In review. Assessing release protocols used for the Canada lynx (Lynx Canadensis) reintroduction in Colorado: Recommendations for future efforts. Journal of Wildlife Management. Devineau, O., T. M. Shenk, P. F. Doherty, Jr., et al. In review. Modeling known-fate and nest survival data within the multistate framework: increased flexibility for telemetry studies. Journal of Applied Ecology. Wolfe, L. L., T. M. Shenk, B. Powell, and T. E. Rocke. In review. Safety of and serum antibody responses to a recombinant F1-V fusion protein vaccine intended to protect Canada lynx (Lynx Canadensis) from plague. Journal of Wildlife Diseases. Fanson, K., T. M. Shenk, et al. In review. Patterns of testicular activity in captive and wild Canada lynx. General and Comparative Endocrinology. Fanson, K., T. M. Shenk, et al. In review. Patterns of ovarian and luteal activity in captive and wild Canada lynx. General and Comparative Endocrinology. One paper is in the process of being submitted for publication and requires no additional work from CDOW personnel: Fanson, K., T. M. Shenk, et al. In prep. Patterns of stress physiology in reintroduced Canada lynx and implications for reintroduction success. General and Comparative Endocrinology. 12 �Six publications are currently in preparation and require the continued efforts of Tanya Shenk and/or Jake Ivan to complete: Theobald, D., and T. M. Shenk. In prep. Lynx habitat use at site-specific and landscape scales. Shenk, T. M. In prep. Lynx denning habitat and reproduction in Colorado. Ivan, J. S., G. C. White, and T. M. Shenk. In Prep. Using telemetry to correct for bias: an approach to estimating density from trapping grids. Ecology. Ivan, J. S., G. C. White, and T. M. Shenk. In Prep. Comparison of methods for estimating density from capture–recapture data. Journal of Applied Ecology. Ivan, J. S., G. C. White, and T. M. Shenk. In Prep. Density and demography of snowshoe hares in westcentral Colorado. Ecological Monographs. Ivan, J. S., G. C. White, and T. M. Shenk. In Prep. Daily and seasonal movements of snowshoe hares in west-central Colorado. Journal of Mammalogy. 13 � https://cpw.cvlcollections.org/files/original/75bc975c72c10bed56af1c21eebf0f92.pdf 534b399d721b0b4f52d6d0629725c6b8 PDF Text Text 35 Colorado Division of Wildlife Wildlife Research Report July 2001 and July 2002 JOB PROGRESS REPORT State of Colorado Division of Wildlife - Mammals Research Work Package No. -~0~6~7~0_ _ _ _ _ __ Task No. 2 Lynx Conservation Ecology of Snowshoe Hares (Lepus americanus) in Colorado Period Covered: July l? 2000 - June 30, 2002 Author: Steven W. Buskirk and Jennifer L. Zahratka Personnel: T. M. Shenk, Ph.D. ABSTRACT Despite what is known about Canada lynx (Lynx canadensis) and snowshoe hare (Lepus americanus) ecology in Canada and Alaska, a paucity of information exists in the contiguous United States. With the listing of the Canada lynx as threatened under the Endangered Species Act in 2000 the need for more knowledge about lynx and their prey become more pressing. The recent reintroduction of Canada lynx to southwestern Colorado (1999) by the state has furthered increased this need. The development of reliable knowledge about snowshoe hare ecology will be key to the recovery and de listing of lynx. Two habitat factors are generally considered overriding in their importance to the abundance and fitness of snowshoe hares: the density of small-diameter (generally< 5 mm) woody stems within reach of the snow surface for food, and the abundance of somewhat larger-diameter woody structure for overhead cover. This project focuses on two central conceptual issues. First, how do site conditions produce woody stems of suitable diameters and heights above the snow for food for snowshoe hares in late winter, and how do site conditions provide overhead cover suitable for hares? Second, do snowshoe hares in fact attain their highest densities in these presumptive high-quality habitats? Ecological information gained about snowshoe hares will be valuable not only to the recovery of Canada lynx in Colorado, but also throughout the range oflynx in the southern U.S. ��37 Ecology of Snowshoe Hares (Lepus Americanus) in Colorado Steven W. Buskirk and Jennifer L. Zahratka Department of Zoology and Physiology, University ofWyoming Laramie, Wyoming 82071 Introduction The snowshoe hare (Lepus americanus) is a widely distributed and well-studied leporid of North American boreal forests. Scientists have long been interested in the snowshoe hare and its cyclic relationship with the Canada lynx (Lynx canadensis). The snowshoe hare is the obligate primary prey item of the lynx, which was listed as threatened under the Endangered Species Act in 2000. Data dealing with the ecology, particularly the habitat ecology, of southern snowshoe hare populations is lacking, especially in the southern Rocky Mountains. Indeed, only a single study (Dolbeer and Clark 1975) described the habitat associations of hares in the southern Rocky Mountains, but only in the most cursory fashion. The reintroduction of Canada lynx to the southern Rocky Mountains in 1999-2000 has further stimulated the need for understanding the habitat requirements of snowshoe hare populations. Therefore, data from the southern Rocky Mountains is critical for understanding the ecology of snowshoe hares in their southern range. The abundance and fitness of snowshoe hares depend on the protection afforded by plants as well as their suitability as foods for hares. Although food is an obvious requirement for snowshoe hare survival, snowshoe hares rarely starve to death. Instead, predation is the overwhelming proximate cause of death for snowshoe hares (Hodges 2000b) and food shortage only predisposes them to predation. The cover afforded by large-diameter woody structure provid~s horizontal and vertical protection from predators (Wolff 1980). Also, small-diameter(< 5-mm) (Grigal and Moody 1980) woody stems< 45 cm from the snow surface (Bider 1961) are an important food source (Hodges 2000a). Whereas largediameter woody stems presumptively provide protection from predation, small-diameter woody sterns presumptively provide nutrition. Therefore, we assume that woody structure in two different size classes meets the needs of snowshoe hares for habitat. Winter is a critical time of year for snowshoe hare survival because fewer woody stems of either size class are available in winter than in other seasons, and herbaceous plants are not available. Understanding how the density of woody stems of different sizes, tree dominants, and successional stage affect densities of snowshoe hares is key to effective management of snowshoe hare habitats in the southern Rocky Mountains. Therefore, we investigated two conceptual issues relating to snowshoe hare habitat in late winter. First, how do site conditions produce woody stems of suitable diameters and heights above the snow surface for food and how do site conditions provide suitable protective cover for hares? Second, do snowshoe hares in fact attain their highest densities in these presumptive high-quality habitats? Study Area Location The study area was a broad area of southwestern Colorado on the Gunnison and Rio Grande National Forests, which we studied during January-April 2002. Within our study area, we established two study sites: one was a 1963-km2 area centered over Taylor Parle Reservoir on the Gunnison National Forest (39° 50' N, 106° 34' W); the second was the Divide District (4,089 km2) of the Rio Grande National Forest (37° 40' N, 106° 40' W) centered directly north of South Fork, Colorado. The Gunnison study area represents the southernmost extent of naturally occurring lodgepole pine. In coniferous forests of the Rocky Mountains, lodgepole pine is an important habitat type for lynx and �38 snowshoe hares. The Rio Grande study area is lower in elevation and contains ponderosa pine, also widely distributed in the Rocky Mountains, and therefore of interest in our study. Topography The landscape of southwestern Colorado is characterized by high, rugged mountains, wide plateaus, and gaping river valleys. Elevation of the Gunnison study site ranged from 2850 m to 3480 m. On the Rio Grande study site the elevation ranged from 2460 m to 2580 m. Our spruce-fir sites occurred at elevations of 32 l 0 - 3480 m, our lodgepole pine sites occurred at 2850 - 3100 m, and our ponderosa pine sites occurred at 2460 - 2580 m. Climate Southwestern Colorado exhibits an arid and temperate climate; strong local variation responds to elevation and aspect. The mean temperature in Gunnison, Colorado from January -April is -7°C. In South Fork, Colorado the mean temperature from January -April is 0°C (National Weather Service, Gunnison, CO, unpublished data). Unlike areas of the Western Slope where more precipitation falls as winter snow than as summer rain, the monsoon season in southwestern Colorado brings most yearly precipitation in late summer. The mean monthly precipitation in Gunnison, Colorado for January -April is 1.6 cm. In South Fork, Colorado the corresponding mean is 1.5 cm (National Weather Service, Gunnison, CO, unpublished data). Methods Trapping grid selection Our study area comprised the Gunnison and Rio Grande National Forests, within which trapping grids were chosen using a GIS database of national forest lands with Common Vegetation Unit (CVU) coverage using the Integrated Resource Inventory protocol (IRI) made available by each of the forests. Two sets of criteria, applied sequentially, were used to select the site of the trapping grids. The first set of criteria was based upon the CVU coverages using GIS: I. Species dominants represented were lodgepole pine, Engelmann spruce, ponderosa pine and riparian (Salix spp.). 2. For forests, structural stage considered was mature (structural stage 4). 3. Vegetation polygons were candidate if ~30 m, but~ 1 km from an improved road. 4. Vegetation polygons were candidate if ~25 ha. 5. Vegetation polygons were candidate if sufficient to admit a 330 m x 550 m (16.5-ha) trapping grid with a 50-m buffer between the edge of ~e trapping-grid an~ the nearest edge ·of the • polygon. Fifteen of the candidate polygons were selected randomly. Within each of these random polygons a 330 m x 550 m rectangle was placed at a randomly generated orientation (0 - 180°). All potential ponderosa pine sites on the Gunnison National Forest were excluded using these criteria. All potential riparian sites on the Rio Grande were excluded using these criteria and no lodgepole pine was available on the Rio Grande to evaluate by CVU layers. Potential sites were visited in random order, at which time we applied the second set of criteria: 1. Forested sites were excluded if ~40% of the trapping grid was dominated by a cover type other than the nominal species dominant. 2. Candidate sites were excluded if inaccessible by snowmobile and snowshoes. 3. Candidate sites were excluded if they held any unmapped roads. 4. Candidate sites were excluded iflogging or thinning had occurred within them. �39 5. Candidate sites were excluded if avalanche danger was present. 6. Candidate sites were excluded if trapping grids were <500 m from a grid that had already been included. The first three from each species dominant to meet these criteria were included as trapping grids. Because of the availability of suitable sites, and for logistical reasons, all three spruce-fir trapping grids, all three lodgepole pine trapping grids and all three riparian trapping grids were selected on the Gunnison National Forest. Only the three ponderosa pine trapping grids were selected on the Rio Grande National Forest. After visiting fourteen sites mapped as lodgepole pine on the Gunnison National Forest, three were found that met our criteria. Fifteen sites mapped as spruce-fir on the Gunnison National Forest were evaluated before three were found that met our criteria. Ten sites tentatively mapped as riparian on the Gunnison were visited, but none were found that met our criteria. Fifteen sites mapped as ponderosa pine-dominant on the Rio Grande National Forest were visited before three were found that met our criteria. Trapping and handling All methods related to trapping and handling were approved by the University of Wyoming Animal Care and Use Committee and by the Colorado Division of Wildlife Animal Care and Use Committee. Snowshoe hares were trapped using Tomahawk Model 204 live traps (18 cm x 18 cm x 51 cm) placed on trapping grids of 84 traps (7 lines of 12 traps each), with 50-m spacing for a trapping grid size of 16.5 ha (Fig. 1). Three replicates for each species dominant were sampled for 6 trap nights, which we assumed to be a closed population for the purposes of mark-recapture models. No reproduction occurred during our winter field season. The trapping grid size and method were developed by Scott Mills and Paul Griffin, University of Montana; we used these methods to maximize comparability between our study and theirs. Upon visiting a suitable site, the trapping grid was flagged and numbered using the UTM coordinates generated by a GPS receiver and a compass bearing (Fig. 1). Traps were placed in suitable habitat within 2 m of the flagging and if necessary, covered with tree branches to provide cover for captured hares. Traps were baited with a mixture of pellets of Timothy grain, alfalfa, com, and oats (TACO), alfalfa pellets and apples (P. Griffin, pers. comm.). Traps were checked as early as possible each morning and re-baited as needed. Once a snowshoe hare was captured, a pillowcase with a drawstring was placed over the front door of the trap. The hare was persuaded into the bag by gently tipping the trap, blowing on the hare, or making noise. Once the hare was in the bag it was immediately weighed using a 2500-g Pesola spring scale. The hare was then carefully placed between the legs of a ~eeling handler with the head facing towards the handler. The second handler marked the hare using a sterile passive-integrated transponq.er (PIT) tag. One tag was injected subcutaneously with a sterile needle between the shoulder blades. Both ears of the snowshoe hare were also marked using a permanent black marker for short-term identification. After the first day of any trapping session (i.e. on traps days 2-6) every snowshoe hare was scanned with a 125-kHz Mini-portable reader to determine whether the hare was a recapture or a new capture. fu the event the snowshoe hare was preyed upon and partially ingested, the earmarks were checked. Each snowshoe hare was sexed by turning the hare on its dorsal side and protruding the genitalia. The forefinger and middle finger were used to apply slight pressure to the vent area just above the anus. Snowshoe hares were then released away from handlers. Snowshoe hares that suffered severe trap or predation injuries were euthanized with a 1-ml intrathoracic injection of sodium pentobarbital. Each carcass was necropsied and the liver and kidneys preserved for metals analysis. After necropsy and tissue collection, euthanized animals were disposed of by cremation or deposited in a landfill. Any non-target species caught-in traps were immediately released. �40 Diet Within spruce-fir stands, where captures were expected to be more common, we marked trap bait in order to determine whether feces collected from traps contained any bait. TACO and alfalfa pellets were marked with a light dusting of fluorescent, non-toxic powder (DayGlo). Fecal pellets were then collected from the inside of each live-trap where a snowshoe hare was captured. Every fecal pellet within the trap was collected, placed in a brown paper bag and allowed to dry at room temperature. After collection the fecal pellets were placed under an ultraviolet light to show presence or absence of any fluorescent marker ingested by the hares. Samples will be submitted to the Wildlife Habitat and Nutrition Lab at Washington State University, Pullman, WA for analysis. Measurement offecal pellets Each fecal pellet was measured to the nearest 0.1 mm using SPI dial calipers. The sizes of all fecal pellets were recorded, and the mean pellet size for each individual hare was calculated. Vegetation measurements Habitat attributes were estimated at two levels: at each trap site and at each trapping grid. Within each trapping grid, vegetation plots were sampled from 15 trap sites, similar to the design of S. Mills (Fig. 1). Methods developed by T. Shenk (Colorado Division of Wildlife) to monitor habitat use by reintroduced lynx to Colorado were followed, with some minor modifications (Fig. 2). Accordingly, a 12-m x 12-m square of 25 points was placed in 5 rows of 5 (3 m apart), centered over the trap location (Fig. 2). The measurements taken at each of the 25 points included: 1. Snow depth (cm), as measured by a calibrated avalanche probe. 2. Understory measured in a column of 3-cm radius around an avalanche probe. a. All live or dead stems and coarse woody debris (CWD) that fall within the 3-cm radius column using the standardized four-letter genus-species code at 3 height categories (00.5 m, 0.51-1.0 m, 1.01-1.5 m) above the snow surface. b. Each of the above stems classified in 3 different diameter categories(< 5 mm, 5.1-10 mm, 10.1-15 mm) measured at the point where the stem hit the avalanche probe. 3. Overstory was measured using a sighting tube ("moosehom') attached to the avalanche probe. a. Species that hit the crosshairs inside the sighting tube were recorded. Multiple hits by the same species were only recorded once. 4. Every shrub within the plot along with its species and diameter at breast height was recorded (dbh). 5. Every tree within the plot along with its species and dbh was recorded. 6. Every snag within the plot along with its dbh was recorded. 7. Every sapling within the plot along with its species was counted. 8. All coarse woody debris (CWD) deemed usable by snowshoe hare for cover or food (i.e., available above the snow) was recorded along with its diameter. At all of the 84 trap sites within the trapping grid, including the 15 trap sites sampled as described above, the following data were measured: 1. Snow depth (cm), as measured by a calibrated avalanche probe. 2. Species of, dbh of, and distance to the closest woody stem in two categories: ~ 1.0 cm - 7.0 and ~ 7 .1 cm at the snow surface. 3. Canopy cover for the center of the trap site, as estimated by the use of a spherical densiometer, in the four cardinal quadrants (NW, NE, SE, SW). The following rules were used for unusual events: 1. If a point in a vegetation plot lay within a tree bole, the tree species and the dbh was written on the data form. �41 2. A snag was defined as any dead tree bole >45 ° from the horizontal. Dead boles <45 ° vertical angle were considered CWD. 3. The mid-point diameter was measured of exposed CWD partially covered by snow. 4. If a leaning tree fell partially outside the 12 m x 12 m sampling plot it was included if >50% of the tree lay within the sampling plot. Results Captures of snowshoe hares by trapping grid and tree species dominant are summarized in Table 1. A total of 28 hares were captured in 4620 trap nights of effort. Mean dbh and density of trees, by species, for the nine trapping grids in three tree species domiriant categories are summarized in Tables 23. Snow depths, and densities of various vegetative structures for the nine trapping grids ( 15-point protocol) in three tree species dominant categories are summarized in Table 4. Corresponding data for the 84-point protocol are summarized in Table 5. Literature Cited Bider, J. R. 1961. An ecological study of the hare Lepus americanus. Canadian Journal of Zoology 39:81-103. Dolbeer, RA., and W.R. Clark. 1975. Population ecology of snowshoe hares in the central Rocky Mountains. Journal of Wildlife Management 39:535-549. Grigal, D. F., and N. R. Moody. 1980. Estimation of browse by size classes for snowshoe hare. Journal of Wildlife Management 44:34-40. Hodges, K. E. 2000a. The ecology of snowshoe hares in northern boreal forests. Pages 117-161 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation oflynx in the United States. University Press of Colorado, Boulder, Colorado. Hodges, K. E. 2000b. Ecology of snowshoe hares in southern boreal and montane forests. Pages 163206 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R Squires, editors. Ecology and conservation oflynx in the United States. University Press of Colorado, Boulder, Colorado. Hoover, R. L., and D. L. Wills. 1987. Managing forested lands for wildlife. Pages 455-477. Colorado Division of Wildlife, Eastwood Printi~g and Publishing, Denver, Colorado. Wolff, J. 0. 1980. The role of habitat patchiness in the population dynamics of snowshoe hares. Ecological Monographs 50:111-130. �42 Figure 1. Schematic of 300 m x 550 m trapping grid to be used for estimating population density of snowshoe hares in southern Colorado. Asterisks(*) indicate the location of the 15 vegetation plots centered on trapping points. Pound signs(#) indicate where the point-quarter method will be used on all other trap locations. 1# 2# 3# 4# 5# 6# 7# 8# 9* 10# 11* 12# 13* 14# 15# 16# 17# 18# 19# 20# 21# 22# 23* 24# 25* 26# 27* 28# 29# 30# 31# 32# 33# 34# 35# 36# 37* 38# 39* 40# 41* 42# 43# 44# 45# 46# 47# 48# 49# 50# 51* 52# 53* 54# 55* 56# 57# 58# 59# 60# 61# 62# 63# 64# 65* 66# 67* 68# 69* 70# 71# 72# 73# 74# 75# 76# 77# 78# 79# 80# 81# 82# 83# 84# �43 Figure 2. Schematic of 12 m x 12 m vegetation plot centered on each of the 15 trap sites (Fig. 1) used in measuring habitat variables for snowshoe hares in southwestern Colorado, late winter 2002. The trap location is at the center of the vegetation plot. II 12m �44 Table I. The number of snowshoe hare captures (1st, 2nd, and 3rd), and total captures on nine trapping grids in three species dominant categories, and trapping effort (trap-nights), southwestern Colorado, late winter 2002. 3rd 1st 2nd Total TrapTrapping grid number LPl LP2 LP3 SFl SF2 SF3 PPl PP2 PP3 Tree species dominant Pinus contorta Pinus contorta Pinus contorta Picea engelmannii, Abies Iasiocarpa Picea engelmannii, Abies lasiocarpa Picea engelmannii, Abies lasiocarpa Pinus ponderosa Pinus ponderosa Pinus ponderosa capture 1 3 1 5 3 15 0 0 0 capture 0 0 0 2 0 6 0 0 0 capture 0 0 0 0 0 2 0 0 0 capture 1 3 1 7 3 23 0 0 0 nights 504 504 504 504 588 504 504 504 504 Table 2. Mean diameter at breast height (dbh) by tree species for 15 trap locations on nine trapping grids in three species dominant categories, southwestern Colorado, late winter 2002. All measurements are in cm± SE (where n> 1). Abies lasiocarpa Pinus contorta Pinus ponderosa Populus tremuloides NA NA NA NA NA NA 14±1 14±1 15±1 23±3 14±1 20±2 13±2 10±1 11±2 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA 27±3 25±3 21±2 11±1 6 3 21±2 18±1 20±3 9±2 Picea Trapping grid number engelmannii LPl LP2 LP3 SFl SF2 SF3 PPl PP2 PP3 MA Psuedotsuga Juniper-us menziesii scopulorum 8 NA Table 3. Mean density by tree species for 15 trap locations on nine trapping grids in three species dominant categories, southwestern Colorado, late winter 2002. All measuremerits are in trees ha- 1•. Picea Trapping grid engelmannii number LPl LP2 LP3 SFl SF2 SF3 PPl PP2 PP3 NA NA 5±5 704±149 1231±159 1194±178 NA NA MA Abies lasiocarpa Pinus contorta NA NA NA 1273±171 1218±231 1310±313 227±64 449±171 449±112 NA NA NA NA NA NA NA NA NA Pinus ponderosa Populus tremuloides NA NA NA NA NA NA 46±19 93±24 120±27 NA NA NA NA NA NA 28±19 5±5 5±5 Psuedotsuga Juniper-us menziesii scopulorum NA NA NA NA NA NA 125±15 14±7 69±22 NA NA NA NA NA NA 51±27 5±5 NA �45 Table 4. Mean snow depth, tree density, sapling density, shrub density, and snag density for 15 trap locations on Colorado, late winter, 2002. All measurements are in cm± SE. Species dominant categories are listed in Table 1. Trapping grid number Snow depth at time of sanipling ± SE Tree density (ha.J) Sapling density Shrub density (ha-I) (ha-I) Snag density (ha- 1) LP I LP2 LP3 SF 1 SF2 SF 3 pp 1 PP2 PP3 37±1 38±2 32±1 70±4 70±3 75±3 0 0 0 1273±171 1218±231 1314±312 931±150 1680±215 1643±180 250±76 116±28 194±31 569±203 333±148 759±203 546±134 749±254 630±154 273±165 481±161 148±67 NA NA NA NA NA NA 315±114 722±305 921±434 620±139 278±95 431±98 162±50 282±44 417±108 162±67 379±125 277±133 Table 5. Mean snow dept~ mean canopy cover, mean diameter at breast height (dbh), and mean distances to nearest stem in two diameter categories for 84 trap locations on nine trapping grids in three species dominant categories, southwestern Colorado, late winter 2002. All measurements are in cm± SE, except canopy cover, which is% ±SE. Trapping grid number LP 1 LP2 LP3 SF 1 SF2 SF 3 pp 1 PP2 pp 3 Species dominant Pinus contorta Pinus contorta Pinus contorta Picea engelmannii, Abies Iasiocarpa Picea engelmannii, Abies Iasiocarpa Picea engelmannii, Abies Iasiocarpa Pinus ponderosa Pinus ponderosa Pinus ponderosa Snow depth at time of sampling Canopy cover dbh of woody stems> 7 cm Distance to nearest stem 1-7 cm dbh Distance to nearest stem > 7 cmdbh 40±1 38±2 40±1 73±2 73±2 79±1 69±2 79±2 15±1 17±1 18±1 28±2 352±35 303±26 351±52 313±38 116±7 146±11 163±14 216±19 66±1 75±2 24±1 238±21 165±16 75±2 85±2 21±1 153±13 145±11 0 0 0 37±4 24±3 48±3 27±1 25±1 26±1 490±41 382±35 479±46 422±37 551±43 291±32 �5 JOB PROGRESS REPORT State of Division of Wildlife- Mammals Research Colorado Work Package No. _ _----"-0-=-67-'--'0"------- Lynx Conservation Task No. Ecology of Snowshoe Hares (Lepus americanus) in Colorado 2 Period Covered: July 1, 2002 - June 30, 2003 Author: Steven W. Buskirk and Jennifer L. Zahratka Personnel: T. M. Shenk Interim Report - Preliminary Results This work continues, and precise analysis of data has yet to be accomplished. Manipulation or interpretation of these data beyond that contained in this report should be labeled as such and is discouraged. ABSTRACT How the densities of woody stems of different sizes, tree dominants, and successional stage affect densities of snowshoe hares is key to effective management of snowshoe hare habitats in the southern Rocky Mountains. Therefore, we investigated two conceptual issues relating to snowshoe hare habitat in late winter. First, how do site conditions produce woody stems of suitable diameters and heights above the snow surface for food and how do site conditions provide suitable protective cover for hares? Second, do snowshoe hares in fact attain their highest densities in these presumptive high-quality habitats? The results in this progress report are preliminary and subject to revision based upon continuing analyses of data. Still, some patterns in the data are apparent. Temperature appeared to have an effect on capture success whereas moon phase, although it has been reported to have an effect, did not. Our preliminary analysis of vegetation data suggests that canopy cover and distance to the nearest 1-7 cm stem also affect capture success. A resource selection model will be generated in the next phase to determine habitat predictors of capture success. Our comparison of diameters of fecal pellets of snowshoe hares and mountain cottontails suggests a difference in size of pellets between the two sympatric lagomorphs that should be useful for identification of pellets to species in the field. �6 Ecology of Snowshoe Hares (Lepus americanus) in Colorado Steven W. Buskirk and Jennifer L. Zahratka Department of Zoology and Physiology University of Wyoming INTRODUCTION The snowshoe hare (Lepus americanus) is a widely distributed and well-studied leporid of North American boreal forests. Scientists have long been interested in the snowshoe hare, its cyclic population fluctuations at high latitudes, and its ecological relationship with the Canada lynx (Lynx canadensis). The snowshoe hare is the obligate primary prey item of the lynx, which was listed as threatened under the Endangered Species Act in 2000 (U.S Fish and Wildlife Service 2000). Data dealing with the ecology, particularly the habitat ecology, of southern snowshoe hare populations are lacking, especially in the southern Rocky Mountains. Indeed, only a single study (Dolbeer and Clark 1975) described the habitat associations of hares in the southern Rocky Mountains, but only in the most cursory fashion. The reintroduction of Canada lynx to the southern Rocky Mountains in 1999-2003 has stimulated the need for understanding the habitat requirements of snowshoe hare populations. Data from the southern Rocky Mountains are critical for managing habitats to conserve lynx and other boreal forest predators at their southernmost limits in the southern Rocky Mountains. The abundance and fitness of snowshoe hares depend on the protection afforded by plants as well as their suitability as food for hares. Although food is an obvious requirement for snowshoe hare survival, snowshoe hares rarely starve to death. Instead, predation is the overwhelming proximate cause of death for snowshoe hares (Hodges 2000a) and food shortage only predisposes them to predation. Largediameter woody structure provides horizontal and vertical protection from predators (Wolff 1980). Also, small-diameter (< 5-mm) (Grigal and Moody 1980) woody stems < 45 cm from the snow surface (Bider 1961) are an important food source (Hodges 2000b). Whereas large-diameter woody stems presumably provide protection from predation, small-diameter woody stems are believed to provide nutrition, particularly in winter. Therefore, we assume that woody structure in two different size classes meets two distinct habitat needs of snowshoe hares. Winter is a critical time of year for snowshoe hare survival because fewer woody stems, large or small, are available than in other seasons, and herbaceous plants are not available. How the densities of woody stems of different sizes, tree dominants, and successional stage affect densities of snowshoe hares is key to effective management of snowshoe hare habitats in the southern Rocky Mountains. Therefore, we investigated two conceptual issues relating to snowshoe hare habitat in late winter. First, how do site conditions produce woody stems of suitable diameters and heights above the snow surface for food and how do site conditions provide suitable protective cover for hares? Second, do snowshoe hares in fact attain their highest densities in these presumptive high-quality habitats? These general questions subsumed more specific ones. 1. In order to understand the links between diet and habitat use in winter, and because diets of snowshoe hares have not been studied in the southern Rocky Mountains, we studied diets of snowshoes hares. 2. Captures of snowshoe hares and non-target leporid species allowed us to collect fecal pellets of known species origin. Because the size of leporid pellets has been used to identify their source to species in the southern Rocky Mountains (Dolbeer and Clark 1975, Hartmann and Byrne 2001) where leporid species are sympatric, we characterized the sizes of fecal pellets of sympatric leporid species, specifically of snowshoe hares and mountain cottontails (Sylvilagus nuttallii). 3. Because various abiotic factors (e.g. air temperature, moon phase) have been reported in the literature (Gilbert and Boutin 1991) or anecdotally to affect capture success of snowshoe hares, we �7 tested for these influences in our data, and accounted for them in our analyses of major treatment effects (e.g. stand type). STUDY AREA Location The study area was a broad area of southwestern Colorado on the Gunnison and Rio Grande National Forests, which we studied during January- April 2002 and January- March 2003. Within our study area, we established two study sites: one was a l 963-km2 area centered over Taylor Park Reservoir on the Gunnison National Forest (39° 50' N, 106° 34' W); the second was the Divide District (4,089 km2) of the Rio Grande National Forest (37° 40' N, 106° 40' W) centered directly north of South Fork, Colorado (Figure 1). Spruce-fir is an important habitat for snowshoe hares throughout its temperate range (Hodges 2000a) and it is the most widely distributed stand type in coniferous forests of Colorado. Approximately 48% of the coniferous forests of Colorado are dominated by spruce-fir (Buttery and Gillam 1987). In Colorado, lodgepole pine accounts for 16% of the coniferous forests (Buttery and Gillam 1987); our Gunnison study area represents the southernmost natural extent of this species. Lodgepole pine is an important habitat type for snowshoe hares in other coniferous forests of the Rocky Mountains (Koehler 1990a, b) and reintroduced lynx have been documented in the Gunnison study area. Therefore, lodgepole pine was included in our study. The Rio Grande study area, although lower in elevation, contains ponderosa pine, also widely distributed in Colorado. About 24% of coniferous forests in Colorado are dominated by ponderosa pine (Buttery and Gillam 1987). Bartmann and Byrne (2001) reported some of their highest densities of lagomorph pellets in ponderosa pine stands. Therefore, it was important for our study to examine the suitability of ponderosa pine stands for snowshoe hares. Topography Southwestern Colorado is characterized by wide plateaus, river valleys, and rugged mountains that reach elevations over 4200 m. Elevations of our Gunnison study site ranged from 2850 m to 3480 m. The Rio Grande study site ranged in elevation from 2460 m to 2580 m. Our spruce-fir sites occurred at elevations of 3210 - 3480 m, our lodgepole pine sites occurred at 2850 - 3100 m, and our ponderosa pine sites occurred at 2460- 2680 m. The overall aspect of each trapping grid varied (Table 1). Climate Southwestern Colorado exhibits an arid and temperate climate; strong local variation reflects elevation and aspect. The mean temperature in Gunnison, Colorado from January - April 2002 was -7°C and in South Fork, Colorado the corresponding mean was 0°C. In 2003 the corresponding mean in Gunnison, Colorado was -5°C and in South Fork was -1°C (Weather Channel web site, unpublished data). Unlike northern Colorado, where more precipitation falls as winter snow than as summer rain, the monsoon season in southwestern Colorado brings most yearly precipitation in late summer. The mean monthly precipitation in Gunnison, Colorado for January - April 2002 was 1.6 cm, whereas in the monsoon months of July and August 2002 the mean was 3.8 cm. In South Fork, Colorado the corresponding means were 1.5 cm and 4.5 cm. �8 METHODS Trapping Grid Selection Our study area comprised the Gunnison and Rio Grande National Forests, within which trapping grids were chosen using a GIS database of national forest lands with Common Vegetation Unit (CVU) coverage using the Integrated Resource Inventory protocol (IRI) made available by each of the forests . Two sets of criteria, applied sequentially, were used to select the site of the trapping grids. The first set of criteria was based upon the CVU coverages using GIS: 1. Stand types included were Engelmann spruce-subalpine fir, lodgepole pine, ponderosa pine and riparian (Salix spp.). 2. Structural stage was mature with canopy cover 2: 40% (SS 4b, 4c) (Buttery and Gillam 1987). 3. Vegetation polygons were considered if 2: 30 m, but :S 1 km from a mapped road, i.e. a highway, paved, graded or gravel road, or a 4-wheel drive road. 4. Vegetation polygons were considered if 2: 25 ha. 5. Vegetation polygons were considered if shaped so as to admit a 330 m x 550 m (16.5-ha) trapping grid with a 50-m buffer between the edge of the trapping grid and the nearest edge of the polygon. 6. Fifteen of the candidate polygons were selected randomly. Within each of these random polygons a 330-m x 550-m rectangle was placed at a randomly generated orientation (0 - 180°). All potential ponderosa pine sites on the Gunnison National Forest were excluded using these criteria. All potential riparian sites on the Rio Grande were excluded using these criteria and no lodgepole pine sites were available on the Rio Grande to evaluate by CVU layers. Potential sites were visited in random order, at which time we applied the second set of criteria: 1. Forested sites were excluded if 2: 40% of the trapping grid was dominated by a cover type other than the nominal species dominant. 2. Candidate sites were excluded if inaccessible by snowmobile and snowshoes. 3. Candidate sites were excluded if they held any unmapped roads. 4. Candidate sites were excluded if logging or thinning had occurred within them. 5. Candidate sites were excluded if avalanche danger was present. 6. Candidate sites were excluded if trapping grids were< 500 m from a grid that had already been included. The first three ~ites from the list of candidates for each stand type to meet these criteria were included as trapping grids. Because of the availability of suitable sites, and for logistical reasons, all spruce-fir trapping grids, all lodgepole pine trapping grids and all riparian trapping grids were evaluated on the Gunnison National Forest. Only the ponderosa pine trapping grids were evaluated on the Rio Grande National Forest. After visiting 14 sites mapped as lodgepole pine on the Gunnison National Forest, three were found that met our criteria. Fifteen sites mapped as spruce-fir on the Gunnison National Forest were evaluated before three were found that met our criteria. Ten sites tentatively mapped as riparian on the Gunnison were visited, but none were found that met our criteria. Fifteen sites mapped as ponderosa pine on the Rio Grande National Forest were visited before three were found that met our criteria. Trapping and Handling All methods related to trapping and handling of animals were approved by the University of Wyoming Animal Care and Use Committee and by the Colorado Division of Wildlife Animal Care and Use Committee. Snowshoe hares were trapped using Tomahawk Model 204 live traps (18 cm x 18 cm x 51 �9 cm) placed on trapping grids of 84 traps (7 lines of 12 traps each), with 50-m spacing for a trapping grid size of 16.5 ha (Figure 2). Three replicates for each stand type were sampled for 6 trap nights, which we assumed to be a closed population for the purposes of mark-recapture models. No reproduction occurred during our winter field season. The trapping grid size and method were developed by Scott Mills and Paul Griffin, University of Montana; we used these methods to maximize comparability between our study and theirs. Upon visiting a suitable site, the trapping grid was flagged and numbered using the UTM coordinates generated by a GPS receiver and a compass bearing (Figure 2). Traps were placed in suitable habitat within 2 m of the flagging and if necessary, covered with tree branches to provide cover for captured hares. Traps were baited with a mixture of pellets of Timothy grain, alfalfa, corn, and oats (TACO), alfalfa pellets and apples (P. Griffin, pers. commun.). Traps were checked as early as possible each morning and re-baited as needed. Once a snowshoe hare was captured, a pillowcase with a drawstring was placed over the front door of the trap. The hare was moved into the bag by gently tipping the trap, blowing on the hare, or making noise. Once the hare was in the bag it was immediately weighed using a 2500-g Pesola spring scale. The hare was then placed between the legs of a kneeling handler with the head facing towards the handler. The second handler marked the hare using a sterile passive-integrated transponder (PIT) tag. One tag was injected subcutaneously with a sterile needle between the shoulder blades. Both ears of the snowshoe hare were also marked using a permanent black marker for short-term identification. After the first day of any trapping session (i.e. on traps days 2-6) every snowshoe hare was scanned with a 125-kHz Miniportable reader to determine whether the hare was a recapture or a new capture. In the event the snowshoe hare was preyed upon and partially ingested, the earmarks were checked. Each snowshoe hare was sexed by turning the hare on its dorsal side and protruding the genitalia. The forefinger and middle finger were used to apply slight pressure to the vent area just above the anus. Snowshoe hares were then released away from handlers. Snowshoe hares that suffered severe trap or predation injuries were euthanized with a 1-ml intrathoracic injection of sodium pentobarbital. Each carcass was necropsied and the liver and kidneys preserved for analysis of metals concentrations. After necropsy and tissue collection, euthanized animals were disposed ofby cremation or deposited in a landfill. Any non-target species caught in traps were immediately released; whole specimens from any mortality of non-target species were donated to the Denver Museum of Nature and Science. Diet In 2003, fecal pellets were collected from the inside of each live-trap where a snowshoe hare was captured and allowed to air dry in kraft brown-paper bags. Fecal pellet samples were randomly selected for diet analyses from 24 individual snowshoe hares: four from each of the three spruce-fir grids and four from each of the three lodgepole pine grids. To reduce the possibility of finding TACO and alfalfa in the diet analyses, only first captures of snowshoe hares were used. Where < 4 snowshoe hares were captured on a trapping grid (e.g. SF 1, LP 2), fecal pellets were collected from fresh snowshoe hare tracks two days after snowfall. Fifteen fecal pellets were required for diet analyses (Bruce Davitt, Washington State University, pers. comm.). If< 15 fecal pellets were collected, a new random sample was chosen. For this reason, one sample (LP 1) was taken from a recaptured snowshoe hare three nights after the initial capture. Fifteen fecal pellets were arbitrarily chosen from each paper bag and transferred to a labeled envelope. Samples were submitted to the Wildlife Habitat and Nutrition Laboratory at Washington State University, Pullman, WA for analysis of diet. �Size of Fecal Pellets We measured snowshoe hare fecal pellets collected in 2002 to 0.1 mm using SPI dial calipers. Fecal pellets were also collected from every mountain cottontail incidentally captured in 2002 and 2003 and measured in the same way. Partial or damaged fecal pellets were eliminated from measurement. We measured the longest diameter for any non-spherical pellets. For snowshoe hares, 32 samples from 23 animals (n = 2374 fecal pellets) were measured. Ten samples from 10 mountain cottontails (n = 655 pellets) were measured. Vegetation Habitat attributes were estimated at two levels: at each trap site and for each trapping grid (Table 2). Within each trapping grid, vegetation was sampled from 15 trap sites, similar to the design of Scott Mills (Figure 2). Methods developed by Tanya Shenk (Colorado Division of Wildlife) to monitor habitat use by reintroduced lynx to Colorado were followed with modification (Figure 3). Accordingly, a 12-m x 12m square of 25 points was placed in 5 rows of 5 (3 m apart), centered over the trap location (Figure 3). The measurements taken at each of the 25 points included: 1. Snow depth (cm), as measured by a calibrated avalanche probe, from the center of each trap location. 2. Understory "hits" measured in a column of 3-cm radius around an avalanche probe. a. All live or dead stems and coarse woody debris (CWD) that fall within the 3-cm radius column using the standardized four-letter genus-species code at 3 height categories (0-0.5 m, 0.51-1.0 m, 1.01 - 1.5 m) above the snow surface. b. Each of the above stems classified in 3 different diameter categories (< 5 mm, 5 .1 - 10 mm, 10.1 - 15 mm) measured at the point where the stem hit the avalanche probe 3. Overstory was measured using a densitometer attached to the avalanche probe. a. Species that hit the crosshairs inside the sighting tube were recorded. Multiple hits by the same species were only recorded once. 4. Every shrub within the plot along with its species and diameter at breast height was recorded (dbh). 5. Every tree within the plot along with its species and dbh was recorded. 6. Every snag within the plot along with its dbh was recorded. 7. Every sapling within the plot along with its species was counted. 8. All coarse woody debris (CWD) deemed usable by snowshoe hare for cover or food (i.e. available above the snow) was recorded along with its diameter. At all of the 84 trap sites within the trapping grid, including the 15 trap sites sampled as described above, the following data were measured: I. Snow depth (cm), as measured by a calibrated avalanche probe. 2. Species of, dbh of, and distance to the closest woody stem in two categories:~ 1.0 cm - 7.0 and~ 7. I cm at the snow surface. 3. Canopy cover for the center of the trap site, as estimated by the use of a spherical densiometer, in the four cardinal quadrants (NW, NE, SE, SW). The following rules were used for unusual events: I. If a point in a vegetation plot lay within a tree bole, the tree species and the dbh was written on the data form. �11 2. A snag was defined as any dead tree bole >45° from the horizontal. Dead boles <45° vertical angle were considered CWD. 3. The mid-point diameter was measured of exposed CWD partially covered by snow. 4. If a leaning tree fell partially outside the 12 m x 12 m sampling plot it was included if>50% of the tree lay within the sampling plot. Temperature and Moon Phase We used daily minimum temperatures recorded by the National Weather Service in Gunnison, Colorado in 2002 and 2003 for each night of trapping. This temperature was intended to represent general weather in the region rather than exact conditions at each trapping grid. We estimated the amount of moonlight for each night of trapping as the percentage of the moon's surface illuminated (Astronomical Applications Department, U.S. Naval Observatory, unpublished data). STATISTICAL ANALYSIS AND PRELIMINARY RESULTS We initially examined our preliminary data for distributional properties and homoscedasticity using SPSS 11. 0. These properties are not important in predictors used in binary logistic regression, but are important in comparisons of means. Where we found substantive violations of assumptions regarding distributional properties, we used the appropriate non-parametric test. Our basic study design involved three stand types as represented by tree species dominants (spruce-fir, lodgepole pine, ponderosa pine). Other predictor variables (e.g. elevation, air temperature, habitat attributes) were highly co-linear with tree stand type and each other (Table 1). Trapping grids in spruce-fir tended to be at higher elevations, and have deeper snow and lower air temperatures (Table 1). Because air temperature (Paul Griffin, University of Montana, pers. comm.) and moon phase (Gilbert and Boutin 1991) have been reported to affect captures of snowshoe hares, we also explored these possible relationships and their relationship to other predictors. We first used binary logistic regression to identify factors measured at the scale of the trapping grid (grid-night= unit of replication) that predict capture success. We included stand type (to include the covariates, elevation and snow depth), percent moon phase, temperature and year as candidate predictors of capture success. In this preliminary analysis stand type and temperature were significant in predicting capture success (Table 3). We then tested how air temperature in Gunnison was related with capture success in spruce-fir and lodgepole pine stands. Although there was no confounding variation with air temperature and stand type (ANOVAF= 98.8, d.f = 2, P = 0.19; spruce-fir = -14°C, lodgepole pine = -I3°C, ponderosa pine = - l l 0 C) we chose to exclude ponderosa pine from this analysis because no hares were captured on the ponderosa pine trapping grids. The relationship between air temperature and captures was significant (t = -3.9, d.f = 45, P < 0.001), with grid-nights for which captures were recorded having mean minimum temperatures of -11 °C, and those for which no captures were recorded having temperatures of -l 8°C. We also examined patterns of captures of snowshoe hares within trapping grids using the response variable of whether a trap location recorded a snowshoe hare capture during either 2002 or 2003. We examined patterns of independence of trap locations within a trapping grid by examining the distribution of trap locations where snowshoe hares were captured, versus those where they were not (Figure 4). We observed no obvious pattern of clumping of successful trap locations, and therefore assumed independence of individual traps as sampling units. When we used the grid-night as the unit of replication (n = 108), and included ponderosa pine trapping grids, trapping success did not differ between years (t = 1.57, df = 106, P = 0.12). However, when trap-night was used as the unit of replication (n = 1512), trapping success did differ between years (t = -3.14, df = 1395, P = 0.002), with more captures in 2003 than 2002. x x x �12 We used binary logistic regression to identify vegetation attributes that predicted capture success for snowshoe hares in an individual trap (trap-night= unit of replication). In this preliminary analysis we found that canopy cover was a significant predictor of capture success at trap locations (Table 5) with successful trap locations ( x = 84% cover) having canopy cover 40% greater than that for unsuccessful trap locations ( x = 60%, Mann-Whitney U = 14086, P < 0.001). The other significant predictor was distance to the nearest woody stem 1-7 cm in diameter, with successful trap locations ( x = 2.0 m) having nearest stems only 56% as far away as unsuccessful trap locations ( x = 3.6 m, M-W U = 21897, P < 0.001). We measured the mean sizes of fecal pellets of snowshoe hares (.x = 8.4 mm) and mountain cottontails ( x = 7 .2 mm) from known species origin and found the means differed (Mann-Whitney U = 26.5, P = 0.001) and 95% confidence intervals did not overlap (Figure 5). DISCUSSION The results in this progress report are preliminary and subject to revision based upon continuing analyses of data. Still, some patterns in the data are apparent. Temperature appeared to have an effect on capture success whereas moon phase, although it has been reported to have an effect, did not. Our preliminary analysis of vegetation data suggests that canopy cover and distance to the nearest 1-7 cm stem also affect capture success. A resource selection model will be generated in the next phase to determine habitat predictors of capture success. Our comparison of diameters of fecal pellets of snowshoe hares and mountain cottontails suggests a difference in size of pellets between the two sympatric lagomorphs that should be useful for identification of pellets to species in the field. �13 LITERATURE CITED Bartmann, R. M. and G. Byrne. 2001. Analysis and critique of the 1998 snowshoe hare pellet survey. Colorado Division of Wildlife, unpublished report no. 20. Bider, J. R 1961. An ecological study of the hare Lepus americanus. Canadian Journal of Zoology 39:81-103. Buttery, R. F. and B. C. Gillam. 1987. Managing forested lands for wildlife. Pages 43-71. Colorado Division of Wildlife, Denver, Colorado. Dolbeer, R. A. and W.R. Clark. 1975. Population ecology of snowshoe hares in the central Rocky Mountains. Journal of Wildlife Management 39:535-549. Gilbert, B. S. and S. Boutin. 1991. Effect of moonlight on winter activity of snowshoe hares. Arctic and Alpine Research. 23:61-65. Grigal, D. F. and N. R Moody. 1980. Estimation of browse by size classes for snowshoe hare. Journal of Wildlife Management 44:34-40. Hodges, K. E. 2000a. The ecology of snowshoe hares in northern boreal forests. Pages 117-161 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation oflynx in the United States. University Press of Coloi,:ado, Boulder, Colorado. Hodges, K. E. 2000b. Ecology of snowshoe hares in southern boreal and montane forests. Pages 163206 in L. F. Ruggiero, K. B. Aubry, S. W. Buskirk, G. M. Koehler, C. J. Krebs, K. S. McKelvey, and J. R. Squires, editors. Ecology and conservation of lynx in the United States. University Press of Colorado, Boulder, Colorado. Hoover, R L. and D. L. Wills. 1987. Managing forested lands for wildlife. Pages 455-477. Colorado Division of Wildlife, Eastwood Printing and Publishing, Denver, Colorado. Koehler, G. M. 1990a. Population and habitat characteristics oflynx and snowshoe hares in north-central Washington. Canadian Journal of Zoology 68:845-851. Koehler, G. M. 1990b. Snowshoe hare, Lepus americanus, use of forest successional stages and population changes during 1985-1989 in north-central Washington. Canadian Field-Naturalist 105:291-293. Lemmon, P. E. 1957. A new instrument for measuring forest overstory density. Journal of Forestry 55:667-668. U.S. Fish and Wildlife Service. 2000. Determination of threatened status for the contiguous U.S. distinct population segment of the Canada lynx and related rule; final rule. U.S. Federal Register 65: 1605116086. Wolff, J. 0. 1980. The role of habitat patchiness in the population dynamics of snowshoe hares. Ecological Monographs 50: 111-130. �Table I. Abiotic characteristics of nine trapping grids in three stand types, southwestern Colorado, late winter 2002 and 2003. Snow depth (cm) is the mean (SE) measured at 84 trap locations at each trapping grid. Temperature (0 C) (SE) is the mean low temperature recorded in Gunnison for each grid-night. The aspect of each trapping grid is shown in degrees in their respective order. Trapping grids Stand Type Snow depth Elevation Temperature Aspect PPl,PP2,PP3 Pinus ponderosa 2 (0.4) 2600 -11 (I) 50°, 130°, 130° LP 1, LP 2, LP 3 Pinus contorta 45 (1) 3000 -13 (I) 230°, 90°, 130° SF 1, SF 2, SF 3 Picea engelmanii, Abies lasiocarpa 74 (I) 3400 -14 (I) 310°, 150°, 110° Table 2. Vegetation characteristics for nine trapping grids (see Table I) in three stand types (n = 3 each), southwestern Colorado. Mean tree density, mean sapling density, and mean snag density (number ha- 1) (SE) were counted at 15 trap locations on each grid, late winter 2002. Mean canopy cover(%) (SE) was measured at 84 trap locations on each grid using a densiometer, late winter 2002. The median horizontal cover(%) was measured at 15 trap locations on each grid using a horizontal profile board, late winter 2003. Stand Type Tree density Sapling density Snag density Canopy cover Horizontal cover Pinus ponderosa 187 (29) 301 (81) 273 (65) 36 (2) 0 Pinus contorta 1268 (138) 554 (I 09) 443 (67) 73 (1) Picea engelmanii, Abies lasiocarpa 1418 (116) 642 (107) 287 (44) 79 (1) 65 �Table 3. Preliminary results of binary logistic regression using stand type (excluding ponderosa pine) and abiotic factors as variables to predict capture success (n = 72). Variables are described fully in the methods section. 95% C.I. Coefficient z Temperature 0.169 3.3 Stand type 1.794 2.7 Year -0.210 Moonlight Constant Variable p Odds Ratio Lower Upper 0.001 1.184 1.071 1.309 1 0.006 6.012 NA -0.3 1 0.771 0.811 NA -0.003 -0.2 1 0.808 0.997 -1.146 -0.6 1 0.541 0.318 d.f 0.997 0.808 NA Table 4. Preliminary results of binary logistic regression using vegetation characteristics to predict capture success at trapping locations within trapping grids (n= 108). Variables are described fully in the methods section. 95% C.I. Coefficient -z- d.f p Odds Ratio Lower Upper Canopy cover 0.061 6.10 1 < 0.001 1.063 1.043 1.084 Diameter 1-7 cm -0.005 0.08 0.942 0.995 0.880 l.126 Distance 1-7 cm -0.002 -2.00 0.002 0.998 0.997 0.999 Diameter >7 cm -0.014 -1.17 0.235 0.986 0.963 l.009 Distance >7 cm 0.001 1.00 0.148 1.001 0.999 1.003 Constant -6.058 -6.44 < 0.001 0.002 Variable NA Vl �SF 3 ' LP3 f SF1 Taylor Park Reservoir PP 1-- ', Rio Grande R. PP 2 • r - -- - - - - - - - -- - PP31mil! \~.,., . ------=:~~ <1.~'t ., ., i•., A ~ ~Ol,.. s.fQI" " South Fork 1 ~ I Paved highways 1-- -- - - -I Rivers and streams I I 0 I km N * I I 10 Figure 1. Location of nine trapping grids ( B), southwestern Colorado, 2002 - 2003. SF is spruce-fir, LP is lodgepole pine, and PP is ponderosa pine. Trapping grids are not to scale. �17 1# 2# 3# 4# 5# 6# 7# 8# 9* 10# 11* 12# 13* 14# 15# · 16# 17# 18# 19# 20# 21# 22# 23* 24# 25* 26# 27* 28# 29# 30# 31# 32# 33# 34# 35# 36# 37* 38# 39* 40# 41* 42# 43# 44# 45# 46# 47# 48# 49# 50# 51* 52# 53* 54# 55* 56# 57# 58# 59# 60# 61# 62# 63# 64# 65* 66# 67* 68# 69* 70# 71# 72# 73# 74# 75# 76# 77# 78# 79# 80# 81# 82# 83# 84# Figure 2. Schematic of 300 m x 550 m trapping grid for estimating population density of snowshoe hares in southern Colorado. Asterisks("') indicate the location of the 15 vegetation plots centered on trapping points. Pound signs (#) indicate where the point-quarter method will be used on all other trap locations. �18 11111111111111111:1:111 111111111111 111111111 1 ■ ■ ■ ■ ■ ■ ■ 1 111111111111:1 1::1:1111 1 11 11111:11~1111111:11111 111 1 1 1 1 111 11~1 ~111111111111 trap 11111111:11:~1111111:111 ■ 1 11111111111~11111111111 1 1 1111:1111 :~111111111111 l:lliiii:li~:illllJIIIII ■ ■ 1 11111:1::11~:111 11:1:11 ■ 12 m Figure 3. Schematic of 12 m x 12 m vegetation plot centered on each of the 15 trap sites (Figure 2) used in measuring habitat variables for snowshoe hares in southwestern Colorado, late winter 2002. The trap location is at the center of the vegetation plot. �19 SF 1 •0 00 •• •0 •0 00 • • 0 0 0 0 0 • • 0 0 0 0 0 0 • 0 0 0 0 • • • 0 0 0 0 • • • 0 0 0 0 • • • 0 0 0 0 •0 • 0 0 0 •0 •0 0 0 0 0 0 • • 0 0 0 • • 0 0 0 • • •• •• 0 0 0 0 0 0 • 0 0 0 0 0 0 0 0 0 0 0 0 •0 0 0 0 0 0 0 0 •0 00 00 00 00 00 00 LP 1 •0 00 00 00 00 00 00 0 0 0 •0 •0 0 0 n= 33 SF 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 • 0 0 0 0 0 0 0 0 0 0 •0 • 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 • 0 0 0 0 0 •0 ••• 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 n = 11 0 0 0 0 0 0 0 0 0 0 0 0 0 0 n=4 • LP 2 n=4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 •0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 •0 •0 •0 0 0 0 0 0 0 SF 3 • 00 00 • 00 • • • 0 0 •0 0 •0 • •0 • 0 0 0 •0 •0 •0 0 • •• 0 0 0 0 0 0 0 • 00 00 0 00 •0 •0 •0 0 •0 0 • 0 0 •0 0 • •0 0 0 0 •0 0• • 0 0 0 0 0 • • LP 3 0 • 0 0 0 0 0 0 0 •0 •0 •0 00 •0 00 0 n= 30 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 •0 ••0 • n=9 Figure 4. 12 x 7 trapping grid schematic representing snowshoe hare trap successes(•) for 6 trapping grids in two stand types, spruce-fir and lodgepole pine (snowshoe hares were absent from all pondersosa pine trapping grids), southwestern Colorado, late winter 2002 and 2003. �20 10 9 8 - E E 'Q)"' ..... Q) E ~ ~ 7 6 5 4 3 2 o....__-------~--------.---------' L. americanus S. nuttallii Figure 5. Mean diameters of fecal pellets of Lepus americanus (n = 23 animals) and Sylvilagus nuttallii (n = 10 animals). Error bars indicate 95% confidence interval. �Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Project No. Colorado : : Cost Center 3430 Mammals Research Work Package No. Task No. 0670 2 : Lynx Conservation : Ecology of Snowshoe Hares (Lepus americanus) in Colorado Federal Aid Project: N/A : Period Covered: July 1, 2003 – June 30, 2004 Author: Jennifer L. Zahratka Personnel: S. W. Buskirk , T. M. Shenk All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT A thesis, entitled ‘The population and habitat ecology of snowshoe hares (Lepus americanus) in the southern Rocky Mountains’ was completed and submitted to the University of Wyoming in partial fulfillment of a Master of Science degree. The thesis is available from The Colorado Division of Wildlife Library or the University of Wyoming Library. Included in this report is an abstract of the thesis. 15 �JOB PROGRESS REPORT THE POPULATION AND HABITAT ECOLOGY OF SNOWSHOE HARES (Lepus americanus) IN THE SOUTHERN ROCKY MOUNTAINS Jennifer L. Zahratka ABSTRACT To better understand the population ecology and habitat associations of snowshoe hares (Lepus americanus), I studied snowshoe hares in southwestern Colorado in winters 2002 and 2003. I estimated densities from mark-recapture data and compared vegetative attributes in the mature structural stage (SS 4) among three stand types: Engelmann spruce (Picea engelmannii)–subalpine fir (Abies lasiocarpa), lodgepole pine (Pinus contorta), and ponderosa pine (Pinus ponderosa). I used three methods to calculate a boundary strip width (W) to estimate the effective area trapped Â(Ŵ) in order to illustrate the effect of different methods of estimating Â(Ŵ) on estimates of density. Density estimates [ D̂ = N̂ / Â(Ŵ)] in mature spruce-fir ranged from 0.1 ± 0.03 (SE) hares/ha to 0.9 ± 0.1 hares/ha in 2002 and 0.3 ± 0.05 to 1.0 ± 0.1 hares/ha in 2003. I report only minimum number alive (MNA) in lodgepole pine due to too few captures to estimate density. No snowshoe hares were captured in mature ponderosa pine stands. Model selection based upon the corrected Akaike’s Information Criterion (AICc) showed a strong relationship between MNA and understory cover, density of woody stems 1-7 cm in diameter, and the availability of suitable woody stems for food among the mature stand types I studied (R = 0.91, df = 8, P = 0.008). My empirical data support the assumption that snowshoe hares select habitat with protection from predation. However, the availability of suitable woody stems for food is also an important vegetative attribute for hare habitat. Snowshoe hares selected for spruce-fir among the mature stand types I studied. Mature spruce-fir provided more understory, greater density of woody stems 1-7 cm in diameter, and more woody stems (<1.5 cm) for food. In my study, the winter diet of snowshoe hares was overwhelmingly gymnosperms. Extremely low temperatures affected capture success, but moon phase did not. Counts of fecal pellets are an attractive tool to estimate densities of snowshoe hares because they are less costly and less labor-intensive than conventional mark-recapture techniques. In the southern Rocky Mountains, snowshoe hares and mountain cottontails (Sylvilagus nuttallii) are syntopic. Indeed, I captured two mountain cottontails in two traps in which I captured snowshoe hares. Therefore, distinguishing between fecal pellets is necessary for making inferences specific to these species. Methods to distinguish between the two leporid species have been developed based upon the assumption that the larger snowshoe hare produces larger fecal pellets than the smaller mountain cottontail. In this study, I measured 655 fecal pellets from 10 individual mountain cottontails and 2,374 fecal pellets from 23 individual snowshoe hares: I found no apparent relationship between the body weight of mountain cottontails or snowshoe hares and the size of their fecal pellets (mountain cottontails: r = 0.04, F = 0.01, P = 0.91; snowshoe hares: r = 0.48, F = 9.3, P = 0.005). Although the two species differed in the size of their fecal pellets, the difference (1.2 mm) would be indistinguishable without measuring equipment and is only applicable to adults. While fecal pellet counts may be accurately used to estimate densities of snowshoe hares in the boreal forests of Canada, in the southern Rocky Mountains where leporid species are potentially syntopic, this method may yield misleading results. 16 � https://cpw.cvlcollections.org/files/original/78a2e6cd6a7e6aff02962f3b526d9cfd.pdf 2981cc3bb9387fd35c3e0cb87d77d25d PDF Text Text 115 Colorado Division of Wildlife Wildlife Research Report July 2000 JOB PROGRESS REPORT State of ----~~=~-----Colorado Project No. W-153-R-13 Mammals Research Work Package No. ___ 0_88_0_ _ _ _ _ __ Task No. 1 Black-footed Ferret Recovery Monitoring and Managing Disease in Black-footed Ferrets Period Covered: July. 1, 1999 - June 30, 2000. Authors: M.A. Wild and K. T. Castle Personnel: E. Wheeler, E. Schmal, and S. Kasven. \ ) ABSTRACT \ ) The black-footed ferret is a federally listed endangered species in the United States. Black-footed ferrets have been extirpated from Colorado, but were scheduled to be reintroduced to Moffat County, Colorado in 1999. However, due to high plague activity and low prairie dog densities at the Little Snake Management Area (LSMA), black-footed ferret reintroduction was postponed indefinitely at the site. The secondary site at Coyote Basin, Utah was readied and reintroduction of ferrets was performed there in 1999. The Wolf Creek site in Moffat County, Colorado is also being readied for reintroduction of ferrets, likely in 2001. Our work in support of black-footed ferret reintroduction can be sub-divided into three broad sections: disease monitoring in the proposed release areas in Colorado, care of black-footed ferrets, and flea control as a tool to manage sylvatic plague in prairie dogs and black-footed ferrets. Disease monitoring was performed using collection of coyotes from LSMA in July 1999 and from the Wolf Creek site in February 2000. Two potentially devastating diseases, canine distemper and plague, are present at LSMA. Although prevalence of positive titers to canine distemper virus (CDV) in coyotes ·have been relatively low over the last 3 years (:S33%), the prevalence of positive titers to plague (Yersina pestis) have been high (up to 89% positive). Titers were present in both adult and juvenile animals, suggesting ongoing plague activity in at least some sections of the management area. Samples collected from coyotes at the Wolf Creek site indicated substantially lower disease activity than at LSMA. Prevalence of positive titers to plague and to CDV was 7% of samples collected in February 2000. In general, captive black-footed ferrets maintained at the LSMA breeding and preconditioning pens remained healthy. One ferret died and another was presumed dead in the burrow system after a severe hailstorm. Of 13 kits born at the pens in spring 1999, 12 survived to weaning in fall 1999. These 12 kits in addition to 50 other black-footed ferrets were released at the Coyote Basin site in fall 1999. Twenty ferrets were maintained in the pens overwinter and produced 11 kits in spring 2000. We summarized research results and presented a paper titled "Dose titration and safety of lufenuron fed to captive white-tailed prairie dogs (Cynomys leucurus)" at the 2000 meeting of the �116 Wildlife Disease Association, Jackson, Wyoming. We also performed a trial to test' the efficacy of lufenuron in controlling fleas on captive prairie dogs. One week post-dosing, fleas that fed on prairie dog treated with 500 mg/kg lufenuron prcxiuced a lower proportion (p < 0.05) of viable rggs than those fed on control prairie dogs (mean 0.16 vs. 0.39, respectively). However, the proportion of viable eggs prcxiuced during week 2 post-dosing was low for each group (0.24 treatment vs. 0.17 control). By week 4 postdosing, egg prcxiuction had fallen to nearly zero in each group; however, the three eggs produced by fleas fed on treated prairie dogs were all viable. Although further, similarly controlled investigation is warranted, preliminary results suggest that given current techniques, the likelihood of lufenuron limiting fleas populations and breaking the plague cycle is not extremely promising. �117 MONITORING AND MANAGING DISEASE IN BLACK-FOOTED FERRETS Margaret A. Wild and Kevin T. Castle P. N. OBJECTIVES 1. Monitor disease activity threatening survival of black-footed ferrets reintroduced into the Little Snake Management Area (LSMA). 2. Develop techniques to manage plague in the LSMA using insect growth regulators applied orally to prairie dogs. SEGMENT OBJECTIVES 1. Provide ;ete~ care to captive and reintroduced black-footed ferrets. 2. Monitor and manage plague activity in LSMA. METHODS AND MATERIALS Carnivore Disease Survey Infectious diseases can severely impact the success of black-footed ferret (}Juste/a nigripes) reintroduction efforts. As part of the black-footed ferret reintroduction protocol, we monitored disease activity in carnivores at proposed ferret reintroduction sites: in July 1999 at the Little Snake Management Area (LSMA), Colorado and in February 2000 at the Wolf Creek Management Area (WCMA), Colorado. Coyotes (Canis latrans) were collected for post-mortem examination and samples collected as described in the Program Narrative (Wild and Castle 1998). Black-footed Ferret Reintroduction and Veterinary Care I assisted in preparation of the black-footed ferret allocation request submitted to the US Fish and Wildlife Service by the Colorado-Utah black-footed ferret recovery working team in 2000. I provided veterinary care and consultation on health matters for captive black-footed ferrets and black-footed ferret releases. Animal care was performed in accordance with the established protocol (Wild and Castle 1999). Flea Control In Prairie Dogs We completed the experiment "Bioavailability of lufenuron administered orally to captive white-tailed prairie dogs (Cynomys leucurus)" and performed a follow-up experiment "Efficacy of lufenuron for the control of fleas in white-tailed prairie dogs (Cynomy leucurus)". These studies were outlined in Wild and Castle 1998. The detailed study plan for the lufenuron bioavailability study and e:fp.cacy trial are reported in Wild and Castle 1999 and in Attachment 1, respectively. �118 RESULTS AND DISCUSSION Carnivore Disease Survey With the assistance of USDA Wildlife Services and the Bureau of Land Management (BLM) we collected 16 coyotes from the LSMA between 26-30 July 1999 and 14 coyotes from the Wolf Creek site on 8-9 February 2000. Coyotes were collected using a combination of calling and aerial gunning. Further collections were not possible due to weather constraints and aircraft availability. No lesions indicative of active disease were noted on gross examination of carcasses. Of the coyotes collected from the LSMA, 31 % (5/16) had positive titers to plague (Fig. I) using the standard HA/HI test while one additional coyote was positive using the ELISA test. Interestingly, all juveniles sampled this summer (n = 9) were negative to plague. Because the juvenile coyotes have consumed (sampled) prairie dogs from this summer only, lack of exposure may indicate that the prevalence of plague in prairie dogs in the area is declining. If this is the case, titers in adults would likely be from exposure in previous years. Alternatively, the prey base of the coyotes may have shifted to species that are less commonly infected with plague (e.g., rabbits) if the density of prairie dogs has been greatly reduced by plague. Forty-three percent of coyotes sampled (6/14 usable samples) had positive titers to tularemia. In contrast to plague, all positive titers to tularemia were observed in juvenile coyotes. As previously observed, tularemia appears to elicit a serologic response of short duration in coyotes. The impact of tularemia on black-footed ferrets and prairie dogs is unknown and warrants investigation. A titer to canine distemper virus (CDV) was found in only one of 14 serum samples tested (Fig. 2). Samples collected from coyotes at the Wolf Creek site indicate substantially lower disease activity than at LSMA. A 9-year-old male coyote had positive titers to plague and to CDV, but all other coyotes (n = 13) were negative to plague and CDV (7% positive; Fig. 3). The adult male and one additional juvenile coyote also had titers to tularemia (14% positive). Black-footed Ferret Reintroduction and Veterinary Care In general, captive black-footed ferrets maintained at LSMA remained healthy. One adult female was treated for an apparent mite infection and secondary bacterial pyoderma. Unfortunately, samples to confirm this diagnosis could not be collected prior to treatment; however, the ferret responded quickly and completely to therapy with ivermectin and amoxicillin. Additional cases of crusty skin have been successfully treated with ivermectin but confirmation of the etiology has not yet been made. A I-yr-old male was found dead and a 4-yr-old female was missing and presumed dead after a severe hailstorm. Necropsy results revealed death from blunt trauma. The death(s) occurred in the new pen complex where shallow burrow systems may have been flooded forcing ferrets to the surface in the severe weather. To avoid this problem in the future, additional above ground shelter will be provided. Of 13 kits born at the pens in spring 1999, 12 survived to weaning in fall 1999. One kit disappeared and is presumed dead in the burrow system. Based on results of carnivore disease monitoring over the past 3 years and prairie dog inventories performed in summer 1999, LSMA was determined to be currently unsuitable habitat for the release of black-footed ferrets. Prairie dog inventories performed by Utah Division of Wildlife showed insufficient densities of prairie dogs to support black-footed ferret reintroduction. As a result, ferrets were not released into LSMA but instead were released into Coyote Basin, Utah. Continued monitoring will determine when (if) LSMA can support reintroduction of black-footed ferrets. A secondary site in Colorado, (Wolf Creek) is also being readied for reintroduction of black-footed ferrets in fall 2000 or 2001. _ _, �119 The 12 kits produced onsite, in addition to three 1-yr-old males, and five 3-yr-old females from the LSMA pens were released at Coyote Basin in November 1999. Additionally, 19 kits and five 3-yr-old females from other captive breeding sites were pre-conditioned at the LSMA prior to release at Coyote Basin. The reintroduction was further supplemented with 28 ferrets released immediately upon arrival from other captive breeding sites without pre-conditioning at LSMA. Prior to release, ferrets were trapped for routine examination, treatment, and identification (Wild and Castle 1999) and a health certificate was issued for each individual. All appeared healthy except for the presence of ectoparasites (ticks, mites, fleas). Individual ferrets were treated with ivermectin and pen dusting was advised. In an attempt to meet our ideal age and sex structure of captive black-footed ferrets (Wild and Castle 1999), we retained seven ferrets and supplemented the population with 13 additional ferrets from other captive breeding facilities. Seven black-footed ferrets (five 1-yr-old females and two males) were retained in captivity at the LSMA pens. In November 1999, six adult females,'three female kits, and four male kits were added to this breeding group bringing the total number of ferrets at the LSMA pens to 20. Ferrets were maintained under the standard care protocol (Wild and Castle 1999). Females were paired with males in spring 2000, and four litters resulted. One litter was apparently consumed by the female when about 1 day of age. The other three litters yielded 11 kits. Flea Control In Prairie Dogs We presented results of the lufenuron dose titration study at the 2000 meeting of the Wildlife Disease Association, Jackson, Wyoming. The abstract of that presentation read: DOSE-TITRATION AND SAFETY OF LUFENURON FED TO CAPTIVE WHITE-TAILED PRAIRIE IX>GS (CYNOMYS LEUCURUS) KEVIN T. CASTLE Colorado Division of Wildlife, 317 W. Prospect St. Fort Collins, CO, 80521, MARGARET A. WILD, Colorado Division of Wildlife, 317 W. Prospect St. Fort Collins, CO, 80521, and S. CRAIG PARKS, Novartis Animal Health, P.O. Box 26402, Greensboro, NC 27404. Plague is a zoonotic disease that impacts populations of prairie dogs (Cynomys spp.) and other species, such as black-footed ferrets, that rely on them for food and shelter. Yerstnia pestis, the etiological agent of plague, is transmitted primarily by the bite of an infective flea. Recently developed compounds used to control fleas in pet animals offer a promising alternative to insecticide dusts for the control of fleas in wild rodents. Lufenuron is a lipid-soluble insect growth regulator with ovicidal and larvicidal activity. Lufenuron is efficacious for controlling fleas in cats and dogs at blood concentrations above 50-100 parts per billion (ppb). A single oral dose oflufenuron has been shown to be effective in controlling the cat flea (Ctenocephaltdes felts felts) for at least 30 days in treated cats and dogs. To date, there have been no studies conducted to detennine the duration oflufenuron blood concentrations in any prairie dog species. We compared lufenuron blood concentrations in white-tailed prairie dogs (C. leucurus) during periods of activity (nontorpid group) and hibernation (torpid group) during January-March 1999. We hypothesized that if high serum concentrations of lufenuron could be maintained over winter during hibernation or for > 1 mo in active prairie dogs, the compound may be effective for use in breaking the plague cycle. Thirty captive WTPD were fed 300 rrig/kg lufenuron; half the animals were allowed to become torpid, while the other half were kept awake. All animals remained healthy throughout the 9 week study period. Prairie dogs in the active group gained weight, while those in the torpid group lost weight over the 9 weeks. Blood was drawn from each animal prior to dosing, one week after dosing, then every other week until week 9 post-dosing. Serum was harvested and tested by HPLC for lufenuron concentration. Blood lufenuron concentration did not differ between the groups one week post-dosing. Concentration in both groups decreased over time, but the �120 concentration in torpid animals declined at a more gradual rate; after weeks 3, 5, and 7, lufenuron levels in torpid WfPD were significantly higher than levels in nontorpid WfPD. After nine weeks, blood levels were again similar, and had approached the limit of detection (10 ppb). Blood levels in nontorpid WfPD declined to <50 ppb after 3 weeks, while levels in torpid WfPD declined to <50 ppb after 7 weeks. Future studies will be required to determine efficacy of lufenuron in controlling fleas on WfPD. If effective blood concentrations are similar to dogs and cats, however, frequent dosing would be required to control flea numbers on prairie dogs and thus break the plague cycle. Results from this experiment indicated that blood concentrations of lufenuron did not reach initial levels as high as anticipated, nor were the concentrations maintained above 50 ppb for as long as anticipated (Fig. 4). The most likely cause of these low concentrations was poor absorption of the drug by prairie dogs. lbis may be due to the difference in gut morphology between rodents and carnivores. Alternatively, other aspects of pharmacokenetics may have been responsible for the low serum levels in prairie dogs despite dosing at rates 10-30 times higher than those recommended for cats and dogs, respectively. We followed up the dose titration experiment with an experiment to test the efficacy of oral lufenuron to control fleas on captive prairie dogs. Based on data from the bioavailability study, we increased the lufenuron dose to 500 mg/kg. Serum samples from the study have been submitted for lufenuron assay. Results are pending. Interpretation of efficacy results will rely on these serum lufenuron levels; however, we were able to make some preliminary comparisons between performance of fleas fed on treatment and control prairie dogs. One week post-dosing, fleas that fed on treated prairie dog produced a lower proportion (p < 0.05) of viable eggs than those fed on control prairie dogs (mean 0.16 vs. 039, respectively). Unfortunately, the proportion of viable eggs produced during week 2 post-dosing was low for each group (0.24 treatment vs. 0.17 control). By week 4 post-dosing, egg production had fallen to nearly zero in each group; however, the three eggs produced by fleas fed on treated prairie dogs were all viable. We are uncertain of the cause of the dramatic reduction in egg production and viability observed during the course of the experiment. Anecdotal reports suggest that flea production may be influenced seasonally (or at least cyclically) despite attempts to maintain controlled environmental conditions in the insectary (Metzger, Pers. Comm.). Regardless, it is unfortunate that serum levels oflufenuron decreased more rapidly than we had expected and that data did not support the hypothesis that lufenuron would be an effective means to significantly reduce flea production over the summer. Although further, similarly controlled investigation is warranted, preliminary results suggest that given current techniques, the likelihood of lufenuron limiting fleas populations and breaking the plague cycle is not extremely promising. Therefore, experiments into efficacy of controlling flea infestations in simulated burrow environments and in the field (described in Wild and Castle 1998) will not be performed. LITERATURE CITED Wild, M.A. and K. T. Castle. 1998. Monitoring and managing disease in black-footed ferrets. Colorado Div. Wildl. Res. Rep., 0880-1, Jul 1997 - Jun 1998, Fort Collins. Wild, M.A. and K. T. Castle. 1999. Monitoring and managing disease in black-footed ferrets. Colorado Div. Wildl. Res. Rep., 0880-1, Jul 1998 - Jun 1999, Fort Collins. �121 Attachment l Kevin T. Castle, Margaret A. Wild, and S. Craig Parks Colorado Division of Wildlife Foothills Wildlife Research Facility (KTC and MAW) Novartis Animal Health, Greensboro, NC (SCP) Introduction Black-footed ferret (Mustela nigripes) recovery plans call for the reintroduction of ferrets to sites characterized by the presence of viable populations of prairie dogs (Cynomys spp.), which provide food and shelter for ferrets. Unfortunately, prairie dogs inhabiting many potential reintroduction sites carry fleas that can serve as vectors ofYersinia pestis, the causative agent of plague (Ubico et al, 1988). Prairie dogs and ferrets are both highly susceptible to plague (Barnes, 1993; Williams et al., 1994) so the chances of successful ferret reintroduction will be enhanced if the numbers of fleas infesting a prairie dog colony can be significantly reduced. Lufenuron is a benzoylphenylurea derivative which inhibits formation of chitin in the exoskeleton of insects (Cohen, 1987). A single oral dose oflufenuron has been shown effective in controlling the cat flea (Ctenocephalides felts felis) for at least 30 days in treated cats (Blagburn et al., 1994) and dogs (Hink et al., 1994; Blagburn et al., 1995). No studies on the efficacy oflufenuron have been conducted with prairie dogs; however, Davis ( 1997) reported a significant reduction in fleas on free-ranging ground squirrels (Spermophilus beecheyi) that had been treated with lufenuron. We are performing a series of experiments to test the applicability oflufenuron to control fleas in captive, and ultimately wild, white-tailed prairie dogs (Cynomys leucurus). Thus far in pilot studies we have determined standard husbandry, maintenance, and handling protocols for captive prairie dogs and determined a test dose oflufenuron based upon a bioavailability study. We have also attempted to establish an insectary colony of a flea species (0ropsylla tuberculata) that naturally infests prairie dogs and their burrows. Oropsylla fleas are among the most common prairie dog fleas, and have been implicated in the transmission of plague in prairie dogs (Ubico et al., 1988). Fleas of this genus are nest fleas that infest the host only to obtain a blood meal. At other times the fleas select microenvironments within in the burrow system that are conducive to successful reproduction and survival. In pilot studies we were unable to develop methods to successfully maintain and produce self-sustaining populations of 0. tuberculata in an insectary. However, a related species, 0. montana, which naturally infests ground squirrels (e.g. Spermophilus beechyi) has been successfully maintained in the insectary using methods of M. Metzger and K. Gage (pers. comm). These fleas will feed on prairie dogs and successfully reproduce after ingesting a blood meal. Because of its similarity to 0. tuberculata, we will use 0. montana as a model to determine if flea reproduction can be controlled in lufenuron-treated prairie dogs. A chambered-flea system has been used to test on-host viability and fecundity of fleas on cats (Thomas et al. 1996) and laboratory mice (D. Engelthaller, pers. comm.). In this system, fleas are contained in a chamber attached to the host, and can obtain a blood meal through a mesh screen. This system has several advantages over other methods of artificial infestation. First, because a known number of adults can be placed into the chamber, flea mortality is easily noted, and survivorship can be determined. Second, sex ratios can be adjusted to maximize egg production. Third, eggs can be recovered readily, to determine viability. Fourth, the environment in the chamber will be warm and humid, and therefore conducive to adult and egg survival. Finally, fleas will not be able to leave the prairie dog, and therefore will not be able to infest caretakers or other animals. �122 In this study we will: 1) develop techniques for artificially infesting white-tailed prairie dogs with chambered populations of fleas, 2) test the flea control efficacy oflufenuron fed to white-tailed prairie dogs artificially infested with fleas, and 3) compare efficacy of flea control with serum lufenuron levels. Our working hypothesis is that fleas which feed on lufenuron-dosed prairie dogs will have reduced survival of eggs and larvae compared to fleas which feed on control animals that receive no lufenuron. Methods Prairie Dog Maintenance We will use 31 captive white-tailed prairie dogs maintained at the Foothills Wildlife Research Facility in Ft. Collins in our experiment. Prairie dogs will be housed singly (if used in experiments) or in pairs (if used for flea colony maintenance) in custom-designed cages (100 cm x 500 cm x 600 cm; Wild and Castle 1998). Prairie dogs will be observed daily, and will have ad libitum access to Teklad Rodent Blocks and water. Windows will provide a natural photoperiod, and a combination of timed heaters and air conditioners will be used to keep ambient temperature between 10 and 25° C. Twenty prairie dogs (10 males and 10 females) will be blocked by sex and randomly divided equally into a treatment group and a control group for the experiment. The treatment group will receive lufenuron while the control group will not. Because all animals cannot be housed in one building due to space limitations, the groups will be split evenly between two separate but similar buildings, to help control for any interbuilding differences (e.g. temperature, humidity) that may exist. The sample size of20 prairie dogs will allow us to detect a ~91 % reduction in the number of eggs that successfully hatch in the treated group given alpha = 0.10 and beta = 0 .90. An additional 11 captive prairie dogs will be maintained under similar conditions, but will not participate in the experiment. Instead, they will be used in the maintenance of the flea colony (see below). Lufenuron Dosing Prairie dogs in the experimental group will be fasted for 24 h prior to dosing; water will be available during the fast. After the fasting period (day 0), each experimental animal will be weighed, then offered a bolus dose oflufenuron (300 mg/kg body mass). Lufenuron will be mixed thoroughly in approximately 10 g of highly palatable bait (ground rat chow and molasses); control animals will receive bait only. We previously determined that a majority of fasted prairie dogs consume about 90-95% of their dose in less than 12 h. We will observe the progress of bait ingestion in each animal to determine when the dose is ingested. Remaining bait will be removed after 24 h. We will weigh the bait/lufenuron mixture before and after the prairie dogs are dosed, in order to calculate the actual dose ingested. After dosing, normal feeding will resume. Flea Infestation An initial stock of laboratory-reared, disease-free fleas (0. montana) will be obtained from Dr. Kenneth Gage at the Centers for Disease Control and Prevention (CDC) in Ft. Collins. Fleas will be housed in 500 ml glass jars containing larva-rearing media, to allow the population to be self-sustaining . Rearing media consists ofwheast (Red Star Biologicals), dried beef blood (Monfort Biologicals), powdered dog chow, and sand. Fleas will be maintained in an incubator at about 22-23° C and ~70% relative humidity (M. Metzger, pers. comm.) to ensure optimal reproduction. A natural photoperiod will be approximated within the incubator using fluorescent lights and a timer. Adult fleas must ingest a blood meal in order to reproduce. We will use two methods to provide blood meals to fleas held in the insectary for propagation of the colony. The first 1:11ethod will follow an established protocol (Castle and Wild 1998) to provide blood to the adult fleas in each rearing chamber, using neonatal rodents. Briefly, when fleas are in need of a blood meal (1-2 times per week), we will obtain �123 neonatal rodents from a private colony. The neonates will be placed into the insectaries with fleas for up to 24 h; previous work has shown that over 80% of the neonates are alive after 24 h. No food or water will be provided for the neonates, as they are strictly dependent on nursing. After feeding by the fleas, neonates will be euthanized by an overdose of inhalant anesthetic. Once per week, for 7 weeks, we will utilize the 11 non-experimental prairie dogs as blood sources, using the chambered flea technique described below. While the use of neonatal rodents is an efficient, approved method of providing blood to fleas, neonatal rodents are not always available from private colonies. Prairie dogs will therefore serve the dual purposes of providing blood meals when neonates are unavailable, and minimizing the number of neonatal rodents sacrificed. One week prior to study initiation, and on study days 2, 7, 14, 21, 28, 35, and 42 we will place flea chambers on experimental prairie dogs. Fleas feeding on treated prairie dogs will potentially be exposed to lufenuron from this blood meal. W.e will collect 50 adult female and 30 ad~lt male fleas from the insectary and place them"into a chamber (2.5 cm diameter). Each chamber will be-attached to a prairie dog so the fleas can obtain a blood meal. To attach the chambers, prairie dogs will be anesthetized using isoflurane delivered by a vaporizer. Respiration and depth of anesthesia will be monitored. A patch of fur on the dorsal thorax caudal to the shoulders will be shaved. A flea chamber will be placed on the skin, and taped into place using Elastikon and Vet-rap. The prairie dog will then be placed in a 30 cm x 20 cm x 20 cm holding box to recover from anesthesia, and will be monitored while the chamber is attached. Chambers will remain on each prairie dog for 30 min. At the end of the feeding time, the prairie dog will again be anesthetized with isoflurane, and the tape and chamber will be removed. Prairie dogs will be returned to their cages after recovery from anesthesia. Fleas will be observed for evidence of feeding by observation under a 10-20x ,;nicroscope, and blood-filled fleas will be placed in plastic vials and put into our insectary for egg recovery. 0. montana fleas typically lay eggs within 2-3 days of a blood meal (M. Metzger, pers. comm.), so adult dishes will be monitored every day for 7 days to monitor egg production. Eggs will be removed from the adult dish and placed in new vials inside the insectary; they will be observed every day for larval emergence. The total number of eggs produced by fleas from each prairie dog will be recorded, as will the number of larva that emerge each ·day. Unfed adult fleas will be returned to the insectary (prior to lufenuron dosing) or preserved in alcohol (after lufenuron dosing). Blood Collection Concentration of lufenuron in the blood may be an indicator of efficacy of flea control. To determine the relationship between blood lufenuron concentration and egg viability and larval development, we will collect 3 ml of blood from each anesthetized prairie dog prior to chamber attachment on each sampling day. Blood will be collected by jugular venipuncture and placed into a glass vacutainer blood tube without anticoagulant. Alternatively, ifwe are unable to collect an adequate blood sample peripherally, we will collect blood from the vena cava or directly from the heart. Although cardiac puncture is considered to be a safe procedure for collection of large volumes of blood from laboratory rodents (CCAC, 1984), due to the increased risks involved with cardiac puncture, we will use this technique only to obtain critical samples. Serum will be harvested and frozen within 4 h of collection. Serum lufenuron concentration will i>C? determined by HPLC at Novartis Laboratories. Based on our pilot studies and other literature (Blagbum et al. 1994, 1995; Thomas et al. 1996) we do not anticipate any prairie dog mortality due to the flea infestations, lufenuron dosing regime or blood collection, but if a severe allergic reaction or other health problem associated with our procedures occurs, the prairie dog will be removed from the study and provided veterinary care or euthanized with an overdose of inhalant anesthetic or barbiturate. A complete post-mortem examination will be performed on any animal that may die during the experimental period. �124 Data Analysis Efficacy of lufenuron will be determined by comparing developmental success of eggs produced by fleas exposed to treated prairie dogs vs. eggs produced by fleas exposed to control group prairie dogs. These formulas will be used in efficacy determination: Developmental success = number of larvae hatched x 100 number of eggs collected Percentage efficacy=mean developmental success (control) - mean developmental success (treated) x 100 mean developmental success (control) Mean developmental success values for eggs collected from prairie dogs at each sampling period will be analyzed using analysis of covariance, using blood lufenuron concentration at each time step as the covariate. Group responses will be considered significantly different ifp < 0.10. Literature Cited Barnes, A. M. 1993. A review of plague and its relevance to prairie dog populations and the black-footed ferret. Proceedings of the symposium on the management of prairie dog complexes for the reintroduction of the black-footed ferret. J. L. Oldemeyer, D. E. Biggins, B. J. Miller, and R. Crete, eds. 96pp. Blagburn, B. L., J. L. Vaughan, D.S. Lindsay, and G. L. Tebbitt. 1994. Efficacy dosage titration of lufenuron against developmental stages of fleas (Ctenocephalides felis felis) in cats. American Journal of Veterinary Research 55: 98-101. Blagburn, B. L., C. M. Hendrix, J. L. Vaughan, D.S. Lindsay, and S. H. Barnett. 1995. Efficacy of lufenuron against developmental stages of fleas (Ctenocephalides felis felis) in dogs housed in simulated home environments. American Journal of Veterinary Research 56: 464-467. Castle, K. T. and M.A. Wild. 1998. Protocol for the use of neonatal rodents as a blood source for insectary-reared fleas. Colorado Division of Wildlife Animal Care and Use Committee Study Plan. Canadian Council on Animal Care (CCAC). 1984. Guide to the care and use of experimental animals, Vol. 2. Ottawa, Ont., Canada. Cohen, E. 1987. Interference with chitin biosynthesis in insects. In Chitin and benzoylphenyl ureas, series entomologica, vol. 38, J.E. Wright and A. Retnakaran (eds), Dr. W. Junk, Publishers, Boston, pp.3342. Davis, R. M. 1997. Use of an orally administered insect development inhibitor (lufenuron) as a flea control agent in the California ground squirrel, Spermophilus beecheyi. Fourth International Symposium on Ectoparasites of Pets, pp. 31-35. Hilton, D. F. J. 1971. A method for rearing fleas of ground squirrels. Transactions of the Royal Society of Tropical Medicine and Hygiene. 66: 188-189. Hink, W. F., M. Zakson, and S. Barnett. 1994. Evaluation of a single oral dose oflufenuron to control flea infestations in dogs. American Journal of Veterinary Research 55: 822-824. Thomas, R. E., L. Wallenfels, and I. Popeil. 1996. On-host viability and fecundity of Ctenocephalides felis (Siphonaptera: Pulicidae), using a novel chambered flea technique. Journal of Medical Entomology 33: 250-256. Ubico, S. R., G. 0. Maupin, K. A. Fagerstone, and R. G. McLean. 1988. A plague epizootic in the whitetailed prairie dogs (Cynomys leucurus) of Meeteetse, Wyoming. Journal of Wildlife Diseases 24: 399-406. Williams, E. S., K. Mills, D.R. Kwiatkowski, E.T. Thome, and A. Boerger-Fields. 1994. Plague in a black-footed ferret (Mustela nigripes). Journal ofWildlife Diseases 30:581-585. Wild, M.A. and K. T. Castle 1998. Monitoring and managing disease in black-footed ferrets. Colorado Division of Wildlife Program Narrative, Project# W-153-R. \ ·\,,,_) �125 20 ~ - - - - - - - - - - - - - i ■ Plague-Neg>----------~ □ Plague-Pos 15 L. (1) ..0 E 10 ::::, z 5 1997-W 1997-S 1998-W 1998-S 1999-W 1999-S Fig. la. Prevalence of exposure to plague in juvenile coyotes from the Little Snake Management Area, Colorado, from winter 1997 through summer 1999. II Plague-Neg 20 □ Plague-Pas 15 Q) .0 E 10 :::J z 5 0 -+------'----'-~--'----'----'----------'-----.J'---..L.---'-----'-----'-----'---J 1997-W 1997-S 1998-W 1998-S 1999-W 1999-S Fig. lb. Prevalence of exposure to plague in adult coyotes from the Little Snake Management Area, Colorado, from winter 1997 through summer 1999. Data from winter 1997 (age class unknown) provided by M Albee. �126 25 - . - - - - - - i □ CDV positive II CDV negative 20 t> 15 J 10 5 0 1997-W 1997-S 1998-W 1998-S 1999-W 1999-S Fig. 2. Prevalence of exposure to canine distemper virus (CDV) in coyotes from the Little Snake Management Area, Colorado,' from winter i 997 through smnmer 1999. All positive coyotes were adults with the exception of one juvenile in summer 1998. Data from winter 1997 (age class unknown) provided by M. Albee. 20 , - - - - - - - - - - - , Ill Negative f - - - - - - - , □ Positive 15 .BS 10 z= 5 Plague Fig. 3. Prevalence of exposure to plague and canine distemper virus in adult coyotes from the Wolf Creek site, Colorado in winter 2000. 200 C: 0 175 "ia = 8 1-, 150 . _J25 C: .0 8 ~()() e= 75 C: 50 = 25 ca ...:I • . • • •• • 0 -t-----,--------.-------.----,,---,----,----r----r---, 2 3 4 5 6 7 8 9 10 Week Post-Dosing Fig. 4. Mean serum lufenuron concentrations of active ( ♦) and torpid (■) prairie dogs orally dosed with 300 mg/kg lufenuron. �31 JOB PROGRESS REPORT Stateof _ _ _ _ _ _C.c;c..=ol=o=ra=d=o_ _ Work Package No. 0880 T~k _ _ _ _ _ _ _ _~l_ __ Division of Wildlife - Mammals Research Black-footed Ferret Conservation Disease Monitoring & Management Period Covered: July I 2002 through June 30, 2003 Author: L. L. Wolfe and L.A. Baeten Personnel: D. Finley, P. Schnurr, K. Cramer, H. Edwards, E. Knox, C. T. Larsen, N. Mier, M. W. Miller, K. Taurman, E. S. Williams Interim Report - Preliminary Results This work continues, and precise analysis ofdata has yet to be accomplished. Manipulation or interpretation of these data beyond that contained in this report should be labeled as such and is discouraged. ABSTRACT We continued monitoring carnivores at proposed black-footed ferret reintroduction sites for serological evidence of select disease epidemics. Sampling at the Wolf Creek Management Area (WCMA) in August 2003 revealed little evidence of ongoing epidemics that could impede black-footed ferret restoration efforts. Serology data from culled coyotes showed no evidence of active canine distemper or plague epidemics in the WCMA vicinity. In contrast, serologic evidence of exposure to tularemia w~ relatively high (~30%), consistent with previous observations in this and other monitored areas. We will continue this work as part of the ongoing Colorado-Utah black-footed ferret reintroduction protocol. �32 INTRODUCTION As part of the Colorado-Utah black-footed ferret reintroduction protocol, we continued monitoring carnivores at proposed ferret reintroduction sites for serological evidence of select disease epidemics. Originally, we monitored coyote (Canis latrans) populations at two Colorado sites: the Little Snake Management Area (LSMA) and the Wolf Creek Management Area (WCMA), Colorado. Under this program, >200 coyotes have been collected for post-mortem examination and samples collected as described in established protocols since March 1997. Monitoring has been accomplished via cooperative efforts of Colorado Division of Wildlife, USDA Wildlife Services, and Bureau of Land Management (BLM) personnel. To date, no lesions indicative of active infections with any of the select pathogens (Francisellia tularensis, Yersinia pestis, canine distemper virus [CDV]) have been noted on gross examinations of carcasses. However, relatively high proportions (31-89%) of the coyotes collected from the LSMA had positive titers to plague between March 1997 and July 1999. Although the proportion of plague-positive coyotes declined during the sampling period, evidence of continued exposure and perhaps declining prairie dog abundance led to abandonment of surveillance at LSMA after 1999. Monitoring at the WCMA has continued, and black-footed ferrets were reintroduced at this site in 200 I. METHODS Coyotes were collected using a combination of calling and aerial gunning (USDA-API-IlS-Wildlife Services). In light of ambiguity in results from mid-winter sampling attributable to the inability to accurately estimate ages of coyotes in the field, we began focusing on late summer sampling to monitor epidemic trends. Postmortem examination, sampling, and serological methods were as described previously (Colorado Division of Wildlife, Disease Survey of Carnivores in the Little Snake Area, ACUC 1997-3). RESULTS AND DISCUSSION Initial sampling (February 2000) at WCMA indicated substantially lower exposure rates to select pathogens than observed at LSMA. Data from 200 I surveys indicated a relatively high proportion of adult coyotes exposed to canine distemper virus (CDV)(Figure 1): in February 200 I, about 79% of the coyotes sampled had serum neutralizing titers :::I: 16. Recent sampling revealed lower proportions of CDV-positive coyotes, similar to the initial sampling periods. In contrast to canine distemper, exposure to plague appears relatively rare among coyotes sampled from WCMA (Figure 1). As tularemia is commonly found in rodents in Colorado, a seroprevalence of 20-40% is not surprising in WCMA Date/pathogen ■ Positive El Negative Figure 1. Seroprevalence of presumed tularemia, plague, and canine distemper exposure among coyotes sampled from the Wolf Creek Management Area, Colorado, during February 2000 to August 2003. �Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Project No. Work Package No. Colorado 0880 : Task Federal Aid Project : : : N/A Cost Center 3430 Mammals Research Black-Footed Ferret Conservation Black-Footed Ferret Recovery Program Disease Monitoring & Management : Period Covered: July 1 2003 through June 30, 2004 Author: L. L. Wolfe Personnel: L. A. Baeten, H. Edwards, C. T. Larsen, M. W. Miller, K. Taurman, E. S. Williams All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT We continued monitoring carnivores at proposed black-footed ferret reintroduction sites for serological evidence of select disease epidemics. Sampling at the Wolf Creek Management Area (WCMA) in August 2004 revealed little evidence of ongoing epidemics that could impede black-footed ferret restoration efforts. Serology data from culled coyotes showed no evidence of active canine distemper or plague epidemics in the WCMA vicinity. In contrast, serologic evidence of exposure to tularemia continues to be relatively high (~30%), consistent with previous observations in this and other monitored areas. We will continue this work as part of the ongoing Colorado−Utah black-footed ferret reintroduction protocol. 17 �JOB PROGRESS REPORT BLACK-FOOTED FERRET RECOVERY PROGRAM DISEASE MONITORING & MANAGEMENT LISA L. WOLFE INTRODUCTION As part of the Colorado−Utah black-footed ferret reintroduction protocol, we continued monitoring carnivores at proposed ferret reintroduction sites for serological evidence of select disease epidemics. Originally, we monitored coyote (Canis latrans) populations at two Colorado sites: the Little Snake Management Area (LSMA) and the Wolf Creek Management Area (WCMA), Colorado. Under this program, >200 coyotes have been collected for post-mortem examination and samples collected as described in established protocols since March 1997. Monitoring has been accomplished via cooperative efforts of Colorado Division of Wildlife, USDA Wildlife Services, and Bureau of Land Management (BLM) personnel. To date, no lesions indicative of active infections with any of the select pathogens (Francisellia tularensis, Yersinia pestis, canine distemper virus [CDV]) have been noted on gross examinations of carcasses. However, relatively high proportions (31-89%) of the coyotes collected from the LSMA had positive titers to plague between March 1997 and July 1999. Although the proportion of plague-positive coyotes declined during the sampling period, evidence of continued exposure and perhaps declining prairie dog abundance led to abandonment of surveillance at LSMA after 1999. Monitoring at the WCMA has continued, and black-footed ferrets were reintroduced at this site in 2001. RESULTS AND DISCUSSION Disease surveillance As part of the Colorado-Utah black-footed ferret reintroduction protocol, we monitored serological evidence of exposure to select infectious diseases in coyotes at Wolf Creek Management Area (WCMA), Colorado; our strategy was to use exposed coyotes as sentinels for detecting epidemics at restoration and prospective release sites. Over 350 coyotes (Canis latrans) have been collected for postmortem examination and samples collected as described in established protocols since March 1997 via cooperative efforts of Colorado Division of Wildlife, USDA Wildlife Services, and Bureau of Land Management (BLM) personnel. Coyotes were collected using a combination of calling and aerial gunning. In 2004, 20 coyotes were sampled (5 pups, 6 juvenile, 6 adult, 3 not aged) in late July. (Because data from juveniles are most useful in detecting evidence of recent epidemics, we discontinued mid-winter sampling in 2002.) No lesions indicative of active infections with select pathogens (Francisellia tularensis, Yersinia pestis, canine distemper virus [CDV]) were noted on gross examinations of carcasses through 2002; in the absence of meaningful necropsy findings, we discontinued gross examinations of carcasses in 2003. Initial sampling (February 2000) at WCMA indicated substantially lower exposure rates to select pathogens than observed at another site (Little Snake Management Area) monitored in earlier years of this survey. Initial sampling demonstrated 24 percent of the coyotes surveyed with antibody titers suggestive of exposure to CDV. Although seroprevalence increased slightly in 2001, sampling since 2002 revealed much lower proportions of CDV-positive coyotes (Figure 1). There was no serologic evidence of CDV exposure in 2002, and only 1 case each in 2003 and 2004. Exposure to plague still appears relatively rare among coyotes sampled from WCMA (Figure 1). In 2004 one juvenile, out of 21 coyotes sampled, was 18 �“moderately positive” antibody titer to plague. The most significant pathogen exposure noted by seroprevalence is for tularemia. As tularemia is commonly found in rodents in Colorado, seroprevalence of 20–40% is not surprising in carnivores, and very little change in tularemia seroprevalence has been seen over the 5-year sampling period. Prepared by ________________________ Lisa L. Wolfe, Veterinarian 100% 80% 60% 40% 20% Y. pestsis F. tularensis CDV Date/pathogen 08/04 (n=20) 08/03 (n=11) 07/02 (n=13) 07/01 (n=16) 07/00 (n=20) 08/04 (n=20) 08/03 (n=11) 07/02 (n=13) 07/01 (n=16) 07/00 (n=20) 08/04 (n=20) 08/03 (n=11) 07/02 (n=13) 07/01 (n=16) 07/00 (n=20) 0% ■ Positive □ Negative Figure 1. Seroprevalence of presumed tularemia (F. tularensis), plague (Y. pestis), and canine distemper virus (CDV) exposure among coyotes sampled from the Wolf Creek Management Area, Colorado, during summer sampling 2000−2004. 19 � https://cpw.cvlcollections.org/files/original/188a97d1068aa14770546b8a69927369.pdf a9101ca9d2887885c3c9b6fc98a1e885 PDF Text Text 245 Colorado Division of Wildlife Wildlife Research Report July 200 l and July 2002 JOB PROGRESS REPORT State of - - - - - - - - - - -Colorado "-==~----- Mammals Research Work Package No. - - - - ~ l ~ A ~ - - - - - Multispecies Investigations Task No. - - - - - - - - - = 5_ _ _ _ _ __ Consulting Services for Mark-Recapture Analvsis Federal Aid Project No. __W~-1_5_3~-R~-~2_ _ _ __ Research and Development 1 Period Covered: July l, 2000 - June 30, 2001 Author: G C. White, Ph.D. Personnel: C. Bishop, R. B. Gill, D. C. Bowden, R. M. Bartmann, D. J. Freddy, T. M. Shenk, M. M. Conner, M. Post Vieira, A. Dharman, B. Lubow ABSTRACT Progress towards the objectives of this job include: Consulting assistance to CDOW on harvest surveys, terrestrial inventory systems, and population modeling procedures was provided. Estimates of spring and fall turkey, spring snow goose, sharp-tailed and sage grouse, chukars, ptarmigan, Abert's squirrels, and general small game harvest were computed from survey data, and programs and harvest estimates provided to CDOW via email and CD ROM. Input on the design and analysis of the Harvest Information Program was provided on several occasions. The DEAMAN software package for the storage, summary, and analysis of big game population and harvest data was revised further as a Windows 95/98/NT/2000/ME program. The capability to incorporate data on radio-collared animals to estimate survival with the Kaplan-Meier estimator and display movement data was added, and distributed to terrestrial biologist via the WWW at http://www.cnr. colostate. edu/-gw-hite/deaman. A 1-day workshop was conducted with NE region personnel in the use of DEAMAN and population modeling procedures, mainly to instruct region personnel on the use of spreadsheet models for ungulate population dynamics. In addition, numerous questions were answered via meetings with biologists, and via email. A preliminary analysis of the survival rates from the mule deer monitoring data was completed. However, I have not received final data from some of the biologists, so have not been able to complete this analysis. A paper, coauthored with Bruce Lubow, was submitted for publication to the Journal of Wildlife Management on past efforts to develop a realistic mule deer population model based on data collected with current CDOW procedures. Data from the Piceance Basin were used to illustrate the modeling I '; li~ri~r BD0WD16832 �246 technique. In addition, a book chapter on modeling big game populations appeared in print: White, G. C. 2000. Modeling Population Dynamics. Pages 84-107 in S. Demarais and P.R. Krausman, eds. Ecology and Management of Large Mammals in North America. Prentice-Hall, Upper Saddle River, New Jersey, USA. A paper on optimal allocation of resources to sample Colorado mule deer populations was published in the Journal of Wildlife Management: Bowden, D. C., G. C. White, and R. M. Bartmann. 2000. Optimal allocation of sampling effort for monitoring a harvested mule deer population. Journal of Wildlife Management 64:1013-1024. A paper on trends in Colorado mule deer age and sex ratios was published in the Journal of Wildlife Management: White, G. C., D. J. Freddy, R. B. Gill, and J. H. Ellenberger. 2001. Effect of adult sex ratio on mule deer and elk productivity in Colorado. Journal of Wildlife Management 65:436-444. Assistance in the design and analysis of candidate systems to estimate deer abundance in GMU 10 was provided. A research study to examine the impact of nutrition on the decline of mule deer fecundity during the last 20 years was initiated. I have provided input on estimation of the number of deer on the feed sites, and developed an estimator of fawn survival rates based on radio-collared does and fall and spring fawn:doe ratios. Data were collected and analyzed on spatial distribution, movement of radio-collared animals, and population sizes related to estimating the spread and impacts of chronic wasting disease in deer populations. A report summarizing these findings was provided to CDOW personnel involved with the study. A final report on the response of elk to lower numbers of archery licenses in the White River Data Analysis Unit was prepared and submitted to CDOW personnel involved with the project. A paper reporting results of the earlier experiment to detect elk response to the opening of archery season has been accepted for publication in the Journal of Wildlife Management: Conner, M. M., G. C. White, and D. J. Freddy. 2001. Elk movement in response to early-season hunting in Colorado. Journal of Wildlife Management 65. In Press. A graduate research project to evaluate the movements of Preble's meadow jumping mouse populations away from riparian areas was completed. A final report of this project was submitted to CDOW personnel involved with the project. • • In cooperation with CDOW personnel, I developed an analysis of survival of lynx released as part of the reintroduction program. An analysis to estimate the effort required to estimate the percent of eastern Colorado inhabited by blacktailed prairie dogs was completed and results provided to CDOW personnel involved with the effort. �247 CONSULTING SERVICES FOR MARK-RECAPTURE ANALYSES G. C. White P, N. OBJECTIVES Design a sampling scheme to estimate the area of black-tailed prairie dog colonies in eastern Colorado. SEGMENT OBJECTIVES I. Develop a sampling scheme to estimate the area of black-tailed prairie dog colonies in eastern Colorado. 2. Develop estimates of the cost of this sampling scheme as a function of the expected precision. RESULTS AND DISCUSSION Area of black-tailed prairie dog colonies in Wyoming, North and South Dakota, and Nebraska have been sampled successfully with aerial line intercept sampling techniques (Sidle et al. In Press). CDOW is interested in applying this technique to eastern Colorado, and obtaining estimates of the areas occupied by prairie dogs by county. However, there are concerns regarding the cost of the survey and expected precision. In the following, I present an analysis of the expected cost as a function of the precision of the estimates of area of black-tailed prairie dog colonies. To compute the expected precision as a function of the cost of the aerial survey for black-tailed prairie dogs, I went through the following steps. I assumed that the lines to be flown would be stratified by county. Because an infinite number of lines can be flown for each county, the sampling scheme can be viewed as sampling with replacement, and hence, no finite population correction is allowed. Further, I treated each line as providing an estimate of the proportion of the county area in active prairie dog towns, and not as a ratio estimator. Sidle et al. (In Press) compared both types of estimators, and developed a composite of the 2. However, design of a new survey seemed easier to conceptualize with the approach taken. This approach allowed the use of the formulas on pages 341-342 of Thompson et al. ( 1998), ignoring the finite population correction. Other pertinent references are Thompson (1992) and Cochran (1997). From the area of each county and the estimated area of prairie dog towns within the county provided by the EDAW survey, I predicted the proportion of the line in each county that would intersect dog towns (r) as: ActiveTownsArea C r = ---------County Area A From Table 1 of Sidle et al. (In Press) I computed the relationship between the standard deviation of r [SD(r)] and the value of r. To compute SD(r) from Table 1 of Sidle et al., I got the nwnber of lines flown for each of the 8 surveys from Douglas Johnson, Northern Prairie Wildlife Research Center, USGS. A linear relationship of SD(-)= 0.0087 + I.0804r was found to provide a decent fit to the data (Figure 1). �248 0.25 0.2 -- "t::' -..;;;.. 0.15 - Q Cl) 0.1 0.05 0 0 0.05 0.1 0.15 0.2 r I assumed that each county was to be estimated with some relative precision (e) such that the 95% confidence interval for each county would be ± eC, where C is the estimated acres of active towns. This approach is not the optimal for the estimate of the total active town area in the state, but would provide good estimates (i.e., estimates of quality e) for each county. For each county, the standard error of the estimate of the active prairie dog town area [SE(C)] was computed based on the SD(r) and the estimated number oflines (n) to be flown, where SE(r) = SD(r)/ ✓ n, and SE(C) 2 = A SE(r). Given the desired level of precision, I computed the number oflines to fly in a county as: Because some counties had values of C = 0 (and hence the above equation i; undefined), and others have very small values, I assumed that all counties had at least 0.5% area in active towns to compute these samples sizes, although the actual value of r was used to compute the standard deviation. The total length of the lines to be flown in the county is the square root of the county area multiplied by the number of lines to be flown. Cost of the survey for a county was figured as the length of line to be flown plus 2 times the square root of county area in miles (to account for ferry time), all divided by a flight speed of 90 mph, times $180 per hour of flight time. The total acreage of prairie dog towns (Cr) is the sum of the county estimates, with the variance computed as the sum of the variances across the counties. �189 JOB PROGRESS REPORT Stateof _ _ _---'C=o=l=o=ra=d=o_ _ _ _ _ __ Mammals Research Program Work Package No. _ _ _ _ _ _ _ _ _ __ Multispecies Investigations Task No. --~5_ _ _ _ _ _ _ _ __ Consulting Service for Mark-Recapture Analysis Federal Aid Project No. W-153-R-2 Period Covered: July 1, 2002 - June 30, 2003 Author: G. C. White Personnel: C. Bishop, G. Miller, T. E. Remington, D. J. Freddy, T. M. Shenk, L. Stevens, J. Craig, R. Kahn, D. C. Bowden, F. Pusateri, J. Dennis, P. Schnurr, B. Andelt, A Seglund, D. Finley, A Linstrom, D. Walsh, K. Strohm. ABSTRACT Progress towards the objectives of this job include: 1. Consulting assistance to CDOW on harvest surveys, terrestrial inventory systems, and population modeling procedures was provided. Estimates of spring and fall turkey, spring snow goose, sharp-tailed and sage grouse, chukars, ptarmigan, Abert's squirrels, and general small game harvest were computed from survey data, and programs and harvest estimates provided to CDOW via email and CD ROM. Computer code written in SAS to compute these estimates and display results graphically was also provided. Computer code was also written in SAS to estimate the compliance rate of Colorado small game license holders with the Harvest Information Program. 2. The DEAMAN software package for the storage, summary, and analysis of big game population and harvest data was revised further as a Windows 95/98/NT/2000/ME/XP program. A User's Manual was provided to terrestrial biologists on CD and also distributed via the WWW at http://www.cnr.colostate.edu/~gwhite/deaman. 3. Consultation with CDOW Terrestial Biologists in the use of DEAMAN and population modeling procedures continued. Numerous questions were answered via meetings with biologists, and via email. 4. A paper, coauthored with Marilet Zablan and Clait Braun, was published in the Journal of Wildlife Management on past efforts to estimate survival rates of sage grouse in North Park from CDOW banding records. The full citation is: Zablan, M. A, C. E. Braun, and G. C. White. 2003. Estimation of northern sage-grouse survival in North Park, Colorado. Journal of Wildlife Management 67:144-154. �190 5. A paper on the estimation of population size from correlated sampling unit estimates of the variable of interest was published in the Journal of Wildlife Management. The methodology developed in this paper is proposed for use in a joint Colorado/Utah survey of the colony area of whitetailed and Gunnison prairie dogs in western Colorado and eastern Utah. The full citation is: Bowden, D. C., G. C. White, A. B. Franklin, and J. L. Ganey. 2003. Estimating population size with correlated sampling unit estimates. Journal of Wildlife Management 67: 1-10. 6. A paper on the use of lek counts to index prairie grouse populations was published in the Wildlife Society Bulletin: Walsh, D. P., G. C. White, T. E. Remington, and D. C. Bowden. 2003. Evaluation of Lek Count Index for Prairie Grouse Wildlife Society Bulletin. 32:56-68. 7. A paper on the estimation of sage grouse populations was submitted to the Journal of Wildlife Management: Walsh, D. P., G. C. White, T. E. Remington, and D. C. Bowden. 2003. Population Estimation of Greater Sage-Grouse. Journal of Wildlife Management. Submitted. 8. A paper on the effects of early season hunter numbers on elk movement was published in the Journal of Wildlife Management: Vieira, M. E. P., M. M. Conner, G. C. White, and D. J. Freddy. 2003. Relative effects of early season hunter numbers and opening date on elk movement in northwest Colorado. Journal of Wildlife Management. 67:717-728. 9. A paper on the impact of limited antlered harvest on mule deer sex and age ratios was submitted to the Wildlife Society Bulletin: Bishop, C. J., G. C. White, D. J. Freddy, and B. E. Watkins. 2003. Effect of limited antlered harvest on mule deer sex and age ratios. Wildlife Society Bulletin. Submitted. 10. A paper on the survival and recruitment of peregrine falcons was published in the Journal of Wildlife Management: Craig, G. R., G. C. White, and J. H. Enderson. 2004. Survival, recruitment, and rate of population change of the Colorado peregrine falcon population. Journal of Wildlife Management. In Press. 11. A research study to examine the impact of nutrition on the decline of mule deer fecundity during the last 20 years was continued. I have provided input on estimation of the number of deer on the feed sites, and developed an estimator of fawn survival rates based on radio-collared does and fall and spring fawn:doe ratios. 12. A graduate research project by Dan Walsh to evaluate utility oflek counts of Greater Sagegrouse in Middle Park was completed. Mark-resight methods are being used to estimate lek attendance and population size. The thesis citation is: Walsh, D. P. 2002. Population Estimation Techniques for Greater Sage-grouse. M. S. Thesis, Colorado State University, Fort Collins. USA. 158pp. 13. A graduate research project to develop a sage grouse population model, using North Park sage grouse data to develop parameter estimates, was initiated. The graduate student is Kristen Strohm. 14. An analysis to estimate the estimate the percent of eastern Colorado inhabited by black-tailed prairie dogs was completed and results provided to CDOW personnel involved with the effort. Estimates were computed in an Excel spreadsheet, and also verified through a program written in SAS to be sure that no errors in the calculations would be found when the spreadsheet is distributed to interested stakeholders. �191 15. Development of the design of a monitoring system for white-tailed prairie dogs in western Colorado and eastern Utah was started. This effort is in cooperation with Pam Schnurr, Bill Andelt, and Amy Seglund. 16. Development of the design of a monitoring system for swift fox in eastern Colorado was started. This effort is in cooperation with Francie Pusatari and Darby Finley. 17. Two new graduate students have been accepted for my supervision in the Department of Fishery and Wildlife Biology at Colorado State University. Chad Bishop will start a Ph.D. program in Fall, 2003, and Aaron Linstrom will start an M.S. program in Fall, 2003. �192 CONSULTING SERVICES FOR MARK-RECAPTURE ANALYSES G. C. White P. N. OBJECTIVES Assess the status of Colorado swift fox population through an occupancy monitoring approach. SEGMENT OBJECTIVES 1. Develop a monitoring scheme to estimate the occupancy rate of swift fox in eastern Colorado. 2. Determine necessary sample sizes to obtain adequate statistical power to detect biologically important changes in the occupancy rate. RESULTS AND DISCUSSION Estimation of occupancy rate for Swift Foxes (Vulpes velox) in eastern Colorado was based on trapping data provided by Finley (1999). The data consist of 72 randomly selected trapping grids 4 miles by 5 miles in area, with 20 traps set at 1 mile intervals. METHODS The occupancy model of MacKenzie et al. (2002) was fit to the 72 trapping grids using Program MARK (White and Burnham 1999). The model fit included 8 detection probabilities (p) for the 8 trapping occasions plus the probability of occupancy ( f// ). Detection probabilities were predicted with the month that a grid was trapped. Month was modeled with trigometric functions; sin(Monthx2n/12) and cos(Monthx2n/12), and powers of these functions. By using these sin and cosine functions, I can make the capture probability continuous across the December to January interval. Trend models were also used to model capture probabilities across occasions, forcing a linear trend on a logit scale in the capture probabilities. The percentage of each trapping grid comprised of short grass prairie was used as an additional covariate to predict both detection probability and probability of occupancy on a logit scale. Model selection was performed with AICc (Burnham and Anderson 1999). RESULTS Model selection results (Table l)suggest that month is an important predictor of the probability of detecting foxes on a grid. In addition, the top-ranked AICc includes a positive trend effect in the detection probabilities across the occasions, consistent with the results from the population estimation models. Model selection results also suggest that short grass prairie vegetation affects both the detection probability as well as the probability of occupancy. Detection probability is affected by the density of animals on the grid, and the percentage of short grass prairie on a trapping grid correlates (r = 0.375) with estimated population sizes provided in Table 5 of my September 23rd memo. �193 Table 1. Model selection results from fitting the occueancy estimation model of MacKenzie et al. (2002). AJCc ~AJCc AICcNum Weights Par. Deviance {p(T+cosMonth+cosMonth/\2) psi(SGPProp)} 318.6146 0.0000 0.31440 6 305.3223 {p(T+cosMonth+cosMonth/\2+SGPProp) psi(SGPProp)} 319.0596 0.4450 0.25168 7 303.3096 {p(T+cosMonth+cosMonth/\2+SGPProp) psi} 320.1094 1.4948 0.14890 6 306.8171 {p(cosMonth+cosMonth/\2) psi(SGPProp)} 321.2674 2.6528 0.08345 5 310.3583 {p(cosMonth+cosMonth/\2+SGPProp) psi(SGPProp)} 321.3341 2.7195 0.08071 6 308.0418 {p(T+cosMonth+cosMonth/\2) psi} 322.6843 4.0697 0.04109 5 311.7753 {p(cosMonth+cosMonth/\2+cosMonth/\3) psi} 322.7973 4.1827 0.03884 5 311.8882 {p(cosMonth) psi(SGPProp)} 323.6307 5.0161 0.02560 4 315.0337 {p(cosMonth+cosMonth/\2) psi} 325.5083 6.8937 0.01001 4 316.9113 {p(cosMonth) psi} 328.1180 9.5034 0.00272 3 321.7651 {p(cosMonth+sinMonth) psi} 329.1845 10.5699 0.00159 4 320.5875 {p(t+cosMonth+cosMonth/\2) psi} 330.1385 11.5239 0.00099 11 303.7385 {p(.) psi} 340.3683 21.7537 0.00001 2 336.1944 {p(sinMonth) psi} 342.5446 23.9300 0 3 336.1917 {p(t) psi} 343.2296 24.6150 0 9 322.3264 Model Parameter values for the top-ranked AJC 0 model (Table 2) demonstrate the increasing detection probability with occasion. In addition, the estimate of 1// of 0.821 suggests that 59.1 of the 72 grids trapped contained foxes, in contrast to the 51 grids that were observed to have foxes. �194 Table 2. Parameter estimates for the month of March from the top-ranked AI Cc occupancy model {p(T + cos(Month) + cos2(Month)) psi(SGP Proportion)}, where month was set to March (3), and the short grass prairie habitat proportion for the trapping grid was set to 66.9%, the mean of the grids trapped. Parameter Estimate SE LCI UCI P1 0.611647 0.083074 0.442448 0.757627 P2 0.675704 0.066681 0.534363 0.790928 p3 0.733793 0.060627 0.600043 0.835107 p4 0.784792 0.061708 0.640538 0.881835 Ps 0.828306 0.064248 0.665567 0.921227 P6 0.864541 0.065008 0.682552 0.949861 P1 0.894106 0.063204 0.695294 0.968985 pg 0.917831 0.059218 0.705628 0.981150 \jl 0.820811 0.065876 0.655653 0.916806 The effect of month in the top-ranked AI Cc model (Figure 1) is significant, and somewhat consistent with the results obtained with the population estimation models that included the variable month (reported in the memo of September 23). That is, the lowest detection probabilities are during summer. However, the occupancy model results suggest that September through March have the highest detection probabilities. The impact of the percentage of short grass prairie habitat on the estimates of occupancy is strong (Figure 2), with the probability of occupancy estimated at 34% for trapping grids with no short grass prairie habitat up to 93% for grids consisting of 100% short grass prairie. �195 0.9 0.8 >- ~ 0.7 :cn:s 0.6 .c e o.s a. g 0.4 0 0.3 .....Q) cosine quadratic --t--Trend+cosine quadratic '3 0.2 0.1 0 1 2 3 4 5 7 6 8 9 10 11 12 Month Figure 1. Effect of month in the 3 of the models of occupancy considered for detection probability: {p(cosMonth+cosMonth"2+cosMonth1''3) psi}=cosine cubic, {p(cosMonth+cosMonth"2) psi}=cosine quadratic, and {p(T + cos(Month) + cos2(Month)) psi}=Trend+cosine quadratic. The values shown for {p(T + cos(Month) + cos\Month)) psi} are for p 1_so estimates for p 2 throughp 8 increase monotonically from this value. >, 1 g 0.9 +-------- :g_ 0.8 ~ 0.7 + - - - - - - - - - - - - - - - - = ~ - - = - - - - - - - - - - - - - - - - - - - - - - - - - - - - c : ou 0.6 -t------------=_..,...,__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _----J 0 0.5 - j - - - - - - - - - = ~ , . . . - C - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - c : £ 0.4 +-~-~--------------------------------': .o 0.3 - + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - c : ~ 0.2 + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ' : ~ 0--t-----.......,..-----..........,--------r-------,--------....-.....,------------i e 0.1 -+-----------------------------------, 0 20 40 60 80 100 Short Grass Prairie (%) Figure 2. Effect of the percentage of the grid consisting of short grass prairie habitat on the probability of occupancy for the top-ranked AICc model {p(T + cos(Month) + cos2(Month)) psi(SGP Proportion)}. 120 �196 DISCUSSION The high detection probabilities during the September through March period suggests that swift fox monitoring should take place during this period. The increasing detection probability with trapping occasion also suggests that increasing the number of occasions will result in higher detection probabilities on each succeeding occasion. However, this trend effect is relatively minor. That is, the probability of not detecting foxes on a grid with 2 occasions trapped during March with the trend model estimates is (I - 0.610297) □ (1 0.6749937) = 0.126656. With the cosine quadratic model that does not include a trend across occasions, the probability of not detecting foxes is 0.085212. With 3 trapping occasions in March, the corresponding probabilities are 0.041164 and 0.024874, respectively. The strong relationship between the probability of occupancy and the short grass prairie habitat variable suggests that the design of an occupancy monitoring scheme should include this covariate. In particular, a ratio estimator can be developed that predicts the probability of occupancy based on the relationship in Figure 2. FURTHER WORK A reasonable estimate of the number of swift foxes in eastern Colorado can be obtained from the grid trapping scheme analyzed here. The population estimate for each trapping grid within a strata can be used to obtain a naive estimate of population density that will be biased high. However, through the use of radio collars, the proportion of time that marked animals spend on the trapping grid where they were initially captured can be used to correct these naive estimates. That is, the naive estimate multiplied by the proportion of radio locations on the trapping grid gives an unbiased estimate of density. Such a procedure has been used by White and Shenk (2001) to estimate population sizes for Preble's Meadow Jumping Mice, and details are provided in that article. Thus, to obtain an unbiased population estimate, radio-collared animals would be required. LITERATURE CITED Burnham, K. P., and D.R. Anderson. 1998. Model selection and inference: a practical information-theoretic approach. Springer-Verlag, New York, New York, USA. 353 pp. Finley, D. J. 1999. Distribution of the swift fox (Vulpes velox) on the eastern plains of Colorado. M. S. Thesis, University of Northern Colorado, Greeley, USA. 96pp. MacKenzie, D. I., J. D. Nichols, G. B. Lachman, S. Droege, J. A. Royle, and C. A. Langtimm. 2002. Estimating site occupancy when detection probabilities are less than one. Ecology 83:2248-2255. White, G. C., and K. P. Burnham. 1999. Program MARK: survival estimation from populations of marked animals. Bird Study 46 Supplement: 120-138. White, G. C., and T. M. Shenk. 2001. Population estimation with radio-marked animals. Pages 329-350 in J. J. Millspaugh and J.M. Marzluff, editors. Design and Analysis of Wildlife Radiotelemetry Studies. Academic Press, San Diego, California, USA. �Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Colorado Project No. Work Package No. Task No. 5 Federal Aid Project: 3001 W-185-R : : : : Cost Center: 3430 Mammals Research Multispecies Investigations Consulting Services for Mark-Recapture Analysis : Period Covered: July 1, 2003 - June 30, 2004 Author: G. C. White Personnel: C. Bishop, G. Miller, T. E. Remington, D. J. Freddy, T. M. Shenk, L. Stevens, J. Craig, R. Kahn, D. C. Bowden, F. Pusateri, J. Dennis, P. Schnurr, B. Andelt, A. Seglund, D. Finley, A. Linstrom, K. Strohm, P. Conn. All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT Progress towards the objectives of this job include: 1. 2. 3. 4. 5. Consulting assistance to CDOW on harvest surveys, terrestrial inventory systems, and population modeling procedures was provided. Estimates of spring and fall turkey, spring snow goose, sharp-tailed and sage grouse, chukars, ptarmigan, Abert’s squirrels, and general small game harvest were computed from survey data, and programs and harvest estimates provided to CDOW via email and CD ROM. Computer code written in SAS to compute these estimates and display results graphically was also provided. Computer code was also written in SAS to estimate the compliance rate of Colorado small game license holders with the Harvest Information Program. The DEAMAN software package for the storage, summary, and analysis of big game population and harvest data was revised further as a Windows 95/98/NT/2000/ME/XP program. A User’s Manual was provided to terrestrial biologists on CD and also distributed via the WWW at http://www.cnr.colostate.edu/~gwhite/deaman. Consultation with CDOW Terrestrial Biologists in the use of DEAMAN and population modeling procedures continued. Numerous questions were answered via meetings with biologists, and via email. A paper on the use of lek counts to index prairie grouse populations was published in the Wildlife Society Bulletin: Walsh, D. P., G. C. White, T. E. Remington, and D. C. Bowden. 2004. Evaluation of the lek count index for greater sage-grouse. Wildlife Society Bulletin 32: 56-68. A paper on the effects of early season hunter numbers on elk movement was published in the Journal of Wildlife Management: Vieira, M. E. P., M. M. Conner, G. C. White, and D. J. 151 �6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. Freddy. 2003. Effects of archery hunter numbers and opening dates on elk movement. Journal of Wildlife Management. 67:717-728. A paper discussing the implications of the GMU 10 special mule deer surveys was accepted for publication in the Wildlife Society Bulletin: Freddy, D. J., G. C. White, M. C. Kneeland, R. H. Kahn, J. W. Unsworth, W. J. deVergie, V. K. Grahm, J. H. Ellenberger, and C. H. Wagner. 2004. How many mule deer are there? Challenges of credibility in Colorado. Wildlife Society Bulletin. In Press. A paper on the impact of limited antlered harvest on mule deer sex and age ratios was accepted for publication in the Wildlife Society Bulletin: Bishop, C. J., G. C. White, D. J. Freddy, and B. E. Watkins. 2004. Effect of limited antlered harvest on mule deer sex and age ratios. Wildlife Society Bulletin. In Press. A paper on the survival and recruitment of peregrine falcons was accepted for publication in the Journal of Wildlife Management: Craig, G. R., G. C. White, and J. H. Enderson. 2004. Survival, recruitment, and rate of population change of the Colorado peregrine falcon population. Journal of Wildlife Management. In Press. A paper on the estimation of the area of black-tailed prairie dog colonies in eastern Colorado was submitted to the Wildlife Society Bulletin: White, G. C., J. R. Dennis, and F. M. Pusateri. 2004. Area of black-tailed prairie dog colonies in eastern Colorado. Wildlife Society Bulletin. Submitted. A paper on methodologies to obtain more rigorous population monitoring data was submitted to Wildlife Research: White, G. C. 2004. Correcting counts: techniques to de-index. Wildlife Research. Submitted. A paper evaluating methods of estimating the impact of harvest on survival rates was submitted to Animal Diversity and Conservation: Otis, D. L., and G. C. White. 2004. Evaluation of ultrastructure and random effects band recovery models for estimating relationships between survival and harvest rates in exploited populations. Animal Biodiversity and Conservation. Submitted. A paper on the procedures to monitor swift fox populations in eastern Colorado was submitted to the Journal of Wildlife Management: Finley, D. J., G. C. White and J. P. Fitzgerald. 2004. Estimation of swift fox population size and occupancy rates in eastern Colorado. Journal of Wildlife Management. Submitted. A research study to examine the impact of nutrition on the decline of mule deer fecundity during the last 20 years was continued in cooperation with Chad Bishop. Portions of this work will serve as his doctoral dissertation. A graduate research project (M. S.) to develop a sage grouse population model, using North Park sage grouse data to develop parameter estimates, was continued. The graduate student is Kristen Strohm. A graduate research project (M. S.) To evaluate line transect methodology for estimating pronghorn populations in eastern Colorado was initiated. The graduate student is Aaron Linstrom, and the project is in addition to his full-time duties as a terrestrial biologist with CDOW. A graduate research project (Ph. D.) to develop statistical models to monitor puma and black bear populations in Colorado based on checks of harvested animals and DNA and/or radiotracking data was initiated. The graduate student is Paul Conn. Development of the design of a monitoring system for white-tailed prairie dogs in western Colorado and eastern Utah was continued. This effort is in cooperation with Pam Schnurr, Bill Andelt, and Amy Seglund. Development of the design of a monitoring system for swift fox in eastern Colorado was continued. This effort is in cooperation with Francie Pusatari and Darby Finley. 152 �19. A workshop on use of the DEAMAN software for data entry, data summaries, and population modeling was presented to CDOW Terrestrial Biologists on May 20, 2004. A revised edition of the DEAMAN User’s Manual was provided on a CD. 153 �JOB PROGRESS REPORT CONSULTING SERVICES FOR MARK-RECAPTURE ANALYSES G. C. White P. N. OBJECTIVES Extend existing methods to better provide rigorous population monitoring systems. SEGMENT OBJECTIVES 1. 2. Extend a mark-recapture monitoring scheme to estimate population sizes with inadequate data per site to estimate encounter probabilities. Contrast line transect distance sampling approaches with mark-recapture approaches for monitoring populations with inadequate data per site to estimate encounter probabilities. ABSTRACT One of the most pervasive uses of indices of wildlife populations is uncorrected counts of animals. Two examples are the minimum number known alive from capture and release studies, and aerial surveys where the detection probability is not estimated from a sightability model, marked animals, or distance sampling. Both the mark-recapture and distance sampling estimators are techniques to estimate the probability of detection of an individual animal (or cluster of animals), which is then used to correct a count of animals. However, often the number of animals in a survey is inadequate to compute an estimate of the detection probability, and hence correct the count. Modern methods allow sophisticated modeling to estimate the detection probability, including incorporating covariates to provide additional information about the detection probability. Examples from both distance and mark-recapture sampling are presented to demonstrate the approach. RESULTS AND DISCUSSION The practice of using raw counts of animals as an index of the population size is one of the most pervasive uses of indices in wildlife management (Anderson 2003; Engeman 2003; Anderson 2001). Two examples include aerial surveys where no probability of detection is used to correct the count of observed animals, and the use of the minimum number known alive (MNKA) in animal (particularly small mammal) trapping studies (Slade and Blair 2000; McKelvey and Pearson 2001). These counts are known to be biased estimates of population size, and when used as an index, are assumed to be proportional to population size. These uses of uncorrected counts are some of the most perilous uses of an index in the practice of wildlife management because this assumption of proportionality is seldom verified, and is often false. Nichols (1992) appealed to researchers to incorporate capture-recapture estimators into small mammal studies. However, various reasons are given to explain why indices are used in place of more rigorous capture-recapture estimators. The most common reasons (Slade and Blair 2000) include fear of violating assumptions basic to mark-recapture models (Nichols and Pollack 1983), failure to recognize that some of these models are relatively robust to heterogeneity of capture probability and trap response (Carothers 1979), mistaken belief that MNKA suffers less than other models from problems of differential probabilities of capture and survival when capture probabilities are high (Nichols and Pollack 1983; Montgomery 1987), and prevalence of protocols involving fewer than the 5 to 7 trapping occasions recommended for model selection and population estimation (Otis et al. 1978). 154 �McKelvey and Pearson (2001) found that 98% of the samples collected in studies published from 1996 through 2000 they reviewed were too small for reliable selection among models of population estimation. However, their results do not take into account improved model selection methodologies, and new software and estimators that allow combining data across multiple studies and/or sites to provide more reliable model selection and estimation of the nuisance parameters. Their results mainly reflect the capabilities of CAPTURE (Otis et al. 1978; White et al. 1982), a software package developed in the late 1970's. More recent developments are available. The purpose of this presentation is to present the advantages of modern methods of analysis that allow combining data from multiple studies into a type of meta-analysis. Although resulting estimates of population size may not be completely unbiased, these estimates will certainly have less bias than MNKA. As discussed by Eberhardt et al. (1999), status of endangered large-mammal populations may have to depend on indices of population trend, and such indices may be improved by using auxiliary variables. However, in this paper, I go beyond just trying to standardize the counts with auxiliary information, as done by Eberhardt et al. (1999), and present methods incorporating auxiliary covariates that provide estimates of the population size. Modern Methods Correcting counts to produce estimates of population size. Estimators of population size based on counts of animals share a common form. A count, C is corrected for the detection probability, p, to give the population size. Because the detection probability must be estimated as p̂ (or otherwise N would be known), the result is an estimate of population size C Nˆ = (Nichols 1992). The standard methods used in wildlife studies to estimate p are markˆ p encounter methods and distance sampling. Both these seemingly diverse methodologies perform the same function: to correct a count of animals by the probability of detecting an animal. To illustrate, consider the simple Lincoln-Petersen estimator: Nˆ = n1n 2 n n = 2 = 2 , m2 ˆ m2 p n1 where n1 and n2 are the numbers of animals captured on occasions 1 and 2, and m2 is the number of animals marked on occasion 1 that are recaptured on occasion 2. Thus, m2 is an estimate of the capture n1 probability on the second occasion (Nichols 1992), because we know that n1 animals are available for capture on the second occasion, of which m2 were captured. For distance sampling, the estimate of density (Buckland et al. 1993) is: -/ nfˆ(0) n n = = Dˆ = A, ˆ2LW ˆ 2LW p p where n is the count of animals, 2LW is the area surveyed (both sides of the transect line of length L out to ˆ(0) is equivalent to 1/ p̂ . So, the right-hand side of the equation is just the a strip width W), and f ⎛ corrected count ⎜⎜ Nˆ = ⎝ n⎞ ⎟ divided by the area counted to give density. The sightability correction ˆ ⎟⎠ p models of Samuel et al. (1987) also use an equivalent approach. The probability of sighting an animal is computed for each of the groups of animals sighted, and then the number of animals in the group is divided by the estimated sighting probability to estimate the number of animals under the observed conditions that were missed. When these estimates are summed across all groups, an overall estimate of population size is obtained. Although at first glance this estimator appears to be different than the forms shown above, in fact, it is exactly the same idea. Counts are corrected by an estimate of sightability. 155 �However, because the sightability models of Samuel et al. (1987) and their extensions require first developing a model that is then applied to multiple surveys, the protocol deviates from what is the focus of this paper. That is, this paper centers on the idea of combining a number of sparse datasets into one analysis to achieve better inferences. Therefore, sightability models do not particularly fit into this approach, and so will not be discussed further here. The take-home message of this section is that counts are corrected by some probability to achieve an estimate of population size. If this correction is the same when comparing results from two surveys, then comparing just the counts will result in the same proportional change. However, without verification of the assumption that the correction is the same for both surveys, erroneous results may ensue (Nichols 1992). Consider two counts of C1 = C2 = 100, but p̂1 = 0.5 and p̂ 2 = 0.25, resulting in N̂ 1 = 200 and N̂ 2 = 400. Without knowledge of the detection probabilities, the erroneous conclusion that the population had not changed would have been made. Of course, the opposite situation can also occur. Suppose C1 = 200 and C2 = 100, with p̂1 = 0.5 and p̂ 2 = 0.25, resulting in both N̂ 1 = N̂ 2 = 400. Just comparing the counts results in the erroneous conclusion that the population has changed, when in fact, only the detection probability has changed. Thus, comparing counts is dangerous without knowledge of the underlying detection probabilities. In the next section, more advanced approaches to estimation of the detection probability are presented. Improved modeling of data to produce estimates. Earlier approaches to estimation of the size of a closed population only used the information available from the data at hand, e.g., Otis et al. (1978). Program CAPTURE (White et al. 1982) produced separate analyses for each species, sex- and age-class, and trapping grid. However, newer software packages, such as Program MARK (White and Burnham 1999) allow the user to model parameters in user-defined models. As a result, the detection probability for a population estimator can be modeled with group-specific, time-specific, and even individual-specific covariates. These covariates provide additional information with which to improve the estimates of the detection parameters. With this kind of model building capability, multiple sparse (but related) datasets can be combined into models to generate more precise estimates of p. These estimates of detection parameters and population size do not necessarily have to be the same for each of the datasets included in the analysis. For example, suppose capture probabilities are related to habitat quality, with animals in high quality habitat having smaller home ranges, and hence less probability of encountering traps. The approach advocated here is to build a model of detection probabilities by combining the data from multiple study areas and using the information about habitat quality in the model. For example, ˆ) = β 0 + β1Habitat Quality , logit(p ⎛ p ⎞ ⎟⎟ ]. The result is a model predicting ⎝1 − p ⎠ where logit is the logit transformation [logit(p) = log⎜⎜ capture probability as a function of habitat quality, where the parameters β 0 and β1 are estimated from the data and the habitat quality values provided by the user. Instead of estimating a separate value of p for each of the study areas, and likely encountering problems with too small of sample sizes, the researcher obtains an estimate of p specific to each study area based on habitat quality. Besides incorporating covariates into the model, less parameter-rich models that are still biologically realistic can be fitted to the observed data. A more extensive example is provided in White (2001), where models are combined across day- and night-time trapping occasions, gender and age-class. That example demonstrates the capability of additive models, where additive effects in the model (e.g., an effect representing the difference between day and night capture probabilities) provide differences, yet maintain a parallelism between the estimates across time or other categories. Additive models provide a useful alternative to the full multiplicative model. For example, suppose there are 5 trapping grids, each 156 �trapped for 5 nights. The full multiplicative time-specific model would have 5 × 5 = 25 parameters. In contrast, the additive model would still have 5 time parameters, but with only 4 additional parameters to represent the differences between study areas, resulting in only 9 parameters being estimated from the data. Hence, at the expense of possible bias, much improved precision of the estimates will be obtained. Finally, modern approaches provide much more flexibility in exploring alternative models. With CAPTURE, it was all or none. Only 8 models from the Mtbh set were defined, and these were cast in concrete. Only 7 models of this set had estimators (with only 5 estimators in the original program). However, with Program MARK (White and Burnham 1999), all the likelihood models from CAPTURE can be reproduced, plus many additional possibilities are provided by variations of the original 8 models. Included in MARK are the mixture models of Pledger (2000) for modeling individual heterogeneity, and Huggins (1989, 1991) and Alho (1990) versions of the closed capture estimators that allow individual covariates to be used to model initial and recapture probabilities. Thus, approaches available in MARK provide 3 areas of improvement to handle sparse markrecapture datasets. First, covariates can be incorporated into the analysis, bring additional information. Second, flexible modeling structures can provide biologically reasonable models to combine sparse datasets. Third, more flexibility is provided to construct capture-recapture models of individual datasets. Program DISTANCE 3.5 (Thomas et al. 1998; Buckland et al. 2001), and now updated to DISTANCE 4.1, provides similar capabilities for distance sampling data as does MARK for markencounter data. Data are presented to the program in strata, with stratum-specific estimates of density provided from detection probabilities estimated across strata. Although not as flexible at this time in terms of building complex models that incorporate covariates, these capabilities are forthcoming in newer versions of the program. The essence of the approach advocated here is currently available in DISTANCE – multiple datasets can be combined to estimate the sightability parameter, yet stratumspecific estimates of density are achieved. Model selection methods. Another feature of modern methods is that the estimate from a single model is not accepted as the best estimate available from the data. Burnham and Anderson (2002) describe information-theoretic model selection methods, leading to model averaging, where estimates from multiple models are combined to obtain an estimate that is an improvement over estimates from single models. The traditional approach was to find the “best” model, and use that model to make inferences from the data. However, the process of sorting through the available models carries some baggage – multiple decisions are required to decide which model is most appropriate. As a result, a source of variation in the data analysis process is ignored – model selection uncertainty. Simulations have shown that the estimates and their confidence intervals from the “best” model do not perform as hoped (Burnham et al. 1995). In particular, confidence intervals do not cover the parameter value for the expected 95% with α = 0.05. Rather than accept the poor performance because of ignoring model selection uncertainty from using the “best” model, the model-averaging methodology provided by Burnham and Anderson (2002) incorporates the model-selection uncertainty into the estimates and associated confidence intervals. Further, the approach is more biologically satisfying. For example, who really believes that the “best” model for making inferences from a capture-recapture study is something simple like Mt? Rather, we would all suspect some individual heterogeneity to be present, as in Mh. Yet, with the traditional approach of just making inferences from the “best” model, the individual heterogeneity aspect would be completely ignored if Mt was determined to be the “best” model. With model averaging, we incorporate information from all the models that have weight associated with them, with the information provided by each model proportional to its weight. The result is an estimate that reflects more accurately what we know from the data, and that a single model is inadequate for making inferences from the data. 157 �A key part of model averaging is estimating the weight to be associated with each model. The information-theoretic approach presented by Burnham and Anderson (2002) is based on Akaike’s Information Criterion (AIC). Without going into the mathematics (details are presented in Burnham and Anderson 2001; Burnham and Anderson 2002), the general idea behind AIC model selection is to rank models based on the trade-off between bias versus precision of the estimates (Figure 1). Simple models, i.e., models with small numbers of parameters, produce more precise estimates at the expense of potentially biased estimates. In contrast, complex models, i.e., models with large numbers of parameters, will produce generally unbiased estimates, but at the cost of poor precision. That is, the sampling variance of the parameter estimates from complex models will be large compared to simple model. From the AIC value for each model, a weight is computed for each model. These weights are standardized to sum to 1, so that the weight of a model reflects the likelihood of the model. From these weights, the model-averaged estimate of the parameter across all the models considered is computed. Program MARK includes information-theoretic (i.e., AIC) model selection criteria (White and Burnham 1999) and the capability for model averaging population estimates (White et al. 2001). Although DISTANCE includes AIC model selection, the capability to model average is not presently available. However, the calculations for model averaging are simple, and can be easily performed in a spreadsheet given the estimates, standard errors, and AIC values. So what’s the price for using the approach advocated here? For the user, likely a fairly steep learning curve must be climbed. More statistical and computer expertise is required to conduct the analyses described than with traditional approaches. Although likely an excuse, competent scientists will not let this reason keep them from applying better methodology to more fully interpret their data. Data are hard to come by, and deserve full treatment once acquired. Quadrat sampling example I now present an example of estimating the population size of the Mexican spotted owl in the Upper Gila Mountains Recovery Unit. Twenty-five quadrats 50–75 km2 were sampled for owls with a 4pass removal sampling scheme (Ganey et al. 1999). When an owl was detected through night-time calling, it was located the next day and leg banded to individually identify it. Recaptures were obtained when a marked owl was located during a latter pass. These capture-recapture data from banded owls on the 11 quadrats where owls had been banded and subsequently resighted were used to estimate p, the probability of capture on a given trapping occasion (Huggins 1989). To estimate p, a closed capturerecapture modeling procedure developed by Huggins (1989, 1991) that was implemented in Program MARK (White and Burnham 1999) was used. The goal was to estimate p as precisely as possible because the sampling variances of the p’s contribute to the sampling variances of the estimated N’s. In addition to p, the probability of recapture (c) can also be estimated, adding an additional parameter to be modeled. In standard closed capture-recapture models, maximum likelihood estimation is used to estimate both p and N, simultaneously (Otis et al. 1978), i.e., the resulting estimates from standard closed capture-recapture models represent the joint maximum likelihood estimates. The Huggins models differ from the standard models in that only p and c are modeled with N being estimated as a derived parameter (i.e., N is computed algebraically from p). Thus, our initial efforts centered about modeling the capturerecapture data to obtain parsimonious estimates of p. The key point relevant to this paper is that no one quadrat had adequate data to estimate p and/or c. Data were pooled across quadrats to obtain these estimates of detection probabilities, and then used to generate an estimate of N for each of the quadrats. To estimate p, 26 closed-capture models were run in program MARK. The notation used to describe these models follows Lebreton et al. (1992). In this set of models, the effects on p were modeled by sex, road access to the quadrat, occasion-specificity, and behavioral response to initial capture (i.e., inclusion of the recapture parameter c in the model). A bias-corrected version of Akaike’s Information 158 �Criteria, AICc (Burnham and Anderson 2002) was used to rank models with the best model having the lowest AICc. The best model was p = cT+roadless+sex, which constrained p’s equal to c’s, and had a linear occasion effect (T), an effect of roadless quadrats versus non-roadless quadrats and a sex effect on the p’s. The linear occasion, roadless and sex effects were all negative and different from zero (βT = -0.350, 95% CI = -0.637, -0.063; βroadless = -1.614, 95% CI = -2.742, -0.486; βsex = -0.983, 95% CI = -1.764, -0.203). This model indicated that capture probabilities declined over occasions in a linear fashion, roadless quadrats had lower capture probabilities than roaded quadrats, and that females had lower capture probabilities than males. Rather than using the p’s solely from this model, Akaike weights were estimated for each model (Buckland et al. 1997; Burnham and Anderson 2002) which represented the likelihood of a specific model as the best model to explain this particular data set, relative to the other models examined in our set of models. Akaike weights were then used to derive a weighted mean estimate of capture probabilities (pi) (i.e., the pi were “model averaged”) for each occasion for each sex and within roaded and unroaded quadrats across all models (see Stanley 1998a, 1998b). These weighted estimates of pi had estimated standard errors that included a variance component due to model selection uncertainty, i.e., which model was best for providing an adequate structure on the p’s (Buckland et al. 1997; Burnham and Anderson 2002). Thus, we ended up with 16 estimates of p, one for each of four occasions times two types of quadrats (roaded versus unroaded) and for each sex. Based on these estimates of p, a population estimate for the recovery unit was 2173 with SE 520. Had just the raw counts been used, the estimate would have been 1564 with SE 222. This example illustrates an extreme case where each trapping grid (quadrat) contained so little information about detection probabilities that by individual quadrat, the researcher is left with no choice but to use the MNKA value. However, by combining these sparse data, useful estimates were obtained that corrected for the bias of MNKA. CONCLUSIONS Sparse data need not be an impediment to correcting counts of populations to less biased estimates of population size. Modern methods incorporate information from auxiliary variables, build models from multiple sources of information, and build biologically reasonable models with fewer parameters than older approaches. Thus, past justifications of using counts as indices to population levels because of sparse data are no longer defensible. If biologists do not correct counts, we run the risk of drawing erroneous conclusions from our data, and generally losing credibility with our public critics. 159 �LITERATURE CITED Alho, J. M. (1990). Logistic regression in capture-recapture models. Biometrics 46, 623-635. Anderson, D. R. (2003). Response to Engeman, index values rarely constitute reliable information. Wildlife Society Bulletin 31, 288-291. Anderson, D. R. (2001). The need to get the basics right in wildlife field studies. Wildlife Society Bulletin 29, 1294-1297. Buckland, S.T., Anderson, D. R., Burnham, K. P., Laake, J. L, Borchers, D. L., and Thomas, L. (2001). ‘Introduction to distance sampling.’ (Oxford University Press: London) Buckland, S. T., Burnham, K. P. , and Augustin, N. H. (1997). Model selection: an integral part of inference. Biometrics 53, 603-618. Buckland, S.T., Anderson, D.R., Burnham, K. P., and Laake, J.L. (1993). ‘Distance sampling: estimating abundance of biological populations.’ (Chapman and Hall: New York) Burnham, K. P., and Anderson, D. R. (2002). ‘Model selection and multimodel inference: a practical information-theoretic approach.’ 2nd edition. (Springer-Verlag: New York) Burnham, K. P., and Anderson, D. R. (2001). Kullback-Leibler information theory as a basis for strong inference in ecological studies. Wildlife Research 28, 111-119. Burnham, K.P., White, G. C., and Anderson, D. R. (1995). Model selection strategy in the analysis of capture-recapture data. Biometrics 51, 888-898. Carothers, A. D. (1979). Quantifying unequal catchability and its effect on survival estimates in an actual population. Journal of Animal Ecology 48, 863-869. Eberhardt, L. L., Garrott, R. A., and Becker, B. L. 1999. Using trend indices for endangered species. Marine Mammal Science 15: 766-785. Engeman, R. M. (2003). More on the need to get the basics right: population indices. Wildlife Society Bulletin 31, 286-287. Ganey, J. L., Ackers, S. , Fonken, P., Jenness, J. S., Kessler, C. , Nodal, K., Shaklee, P. , and Swarthout, E. (1999). Monitoring populations of Mexican spotted owls in Arizona and New Mexico: 1999 Progress report. USDA Forest Service, Rocky Mountain Research Station, Flagstaff, Arizona, USA. (available at http://www.rms.nau.edu/lab/4251/spowmonitoring.html). Huggins, R. M. (1989). On the statistical analysis of capture-recapture experiments. Biometrika 76, 133-140. Huggins, R. M. (1991). Some practical aspects of a conditional likelihood approach to capture experiments. Biometrics 47, 725-732. Lebreton, J.-D., Burnham, K. P., Clobert, J., and Anderson, D. R. (1992). Modeling survival and testing biological hypotheses using marked animals: a unified approach with case studies. Ecological Monographs 62: 67-118. Montgomery, W. I. (1987). The application of capture-mark-recapture methods to the enumeration of small mammal populations. Symposia of the Zoological Society of London 58, 25-57. Nichols, J. D. (1992). Capture-recapture models: using marked animals to study population dynamics. Bioscience 42, 94-102. Nichols, J. D., and Pollock, K. H. (1983). Estimation methodology in contemporary small mammal capture-recapture studies. Journal of Mammalogy 64, 253-260. McKelvey, K. S., and Pearson, D. E. (2001). Population estimation with sparse data: the role of estimators versus indices revisited. Canadian Journal of Zoology 79, 1754-1765. Otis, D. L., Burnham, K. P., White, G. C., and Anderson, D. R. (1978). Statistical inference from capture data on closed animal populations. Wildlife Monographs 62, 1-135. Pledger, S. (2000). Unified maximum likelihood estimates for closed capture-recapture models using mixtures. Biometrics 56, 434-442. Samuel, M. D., Garton, E. O., Schlegel, M. W., and Carson, R. G. (1987). Visibility bias during aerial surveys of elk in northcentral Idaho. Journal of Wildlife Management 51, 622-630. 160 �Slade, N. A., and Blair, S. M. (2000). An empirical test of using counts of individuals captured as indices of population size. Journal of Mammalogy 81, 1035-1045. Stanley, T. R., and Burnham, K. P. (1998a). Estimator selection for closed-population capture-recapture. Journal of Agricultural, Biological, and Environmental Statistics 3, 131-150. Stanley, T. R., and Burnham, K. P. (1998b). Information-theoretic model selection and model averaging for closed-population capture-recapture studies. Biometrical Journal 40, 475-494. Thomas, L., Laake, J. L. , Derry, J. F., Buckland, S. T., Borchers, D. L., Anderson, D. R., Burnham, K. P., Strindberg, S., Hedley, S. L., Burt, M. L., Marques, F. F. C., Pollard, J. H., and Fewster, R. M. (1998). Distance 3.5. Research Unit for Wildlife Population Assessment, University of St. Andrews, United Kingdom. http://www.ruwpa.st-and.ac.uk/distance/. White, G. C. (2001). Statistical models: keys to understanding the natural world. In ‘Modeling in Natural Resource Management.’ (Ed. T. M. Shenk and A. B. Franklin). pp 35-56. (Island Press: Washington, D. C.) White, G. C., Burnham, K. P., and Anderson, D. R. (2001). Advanced features of Program Mark. In ‘Wildlife, land, and people: priorities for the 21st century. Proceedings of the Second International Wildlife Management Congress.’ (Ed. R. Field, R. J. Warren, H. Okarma, and P. R. Sievert). pp 368-377. (The Wildlife Society: Bethesda, Maryland) White, G. C., and Burnham, K. P. (1999). Program MARK: survival estimation from populations of marked animals. Bird Study 46 Supplement, 120-138. White, G. C., Anderson, D. R., Burnham, K. P., and Otis, D. L. (1982). Capture-recapture and removal methods for sampling closed populations. LA-8787-NERP, Los Alamos National Laboratory, Los Alamos, New Mexico, USA. 235 pp. _________________________ Gary C. White, CSU Professor Variance Bias Squared Prepared by: Few Number of Parameters Many Figure 1. The trade-off between bias2 and variance as a function of the number of parameters (from Burnham and Anderson 2002; 2001). Models with few parameters produce precise estimates that are biased, whereas models with many parameters produce less biased estimates, but imprecise. 161 �Colorado Division of Wildlife July 2004 – June 2005 WILDLIFE RESEARCH REPORT State of Colorado Cost Center 3430 Work Package 3001 Task No. 5 Federal Aid Project: W-185-R : Division of Wildlife : Mammals Research : Deer Conservaton : Multispecies Investigations Consulting Services for Mark-Recapture Analysis : Period Covered: July 1, 2004 - June 30, 2005 Author: G. C. White Personnel: C. Bishop, G. Miller, T. E. Remington, D. J. Freddy, T. M. Shenk, L. Stevens, J. Craig, R. Kahn, D. C. Bowden, F. Pusateri, J. Dennis, P. Schnurr, B. Andelt, A. Seglund, D. Finley, A. Linstrom, K. Strohm, P. Conn. All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT Progress towards the objectives of this job include: 1. Consulting assistance to CDOW on harvest surveys, terrestrial inventory systems, and population modeling procedures was provided. Assistance with estimation of spring and fall turkey, spring snow goose, sharp-tailed and sage grouse, chukars, ptarmigan, Abert’s squirrels, and general small game harvest was provided, and programs and harvest estimates provided to CDOW via email and CD ROM. Computer code written in SAS to compute these estimates and display results graphically was also provided. Computer code was also written in SAS to estimate the compliance rate of Colorado small game license holders with the Harvest Information Program. 2. The DEAMAN software package for the storage, summary, and analysis of big game population and harvest data was revised further as a Windows XP program. A User’s Manual has been provided to terrestrial biologists via the WWW at http://www.cnr.colostate.edu/~gwhite/deaman. I met with the CDOW software group to discuss conversion of DEAMAN to a central server application. 3. Consultation with CDOW Terrestrial Biologists in the use of DEAMAN and population modeling procedures continued. Numerous questions were answered via meetings with biologists, and via email. 4. A paper on the estimation of mule deer population sizes in GMU 10 was published in the Wildlife Society Bulletin: Freddy, D. J., G. C. White, M. C. Kneeland, R. H. Kahn, J. W. Unsworth, W. J. deVergie, V. K. Grahm, J. H. Ellenberger, and C. H. Wagner. 2004. How many mule deer are there? Challenges of credibility in Colorado. Wildlife Society Bulletin 32:916-927. 5. A paper on the peregrine falcon population dynamics in Colorado was published in the Journal of Wildlife Management: Craig, G. R., G. C. White, and J. H. Enderson. 2004. Survival, recruitment, and rate of population change of the peregrine falcon population in Colorado. Journal of Wildlife Management 68:1032-1038. 67 �6. A paper on the impact of limited antlered harvest on mule deer sex and age ratios was accepted for publication in the Wildlife Society Bulletin: Bishop, C. J., G. C. White, D. J. Freddy, and B. E. Watkins. 2005. Effect of limited antlered harvest on mule deer sex and age ratios. Wildlife Society Bulletin. In Press. 7. A paper on the estimation of the area of black-tailed prairie dog colonies in eastern Colorado was accepted for publication in the Wildlife Society Bulletin: White, G. C., J. R. Dennis, and F. M. Pusateri. 2005. Area of black-tailed prairie dog colonies in eastern Colorado. Wildlife Society Bulletin. In Press. 8. A paper on methodologies to obtain more rigorous population monitoring data was accepted for publication in Wildlife Research: White, G. C. 2004. Correcting counts: techniques to de-index. Wildlife Research. In Press. 9. A paper evaluating methods of estimating the impact of harvest on survival rates was published in Animal Diversity and Conservation: Otis, D. L., and G. C. White. 2004. Evaluation of ultrastructure and random effects band recovery models for estimating relationships between survival and harvest rates in exploited populations. Animal Biodiversity and Conservation 27.1: 157-173. 10. A paper on the procedures to monitor swift fox populations in eastern Colorado was accepted for publication in the Journal of Wildlife Management: Finley, D. J., G. C. White and J. P. Fitzgerald. 2004. Estimation of swift fox population size and occupancy rates in eastern Colorado. Journal of Wildlife Management. In Press. 11. A research study to examine the impact of nutrition on the decline of mule deer fecundity during the last 20 years was continued in cooperation with Chad Bishop. Portions of this work will serve as his doctoral dissertation. 12. A graduate research project (M. S.) to develop a sage grouse population model, using North Park sage grouse data to develop parameter estimates, was completed. The graduate student is Kristen Strohm and her thesis is “Sage Grouse Population Dynamics in North Park, Colorado”. 13. A graduate research project (M. S.) To evaluate line transect methodology for estimating pronghorn populations in eastern Colorado was continued. The graduate student is Aaron Linstrom, and the project is in addition to his full-time duties as a terrestrial biologist with CDOW. 14. A graduate research project (Ph. D.) to develop statistical models to monitor puma and black bear populations in Colorado based on checks of harvested animals and DNA and/or radio-tracking data was continued (with funding for 04-05 through the CSU PRIMES program). The graduate student is Paul Conn. 15. Development of the design of a monitoring system for white-tailed prairie dogs in western Colorado and eastern Utah was continued. This effort is in cooperation with Pam Schnurr, Bill Andelt, and Amy Seglund. 16. Development of the design of a monitoring system for swift fox in eastern Colorado was continued, and data analysis for this project was initiated. This effort is in cooperation with Francie Pusatari and Darby Finley. 68 �WILDLIFE RESEARCH REPORT CONSULTING SERVICES FOR MARK-RECAPTURE ANALYSES GARY C. WHITE P. N. OBJECTIVE Monitor swift fox populations in eastern Colorado. SEGMENT OBJECTIVES 1. Extend a mark-recapture monitoring scheme to estimate occupancy rates of swift foxes (Vulpes velox) on 12-mi2 quadrats in eastern Colorado. 2. Contrast estimates from the current survey with thoses obtained in 1998 and published in Finley et al. (2005). ABSTRACT A randomly selected sample of 15 ~12-mi2 grids in eastern Colorado were trapped with a 4 × 5 grid of traps between August, 2004 and February, 2005. Swift foxes were trapped on 36 of the 51 grids, with 136 total fox captures. Comparison of the estimates of the percent of 12-mi2 grids occupied by swift foxes in eastern Colorado does not appear to have changed since a comparable sample was taken of 72 grids in March, 1995 – January, 1997 (Finley et al. 2005). Using the average percentage of the grids in short grass prairie with the minimum AICc model, the earlier estimate was ψ̂ = 0.821 (SE 0.0659), compared to the current estimate of ψ̂ = 0.777 (SE = 0.0786). The estimated change is −0.044 (SE = 0.103, 95% CI −0.245 – 0.157). Summing the predicted occupancy values across the sampled grids for the respective studies provides a similar conclusion: Finley et al. (2005) found ψ̂ = 0.790 (SE = 0.0574), whereas this study found ψ̂ = 0.742 (SE = 0.0869), providing an estimate of the change of −0.048 (SE = 0.104, 95% CI −0.252 – 0.156). These differences are well within the sampling variation of the estimates, and do not suggest a decline in swift fox populations in eastern Colorado. RESULTS Sample of Grids Finley et al. (2005) found that the covariate percent Short Grass Prairie (SGP) is a good predictor of the presence of foxes in eastern Colorado. The distribution of this covariate is bimodal (Figure 1). To build the best relationship between SGP and fox numbers, we sampled across this continuum of SGP values. Thus, the 2,566 trapping grids considered in the sampling frame of grids (Figure 1) to be trapped were sorted by the percentage of SGP predicted by the CDOW GIS system. Then a random start between 1 and 66 was picked, and every 50th grid was selected to be sampled. This procedure resulted in a sample of 51 blocks. When I multiply the frequency of the sample by 50, I obtain a close relationship between the sampling frame and the grids sampled (Figure 2). Statistical Methods Analysis methods to estimate occupancy rates followed the procedures of Finley et al. (2005), using the occupancy model of MacKenzie et al. (2002) in Program MARK (White and Burnham 1999). I considered a set of a priori models that incorporated month as sine and cosine functions to model detection probabilities (p), and the percentage of short grass prairie on the trapping grid to model both 69 �detection probabilities and probability of occupancy, ψ (psi). Model selection was performed with information-theoretic methods following Burnham and Anderson (2002). Analysis methods to estimate the population of foxes using a trapping grid also followed Finley et al. (2005), using the Huggins estimator (Huggins 1989, 1991) to estimate population size. Model selection was performed with information-theoretic methods following Burnham and Anderson (2002). Occupancy Estimation Model selection results for occupancy estimation are shown in Table 1. The sine and cosine functions for month did not improve model fit of detection probabilities, nor did the percentage of short grass prairie improve estimates of detection probabilities. However the percentage of short grass prairie did provide an important predictor of occupancy (Figure 3) with the logit predictive equation: exp[βˆ 0 + βˆ 1 (SGP%)] Occupancy Probability = , 1+ exp[βˆ 0 + βˆ 1 (SGP%)] where β̂ 0 = -0.287 (SE = 0.624, 95% CI −1.510 – 0.936) and β̂1 = 2.775 (SE = 1.299, 95% CI 0.229 – 5.322). The estimated occupancy rate using the average amount of short grass prairie found on the 51 grids samples was ψ̂ = 0.777 (SE = 0.0786, 95% CI 0.589 – 0.894). When the estimated occupancy was summed across the 51 grids using the observed amount of short grass prairie on each grid, ψ̂ = 0.742 (SE = 0.0869, 95% CI 0.572 – 0.912). Finally, the entire population of grids from which the 51 sampled grids were drawn was used to compute the proportion of eastern Colorado occupied by swift foxes: ψ̂ = 0.711. The amount of short grass prairie for each of the grids in the population was estimated based on a GIS layer. Population Estimation Model selection results for population estimation (Table 2) suggest a behavioral effect in response to initial capture, with capture probabilities a function of month and SGP. Initial capture probabilities (Figure 4) and recapture probabilities (Figure 5) from the minimum AICc model are a function of month through a sin transformation, and SGP. The mean number of animals estimated per grid for all 51 grids was 4.83 (SE = 1.990, 95% CI 0.933 – 8.735), ranging from 0 to 26. DISCUSSION Simulations reported in Finley et al. (2005) reported expected power to detect declines given various combinations of numbers of trapping occasions and numbers of grids trapped. For 50 grids trapped with 3 occasions, their simulation results suggested a SE of about 0.070 for ψ = 0.8. The estimated SEs from this study are slightly greater then this value, likely because of the variation in SGP over the range of the sample. However, the values are close enough to make the simulation results reported in Finley et al. (2005) useful if taken a bit conservatively. The results from this study concerning the importance of SGP in predicting swift fox occupancy compared favorably with the results obtained by Finley et al. (2005) (Figure 6). Basically, the same relationship of SGP to occupancy was found. However, the minimum AICc model for occupancy in this study was much simpler than that of Finley et al. (2005), mainly because grids were trapped only during 70 �the period late August through March when the highest detection probabilities were expected based on Finley et al. (2005) work. When Finley et al. (2005) used the percentage of short grass prairie for each of their sampled grids to estimate a grid-specific ψ value, the sum of ψ̂ values was 56.9 (SE = 4.13), or 56.9 of the 72 ˆ = 0.790, SE = 0.0574). Alternatively, they estimated ψ of 0.821 grids actually contained foxes ( ψ (SE = 0.0659) using the mean (66.9%) of the short-grass prairie habitat for the 72 grids. In either case, their estimates are slightly greater than the values of ψ estimated in this study with the same approaches, but negligibly so when the uncertainty of the estimates is taken into account. As cautioned in Finley et al. (2005), the mean number of animals estimated per grid cannot be extrapolated to a population estimate for eastern Colorado because the grids attract foxes from some unknown distance outside the trapping grid. SUMMARY Comparison of the estimates of the percent of 12-mi2 grids occupied by swift foxes in eastern Colorado does not appear to have changed since a comparable sample was taken of 72 grids in March, 1995 – January, 1997 (Finley et al. 2005). Using the average percentage of the grids in short grass prairie with the minimum AICc model, the earlier estimate was ψ̂ = 0.821 (SE 0.0659), compared to the current estimate of ψ̂ = 0.777 (SE = 0.0786). The estimated change is −0.044 (SE = 0.103, 95% CI −0.245 – 0.157). Summing the predicted occupancy values across the sampled grids for the respective studies provides a similar conclusion: Finley et al. (2005) found ψ̂ = 0.790 (SE = 0.0574), whereas this study found ψ̂ = 0.742 (SE = 0.0869), providing an estimate of the change of −0.048 (SE = 0.104, 95% CI −0.252 – 0.156). These differences are well within the sampling variation of the estimates, and do not suggest a decline in swift fox populations in eastern Colorado. LITERATURE CITED FINLEY, D. J., G. C. WHITE AND J. P. FITZGERALD. 2005. Estimation of swift fox population size and occupancy rates in eastern Colorado. Journal of Wildlife Management. In Press. BURNHAM, K. P., AND D. R. ANDERSON. 2002. Model selection and multimodel inference: a practical information-theoretic approach. 2nd edition. Springer-Verlag, New York, New York, USA. HUGGINS, R. M. 1989. On the statistical analysis of capture-recapture experiments. Biometrika 76: 133– 140. ______________. 1991. Some practical aspects of a conditional likelihood approach to capture experiments. Biometrics 47: 725–732. MACKENZIE, D. I., J. D. NICHOLS, G. B. LACHMAN, S. DROEGE, J. A. ROYLE, AND C. A. LANGTIMM. 2002. Estimating site occupancy when detection probabilities are less than one. Ecology 83: 2248–2255. WHITE, G. C., AND K. P. BURNHAM. 1999. Program MARK: survival estimation from populations of marked animals. Bird Study 46 Supplement: 120–138. Prepared by: ________________________ Dr. Gary C. White, Department Fishery & Wildlife Conservation Biology Colorado State University 71 �Table 1. Occupancy model selection results for 51 swift fox grids trapped in eastern Colorado, August 2004 to February, 2005. AICc ∆AICc 196.785 198.882 198.891 199.133 200.412 201.176 201.242 201.522 203.449 0 2.0969 2.1065 2.3486 3.6277 4.3916 4.4573 4.7372 6.6642 Model {p(.) ψ(SGP)} {p(sinMonth) ψ(SGP)} {p(SGP) ψ(SGP)} {p(cosMonth) ψ(SGP)} {p(.) ψ(.)} {p(sinMonth+cosMonth) ψ(SGP)} {p(T) ψ(.)} {p(cosMonth+cosMonth^2) ψ(SGP)} {p(t) ψ(.)} AICc Model Weights Likelihood 0.39689 1 0.1391 0.3505 0.13843 0.3488 0.12265 0.309 0.0647 0.163 0.04416 0.1113 0.04273 0.1077 0.03715 0.0936 0.01418 0.0357 Num. Deviance Par 3 190.274 4 190.012 4 190.022 4 190.264 2 196.162 5 189.843 3 194.731 5 190.189 4 194.579 Table 2. Closed population estimator model selection results for 51 swift fox grids trapped in eastern Colorado, August 2004 to February, 2005. Model {p(SGP+sinMonth)= c(SGP+sinMonth)+additive effect} {p(SGP+sinMonth+ cosMonth)=c(SGP+ sinMonth+cosMonth)+additiv e effect} {p(SGP)=c(SGP)+ additive effect} {p(sinMonth)= c(sinMonth)+ additive effect} {p(cosMonth)= c(cosMonth)+ additive effect} {p(cosMonth+ sinMonth)= c(cosMonth+ sinMonth)+additive effect} {p(.) c(.)} {p(T)=c(T)} {p(.)=c(.)} {p(T)=c(T)+ additive effect} {p(t)=c(t)+ additive effect} {p(g*t)=c(g*t)} AICc Delta AICc AICc Weights Model Likelihood No. Par Deviance 331.785 0 0.4213 1 4 323.666 333.739 1.9538 0.15861 0.3765 5 323.56 334.006 2.2207 0.13879 0.3294 3 327.935 334.421 2.636 0.11277 0.2677 3 328.35 336.195 4.4094 0.04646 0.1103 3 330.124 336.322 337.474 337.658 338.619 339.211 339.84 537.81 4.5372 5.6892 5.8723 6.8334 7.4255 8.0543 206.024 0.04359 0.0245 0.02236 0.01383 0.01028 0.00751 0 0.1035 0.0582 0.0531 0.0328 0.0244 0.0178 0 4 2 2 1 3 4 108 328.204 333.439 333.622 336.607 333.14 331.721 220.762 72 �350 250 200 150 100 I I I I I I I I I I I I I I 90 -9 5 I 80 -8 5 I 70 -7 5 I 60 -6 5 I 20 -2 5 I 10 -1 5 05 0 - 50 -5 5 - 40 -4 5 50 30 -3 5 Frequency ` 300 Percent Short Grass Prairie Figure 1. Histogram of percentage of short grass prairie on 12-mi2 trapping grids comprising the sampling frame for this study. 350 ■ Frame □ Sample 250 200 150 - ~ 50 - - - ~ ,- - ,- - - ~ r - 50 -5 5 40 -4 5 30 -3 5 20 -2 5 10 -1 5 05 0 - - ,- - - I ~ ,- 90 -9 5 - 80 -8 5 ,- 70 -7 5 100 60 -6 5 Frequency ` 300 Percent Short Grass Prairie Figure 2. Histogram showing the close relationship between the grids included in the sample compared to the sampling frame. A representative sample relative to the availability of the SGP variable was selected. 73 �` Occupancy Probability - -- - --- 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 --- 0% 20% 40% 60% 80% 100% Short Grass Prairie (Percent) Figure 3. Prediction of the probability of occupancy with 95% confidence intervals as a function of the percentage of short grass prairie on the 12-mi2 trapping grid. Ticks on the 0 and 1 lines indicate the status of the 51 trapping grids, with 36 of the grids recording foxes captured. Capture Probability ` 0.6 0.5 0.4 - --- - - 0.3 0.2 - - --------- 0.1 0 0% 20% 40% 60% 80% 100% Short Grass Prairie (Percent) 1- p Sep p Jan -- p Oct p Feb -- p Nov - p Dec Figure 4. Changes in initial capture probability for swift fox trapped in eastern Colorado on 12-mi2 grids, August 2004 – February, 2005. 74 �Reapture Probability 0.25 0.2 0.15 0.1 _,. . - - - - -- - - - . -- - - - - 0.05 0 0% 20% 40% 60% 80% 100% Short Grass Prairie (Percent) c Sep c Jan c Oct c Feb c Nov c Dec Figure 5. Changes in recapture probability for swift fox trapped in eastern Colorado on 12-mi2 grids, August 2004 – February, 2005. - - - .. 1 .- Probability of occupancy 0.9 -. 0.8 0.7 0.6 - . -. - ----- 0.5 - - - - , - -- - ~ - - - - __ __ _ 11111111 _ • - - - - 0.4 # 0.3 - - - -~ - -# -# - - - -# - - - - - - - 0.2 -- 0.1 0 0 20 40 60 80 100 Short grass grairie (%) Figure 6. Effect of the percentage of the 12–mi2 grid consisting of short-grass prairie habitat on the probability of occupancy by swift foxes trapped on 72 grids in eastern Colorado, March, 1995 – January, 1997, for the top-ranked AICc model {p(T + cos(Month) + cos2(Month)) ψ (SGP Proportion)} from Finley et al. (2005). The dashed lines are 95% confidence intervals for the estimated probability of occupancy. Ticks across the 0 and 1 occupancy lines are the observed occupancy values plotted against the percentage of short grass prairie for the 72 grids, with short grass prairie values dithered so that grids would not plot on top of each other. 75 �Colorado Division of Wildlife July 2005 – June 2006 WILDLIFE RESEARCH REPORT State of Cost Center Work Package Task No. Colorado 3430 3001 5 Federal Aid Project: W-185-R : Division of Wildlife : Mammals Research : Deer Conservaton : Multispecies Investigations Consulting Services for Mark-Recapture Analysis : Period Covered: July 1, 2005 - June 30, 2006 Author: G. C. White Personnel: C. Bishop, D. J. Freddy, T. M. Shenk, L. Stevens, R. Kahn, F. Pusateri, E. O’Dell, D. Martin, P. Schnurr, K. Navo, B. Andelt, D. Finley, A. Linstrom, K. Strohm, P. Conn. All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT Progress towards meeting the objectives of this job include: 1. Consulting assistance to Colorado Division of Wildlife (CDOW) on harvest surveys, terrestrial inventory systems, and population modeling procedures was provided. Assistance with estimation of spring and fall turkey, spring snow goose, sharp-tailed and sage grouse, chukars, ptarmigan, Abert’s squirrels, and general small game harvest was provided, and programs and harvest estimates provided to CDOW via email and CD ROM. Computer code written in SAS to compute these estimates and display results graphically was also provided. Computer code was also written in SAS to estimate the compliance rate of Colorado small game license holders with the Harvest Information Program. 2. The CDOW DEAMAN software package for the storage, summary, and analysis of big game population and harvest data was revised further as a Windows XP program. A User’s Manual has been provided to terrestrial biologists via the WWW at http://www.cnr.colostate.edu/~gwhite/deaman. I met with the CDOW software group to discuss conversion of DEAMAN to a central server application. 3. Consultation with CDOW Terrestrial Biologists in the use of DEAMAN and population modeling procedures continued. Numerous questions were answered via meetings with biologists, and via email. 4. A paper comparing the population levels of swift foxes in eastern Colorado to a previous study in cooperation with CDOW was submitted to Southwestern Naturalist: Martin, D. J., G. C. White, and F. M. Pusateri. 2006. Monitoring swift fox populations in eastern Colorado. Southwestern Naturalist. Submitted. 5. A paper on the use of vaginal implant transmitters in cooperation with CDOW was submitted and accepted for publication in the Journal of Wildlife Management: Bishop, C. J., D. J. Freddy, G. C. 91 �White, B. E. Watkins, T. R. Stephenson, and L. L. Wolfe. 2006. Using vaginal implant transmitters to aid in capture of neonates from marked mule deer. Journal of Wildlife Management. In Press. 6. A paper resulting from Dan Walsh’s M.S. project in cooperation with CDOW was submitted to Ecological Applications: Walsh, D. P., G. C. White, T. E. Remington, and D. C. Bowden. 2006. Population Estimation of Greater Sage-Grouse. Ecological Applications. Submitted. 7. A paper resulting from Sherri Huwer’s M.S. project in cooperation with CDOW was submitted to the Journal of Wildlife Management: Huwer, S. L., D. R. Anderson, T. E. Remington, and G. C. White. 2006. Evaluating the importance of forbs to sage-grouse using human-imprinted chicks. Journal of Wildlife Management. Submitted. 8. A paper on mountain sheep populations in Rocky Mountain National Park was submitted and accepted for publication in the Wildlife Society Bulletin: McClintock, B. T., and G. C. White. 2006. Bighorn sheep abundance following a suspected pneumonia epidemic in Rocky Mountain National Park. Wildlife Society Bulletin. In Press. 9. A paper on extending the mark-resight estimator using a beta-binomial distribution was submitted and accepted in the Journal of Agricultural, Biological, and Ecological Statistics: McClintock, B. T., G. C. White, and K. P. Burnham. 2006. A robust design mark-resight abundance estimator allowing heterogeneity in resighting probabilities. Journal of Agricultural, Biological, and Ecological Statistics. In Press. 10. A paper resulting from the May, 2005 Elk and Deer Workshop was submitted and accepted for publication in the Wildlife Society Bulleting: Mason, J. R., L. H. Carpenter, M. Cox, J. C. Devos, J. Fairchild, D.J. Freddy, J. R. Heffelfinger, R. H. Kahn, S. M. McCorquodale, D. F. Pac, D. Summers, G. C. White, and B. K. Williams. 2006. A case for standardized ungulate surveys and data management in the western United States. Wildlife Society Bulletin. In Press. 11. A paper describing the use of closed captures models to estimate population size with Program MARK was submitted and accepted for publication in Environmental and Ecological Statistics: White, G. C. 2006. Closed population estimation models and their extensions in program MARK. Environmental and Ecological Statistics. In Press. 12. A paper on the application of multistate models in Program MARK was submitted and accepted for publication in the Journal of Wildlife Management: White, G. C., W. L. Kendall, and R. J. Barker. 2006. Multistate survival models and their extensions in program MARK. Journal of Wildlife Management. In Press. 13. A paper on the estimation of female grizzly bears was submitted to the Journal of Agricultural, Biological, and Ecological Statistics: Cherry, S., G. C. White, K. A. Keating, M. A. Haroldson, C. C. Schwartz. 2006. Evaluating estimators of the numbers of females with cubs-of-the-year in the Yellowstone grizzly bear population. Journal of Agricultural, Biological, and Ecological Statistics. Submitted. 14. A paper on the survival of mule deer in the Bridger Mountains, Montana, was submitted and accepted for publication in the Journal of Wildlife Management: Pac, D. F., and G. C. White. 2006. Survival and cause-specific mortality of mule deer in the Bridger Mountains, Montana. Journal of Wildlife Management. In Press. 92 �15. A paper on the impact of limited antlered harvest on mule deer sex and age ratios in cooperation with CDOW was published in the Wildlife Society Bulletin: Bishop, C. J., G. C. White, D. J. Freddy, and B. E. Watkins. 2005. Effect of limited antlered harvest on mule deer sex and age ratios. Wildlife Society Bulletin 33: 662–668. 16. A paper on estimation of nest survival was submitted and accepted for publication in Studies in Avian Biology: Heisey, D. M., T. L. Shaffer, and G. C. White. 2006. The ABCs of nest survival: theory and application from a biostatistical perspective. Studies in Avian Biology. In Press. 17. A paper on the estimation of the area of black-tailed prairie dog colonies in eastern Colorado in cooperation with CDOW was published in the Wildlife Society Bulletin: White, G. C., J. R. Dennis, and F. M. Pusateri. 2005. Area of black-tailed prairie dog colonies in eastern Colorado. Wildlife Society Bulletin 33:265–272. 18. A paper in response to a critique by Sterling Miller was published in the Wildlife Society Bulletin in cooperation with CDOW: White, G. C., J. R. Dennis, and F. M. Pusateri. 2005. Response to: Overestimation bias in estimate of black-tailed prairie dog abundance in Colorado. Wildlife Society Bulletin 33:1452–1455. 19. A paper on methodologies to obtain more rigorous population monitoring data was published in Wildlife Research: White, G. C. 2005. Correcting wildlife counts with detection probabilities. Wildlife Research 32:211–216. 20. A paper on the procedures to monitor swift fox populations in eastern Colorado was published in the Journal of Wildlife Management: Finley, D. J., G. C. White and J. P. Fitzgerald. 2005. Estimation of swift fox population size and occupancy rates in eastern Colorado. Journal of Wildlife Management 69:861–873. 21. A research study to examine the impact of nutrition on the decline of mule deer fecundity during the last 20 years was continued in cooperation with Chad Bishop and CDOW. Portions of this work will serve as his doctoral dissertation in addition to his full-time duties as a researcher with CDOW. 22. A graduate research project (M. S.) was continued in cooperation with CDOW to evaluate line transect methodology for estimating pronghorn populations in eastern Colorado. The graduate student is Aaron Linstrom, and the project is in addition to his full-time duties as a biologist with CDOW. 23. A graduate research project (Ph. D.) in cooperation with CDOW to develop statistical models to monitor puma and black bear populations in Colorado based on checks of harvested animals and DNA and/or radio-tracking data was continued. The graduate student is Paul Conn. 24. A graduate research project (M. S.) in cooperation with CDOW to evaluate methods of redistributing elk in and around Great Sand Dunes National Park was started and then discontinued. The student, Greg Davidson, switched his work to evaluate habitat use by elk on the Grand Mesa. 25. Development of the design of a monitoring system for white-tailed prairie dogs in western Colorado and eastern Utah was continued in cooperation with CDOW with P. Schnurr, K. Navo and B. Andelt. 26. Design of a monitoring system for black-tailed prairie dogs in eastern Colorado in cooperation with CDOW was continued. This effort is in cooperation with Francie Pusateri and Eric O’Dell of CDOW. 93 �WILDLIFE RESEARCH REPORT CONSULTING SERVICES FOR MARK-RECAPTURE ANALYSES GARY C. WHITE P. N. OBJECTIVE Provide expert biostatistical and experimental design services to the Colorado Division of Wildlife, Wildlife Programs Branch. SEGMENT OBJECTIVES 1. Provide biostatisitical support to implement and analyze CDOW hunter harvest surveys. 2. Provide professional oversight, critiques, and analytical support to CDOW terrestrial management and avian and mammals research sections. 3. Convey to CDOW research and management sections new and pertinent information obtained in various collaborative projects conducted with other agencies and entities. RESULTS, DISCUSSION, SUMMARY See ABSTRACT for summary of key activities and publications. Prepared by: Dr. Gary C. White, Department of Fish, Wildlife, and Conservation Biology Colorado State University 94 �Colorado Division of Wildlife July 2006 – June 2007 WILDLIFE RESEARCH REPORT State of Cost Center Work Package Task No. Colorado 3430 3001 5 Federal Aid Project: W-185-R : Division of Wildlife : Mammals Research : Deer Conservation : Multispecies Investigations Consulting Services for Mark-Recapture Analysis : Period Covered: July 1, 2006 - June 30, 2007 Author: G. C. White Personnel: C. Bishop, D. J. Freddy, T. M. Shenk, P. Lukacs, R. Kahn, F. Pusateri, E. O’Dell, D. Martin, P. Schnurr, K. Navo, B. Andelt, A. Linstrom, P. Conn, B. McClintock, G. Davidson, and J. Ivan. All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT Progress towards meeting the objectives of this job include: 1. Consulting assistance to Colorado Division of Wildlife (CDOW) on harvest surveys, terrestrial inventory systems, and population modeling procedures was provided. Computer code written in SAS to compute these estimates and display results graphically was provided. Specific input involved estimation of variances for design of the E-2 elk aerial survey. 2. Support for the CDOW DEAMAN software package for the storage, summary, and analysis of big game population and harvest data was provided. I met with the CDOW software group to discuss conversion of DEAMAN to a central server application. 3. Consultation with CDOW Terrestrial Biologists in the use of DEAMAN and population modeling procedures continued. Numerous questions were answered via meetings with biologists, and via email. A workshop on modeling Colorado’s deer, elk, and antelope populations was conducted for biologist in Glennwood Springs, August, 2006. 4. A paper comparing the population levels of swift foxes in eastern Colorado to a previous study in cooperation with CDOW was accepted for publication in Southwestern Naturalist: Martin, D. J., G. C. White, and F. M. Pusateri. 2007. Monitoring swift fox populations in eastern Colorado. Southwestern Naturalist. In Press. 5. A paper on estimation of abundance and demography using age-at-harvest and mark-recovery data was submitted to Environmental and Ecological Statistics: Conn, P. B., G. C. White, and J. L. Laake. 2007. Estimating abundance and demography using age-at-harvest and mark-recovery data: a Bayesian approach. Environmental and Ecological Statistics. Submitted. 97 �6. A paper on Bayesian methods to analyze age-at-harvest data was submitted to Biometrics: Conn, P. B., J. L. Laake, D. R. Diefenbach, G. C. White, and M. A. Ternent. 2007. Bayesian analysis of wildlife age-at-harvest data. Biometrics. Submitted. 7. A paper on the use of vaginal implant transmitters in cooperation with CDOW was submitted and accepted for publication in the Journal of Wildlife Management: Bishop, C. J., D. J. Freddy, G. C. White, B. E. Watkins, T. R. Stephenson, and L. L. Wolfe. 2007. Using vaginal implant transmitters to aid in capture of neonates from marked mule deer. Journal of Wildlife Management 71:945–954. 8. A paper resulting from collaboration with Montana colleagues resulted in a publication in the Journal of Wildlife Management: Pac, D. F., and G. C. White. 2007. Survival and cause-specific mortality of male mule deer under different hunting regulations in the Bridger Mountains, Montana. Journal of Wildlife Management 71:816–827. 9. A paper resulting from Sherri Huwer’s M.S. project in cooperation with CDOW was accepted for publication in the Journal of Wildlife Management: Huwer, S. L., D. R. Anderson, T. E. Remington, and G. C. White. 2007. Evaluating the importance of forbs to sage-grouse using human-imprinted chicks. Journal of Wildlife Management. In Press. 10. A paper on mountain sheep populations in Rocky Mountain National Park was published in the Journal of Wildlife Management: McClintock, B. T., and G. C. White. 2007. Bighorn sheep abundance following a suspected pneumonia epidemic in Rocky Mountain National Park. Journal of Wildlife Management 71:183–189. 11. A paper on extending the mark-resight estimator using a beta-binomial distribution was published in the Journal of Agricultural, Biological, and Ecological Statistics: McClintock, B. T., G. C. White, and K. P. Burnham. 2006. A robust design mark-resight abundance estimator allowing heterogeneity in resighting probabilities. Journal of Agricultural, Biological, and Ecological Statistics 11:231–248. 12. A paper resulting from the May, 2005 Elk and Deer Workshop was published in the Wildlife Society Bulletin: Mason, J. R., L. H. Carpenter, M. Cox, J. C. Devos, J. Fairchild, D.J. Freddy, J. R. Heffelfinger, R. H. Kahn, S. M. McCorquodale, D. F. Pac, D. Summers, G. C. White, and B. K. Williams. 2006. A case for standardized ungulate surveys and data management in the western United States. Wildlife Society Bulletin 34:1238–1242. 13. A paper describing the use of closed captures models to estimate population size with Program MARK remains in press in Environmental and Ecological Statistics: White, G. C. 2007. Closed population estimation models and their extensions in program MARK. Environmental and Ecological Statistics. In Press. 14. A paper on the application of multistate models in Program MARK was published in the Journal of Wildlife Management: White, G. C., W. L. Kendall, and R. J. Barker. 2006. Multistate survival models and their extensions in program MARK. Journal of Wildlife Management 70:1521–1529. 15. A paper on the estimation of female grizzly bears was published in the Journal of Agricultural, Biological, and Ecological Statistics: Cherry, S., G. C. White, K. A. Keating, M. A. Haroldson, C. C. Schwartz. 2007. Evaluating estimators of the numbers of females with cubs-of-the-year in the Yellowstone grizzly bear population. Journal of Agricultural, Biological, and Ecological Statistics 12:195–215. 98 �16. A paper on estimation of nest survival was published in Studies in Avian Biology: Heisey, D. M., T. L. Shaffer, and G. C. White. 2007. The ABCs of nest survival: theory and application from a biostatistical perspective. Studies in Avian Biology 34:13–33. 17. A paper on extending the mark-resight population estimation method was submitted to Environmental and Ecological Statistics: McClintock, B.T., G. C. White, K. P. Burnham, and M. A. Pryde. 2007. A robust design mixed effects mark-resight model for estimating abundance when sampling is without replacement. Environmental and Ecological Statistics. Submitted. 18. A paper on estimation of random effects with Bayesian methods was submitted to Environmental and Ecological Statistics: White, G. C., K. P. Burnham, and R. J. Barker. 2007. Evaluation of some Bayesian MCMC random effects inference methodology applicable to bird ringing data. Environmental and Ecological Statistics. Submitted. 19. A paper on analysis of small count data was submitted to Condor: McDonald, T. L., and G. C. White. 2007. A comparison of regression models for small counts. Condor. Submitted. 20. A paper on the impact of previous capture on sampling probabilities with DNA hair-snag grids for grizzly bear populations was submitted to the Journal of Wildlife Management: Boulanger, J., and G. C. White. 2007. Influence of past live captures on detection probabilities of grizzly bears using DNA hair snagging methods. Journal of Wildlife Management. Submitted. 21. A paper on detecting trends in the Yellowstone grizzly bear population was submitted to Ursus: Harris, R. L., G. C. White, C. C. Schwartz, and M. A. Haroldson. 2007. Population growth of Yellowstone grizzly bears: uncertainty and future monitoring. Ursus. Submitted. 22. A graduate research project (Ph. D.) in cooperation with CDOW to develop statistical models to monitor puma and black bear populations in Colorado based on checks of harvested animals and DNA and/or radio-tracking data was completed. The graduate student is Paul B. Conn. The dissertation is: Conn, P. B. 2007. Bayesian Analysis of Age-at-Harvest Data with Focus on Wildlife Monitoring Programs. Ph. D. Dissertation, Colorado State University, Fort Collins. 184 pp. 23. A research study to examine the impact of nutrition on the decline of mule deer fecundity during the last 20 years was continued in cooperation with Chad Bishop and CDOW. Portions of this work will serve as his doctoral dissertation in addition to his full-time duties as a researcher with CDOW. 24. A graduate research project (M. S.) was continued in cooperation with CDOW to evaluate line transect methodology for estimating pronghorn populations in eastern Colorado. The graduate student is Aaron Linstrom, and the project is in addition to his full-time duties as a biologist with CDOW. 25. A graduate research project (M. S.) in cooperation with CDOW to evaluate methods of redistributing elk in and around Great Sand Dunes National Park was started and then discontinued. The student, Greg Davidson, switched his work to evaluate habitat use by elk on the Grand Mesa. A report on the San Luis Valley elk work is nearly completed. 26. A graduate research project (Ph.D.) in cooperation with CDOW to evaluate snowshoe hare densities relative to lodge pole pine and mixed conifer habitats was continued. The graduate student is Jake Ivan. 27. Development of the design of a monitoring system for white-tailed prairie dogs in western Colorado and eastern Utah was continued in cooperation with CDOW with P. Schnurr, K. Navo and B. Andelt. 99 �A final draft of a manuscript on the use of occupancy monitoring for white-tailed and Gunnison prairie dogs was given to B. Andelt for submission to the Journal of Wildlife Management on 20 February, 2007. 28. Preliminary analysis of monitoring data on black-tailed prairie dogs in eastern Colorado in cooperation with CDOW was continued. This effort is in cooperation with Francie Pusateri and Eric O’Dell of CDOW. 100 �WILDLIFE RESEARCH REPORT CONSULTING SERVICES FOR MARK-RECAPTURE ANALYSES GARY C. WHITE P. N. OBJECTIVE Provide expert biostatistical and experimental design services to the Colorado Division of Wildlife, Wildlife Programs Branch. SEGMENT OBJECTIVES 1. Provide biostatistical support to implement and analyze CDOW hunter harvest surveys. 2. Provide professional oversight, critiques, and analytical support to CDOW terrestrial management and avian and mammals research sections. 3. Convey to CDOW research and management sections new and pertinent information obtained in various collaborative projects conducted with other agencies and entities. RESULTS, DISCUSSION, SUMMARY See ABSTRACT for summary of key activities and publications. Prepared by: ___________________________________ Dr. Gary C. White, Department of Fish, Wildlife, and Conservation Biology Colorado State University 101 � https://cpw.cvlcollections.org/files/original/9ac467a63f9bd69b3c8ddb51c0416381.pdf b9da7ec4817212b47fd148dabe2fefee PDF Text Text Colorado Division of Wildlife July 2005 – June 2006 WILDLIFE RESEARCH REPORT State of Cost Center Work Package Colorado 3430 3001 Federal Aid Project W-185-R : Division of Wildlife : Mammals Research : Deer Conservation Program Final Report Deer Conservation Research For 5-Year Federal Aid Grant W-185-R July 2001 – June 2006 : Period Covered: July 1, 2001 – June 30, 2006 Author: David J. Freddy, Mammals Research Leader, 1 June 2006 Principal Investigators: D. L. Baker, C. J. Bishop, E. J. Bergman, D. J. Freddy, and T. M. Pojar, Colorado Division of Wildlife; W. F. Andelt, N. T. Hobbs, and G. C. White, Colorado State University All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT This report highlights the accomplishments of mule deer research and associated activities conducted by the Colorado Division of Wildlife (CDOW) with the funding support of Federal Aid Grant W-185-R during the 5-year grant segment, July 2001-June 2006. Five major multi-year research projects addressing mule deer population limiting factors, habitat status, and habitat enhancements were designed, implemented, completed, and reported upon during this segment in response to addressing stakeholder interests that influenced the direction of mule deer management and research beginning in the late 1990s. Additionally, funding provided critical scientific and technical expertise quality control oversight for statewide deer hunter harvest surveys, statewide deer population databases, mule deer survival and population estimate management surveys, mule deer population modeling, and mule deer research projects. Funding also partially supported research projects addressing chronic wasting disease and fertility control in mule deer. Research experiments provided strong evidence that habitat nutritional quality had a greater impact on net productivity of mule deer than did existing levels of coyote, cougar, and black bear predation and therefore, future research and management efforts should focus on improving the nutritional capabilities of senescent pinyon-juniper winter ranges for deer. Research also provided strong evidence that the timing and rate of breeding were within normal ranges for mule deer and therefore concerns about the breeding cycle could be dismissed as a major contributor to declining performance of mule deer populations. Comparative assessments of vegetation inside and outside sagebrush and mountain brush exclosures indicated that after 40 to 50 years of protection from ungulate herbivory, woody species increased in cover dominance with only minor changes in herbaceous cover. Increasing plant species diversity in these types of winter ranges will probably not be accomplished by excluding herbivory. In a highly scrutinized public experiment, research and management expertise codemonstrated that methods used by Colorado to estimate mule deer population size and to develop 47 �population management models provided reliable information to adequately guide mule deer harvest management decisions. From activities supported by this Grant during this segment, principal investigators published 15 peer-reviewed scientific articles pertaining to mule deer for prominent wildlife research journals with an additional 4 manuscripts currently in review with journals, provided 18 annual CDOW Wildlife Research Reports summarizing yearly progress of projects, and provided 13 presentations at prominent professional workshops or symposia. The cumulative impact of this programmatic effort provides Colorado the basis to progress and proactively sustain the mule deer resource in an increasingly demanding and complex landscape, social, and political environment. The relative success of mule deer management in Colorado reflects the positive synergy between the terrestrial research and management sections in sharing expertise, financial resources, manpower, and common goals. 48 �WILDLIFE RESEARCH REPORT PROGRAM FINAL REPORTDEER CONSERVATION RESEARCH FOR 5-YEAR FEDERAL AID GRANT W-185-R JULY 2001 – JUNE 2006 DAVID J. FREDDY Mammals Research Leader PROGRAM NEED During the late 1990s, the Colorado Division of Wildlife (CDOW) was challenged by sportsmen and other stakeholders to investigate potential causes of declining numbers of mule deer in Colorado. Additionally, sportsmen were critical of methods used to estimate numbers of mule deer and subsequently did not trust the CDOWs assessment of the overall status of mule deer in Colorado. The concerns of stakeholders gained the attention of the Colorado Legislature which directed CDOW to prepare a document to address causes of the mule deer decline and outline a plan of action to reverse the perceived trend in mule deer populations. That document was prepared for the legislature in 1999 (Gill et al. 2001) and established the direction and objectives for mule deer management and research beginning in 1999. Research objectives and program implementation were outlined and initiated in 1999 with most of the research effort directed at the Uncompahgre Plateau mule deer population which was of high concern to various stakeholders. This Federal Aid Grant Final Report highlights the accomplishments of the research pertaining to the mule deer program that was conducted from July 1, 2001 through June 30, 2006 and wholly or partially supported by Federal Aid Grant funds. PROGRAM NARRATIVE OBJECTIVES The primary Program Narrative research objectives were: I. Identify factors limiting the growth of mule deer populations. II. Assess methods to reduce impacts of limiting factors. III. Improve and evaluate statewide systems and technical methods used to determine status of mule deer populations. IV. Assess the impacts of chronic wasting disease on mule deer populations. V. Develop alternative approaches to control over-abundant urban-exurban mule deer populations. RESULTS Objective I. Factors Limiting Growth of Mule Deer Populations. Initially, stakeholders expressed concern that statewide declines in mule deer populations were caused by low pregnancy rates in adult females due to inadequate numbers of mature bucks to breed females, and by low recruitment of neonatal fawns due to excessive predation on neonates. Two primary projects were funded to focus on: 1) estimating pregnancy and fetal rates in adult female mule deer; and, 2) estimating survival rates of neonate fawn mule deer. 49 �Result Highlights: • Pregnancy and fetal rates were determined with ultrasonography and PSPB blood values to be within normal limits for the Poudre River and Uncompahgre Plateau mule deer in 1998 and 1999. Therefore, numbers of mature mule deer bucks were adequate to assure acceptable rates and timing of breeding for adult female deer. • Survival of radio-collared neonatal fawns from birth in June to December averaged 0.50 during 3 years from 1999 through 2001. This rate of neonate survival was only marginally adequate to assure population growth. Primary cause of death in neonates was sick/starve implicating inadequate nutrition for either adult does or neonates. Predation on neonates by canids, ursids, and felids occurred but not at rates considered to be limiting the population. Coyotes were the primary predator accounting for about 13% of the neonate deaths. Resulting Peer-Reviewed Publications: ANDELT, W.F., T.M. POJAR, AND L.W. JOHNSON. 2004. Long-term trends in mule deer pregnancy and fetal rates in Colorado. Journal of Wildlife Management 68:542-549. POJAR, T.M., AND D.C. BOWDEN. 2004. Neonatal mule deer fawn survival in west-central Colorado. Journal of Wildlife Management 68:550-560. Associated Annual Wildlife Research Progress Reports Available from the Colorado Division of Wildlife Research Library, Fort Collins, Colorado: POJAR, T.M., AND D.C. BOWDEN. 2002. Mule deer life-cycle-neonatal fawn survival. Colorado Division of Wildlife, Wildlife Research Report July: 47-63. POJAR, T.M. 2003. Mule deer life-cycle-neonatal fawn survival. Colorado Division of Wildlife, Wildlife Research Report July: 55. Objective II. Assess Methods to Reduce Impacts of Limiting Factors. A widespread debate throughout the western states in the late 1990s was whether mule deer were declining primarily due to predation from perceived abundant coyote, cougar, and black bear populations or if the decline was due to long-term losses in habitat quality and availability which negatively affected mule deer nutrition and subsequent recruitment and survival. Both predation and habitat quality were judged by various stakeholders to be the ‘cause’ of declining mule deer in Colorado and specifically the Uncompahgre deer population. Although painting the picture that the mule deer decline was caused by one major factor versus another major factor oversimplified the situation, such a dichotomy of thought quickly helped focus thrusts for potential research and management actions. The case for predation was assessed by Ballard et al. (2001) in their influential overview of predation and deer populations. The potential effects of habitat deterioration resulting from successional senescence of important plant communities and direct losses of habitat space due to human encroachment was argued by deVos, Jr. et al. (2003) in their overview of mule deer conservation strategies. As this debate evolved, Colorado was fortunate to have developed a strong working relationship with the Idaho Department of Fish and Game in our mutual attempts to address causes of the mule deer decline. The research sections of these 2 agencies decided to cooperatively investigate whether predation or habitat was the cause of the mule deer decline. Idaho, because of political, social, and legal aspects, was more capable of addressing the impacts of predation on mule deer than was Colorado and therefore, Idaho designed and implemented an intensive experimental reduction of coyote and cougar populations to measure the impacts of such actions on mule deer net recruitment (Hurley et al. 2002, Hurley et al. 2005). To compliment Idaho’s efforts, Colorado designed and implemented a series of experiments to measure the impacts of improving the nutritional quality of habitats on mule deer net recruitment (Bishop and White 2000). 50 �Three primary projects were funded to focus on: Phase 1A) effect of enhanced nutrition on mule deer population parameters; Phase 1B) long-term effects of herbivory on sagebrush and mountain brush winter ranges; and Phase 2A) effects of landscape habitat alterations within senescent old-growth pinyonjuniper winter ranges to enhance mule deer population parameters. Result Highlights: Phase 1A • Survival of fawns receiving an enhanced nutrition treatment from December through April had an over-winter survival rate of 0.89 which was higher (P < 0.001) than the survival rate of 0.65 for control fawns not receiving enhanced nutrition. The over-winter survival period was 15 December to 15 June during 3 years, 2001-02, 2002-03, 2003-04, and survival rates were based on 240 6-month old fawns with 120 fawns captured and radio-collared in each of the control and treatment areas. The effect of enhanced nutrition was highly evident even with the presence of ongoing predation by coyotes and cougars. • Survival rates of fetuses to neonate and through 1-year of age that were born to adult females receiving enhanced winter nutrition were 0.46 and higher (P < 0.001) than survival rate of 0.28 for fetuses born to adult females not receiving enhanced nutrition. Survival rates were over 3years, 2002-2004, and based on 276 fawns monitored across 293 adult females that were radiocollared of which, 154 adults received vaginal implant transmitters to aid in capture and monitoring neonate survival. Ultrasonography was used to determine in-utero fetal rates. • Body condition on about 1 March, as estimated from percent body fat and depth of longissimus dorsi muscle via ultrasonographpy, was better (P < 0.001) in adult females receiving enhanced winter nutrition (n = 78) than for control adult females not receiving enhanced nutrition (n= 76). Serum thyroid hormone levels were also higher in adult females receiving enhanced nutrition compared to control adult females not receiving better nutrition. Pregnancy and fetal rates were similar (0.95 and 1.80 fetuses per female) for adult females receiving and not receiving enhanced nutrition. • The finite rate of population increase, λ, was 1.20 for those deer receiving enhanced nutrition. For those deer not receiving enhanced nutrition, the finite rate of increase was 1.04 indicating a stable or slightly increasing population. The nutrition enhancement therefore, had a dramatic effect on deer population performance, indicating habitat quality was ultimately limiting the population that was also subject to natural levels of predation. In comparison, intensive control of coyote and cougar populations in Idaho had marginal positive impacts on survival rates of neonate fawns, 6-month old fawns, and adult mule deer and ultimately, net population growth (Hurley et al. 2005). • These enhanced nutrition experimental results provided a foundation for focusing deer management efforts on improving habitat quality in Colorado’s pinyon-juniper mule deer winter ranges rather than trying to intensively control or reduce coyote and/or cougar populations. • Phase 1B Excluding herbivory from semi-arid sagebrush and mountain brush plant communities resulted in increased dominance by shrub species and only minor changes in herbaceous species in nongrazed compared to adjacent grazed areas. Comparisons were based on measurements made in summer 2000 at 17 permanently fenced exclosures in western Colorado where ungulate herbivory was excluded for 40 to 50 years. Improving herbaceous and overall species diversity within established shrub dominated habitats will not likely occur by excluding grazing. 51 �• • Phase 2A Evaluating the effects of landscape alterations within senescent old-growth pinyon-juniper winter ranges on mule deer population performance parameters was initiated in 2004-05 as a pilot study and precursor to full-scale study implementation. Over-winter fawn survival was estimated on 2 critical pinyon-juniper winter range habitat treatment evaluation areas on the Uncompahgre Plateau in 2004-05. Both areas were found to be logistically adequate for future work and fawn survival was 0.84 to 0.96 on both sites (n = 25 radio-collared fawns per site). A project study plan for evaluating landscape habitat treatments was completed in 2005. Fullscale 4-year implementation began during winter 2005-06 as over-winter fawn survival, adult female body condition, and mule deer density were estimated among 8 habitat treatment evaluation areas (each 10-20 km2 in size) on the Uncompahgre Plateau. Pinyon-juniper habitat areas were categorized as controls (non-treated and senescent), pre-treatmeant (treated to reduce density of pinyon-juniper during last 10-15 years), and treatment (receiving additional habitat enhancements during this study). Initial survival rate estimates ranging from 0.76 to 0.88 suggest over-winter fawn survival may vary among habitat treatment levels. Estimates of deer density reaffirmed that deer densities in the northern study areas were lower (4-8 deer/km2) than densities in the southern study areas (19-57 deer/km2). Continued estimation of deer performance parameters over the next 3 years should allow detecting whether altering senescent pinyonjuniper habitats improves mule deer net productivity. Resulting Peer-Reviewed Publications: BISHOP, C.J., G.C. WHITE, D.J. FREDDY, AND B.E. WATKINS. 2003. How habitat quality affects hunting. Mule Deer 8:18-20. BISHOP, C.J., D.J. FREDDY, G.C. WHITE, B.E. WATKINS, T.R. STEPHENSON, AND L.L. WOLFE. 2006 In Review. Using vaginal implant transmitters to aid in capture of neonates from marked mule deer. Journal of Wildlife Management. MANIER, D.J., AND N.T. HOBBS. 2006. Large herbivores influence the composition and diversity of shrub-steppe communities in the Rock Mountains, USA. Oecologia 146:641-651. SCHULTHEISS, P.C., H. VAN CAMPEN, C.J. BISHOP, L.L. WOLFE, AND B. PODELL. 2006 In Review. Malignant catarrhal fever associated with ovine herpesvirus-2 in free-ranging mule deer (Odocoileus hemionus) in Colorado. Journal of Wildlife Disease. Associated Annual Wildlife Research Progress Reports Available from the Colorado Division of Wildlife Research Library, Fort Collins, Colorado: BERGMAN, E.J., C.J. BISHOP, D.J. FREDDY, AND G.C. WHITE. 2005. Pilot evaluation of winter range habitat treatments on over-winter mule deer fawn survival. Colorado Division of Wildlife, Wildlife Research Report July: 24-35. BERGMAN, E.J., C.J. BISHOP, D.J. FREDDY, AND G.C. WHITE. 2006 In Press. Evaluation of winter range habitat treatments on over-winter mule deer fawn survival. Colorado Division of Wildlife, Wildlife Research Report July: Available September 2006. BISHOP, C.J. AND G.C. WHITE. 2002. Effect of nutrition and habitat enhancements on mule deer recruitment and survival rates. Colorado Division of Wildlife, Wildlife Research Report July: 65-79. BISHOP, C.J., D.J. FREDDY, AND G.C. WHITE. 2002. Pilot study: use of ultrasound and vaginal implants. Colorado Division of Wildlife, Wildlife Research Report July: 81-92. BISHOP, C.J. G.C. WHITE, D.J. FREDDY, AND B.E. WATKINS. 2003. Effect of nutrition and habitat enhancements on mule deer recruitment and survival rates. Colorado Division of Wildlife, Wildlife Research Report July: 33-54. 52 �BISHOP, C.J. G.C. WHITE, D.J. FREDDY, AND B.E. WATKINS. 2004. Effect of nutrition and habitat enhancements on mule deer recruitment and survival rates. Colorado Division of Wildlife, Wildlife Research Report July: 21-43. BISHOP, C.J. G.C. WHITE, D.J. FREDDY, AND B.E. WATKINS. 2004. Effect of nutrition and habitat enhancements on mule deer recruitment and survival rates. Colorado Division of Wildlife, Wildlife Research Report July: 21-43. BISHOP, C.J. G.C. WHITE, D.J. FREDDY, AND B.E. WATKINS. 2005. Effect of nutrition and habitat enhancements on mule deer recruitment and survival rates. Colorado Division of Wildlife, Wildlife Research Report July: 37-65. BISHOP, C.J. G.C. WHITE, D.J. FREDDY, AND B.E. WATKINS. 2006 In Press. Effect of nutrition and habitat enhancements on mule deer recruitment and survival rates. Colorado Division of Wildlife, Wildlife Research Report July: Available September 2006. Associated Presentations at Professional Workshops/Symposia: BISHOP, C.J., G.C. WHITE, D.J. FREDDY, AND B.E. WATKINS. 2001. Effects of nutrition and habitat enhancements on mule deer fawn recruitment: preliminary results. Fourth Western States and Provinces Deer and Elk Workshop, August 1-3, Wilsonville, Oregon, USA. BISHOP, C.J., G.C. WHITE, D.J. FREDDY, AND B.E. WATKINS. 2003. Effects of enhanced winter nutrition of free-ranging mule deer on fawn recruitment and recruitment. Fifth Western States and Provinces Deer and Elk Workshop, May 21-24, Jackson, Wyoming, USA. BISHOP, C. J., G. C. WHITE, D. J. FREDDY, AND B. E. WATKINS. 2004. The effect of habitat quality on mule deer fawn survival and recruitment: interim report. Society for Range Management 57th Annual Meeting, January 24−30, Salt Lake City, Utah, USA. BISHOP, C. J., G. C. WHITE, D. J. FREDDY, AND B. E. WATKINS. 2004. Effect of enhanced nutrition of free-ranging mule deer on fawn survival and recruitment rates. The Wildlife Society 11th Annual Conference, September 18−22, Calgary, Alberta, Canada. BISHOP, C. J., G. C. WHITE, D. J. FREDDY, AND B. E. WATKINS. 2005. Effect of enhanced nutrition of free-ranging mule deer on population performance. Sixth Western States and Provinces Deer and Elk Workshop, May 16−18, Reno, Nevada, USA. BISHOP, C. J., G. C. WHITE, D. J. FREDDY, AND B. E. WATKINS. 2005. Effect of enhanced nutrition on mule deer population performance in pinyon-juniper habitat. Ecology and Management of Pinyon-Juniper and Sagebrush Communities Workshop, May 16−19, Montrose, Colorado, USA. III. Improve and Evaluate Statewide Systems and Technical Methods Used to Determine Status of Mule Deer Populations. Monitoring the status of mule deer in Colorado has advanced due to the synergy of the research section developing population monitoring systems and terrestrial management section implementing those monitoring systems as appropriate on a statewide basis. Developing, implementing, and maintaining, statistically rigorous systems to estimate statewide hunter harvest of mule deer, population densities and size for selected deer populations, adult female and fawn survival rates for selected populations, and developing future research projects requires scientific input, oversight and periodic evaluations. Additionally, proper evaluation requires a rigorously maintained and updated database containing statewide mule deer population data. As part of a multi-functional quality control process, the CDOW obtains oversight of key statewide mule deer research and management activities via a contract to a qualified individual through this Federal Aid Grant. Result Highlights: 53 �• Provided annual assistance to maintaining and improving statewide deer hunter harvest survey sampling systems and harvest data acquisition. • Provided annual maintenance and oversight of the DEAMAN (Deer-Elk-Antelope-Management) database representing 20 years of statewide data acquisition and storage. Included updating data files, updating user’s manual, converting DEAMAN operating system to Windows 2000 and then Windows XP, and facilitating the conversion of DEAMAN to a server-based operating system. • Provided 1- and 3-day training workshops in 2002 and 2004 in population modeling and use of DEAMAN for up to 18 terrestrial management biologists. Provided annual support to up to 18 management biologists in their day-to-day use of DEAMAN and associated population modeling spreadsheet analyses. • Provided annual assistance to management biologists in analyzing survival rates of adult female and fawn mule deer and estimates of population density and size within 5 key deer populations used to critically assess the statewide trends mule deer. • Provided critical technical expertise and credibility to designing and implementing a public demonstration experiment to evaluate the reliability of Colorado’s methods to estimate mule deer population size and to model mule deer populations. • Provided scientific and technical expertise annually to all facets of the mule deer research program inclusive of experimental designs and approaches to addressing mule deer population estimation techniques, habitat enhancement studies, and spatial analyses of deer as related to the spread of chronic wasting disease. • Senior or co-authored multiple peer-reviewed publications regarding mule deer research and statewide management in Colorado and provided scientific comment and expertise and several professional workshops pertaining to mule deer and other ungulate research and management. Resulting Peer-Reviewed Publications: BISHOP, C.J., G.C. WHITE, D.J. FREDDY, AND B.E. WATKINS. 2005. Effect of limited antlered harvest on mule deer sex and age ratios. Wildlife Society Bulletin 33:662-668. BOWDEN, D.C., G.C. WHITE, A.B. FRANKLIN, AND J.L. GANEY. 2003. Estimating population size with correlated sampling unit estimates. Journal of Wildlife Management 67:1-10. FREDDY, D.J., G.C. WHITE, M.C. KNEELAND, R.H. KAHN, J.W. UNSWORTH, W.J. DEVERGIE, V.K. GRAHAM, J.H. ELLENBERGER, AND C.H. WAGNER. 2004. How many mule deer are there? Challenges of credibility in Colorado. Wildlife Society Bulletin 32:916-927. MASON. R., L.H. CARPENTER, M. COX, J.C. DEVOS, JR., J. FAIRCHILD, D.J. FREDDY, J.R. HEFFELFINGER, R.H. KAHN, S.M MCCORQUODALE, D.F. PAC, D. SUMMERS, G.C. WHITE, AND B.K. WILLIAMS. 2006 In Press. A case for standardized ungulate surveys and data management in the western United States. Wildlife Society Bulletin. WHITE, G.C. 2004 In Press. Correcting counts: techniques to de-index. Wildlife Research. WHITE, G.C., D.J. FREDDY, R.B. GILL, AND J.H. ELLENBERGER. 2001. Effect of adult sex ratio on mule deer and elk productivity in Colorado. Journal of Wildlife Management 65: 436444. WHITE, G.C., AND B.C. LUBOW. 2002. Fitting spreadsheet population models to multiple sources of observed data. Journal of Wildlife Management 66:300-309. 54 �Associated Annual Wildlife Research Progress Reports Available from the Colorado Division of Wildlife Research Library, Fort Collins, Colorado: FREDDY, D.J. 2002. Deer aerial survey population estimation Rangely deer data analysis unit D6, GMU 10. Colorado Division of Wildlife, Wildlife Research Report July: 117-168. WHITE, G.C. 2002. Improved population modeling-DEAMAN system administration. Colorado Division of Wildlife, Wildlife Research Report July: 93-102. WHITE, G.C. 2003. Multispecies Investigations: consulting services for mark-recapture analysis. Colorado Division of Wildlife, Wildlife Research Report July: 189-196. WHITE, G.C. 2004. Multispecies Investigations: consulting services for mark-recapture analysis. Colorado Division of Wildlife, Wildlife Research Report July: 151-161. WHITE, G.C. 2005. Multispecies Investigations: consulting services for mark-recapture analysis. Colorado Division of Wildlife, Wildlife Research Report July: 67-75. Associated Presentations at Professional Workshops/Symposia: FREDDY, D.J., G.C. WHITE, M.C. KNEELAND, V.K. GRAHAM, W.J. DEVERGIE, J.H. ELLENBERGER, J.W. UNSWORTH, C.H. WAGNER, P.M. SCHNURR, V.W. HOWARD, JR., AND T.S. BICKLE. 2001. Estimating mule deer populatin size using Colorado quadrat system corrected for Idaho mule deer sightability: a sportsmen’s issue. Fourth Western States and Provinces Deer and Elk Workshop, August 1-3, Wilsonville, Oregon, USA. FREDDY, D.J. 2005. Moderator: Session on Representative Strategies. International Association of Fish and Wildlife Agencies Ungulate Data Gathering, Analysis, and Use Workshop, 19 May. Reno, Nevada, USA. WATKINS, B.E., J.H. OLTERMAN, AND T.M. POJAR. 2001. Mule deer survival studies on the Uncompahgre Plateau, Colorado 1997-2001. Fourth Western States and Provinces Deer and Elk Workshop, August 1-3, Wilsonville, Oregon, USA. WAGNER, C.H., B.E. WATKINS, J. VAYHINGER, AND S. STEINERT. 2001. Summary of mule deer survival studies in Colorado, 1997-2001. Fourth Western States and Provinces Deer and Elk Workshop, August 1-3, Wilsonville, Oregon, USA. WHITE, G.C. 2001. Effect of adult sex ratio on mule deer and elk productivity in Colorado. Fourth Western States and Provinces Deer and Elk Workshop, August 1-3, Wilsonville, Oregon, USA. WHITE, G.C. 2005. Featured Speaker: Theoretical considerations and practical implications. International Association of Fish and Wildlife Agencies Ungulate Data Gathering, Analysis, and Use Workshop, 19 May. Reno, Nevada, USA. IV. Assess the Impacts of Chronic Wasting Disease on Mule Deer Populations. Chronic wasting disease (CWD) in mule deer has been a focal point of various research efforts within the CDOW since the early 1990s. Research on CWD was proposed to be funded within this Federal Aid 5-Year Grant. Partial funding from Federal Aid occurred during 2001 but after that year, funding for CWD was obtained from sources other than the Federal Aid Grant. As such, research potentially occurring while Federal Aid funding was in effect was limited to supporting activities associated with publications. Resulting Peer-Reviewed Publications: GROSS, J.E., AND M.W. MILLER. 2001. Chronic wasting disease in mule deer: disease dynamics and control. Journal of Wildlife Management 65:205-215. MILLER, M.W., AND E.S. WILLIAMS. 2002. Detecting PrPCWD in mule deer by immunohistochemistry of lymphoid tissues. Veterinary Record 151:610-612. 55 �WILLIAMS, E.S., AND M.W. MILLER. 2002. Chronic wasting disease in deer and elk in North America. Revue Scientifique et Technique Office International des Epizooties 21:305316. WILLIAMS, E.S., M.W. MILLER, T.J. KREEGER, R.H. KAHN, AND E.T. THORNE. 2002. Chronic wasting disease of deer and elk: a review with recommendations for management. Journal of Wildlife Management 66:551-563. WOLFE, L.L., M.M. CONNER, T.H. BAKER, V.J. DREITZ, K.P. BURNHAM, E.S. WILLIAMS, N.T. HOBBS, AND M.W. MILLER. 2002. Evaluation of antemortem sampling to estimate chronic wasting disease prevalence in free-ranging mule deer. Journal of Wildlife Management 66:564-573. Associated Annual Wildlife Research Progress Reports Available from the Colorado Division of Wildlife Research Library, Fort Collins, Colorado: Miller, M.W. 2002. Chronic wasting disease in mule deer; monitoring and management. Colorado Division of Wildlife, Wildlife Research Report July: 113-116. Associated Presentations at Professional Workshops/Symposia: Conner, M.M. 2005. Increasing the efficacy of chronic wasting disease detection via selective and targeted sampling. Sixth Western States and Provinces Deer and Elk Workshop, May 16−18, Reno, Nevada, USA. V. Develop Alternative Approaches to Control Over-abundant Urban-exurban Mule Deer Populations. An increasing problem with mule deer in Colorado and other states is localized over-abundance of deer in urban-exurban areas. Deer have successfully invaded highly developed human habitats where increasing incidences of browsing damage to lawns, ornamentals, and gardens, and vehicle-deer collisions created the need for some form of deer population control. In these urban-exurban situations, traditional hunting or even highly controlled hunting or culling may not be feasible or socially acceptable. The potential to develop and use hormonal fertility control to reduce net recruitment of deer into these localized populations was recognized by CDOW during the 1990s. Research was initiated to test available hormonal therapies using captive mule deer at the CDOW Foothills Wildlife Research Facility. A portion of this fertility control research was supported by this Federal Aid Grant. After late 2002, other sources of funding were applied to continue this research. Resulting Peer-Reviewed Publications: Baker, D.L., M.A. Wild, M.M. Conner, H.B. Ravivarapu, R.L. Dunn, and T.M. Nett. 2004. Gonadotropin-releasing hormone agonist: a new approach to reversible contraception in female deer. Journal of Wildlife Diseases 40:713-724. Associated Annual Wildlife Research Progress Reports Available from the Colorado Division of Wildlife Research Library, Fort Collins, Colorado: Baker, D.L. 2002. Evaluation of GnRH-PAP as a long-term fertility control agent in female mule deer. Colorado Division of Wildlife, Wildlife Research Report July: 103-112. 56 �SUMMARY Five major multi-year research projects addressing mule deer population limiting factors, habitat status, and habitat enhancements were designed, implemented, completed, and reported upon during this segment. Furthermore, funding partially supported research projects addressing chronic wasting disease and fertility control in mule deer. Additionally, funding provided critical scientific and technical expertise quality control oversight for statewide deer hunter harvest surveys, statewide deer population databases, mule deer survival and population estimate management surveys, mule deer population modeling, and mule deer research projects. LITERATURE CITED BALLARD, W.B., D. LUTZ, T.W. KEEGAN, L.H. CARPENTER, AND J.C. DEVOS, JR. 2001. Deer-predator relationships: a review of recent North American studies with emphasis on mule and black-tailed deer. Wildlife Society Bulletin 29:99-115. BISHOP, C.J., AND G.C. WHITE. 2000. Effects of habitat enrichment on mule deer recruitment and survival rates-a program study plan narrative. Colorado Division of Wildlife, Wildlife Research Report July: 135-180. DEVOS, JR., J.C., M.R. CONOVER, AND N.E. HEADRICK (EDITORS). 2003. Mule deer conservation: issues and management strategies. Berryman Institute Press, Utah State University, Logan, USA. GILL, R.B., T.D.I BECK, C.J. BISHOP, D.J. FREDDY, N.T. HOBBS, R.H. KAHN, M.W. MILLER, T.M. POJAR, AND G.C. WHITE. 2001. Declining mule deer populations in Colorado: reasons and responses. Colorado Division of Wildlife Special Report 77. Fort Collins, Colorado, USA. HURLEY, M.A., M. SCOTT, AND J.W. UNSWORTH. 2002. Influence of predators on mule deer populations. Federal Aid in Wildlife Restoration, Job Progress Report, Project W-160-R-28. Idaho Department of Fish and Game, Boise, USA. HURLEY, M.A., J.W. UNSWORTH, P. ZAGER, E.O. GARTON, AND D.M. MONTGOMERY. 2005. Mule deer survival and population response to experimental reduction of coyotes and mountain lions. Sixth Western States and Provinces Deer and Elk Workshop, May 16−18, Reno, Nevada, USA. Prepared by ______________________________________ David J. Freddy, Mammals Research Leader 57 � https://cpw.cvlcollections.org/files/original/73038235ae1bc33fc5e1a6e2dc9a6dbc.pdf 4555251ee4bed8f36481256f0ae47e88 PDF Text Text 47 Colorado Division of Wildlife Wildlife Research Report July 2001 and July 2002 PROGRESS REPORT State of -------=C=o=lo=r-=a=do=-------- Division of Wildlife-Mammals Research ·work Package No. ___3_0-0-1_ _ _ _ __ Deer Conservation Task _ _ _ _ _ _ _ _.o. . 1_ _ _ _ __ Mule Deer Life Cycle - Neonatal Fawn Survival Federal Aid Project_~W~-~18_5~-~R~---- Research and Development Period Covered: July 1, 2001 - June 30, 2002 Author: T. M. Pojar and D. C. Bowden Personnel: W. Andelt, R Arant, D. Baker, T. Baker, B. Banulis, T. Beck, C. Bishop, G. Bock, D. Bowden, P. Burke, T. Burke, M. Caddy, D. Coven, B. Diamond, B. Dreher, J. Ellenberger, M. Farnsworth, J. Foster, V. Graham, J. Griggs, D. Gustine, P. Hayden, B. Hoffner, B. Lamont, M. King, K. Larsen, M. Mclain, H. McNally, G. Miller, M. W. Miller, E. Myers, J. Olterman, M. Potter, J. Risher, D. Schweitzer, D. Steele, J. Skinner, T. Spraker, D. Swanson, B. Watkins, G. White, S. Znamenacek. . The following is an abstract and manuscript now in preparation for submission to the Journal of Wildlife Management describing the neonatal fawn survival study on the Uncompahgre Plateau. Because ofrequests by reviewers or editors some aspects of the presentation and analysis may be modified. Manipulation or interpretation of these data beyond that contained in this report should be labeled as such, and is discouraged. • ,-/ Abstract: Declining mule deer (Odocoi/eus hemionus) populations resulting from apparent low recruitment brought management and political focus on neonatal fawn survival. Mule deer fawns on the Uncompahgre Plateau (5,957 km 2l in west central Colorado were captured at mean age of 3 days (range from newborn to 6 days) and collared with.mortality sensing drop-off radio collars. Two hundred thirty fawns were radioed with samples of 50, 88, ahd 92 during 1999, 2000, and 2001, respectively. Designated neo~ate survival period was from capture to 14 December. Survival was different among years (X/ = 6.160, P = 0:046) with annual survival (Kaplan-Meier, 95% CL) of0.321 (0.125-0.517); 0:589 (0.474-0.703), and 0.594 (0.472-0.716) for 1999, 2000, and 2001, respectively; the 3-year mean survival was 0.501. Combined 3-year cause-specific mortality (95% CL) was sick/starve 0.171 (0.1160.226), Coyote 0.126 (0.078-0.174), bear 0.040 (0.012-0.068), feline 0.032 (0.006-0.057), trauma 0.043 (0.014-0.072), and unknown 0.047 (0.016-0.077). Neither all predation combined (coyote, bear, and feline) (P = 0.379) nor coyote predation alone (P > 0.989) differed among years. Mortality in the sick/starve category is the only source that approached significance among years (P = 0.070). The major difference was in. 1999 with 0.318 mortality due to sick/starve compared to 0.115 and 0.148 in 2000 and 2001, respecti\'ely. Historic December (1990-99) fawns per 100 does ratios (f:d) were significantly . co.rrelated with the preceding June precipitation (P = 0.004) but not with June temperature (P = 0.441). June precipitation for 1999 was 3 .66 cm and was 1.04 and 0.86 cm in 2000 and 2001, respectively, which BDOW016776 �48 may have contributed directly or indirectly to the differences in sick/starve mortality. Three-fourths of mortality from predation (75.0%) and sick/starve (73.7%) had taken place by 31 July with 76.3% of mortality from all sources occurring by 31 July. Mean fawn weights at capture were 4.35 kg, 4.50 kg, and 4.13 kg for 1999, 2000, 2001, respectively and were different among years (P = 0.044). There was also a difference in hind foot length among years (P = 0.002) with mean length of26.14 cm, 26.62 cm, and 25.63 cm for 1999, 2000, and 2001, respectively. Weight and hind foot means were different between 2000 and 2001 (P > 0.017) with 1999 not different from either 2000 or 2001 (P < 0.017) using mean separation procedure controlled with Bonferonni significance level. Mean capture date was 19 June ( 4.83 days SD) and median capture date was 19 June (range 9 June to 6 July) with 94.78% of all captures occurring between 13-30 June. This implies that most does were bred during their first estrous cycle. Neonatal survival through 14 December does not completely account for observed low f:d ratios. Fetus mortality during late pregnancy or mortality of fawns at birth (before they could be detected for capture) is implicated as a potential cause of poor recruitment. [• ! L L I l I I L . / I ' �49 NEONATAL MULE DEER FAWN SURVIVAL AND CAUSE SPECIFIC MORTALITY T. M. Pojar and D. C. Bowden There is evidence that the mule deer populations in Colorado, as well as other western states, have declined during recent years due mostly to low fawn survival and subsequent low population recruitment (Unsworth et al. 1999). December f:d ratios on the Uncompahgre Plateau (5,957 km2l in west central Colorado have declined (t = -3.41, P = 0.004) by an average of 1.8 fawns:100 does per year from 1982-1998 where December ratios ranged from a high of 79f: 100d in 1982 to a low of 32f: 100d in 1997 (White et al. 2001). Declining deer densities and long-term decline in f:d ratios have resulted is much debate and concern among managers, sportsmen, administrators, and politicians. Gill et al. (2001) offers several potential causes for long-term decline in deer density: 1) habitat deterioration, 2) competition, 3) disease, 4) predation, and 5) hunting. Specifically, low December ratios could be the result of: 1) low pregnancy rate, 2) reduced fetal production, 3) prolonged breeding (fawning) season due to low buck to doe ratios, 4) late term mortality of fetuses and weak or stillborn fawns, or 5) low neonatal fawn survival through December. A popular perception is that predation especially by coyotes (Canis /atrans), but also by black bear (Ursus americanus) and felines (Fe/is conco/or and F. rufus), is a major contributing factor in apparent low neonatal fawn survival. Pregnancy and fetal rates have been relatively constant for this species at population densities encountered during recent decades. During the 1960s and 1970s when deer populations were thriving in Colorado pregnancy rates in the following habitats were: 1) front-range foothills 92.0% (n=163) (Medin and Anderson 1979), 2) high mountain park 94.0% (n=134) (Gill 1971), and western slope pinyonjuniper 89.0% (n = 47) during 1973 and 82.0% (n = 83) during 1978 (Bartmann 1998). During 1963-71, a penned herd that was fed alfalfa hay ad libitum and a 16% protein supplement ranging from 0.23 to 0.90 kg per deer-day averaged 94.7% pregnancy (n=135) (Robinette et al. 1973). Fetal rates, as the other major component of reproductive rates, for adult does were 1.83 (n = 41) during 1961-1965 in frontrange foothills habitat (Medin and Anderson 1979) and 1.82 (n = 114) in high mountain park habitat during 1969-1971 (Gill 1971) in Colorado. To determine if both pregnancy and fetal rates of adult does have not changed since the 1960s and 1970s and to have specific information from the Uncompahgre Plateau, a separate preliminary study examined these factors using transrectal ultrasound. A sample of 40 does was examined in February 1999. Pregnancy rate (93%) did not differ from the historic rate of 94%, (n = 328, X/ = 0.07, P = 0.791) and the fetal rate (1.70) was not less than the historic rate of 1.87 (n = 307, Z = 0.412, P = 0.681) (Andelt, Pojar, and Johnson, unpublished data). Density dependent effects of population growth follows the progression ofreducedjuvenile survival, increased age at first reproduction followed by a decline in reproductive rates of mature females late in the progression to carrying capacity and, finally, reduced adult survival (Eberhardt 1977). Juvenile survival is highly variable and sensitive to population density and stochastic environmental factors whereas, adult female survival is robust against most limiting factors (Gaillard et al. 1998, White and Bartmann 1998). Therefore, the most obvious area of investigation was to examine the survival and mortality sources of neonate fawns. The primary objectives of this study were to estimate neonatal fawn survival from birth (time of capture) to 14 December and cause-specific mortality to determine the contribution of summer fawn mortality to December f:d ratios on the Uncompahgre Plateau. Timing of births (captures), as an index of fawning season compression was a secondary objective; this relates to cohort exposure to predation (predator swamping). Extent and timing of the various mortality sources are described. �50 STUDY AREA The Uncompahgre Plateau was formed by a structural up-lift; it runs generally southeast to northwest with a crest, a_ break in the up-lift, roughly bisecting it forming drainages to the northeast and southwest. The small communities of Gateway and Ridgeway are on the northwest and southeast perimeter, respectively, with the larger communities of Montrose and Delta along the east boundary (Figure 1). The terrain slopes generally to the northwest along the crest with the highest point being Horsefly Peak at 3,147 m and the lowest, 1,389 m, near Gateway. Along the crest, the terrain slopes gently to the northeast where many tributaries to the Gunnison and Uncompahgre rivers have worn deep canyons; the Unaweep Canyon along the northwest boundary of the area averages 914 m deep (Young and Young 1984). Southwest of the crest terrain drops abruptly because the plateau cap lifted and tilted to the northeast giving a gentle slope to the northeast and a steep drop-off to the southwest (Marshall 1998) with drainages flowing to the San Miguel and Dolores rivers. The latitude is from 38° 1.23' to 38° 59_05' and the longitude ranges from 107° 45.52' to 108° 58.80'. Normal annual precipitation (1971-2000 mean) is 29.11 cm with most precipitation in late summer and fall (July-October); the 30-year mean low temperature (-2.78° C) was in January and the high (22.86° C) was in July. These weather statistics were taken from 5 stations on the perimeter of the plateau and would best represent winter range; summer range is at higher elevations and would have lower temperatures and higher precipitation. The study area includes Game Management Units 61 and 62 (5,957 km 2) and is designated as Data Analysis Unit D-19. Vegetation types are agricultural along the major river drainages and as elevation increases the following types are encountered: saltbrush-greaswood (Atriplex canescens and Sarcobatus vermiculatus), mature pinyon-juniper (Pinus edulis and Juniperus osteosperma) forest interspersed with big sagebrush (Artimisia tridentata) parks, Gambel oak (Quercus gambelii), ponderosa pine (Pinus ponderosa), and spruce-fir (Picea-Abies). The oak, pine, and spruce-fir communities are interspersed with aspen (Populus tremuloides) with some areas of pure aspen stands bordered by areas of mountain shrub and high mountain grass-forb meadows. There are vast stands of aspens on the rugged southwest slopes of the Uncompahgre. The oak, pine, spruce-fir, and aspen types are the major summer range and fawning habitat for mule deer and are generally above 2,438 m. Winter range can be from the oakmountain shrub community (depending on winter severity) decreasing in elevation to the agricultural lands along major drainages (Figure 1). Livestock grazing began in 1881 immediately after the Ute Tribe was expelled from the Uncompahgre Plateau (Anderson et al. 1992). Large herds of cattle were brought in from Texas, Kansas, and Mexico and within 20 years the town of Placerville (on the south end of the Uncompahgre Plateau) became " ... the largest cattle-shipping point in the world" (Marshall 1998:59). Severe overuse by livestock continued for 70 plus years, at least until 1951, accompanied by extreme overpopulation of deer beginning in the 1940s until liberal harvests during the 1950s and 1960s reduced the deer population (Kufeld 1979). Range condition around 1900 was described as being eaten " ... down to the nub" (Marshall 1998:59). Still, by about 1944 the range was described as appearing as having " ... had a band of sheep "caked" on it" (anonymous, c.a. 1944:8). Sharp stocking rate declines began in 1948 and " ... have remained relatively constant near their lowest rates from 1951 to the present" (Kufeld 1979: 13 ). The overstocking of livestock from 1881 to 1948 and deer from the 1940s to the 1960s has been alleviated; however, reduction in grazing pressure does not necessarily result in range condition improvement because the range may stabilize at a lower successional state (Laycock 1991 ). Timber harvest has been a factor in the past and continues to the present. Roads and rural subdevelopments are expanding especially in the southern area of the plateau. Summer and winter recreation, including sightseeing, camping, biking, hunting, all terrain vehicle, and snowmobiling and cross-country skiing are common human activities (Uncompahgre Plateau Partners 2002). This study was approved by Colorado State University and Colorado Division of Wildlife animal care and use committees under Protocol Number 99-063A-01. �51 METHODS The fawning area generally included the entire summer range of the plateau above 2,438 m. Mature does (n = 74) that had been radioed during the ultrasound reproductive study and during a separate survival study were tracked to more precisely identify fawning habitat during the first year of capture efforts. Fawns ofradioed does were not necessarily targeted for capture. Searches for parturient does were from the ground, either on foot or by vehicle. Behavior and physical features were clues to identifying parturient does. Does that were alone, had udder development, and sunken flanks were prime candidates for initiating an intense search for bedded fawns within a 50± m radius of where the doe was first spotted. Searches were terminated if young were not found within about a half-hour. Bedded fawns were approached from the rear, avoiding eye contact, quickly approaching for the last couple meters. Upon placing a hand (latex gloved) on it's back, the fawn would freeze. Immediately upon capture, it was blindfolded and hobbled; processing took :s_ 5 minutes and included weighing, hind foot measurement, and attaching the radio collar. Age of the fawn was estimated by condition of the umbilical cord, pelage, hoof condition, and behavior. Before release, the fawn was examined for signs of dehydration, sickness (diarrhea or respiratory discharge), or physical deformities. Attempts were made to leave the fawn at the same site where it was captured. Radio collars (weighing< 110 g) were expandable from 22 cm to 33 cm and designed to drop off at approximately 6 months. Mortality sensor was set at 2 hours. Once radioed, each fawn was tracked to determine live status at least twice a day and sometimes as many as 6 times per day through 1 September and then once a day, excluding weekends, through 14 December. Determination oflive status was done from distances (0.5-5.0+ km) sufficient to minimize disturbance to either the fawn or doe. Mortality signals were investigated immediately upon detection. The site and evidence was surveyed and a determination of cause was assigned as best as evidence offered. If no carcass was present and evidence indicated predation or scavenging, criteria offered in the literature was used to help assign a specific predator (White 1973, Wade and Bowns 1984, Acom and Dorrance 1990, and Andelt et al. 1998). Most of these; however, deal with adult or domestic animals (except White 1973) and other geographic areas but were used anyway to help evaluate on-site evidence. Ground cover and vegetation was too thick to find actual tracks so, in addition to the above, the following general criteria were used. Coyote kill/scavenge sites typically had only shards of bones, tufts of hair/hide, usually> 1 feeding site within 30 m, and sometimes fawn parts buried in mineral soil. Bear sites were identified by one relatively large feeding site (1-5 m diameter) with the fawn hide nearby and usually intact and inverted (Schlegel 1976); bear usually defecate near their feeding site. The presence of hooves and small leg and cranial bones typify black bear kill sites (Bertram and Vivion 2002:751). Felines (Mt. Lion and bobcat) drag their prey from the kill site to a protected feeding area and cover any remains with litter and duff. Deaths due to vehicle collisions, entanglement in fences, accidents, or human caused (poaching) were categorized as trauma deaths. The unknown category included collars that were found with no carcass parts or other evidence of the fawn in the vicinity. It is possible the collar was carried from a kill site by an avian or terrestrial predator/scavenger. This category undoubtedly includes some collars that were slipped either by maternal grooming or a lucky stroke by a hind foot. Even if the collar had bite marks or blood on it, it was classified as unknown because assigning it to any of the 3 major terrestrial predators would be quite uncertain. Ballard et al. ( 1999) chose to construe this line of evidence as coyote predation because coyotes are known to prey on fawns in summer. Whole carcasses that were found were examined for external evidence of sickness such as diarrhea or respiratory discharge or external injuries. Proximate site characteristics were described as with other mortalities. Wh(?le carcasses were field classified as sick/starve but the mortality category was changed to trauma if necropsy revealed internal injury from trauma. During the first year, carcasses were frozen and transported to the necropsy laboratory within 5-7 days. Protocol for 2000 and 2001 provided for the carcass to be iced and transported to the laboratory within 2 hours of discovery. A trained pathologist did all necropsy and tissue collections. Details of laboratory procedures, tissue sample collection, diagnostic techniques and disease detection results are found in Myers (2001 ). �52 Thymus glands were not weighed during the first year of our study but the attending pathologist made a subjective judgment on the condition and size of the thymus at necropsy. During the last 2 years, thymus glands were weighed to the nearest hundredth of a gram. Data in terms of survival time were censored if the radio dropped off before 14 December. If a radio was not heard and could not be accounted for, survival time data were censored on the last day the radio was heard. All survival time data associated with radios recovered and assigned to the "unknown" category were censored. The key assumptions of a survival study are: 1) animals are sampled randomly, 2) experimental unit survival times are independent, 3) capture and carrying a radio package does not affect survival, 4) the censoring mechanism is random, and 5) emigration is zero (Pollock et al. 1989, Tsai et al. 1999). Because there is high probability of mortality soon after birth in mule deer and because the fawning season is temporally compressed (about 2 weeks), staggered entry of subjects was not used. The time origin for the nonparametric Kaplan-Meier (Kaplan and Meier 1958) survival estimator was the date when the first fawn was radioed. Staggered exits from mortalities and censoring were incorporated in the calculations of survival and confidence limits following Pollock et al. ( 1989). Large-sample Chi-square tests were used to compare yearly Kaplan-Meier estimates of survival rates. Fisher's Exact Test was used for tests of association and Log Rank statistic was used to compare mortality distributions among years (Cantor 1997). ANOV A and pair-wise mean comparisons were made following the Bonferroni inequality to compare fawn weight and hind foot measurements. The 0.05 significance level was used for all tests. RESULTS During 3 fawning seasons 230 fawns were radio collared with 50, 88, 92 captured during 1999, 2000, and 2001, respectively (Table 1). Mean capture date was 19 June (4.83 days SD) and median capture date was 19 June (range 9 June to 6 July) with 94.78% of all captures occurring between 13-30 June. Mean fawn weights at capture were 4.35 kg, 4.50 kg, and 4.13 kg for 1999, 2000, 2001, respectively and were different among years (P = 0.044). There was also a difference in hind foot length among years (P = 0.002) with mean length of 26.14 cm, 26.62 cm, and 25.63 cm for 1999, 2000, and 2001, respectively. Both weight and hind foot means were different between 2000 and 2001 (P < 0.017) with 1999 not different from either 2000 or 2001 (P > 0.017). During the first year, the attending pathologist diagnosed 8 of 15 (53%) fawns in the sick/starve mortality category with "severe thymic atrophy". Mean thymus weight was 2.62 g (SD 2.90, n = 10) and 1.96 g (SD 2.40, n = 10) for 2000 and 2001, respectively. These means were not different (P = 0.584) and were combined for a 2-year mean of 2.29 g (SD 2.62, n = 20). Mean June precipitation was 3.66 cm in 1999 and 1.04 and 0.86 in 2000 and 2001, respectively. June precipitation on the Uncompahgre Plateau during 1990-1999 was negatively correlated with subsequent December f:d ratios (P = 0.004) but was not correlated with June temperature (P = 0.441). Survival was different (X22 = 6.160, P = 0.046) among years with annual survival (95% CL) of 0.321 (0.125-0.517), 0.589 (0.474-0.703), and 0.594 (0.472-0.716) for 1999, 2000, and 2001, respectively. Sick/starve was the only cause-specific mortality that approached significance among years (P = 0.070) (Table 2). The major difference was in 1999 with 0.318 mortality due to sick/starve compared to 0.115 and 0.148 in 2000 and 2001, respectively. The 3-year combined mortality due to sick/starve was 0.171 (0.116-0.226). Coyote predation was 0.126 for combined 3-year data and was highly consistent (P = 0.989) among years. Likewise, bear and feline caused mortality was consistent among years (Table 2) and was 0.040 (0.012-0.068) and 0.032 (0.006-0.057), respectively. All predation combined did not differ among years (P = 0.379). Trauma, which included roads, fences, injury, etc. was 0.043 (0.014,..0.072) and unknown causes accounted for 0.047 (0.016-0.077) of the 3-year mortalities. The temporal distribution of mortalities was consistent among years (X/ = 0.680, P = 0.712) with 76.3% • of all mortalities occurring by 31 July. This is the result of the major sources of mortality, sick/starve (73.7%) and predation (75.0%), taking place by 31 July. �53 DISCUSSION Fawn capture effort was focused on the high elevation summer range generally above 2,438 m. How well the sample of fawns captured in this area represents the entire Uncompahgre Plateau population is of major importance. Mule deer tend to be seasonally migratory in the mountainous areas of Colorado (Garrott et al. 1987). However, in the Colorado eastern foothills the majority of the herd may rernain at lower elevations yearlong (Kufeld et al. 1989). In Idaho, 26% of a herd that wintered in broad agricultural valleys and low elevation rangelands stayed on the wintering area yearlong (Brown 1992). Wintering area of the Uncompahgre Plateau herd includes some low elevation valleys but raises quickly into sagebrush and pinyon-juniper habitat (Figure 1). To determine how representative the fawns captured at high elevations were of the entire population we used the elevations of winter-captured does (n = 95) during 1997-2000 and their subsequent aerial relocations during mid-May (n = 64) and late-May (n = 144). The mean elevations for capture, mid-May (May 18-21), and late-May (May 28-31) relocations were 1,927 m, 2,551 m, and 2,603 m, respectively with a highly significant difference (P,:::: 0.001); mean of winter capture locations was different (P < 0.05) from the relocations. The 95% kernel home range of all does relocated during midor late-May closely matches the 95% kernel home range of all fawn capture sites (Figure 1). This indicates this population generally fits the near-total migratory pattern described by Garrott et al. (1987). Does had about 3 more weeks to complete their migration to summer/fawning areas, which would have reduced the May relocation home range size because does do not settle on their fawning area and reduce their individual home range until about 3-5 days of parturition (Haegel et al. 1985). The aerial trapping operation did not attempt to capture does among farmsteads along the river bottoms and agricultural lands. These lands compose 6% of the total area (Figure 1). During the 5 years prior to this study, the sex ratio of the Uncompahgre Plateau deer herd averaged 12.3 bucks per 100 does (Colorado Division of Wildlife data). Later mean parturition date, a less synchronized birthing pulse, and lower pregnancy rates resulting in reduced recruitment are some of the postulated consequences of low sex ratios (Squibb 1985, White et al. 2001). In a controlled experiment with elk, the calving season was later by 17 days and extended by 30 days when breeding was done by yearlings compared to 5-year-old bulls; pregnancy rate was 89% with yearlings and 97% with 5year-olds (Noyes et al. 1996). In a free-ranging elk herd, bulls> 1 year old did 76% of the breeding with yearling bulls making an appreciable contribution to successful breeding in this herd and " ... completely compensating for the absence of older bulls" (Squibb 1985 :750). Low sex ratios could have a greater impact on deer compared to elk because deer have a tending-bond breeding system and elk form harems (Kie and Czech 2000). Examination of 20 years of sex ratio data and fawn to doe ratios provided no evidence to indicate that sex ratios observed across the state affected population productivity in Colorado mule deer (White et al. 2001 ). Our data support the contention that low sex ratios did not adversely affect herd productivity. Mean capture date was 19 June in this study, which is similar to the mean fawning date of 18 June (n=215) for a captive herd in Colorado (Robinette-et al. 1973). Ninety-five percent of our captures were within a 2-week period, between 13 and 30 June, providing evidence that most does were bred during their first estrus. Estrous cycle is 23-29 days for mule deer and 97% conceive during their first cycle (Anderson and Wallmo 1984). Survival of individual fawns is related to the interaction of nutrition, cover, and climate Picton (1979). On northern ranges, deer are frequently subjected to rigorous winters resulting in chronic malnutrition of does and subsequent stillbirths, weak fawns, and lactation failure (Verme 1969). Although the 3 winters encountered during this study were milder than normal, this population seems to fit the description of a herd that is stressed during winter. During our searches of fawning areas we discovered a total of 9 fawns that were stillborn or died within minutes/hours of birth. In addition, we found one doe that was so weakened from trying to deliver dead twin fetuses that she was captured, restrained, and the fetuses delivered; the doe apparently survived as she was not found in the vicinity the next day. Two mature does were killed or scavenged by bear; one was prime aged (3-4 years old) and the other was of unknown age. Both were found during the peak of fawning - 18 and 19 June. Bear are �54 known to prey on both fawns and adult deer (Behrend and Sage 1974, Conger and Giusti 1992, Verspoor 1983) but bear rarely prey upon healthy adult deer (Verspoor 1983 ). The best speculation is that the does were in a difficult delivery, as the above mentioned doe, or had died from delivery complieations and were scavenged. • During extended periods of damp cool weather fawns may have a higher incidence of exposurerelated complications and deaths (Ginnett and Young 2000). Cool dry summer weather in the northern regions of mule deer range enhanced fawn survival (Picton 1979). Mortality due to sick/starve during 1999 was 0.312 compared to 0.115 and 0.148 for 2000 and 2001, respectively. In an attempt to discover possible causes for the higher sick/starve mortality in 1999 compared to 2000 and 2001, we examined June precipitation and temperature and fawn weights and skeletal development as gauged by hind foot length. We speculated that fawning-season weather might be a factor or that fawn robustness, as measured by fawn weights and skeletal development, might differ among years. Since the 1999 fawn size indicators were not different from the other 2 years when sick/starve mortalities were much lower, it cannot be concluded that fawn size affected the proportion of fawns dying of sickness or starvation. Mean birth weight of fawns from a captive herd was 3 .69 kg (n= 172) (Robinette et al. 1973) and is less than means we observed. This is expected because these were nearly true birth weights and our measurements were taken at mean age of 3 days and fawns can gain 0.29 kg per day during their first 12 days (Robinette et al. 1973). Given these weight comparisons, there is no evidence to suggest that fawn weight was a factor in the difference in fawn mortality due to sick/starve among years. Thymus gland development in cervids of similar size to mule deer (fallow deer (Dama dama) and sitka deer (Cervus nippon), follows a pattern of growth during fetal development then remains relatively constant from birth to puberty (Chapman and Twigg 1990). In mule deer (as in other cervids) it then declines into adulthood with seasonal peaks and troughs in summer and winter, respectively (Anderson et al. 1974). Measurements of thymus glands of fawn, yearling, and adult mule deer, Browman and Sears (1956) observed annual cycles of highs in summer and lows in winter with fawns having the highest values of the 3 ages. Thymic atrophy is generally the result of chronic stress and can be seasonal (related to photoperiod or climate) or due to immediate stress such as inanition and disease (Chapman and Twigg 1990). White-tailed deer (0. virginianus) on low energy diets had lower (P < 0.05) thymus weights than deer on high energy diets (Lawrence et al. 1986). Ozoga and Verme (1978) conclude that the thymus provides a reliable index to physiological status and Lawrence et al. (1986) suggest thymus weight of adult does could be used in management decisions. Neonatal white-tailed fawns that were dying of disease or malnutrition had "extremely small" thymus glands averaging 1.3 g (range 0.5-3.0g, n = 14) compared to healthy fawns of similar age (X= 9.7 g, range 4.3-23.7 g, n = 7) (Ozoga and Verme 1978:794). Our combined 2-year mean thymus weight of 2.29 g (SD 2&2, n = 20) is similar to the mean of 1.3 g for fawns near death from disease or starvation observed by Ozoga and Verme (1978). In Colorado Anderson et al. (1974) collected 13 mule deer fawns (6 male and 7 female) from 1 to 5 months old; their mean thymus weight was 9.22 g (SD 2.88, n = 13) and was comparable to the healthy fawns sampled by Ozoga and Verme ( 1978). There were· only 3 measurements in our sample of sick/starve fawns that approached the means observed by Anderson et al. (1974) or Ozoga and Verme (1978) for healthy fawns. In 2000 a fawn 68 days old died (6 September) of a hemorrhagic condition and had thymus weight of 8.00 g and a second fawn 69 days old died (8 August) of pneumonia with a thymus weight of 7.95 g. Both had fat reserves but were judged to be less than optimal. In 2001, a fawn 21 days old died (7 July) of a hemorrhagic condition and had a thymus weight of 8.28 g; it was judged to have poor fat reserves. Excluding these 3 values, the mean for fawn thymus weight in this study was 1.26 g (SD 0.85, n = 17). The obvious reduced thymus size of fawns dying of sickness or starvation in this study should serve as a point of concern for managers. The reduced thymus size was probably initiated during the fetal stage of development and would indicate the stress factor was affecting the dam. Inanition has been shown to result in reduced thymus size in deer (Lawrence et al. 1986, Ozoga and Verme 1978) so the nutritional status of does during pregnancy, and especially during the last trimester, should be investigated. Fawns dying of sickness and starvation in 2000 and 2001 was nearly half the mortalities �55 attributed to this cause in 1999. The weather during fawning seasons of 2000 and 2001 was warm and dry possibly allowing fawns that were not robust to stresses to survive. This study was not designed as a manipulative study where some factor or factors were controlled or manipulated and the impact on fawn survival measured. Coyote predation on neonatal fawns was a popular theory and the opportunity arose to examine the effects of coyote control on a small portion of the study area. The area, 130 km (2% of the total area) included 3 sheep operations on private land. These ranches were used as lambing and summer ranges so coyotes were killed before the sheep were moved onto the area. Coyotes were killed from January trough September with most kills during winter and spring mostly by aerial gunning with a few kills from the ground. There was an active predator (coyotes and bear) control program during 1994 through 2001 on this area with a total of 187 coyotes and 17 bear killed (Animal and Plant Health Inspection Service, Wildlife Services records, Grand Junction, Colorado). During the 3 years of the fawn survival study there were 53 coyotes and 11 bear killed in the predator control area. Forty fawns were collared on this corresponding area allowing a comparison of fawn survival on and off the control area. Seven fawns were killed by predators inside the control area (4, coyote; 1, bear; and 2 feline) and 37 outside the area (24, coyote; 8 bear; and 5 feline). Comparison of predator kills inside and outside the area resulted in a Fisher's Exact Test result with P = 0.830; limiting the test to only coyote kills the results offered no evidence of an association between coyote control and fawn survival (P = 0.794). For fawn survival study results to be comparable they should be similar in the following: 1) fawn age at capture, 2) equipment and handling procedures, 3) tracking frequency, 4) nutritional status of does, 5) vegetation and hiding cover, 6) predator density, and 4) mortality identification criteria. Although it is impossible to match all of the above for comparisons, generalizations may be useful to assess the possible impact of the various mortality sources, particularly predators, on neonatal fawns. In Montana Hamlin et al. (1984) radioed 91 fawns over a 6-year period (1976-1981) and tracked them at 2-3-day intervals. Fawns up to 3 weeks old were included in their sample (Riley and Dood 1984). Mortalities were categorized as either "probable or known coyot~ involved deaths" or "other". They found no whole carcasses, which may be the result of tracking them on 2-3 day intervals allowing scavengers (including coyotes) time to find the carcass. Eighteen of 20 deaths (90%) were attributed to coyotes and 2 (2.2%) were listed as "other". Eighteen mortalities of 91 radioed fawns (19.8%) were assumed to be coyote-caused and total survival was 78.0%, which is higher than we observed. Their sample of fawns was most likely older than our sample. They used aerial observers to spot fawns indicating that the fawns were old enough to be trailing the does and ground crews used long-handled hoop nets to capture the fawns indicating the fawns were no longer in the hiding phase. A sample of older fawns would miss mortalities immediately after parturition and result in a higher survival rate compared to a sample of younger fawns such as ours. A fawn survival study on a 51.8 km Steens Mountain study area in Oregon during 1971-74 had a sample of 106 neonate fawns aged 1-14 days old and were monitored every 3-4 days (Trainer 1975). Mortality attributed to coyotes was 10.3% and for all predators it was 15.1 %. Disease and starvation mortality accounted for 9.4% of the total; survival was 72.6%. Preliminary results of an Idaho study with a sample of 69 fawns during 1998-99 exhibited a loss to coyotes of 13% and total predators (coyotes and lions) of32%. Overall survival was 44.9%. These results are not directly comparable to our study because coyotes and lions were controlled on a portion of the area. Given the shortcomings of comparing results of studies where protocol is not similar, neonatal fawn mortality attributed to coyotes is in the range of 10-20% for the various studies. Survival is highly variable ranging from 44.9% to 78.0%; the range in survival is undoubtedly heavily influenced by differences in age of fawns at capture (beginning of monitoring). Neonatal survival through 14 December does not account for observed low f:d ratios. In addition to pregnancy and fetal rates from the preliminary productivity study in February 1999, data available for this herd included random quadrat-based population size and herd structure estimates in December 1999. Survival estimates for bucks, does, and fawns during winter of 1999-2000 based on radioed animals �56 (Bruce Watkins, Colorado Division of Wildlife, Montrose, personal communication) were available. We used this information and incorporated our observed year 2000 summer fawn survival (0.5887) to calculate the expected f:d ratio for December 2000. Our calculations included I 0% lower fetal rates of primaparous does (Robinette et al. 1973, Trainer et al. 1981) and a differential of viable neonates of 96% for multiparous does and 82% for primaparous does (Robinette et al. 1973). They did not have an estimate of fetal rate, but the birth rate of 1.92 fawns per doe is similar to the maximum fetal rate for mule deer (Jensen and Robinette 1955). Hamlin and Mackie (1989) estimated 80% viable neonates for all-age does; this estimate includes fetal and neonate mortality. Using differential viable rates of Robinette et al. (1973)(96% and 82%) the projected December f:d ratio was 75 and using all-age estimated viable rate of Hamlin and Mackie (1989) (80%) the projected December f:d ratio was 64. Both of these projections were higher than the observed f:d ratio of 51 as estimated by a random quadrat helicopter survey in December 2000. Our data are most comparable to those of Hamlin and Mackie ( 1989) because theirs was a wild population. The herd studied by Robinette et al. ( 1973) was a fed captive population but indicates that a proportion of fawns born are not viable for various reasons even in a well nourished herd. Assuming the f:d ratio of 51 from the helicopter survey is unbiased, mortality of 3 7% from February when fetal rates via ultrasound were taken and June when fawns were captured would be necessary to match the observed f:d ratio. This indicates fetal or early neonate mortality that could be caused by inanition of the does, disease, or effects of poisonous plants. The importance of nutrition in reproductive success and recruitment is well documented. However, in the study by Robinette et al. (1973) fawn weights did not vary with nutrition level of does. The fawn weights in our study were comparable to those of other studies (Robinette et al. 1973, Stieigers and Flinders 1980, Trainer et al. 1981, and others). Apparently, fawn weights do not provide a useful index of doe nutritional status. Although fawns are born at relatively uniform weights the nutritional status of the doe can affect fawn survival through indirect effects such as susceptibility to predation and disease. The dam can be directly affected by failure to conceive, resorption of fetuses, and inability to nourish offspring (Dietz and Nagy 1976). What appears to be excessive fetal and neonate mortality from early pregnancy to a few days postparturition and discovery of 9 under-sized (X = 1.67 kg, n .= 7) stillborn fawns may be indicative of an under nourished adult population. Increased loss to sickness and starvation during 1999 when June precipitation was higher that the other 2 years may also indicate that neonates are in a compromised condition and not robust to stresses. Hemorrhagic diseases (HD), bluetongue (BT) and epizootic hemorrhagic disease (EHD), of the genus Orbivirus are present in the Uncompaghre Plateau mule deer herd (Myers 2001). These diseases are capable of causing significant mortality and Howerth et al. (200 I) cite literature documenting many outbreaks in Western North America dating back to 1886. In temperate regions, mortalities from hemorrhagic disease usually occur in late summer, before first frost, and epidemics can develop when conditions are favorable to the vector, Culicoides spp. These outbreaks are usually sporadic (Howerth et al. 200 I) and localized with total mortality estimated at < I% for mule deer (Thome et al. 1988). Infection with BT or EHD in mule deer may be asymptomatic, result in chronic disease, nonfatal infections, or sudden death (Howerth et al. 2001). Fever, internal bleeding, and shock resulting in death characterize hemorrhagic diseases (Shope 1967). Death may happen so suddenly that some animals may die " ... while walking or running" while others struggle in lateral recumbency position (Thome et al. 1988:115). Because this disease strikes quickly, animals in good physical condition may be found dead from HD (Chalmers et al. 1964). Only I positive result based polymerase chain reaction (PCR) test was obtained for HD during our 3-year study. The low detection rate may be because these are RNA viruses and are very unstable in an open environment (Myers 2001). There may have been other deaths from HD based on the time of year, hemorrhagic condition, and the relatively good condition of the fawn at death indicating a sudden death. Five fawns that died between 18 August and 4 October satisfy the above criteria. An adult female found near (100 m) one of the fawns tested positive by PCR for EHD. �57 Hemorrhagic disease is present on the Uncompahgre Plateau but it is hard to assess the impact on the mule deer population. It is unlikely that an epidemic of HD occurred during this 3-year study. There were no reports of numerous dead deer as was the case in other epidemics (Chalmers et al. 1964, Thome et al. 1988) and field personnel did not observe any abnormal concentrations of mortalities of either radioed fawns or unmarked deer during late summer. Diseases that affect the reproductive capacity of a host population are most liable to have a noticeable impact on that population (Anderson and May 1979). Bovine viral diarrhea virus (BVDV) infections produce abortions, fetal malformations, stillbirths, weakened neonates, and immunosuppression in domestic livestock (Baker 1995, Van Campen et al. 2001a). There is a> 60% prevalence ofBVDV titers in the adult population of deer on the Uncompahgre Plateau (Myers 2001). A mule deer population from northwestern Wyoming, USA, also had a 60% prevalence of BVDV titers (Van Campen et al 2001a), and a serological survey of 4 western national parks resulted in 59% prevalence in mule deer (Aguirre et al. 1995). Viral isolation (VI) is the most reliable method to determine exposure to BVDV and isolations from wild ruminants are rare (Van Campen et al. 2001a, Van Campen et al. 2001b). Isolates were obtained from 2 fawns in our study. These fawns died in the same general area(< 500 m apart) and within 2 days of each other, 17 and 19 July. Diseases that have low mortality and produce immunity with exposure are self-limiting (Myers 2001). In closed herds with no previous exposure, introduction ofBVDV can result in the loss of 75% of the first neonate cohort after exposure through abortions, stillbirths, and compromised immune response (Hana Van Campen, personal communication). With a high proportion of the Uncompahgre Plateau deer herd having titers, and presumed immunity to BVDV, this disease should not have a significant impact on the overall herd performance. However, its presence is certainly a depressant to some unknown degree. Other than the 2 fawns that provided VI ofBVDV, there were 2 other fawns with symptoms of being exposed to BVDV in utero. One was hydrocephalic (1.90 kg) and the other had skeletal deformities (2.09 kg) both characteristic of BVDV exposure. BVDV was isolated from a stillborn fawn from northern New Mexico that had an atrophied thymus and weighed 2.3 kg (Hibler 1981). The high prevalence of BVDV in the 2 above mentioned mule deer populations suggests that this virus circulates in these populations (Van Campen et al. 2001b) without exposure to outside sources such as cattle herds. High prevalence of titers to BVDV does not necessarily mean a population suffers significant consequences. In immunocompetent cattle the majority of infections (70-90%) are subclinical (Baker 1995). So unless the immune response of mule deer is compromised from some other cause, the impact of BVDV may not produce significant or detectable manifestations in population performance. Ingestion of poisonous plants can impair reproductive functions of domestic livestock (Panter et al. 2002). Some of the plants poisonous to livestock are found on the Uncompahgre Plateau and could conceivably also affect the wild ruminants of the plateau. Lupines (Lupinus spp.) can cause skeletal defects through the effects of alkaloids that are toxic to fetuses (Panter et al. 2002). We found a stillborn fawn (2.09 kg) with "Multiple congenital skeletal defects, including flexion contraction, limbs, neck, and thoracic spine" (from lab report) with minimally deformed joints and normal limb bones; all of these symptoms fit lupine poisoning as described by (Panter et al. 2002). In the year following our study, another stillborn fawn with severe skeletal deformities was found (Chad Bishop, Colorado Division of Wildlife, Montrose, personal communication). Both of these fawns came from the same general area of the plateau where lupine is common. Livestock loses to poisonous plants is associated with range in poor condition (Ralphs 2002) and this area is heavily grazed. We have observed some possible pre-natal mortality from poisonous plants; it could be one of many mortality sources but it is unlikely this is a major factor in herd recruitment. MANAGEMENT IMPLICATIONS Conditions of Western ranges have changed dramatically from pristine times (Vale 1975) through an era of extensive overgrazing to the current level of management. The era of unsustainable livestock grazing promoted shrub and forb growth to the benefit of deer (Clements and Young 1997) and led to the �58 eruption of mule deer populations (Gruell 1986). Subsequent management and wild fire control has resulted in over-mature shrubs and invasion of woody species into shrub communities reducing carrying capacity for deer (Gruell 1986). This trend continues. Nutrition is key to recruitment. It is a very common conclusion that although predators can cause a short-term effect on a deer population, alternate prey species abundance dictates the density of coyotes (Hamlin et al. 1984). Increases in vegetation production has a positive effect on abundance of alternate prey species (Hamlin et al. 1984) and may explain why Salwasser (1979) observed that coyote densities and fawn survival trend in unison. Most current land use patterns in the West are detrimental to deer and rather than control of hunting pressure or predators " ... deer numbers are ultimately governed by quality and quantity of habitat" (Connolly 1981 :238). Peek et al. (2002) and Salwasser et al. (1978) suggest that long-term decline of deer populations is not the result of predation but the result of deteriorating forage conditions. Given the high reproductive potential of mule deer, it seem reasonable that improving range conditions on the Uncompahgre Plateau and other mule deer ranges of the West would be the most fruitful for increased and long-lasting improved recruitment. Ballard et al. (2001: 112) state that "The relationship between predators and their prey is a very complex issue". They list numerous possible causes for deer declines including habitat loss (i.e. food and cover), disease, predation, competition, and others. The results of the Uncompahgre Plateau study do not provide evidence to suggest that predators are the cause of low recruitment in this particular herd. Coyote predation accounted for 0.126 of the neonatal mortality with bear and feline predation accounting for 0.040 and 0.032, respectively. Whether or not this degree of predation would warrant a control program would be a societal value judgment based on both the cost and the ethics of killing one species to favor another. Although there have been no studies that demonstrate predator reduction resulted in more mule deer in the possession of hunters (Ballard et al. 2001), that is obviously the ultimate objective of predator control. Predator control is a value judgment and has segments of the public sharply divided on the need or desirability for such a program. This study has provided information for a particular study area on the extent of fawn mortality from various causes that should be of assistance to the entities that make management decisions. ACKNOWLEDGMENTS This study was a contribution of Federal Aid Project W-153-R. We especially thank the fawn capture crew members. They exhibited the patience and persistence necessary to capture, handle, and radio track a reasonable sample of fawns, which contributed to the value of the study: T. Baker, B. Banulis, P. Burke, B. Diamond, B. Dreher, J. Foster, J. Griggs, D. Gustine, B. Hoffner, B. Lamont, H. McNally, J. Risher, E. Scott, J. Skinner, D. Swanson, and S. Znamenacek. We thank W. Andelt for input during the early phases of the study. Local Division of Wildlife field personnel were instrumental in gaining permission on private land and for helpful information on field access in addition to helping capture fawns: R Arant, G. Bock, M. Caddy, D. Coven, J. Ellenberger, V. Graham, M. King, M. Mclain, K. Miller. Others that assisted in fawn capture: T. Burke, and D. Steele. We thank D. Moreno for providing predator kill figures for the area of interest. The following were instrumental in establishing field handling procedures of fawn carcasses and necropsy protocol: D. Gould, K. Larsen, G. Mason, M. W. Miller, E. Myers, B. Powers, T. Spraker, and H. Van Campen. D. Schweitzer did most of the necropsies. M. Farnsworth and S. Strain provided assistance in graphic presentation and analysis. Aircraft piloting for radio tracking, field assistance, and general support of the project provided by J. Olterman is appreciated; R. B. Gill provided administrative support and G. Miller provided administrative support and editorial comments. Planning and statistical consultation by G. White improved the overall results. M. Potter provided late season radio tracking and radio retrieval. We thank colleagues T. Beck, and C. Bishop for many helpful suggestions and discussions and for their field assistance. The interest, support, and commitment for the duration of the project by B. Watkins were valuable contributions. We thank the various reviewers of the manuscript for their constructive criticism and suggestions. �59 LITERATIJRE CITED Acom, R. C., and M. J. Dorrance. 1990. Methods of investigating predation of livestock. Alberta Agriculture Protection Branch. Edmonton. Canada. Aguirre, A. A., D. E. Hansen, E. E. Starkey, R. G. McLean. 1995. Serologic survey of wild cervids for potential disease agents in selected national parks in the United States. Preventive Veterinary Medicine 21:313-322. Andelt, W. F., M. Bruscino, and C. Niemeyer. 1998. Interpreting the physical evidence of predation on big game animals and domestic livestock. Unpublished. Anderson, A. E., D. C. Bowden, and D. M. Kattner. 1992. The puma on Uncompahgre Plateau, Colorado. Colorado Division of Wildlife Technical Publication No. 40. Fort Collins, Colorado, USA. _ _, D. E. Medin, and D. C. Bowden. 1974. Growth and morphometry of the carcass, selected bones, organs, and glands of mule deer. Wildlife Monographs Number 39. _ _, and 0. C. Wallmo. 1984. Odocoileus hemionus. Mammalian Species No. 219. American Society of Mammalogists. Anderson, R. M. and R. M. May. 1979. Population biology of infectious diseases: Part I. Nature 280:361-367. Anonymous, c.a. 1944. Range management plan Escalante Unit, Ouray District, Colorado. Bureau of Land Management files, Delta, Colorado, USA. Baker, J.C. 1995. The clinical manifestations of bovine viral diarrhea infection. Veterinary clinics of North America: Food animal practice 11 :425-445. Ballard, W. B., D. Lutz, T. W. Keegan, L. H. Carpenter, and J.C. deVos, Jr. 2001. Deer-predator relationships: a review of recent North American studies with emphasis on mule and blacktailed deer. Wildlife Society Bulletin 29:99-115. _ _, H. A. Whitlaw, S. J. Young, R. A. Jenkins, and G. J. Forbes. 1999. Predation and survival of white-tailed deer fawns in northcentral New Brunswick. Journal of Wildlife Management 63:574-479. Bartmann, R. M. 1998. A prospectus on mule deer in Colorado. Colorado Division of Wildlife, Fort Collins, Colorado, Unpublished report. Bertram, M. R. and M. T. Vivion. 2002. Moose mortality in eastern interior Alaska. Journal of Wildlife Management. 66:747-756. Browman, L. G. and H. S. Sears. 1956. Cyclic variation in the mule deer thymus. Proceeding of the Society of Experimental Biological Medicine 93:161-162. Brown, C. G. 1992. Movement and migration patterns of mule deer in southeastern Idaho. Journal of Wildlife Management 56:246-253. Cantor, A. B. 1997. Extending SAS survival analysis techniques for medical research. SAS Institute, Cary, North Carolina, USA. Chalmers, G. A., H. N. Vance, and G. J. Mitchell. 1964. An outbreak of epizootic hemorrhagic disease in wild ungulates in Alberta. Wildlife Diseases 42: 1-6. Chapman, N. and G. I. Twigg. 1990. Studies on the thymus gland of British cervidae, particularly muntjac, Munticus reevesi, and fallow, Dama dama, deer. Journal of the Zoological Society of London 222:653-675. Clements, C. D. and J. A. Young. 1997. A viewpoint: Rangeland health and mule deer habitat. Journal of Range Management 50:129-138. Connolly, G. E. 1981. Trends in populations and harvest. Pages 225-243 in 0. C. Wallmo, editor. Mule and black-tailed deer of North America. University of Nebraska Press, Lincoln, Nebraska, USA. Dietz, D.R. and J. G. Nagy. 1976. Mule deer nutrition and plant utilization. Pages 71-78 in G. W. Workman and J.B. Low, editors. Mule deer decline in the West a symposium. Utah State University, Logan. �60 Eberhardt, L. L. 1977. "Optimal" management policies for marine mammals. Wildlife Society Bulletin. 5: 162-169 . . Gaillard, J.M., M. Festa-Bianchet, and N. G. Yoccoz. 1998. Population dynamics of large herbivores: variable recruitment with constant adult survival. Trends in Ecology and Evolution 13:58-63. Garrott, R. A., G. C. White, R. M. Bartmann, L. H. Carpenter, and A. W. Alldredge. 1987. Movements of female mule deer in northwest Colorado. Journal of Wildlife Management 51:634-643. Gill, R. B., T. D. I. Beck, C. J. Bishop, D. J. Freddy, N. T. Hobbs, R.H. Kahn, M. W. Miller, T. M. Pojar, and G. C. White. 2001. Declining mule deer populations in Colorado: Reasons and responses. Colorado Division of Wildlife Special Report Number 77. 1971. Middle Park deer study - productivity and mortality. Colorado Division of Wildlife, Federal Aid in Wildlife Restoration Job Progress Report, Project W-38-R-25. July: 189-207. Ginnett, T. F. and E. L. Butch Young. 2000. Stochastic recruitment in white-tailed deer along an environmental gradient. Journal of Wildlife Management 64:713-720. Gruell, G. E. 1986. Post-1900 mule deer irruptions in the Intermountain West: Principle cause and influences. General Technical Report INT-206. Ogden, Utah, U.S. Department of Agriculture, Forest Service, Intermountain Research Station. Huegel, C. N., R. B. Dahlgren, and H. L. Gladfelter. 1985. Use of doe behavior to capture white-tailed deer fawns. Wildlife Society Bulletin 13:287-289. Hamlin, K. L. and R. J. Mackie. 1989. Mule deer in the Missouri River breaks, Montana. Federal Aid Final Report, Project Number W-120-R-7-18. Montana Department of Fish, Wildlife and Parks. _ _, S. J. Riley, D. Pyrah, A. R. Dood, and R. J. Mackie. 1984. Relationships among mule deer fawn mortality, coyotes and alternate prey species during summer. Journal of Wildlife Management 48:489-499. Hibler, C. P. 1981. Diseases. Pages 129-155 in 0. C. Wallmo, editor. Mule and black-tailed deer of North America. University of Nebraska Press, Lincoln, Nebraska, USA. Howerth, E.W., D. E. Stallknecht, and P. D. Kirkland. 2001. Bluetongue, epizootic hemorrhagic disease, and other orbivirus-related diseases. Pages 77-97, in E. S. Williams and I. K. Barker, editors. Infectious diseases of wild mammals. Iowa State University Press, Ames, Iowa, USA. Jensen, W. and W. L:Robinette. 1955. A high reproductive rate for Rocky Mountain mule deer. Journal of Wildlife Management 19:503. Kaplan, E. L. and P. Meier. 1958. Nonparametric estimation from incomplete observations. American Statistical Association Journal 53:457-481. Kie, J. G. and B. Czech. 2000. Mule and black-tailed deer. Pages 629-657 in S. Demarais and P.R. Krausman, editors. Ecology and management of large mammals in North America. Prentice Hall, Upper Saddle River, New Jersey, USA. Kufeld, R. C., D. C. Bowden, and D. L. Schrupp. 1989. Distribution and movements of female mule deer in the Rocky Mountain foothills. Journal of Wildlife Management 53:871-877. _ _. 1979. History and current status of the mule deer population on the east side of the Uncompahgre Plateau. Division Report No. 11. Colorado Division of Wildlife. Lawrence, R. K., S. Demarias, R. D. Brown, M. Abbott. 1986. Nutritional effects on thyroids, ovaries, and thymuses in white-tailed deer. Proceedings of the Annual Conference of Southeast Association of Fish and Wildlife Agencies 40:416-423. Laycock, W. A. 1991. Stable states and thresholds ofrange condition on North American rangelands: A viewpoint. Journal of Range Management 44:427-433. Marshall, M. 1998. Uncompahgre. Western Reflections, Ouray, Colorado. Medin, D. E., and A. E. Anderson. 1979. Modeling the dynamics of a Colorado mule deer population. Wildlife Monographs 68:1-77. Myers, E. P. 2001. Assessing the role of selected infectious disease agents in neonatal mule deer fawn mortality on the Uncompahgre Plateau of Western Colorado. M. S'. Thesis, Colorado State University, Fort Collins, Colorado, USA. �61 Noyes, J. H., B. K. Johnson, L. D. Bryant, S. L. Findholt, J. W. Thomas. 1996. Effects of bull age on conception dates and pregnancy rates of cow elk. Journal of Wildlife Management 60:508-517. Ozoga, J. J. and L. J. Verme. 1978. The thymus gland as a nutritional status indicator in deer. Journal of Wildlife Management 42:791-798. Panter, K. E., L. F. James, D.R. Gardner, M. H. Ralphs, J. A. Pfister, B. L. Stegelmeier, and S. T. Lee. 2002. Reproductive losses to poisonous plants: Influence of management strategies. Journal of Range Management 55:301-308. Peek, J.M., B. Dennis, and T. Hershey. 2002. Predicting population trends of mule deer. Journal of Wildlife Management 66:729-736. Picton, H. D. 1979. A climate index and mule deer fawn survival in Montana. International Journal of Biometerorology 23(2):115-122. Pollock, K. H., S. R. Winterstein, C. M. Bunck, and P. D. Curtis. 1989. Survival analysis in telemetry studies: the staggered entry design. Journal of Wildlife Management 53:7-15. Ralphs, M. H. 2002. Ecological relationships between poisonous plants and rangeland condition: A review. Journal of Range Management 55:285-290. Riley, S. J. and A. R. Dood. 1984. Summer movements, home range, habitat use, and behavior of mule deer fawns. Journal of Wildlife Management 48:1302-1310. Robinette, W. L., C.H. Baer, R. E. Pillmore, and C. E. Knittle. 1973. Effects of nutritional change on captive mule deer. Journal of Wildlife Management 37:312-326. Salwasser, H. J. 1979. Ecology and management of the Devil's Garden Interstate deer herd and range. Dissertation, University of California, Berkeley, USA. _ _, S. A. Holl, and G. A. Ashcraft. 1978. Fawn production and survival in the North Kings River deer herd. California Fish and Game 64:3 8-52. Schlegel, M. 1976. Factors affecting calf elk survival in north central Idaho: A progress report. Western Association of State Game and Fish Commissioners 56:342-355. Squibb, R. C. 1985. Mating success of yearling and older bull elk. Journal of Wildlife Management. 49:744-750. Thome, E.T., E. S. Williams, T. R. Spraker, W. Helms, and T. Segerstrom. 1988. Bluetongue in freeranging pronghorn antelope (Antilocapra americana) in Wyoming: 1976 and 1984. Journal of Wildlife Diseases 24:113-119. Trainer, C. 1975. Direct causes of mortality in mule deer fawns during summer and winter periods on Steens Mountain, Oregon - a progress report. Western Association of State Game and Fish Commissioners 55: 163-170. Tsai, K., K. H. Pollock, and C. Brownie. 1999. Effects of violation of assumptions for survival analysis methods in radiotelemetry studies. Journal of Wildlife Management 63: 1369-13 75. Uncompahgre Plateau Partners (BLM, CDOW, Public Lands Partnership, USFS). 2002. Uncompahge Plateau Project (UP) Plan: ~ collaborative approach to restore and maintain the ecosystem heath of the Uncompahgre Plateau in western Colorado. Available at the Colorado Division of Wildlife Service Center, Montrose, Colorado, USA. Unsworth, J. W., D. F. Pac, G. C. White, and R. M. Bartmann. 1999. Mule deer survival in Colorado, Idaho, and Montana. Journal of Wildlife Management 63:315-326. Vale, T. R. \975. Presettlement vegetation in the sagebrush-grass area of the Intermountain West. Journal of Range Management 28:32-36. Van Campen, H., K. Frolich, and M. Hofmann. 2001a. Pestivirus infections. Pages 232-244 in E. S. Williams and I. K. Barker, editors. Infectious diseases of wild mammals. Iowa State University Press, Ames, Iowa, USA. _ _, J. Ridpath, E. Williams, J. Cavender, J. Edwards, S. Smith, and H. Sawer. 2001b. Isolation of bovine viral diarrhea virus from a free-ranging mule deer in Wyoming. Journal of Wildlife Diseases 37:306-311. Verme, L. J. 1969. Reproductive patterns of white-tailed deer related to nutritional planes. Journal of Wildlife Management 33:881-887. �62 Wade, D. A., and J.E. Bowns. 1984. Procedures for evaluating predation on livestock and wildlife. Texas Agricultural Experiment Station Bulletin B-1429, College Station. White, G. C. and R. M. Bartmann. 1998. Mule deer management- What _should be monitored? Pages 102-116 in J.C. De Vos, Jr., editor. Proceedings of the 1997 deer-elk workshop, Rio Rico, Arizona. Arizona Game and fish Department, phoenix, Arizona, USA. _ _, D. F. Freddy, R. B. Gill, and J. H. Ellenberger. 2001. Effect of adult sex ratio on mule deer and elk productivity in Colorado. Journal of Wildlife Management 65:543-551. • White, M. 1973. Description of remains of deer fawns killed by coyotes. Journal of Mammalogy 54:291-293. Young, R. G. and J. W. Young. 1984. Colorado West land of geology and wildflowers. Artistic Printing, Salt Lake City, Utah, USA. Table 1. Fate of neonatal mule deer fawns from capture (mean 3 days old) to 14 December, Uncompahgre Plateau, west central Colorado, USA, 1999-2001. Mortality Causes Sick/Starve Coyote Bear Feline Trauma Unknown Total mortality Surviving Total sample Censored 1999 2000 2001 Total 15 6 3 4 1 2 31 19 50 12 10 11 3 2 4 13 11 3 1 4 3 35 57 92 20 38 28 9 7 9 10 101 129 230 43 5 35 53 88 11 Table 2. Chi-square comparison of mortality source proportion by year for mule deer neonates from capture (mean 3 ~~ys old) to 14 December on the Uncompahgre Plateau, west central_Co~orado, USA. Kaplan-Mei_er staggered exits to account for censored subjects were used... •• Mortality Causes 1999 2000 Sick/starve Coyote Bear Feline Trauma Unknown 0.318 0.126 0.060 0.083 0.025 0.042 0.115 0.129 0.034 0.023 0.048 0.058 2001 0.148 0.121 0.033 0.012 0.047 0.037 ~ 5.330 0.023 0.416 2.330 0.568 0.388 p 0.070 0.989 0.812 0.312 0.753 0.824 �63 Figure 1. Uncompahgre Plateau mule deer Data Analysis Unit, D-19, is shown outlined in red. The 95% kernel home range for radioed does aerial located on May 18 and May 31 is outlined in black with black dots representing individual locations. Fawn capture locations are seen as white dots and the 95% kernel home range outlined in white. Blue shading shows pinyon-juniper type and higher elevation vegetative types (summer range) are shown in green, yellow, and brown. Agricultural land is in pink. See text for type descriptions. �55 JOB PROGRESS REPORT State of ------=C~o=lo=r=a=do=------- Division of Wildlife- Mammals Research Work Package No. _ _ _3=0"-'0C...:l~---- Deer Conservation Task _ _ _ _ _ _ _ _...,:le.___ _ _ _ __ Mule Deer Life Cycle- Neonatal Fawn Survival Federal Aid Project _W~--18_5_-R _ _ _ _ __ Period Covered: July I 2002 through June 30, 2003 Author: Thomas M. Pojar Personnel: W. Andelt, R Arant, D. Baker, T. Baker, B. Banulis, T. Beck, C. Bishop, G. Bock, D. Bowden, P. Burke, T. Burke, M. Caddy, D. Coven, B. Diamond, B. Dreher, J. Ellenberger, M. Farnsworth, J. Foster, V. Graham, J. Griggs, D. Gustine, P. Hayden, B. Hoffner, B. Lamont, M. King, K. Larsen, M. Mclain, H. McNally, G. Miller, M. W. Miller, E. Myers, J. Olterman, M. Potter, J. Risher, D. Schweitzer, D. Steele, J. Skinner, T. Spraker, D. Swanson, B. Watkins, G. White, S. Znamenacek. The following is the abstract of the manuscript submitted to the Journal of Wildlife Management describing the neonatal fawn survival study on the Uncompahgre Plateau. Because of requests by reviewers or editors some aspects of the presentation and analysis may be modified. Manipulation or interpretation of these data beyond that contained in this report should be labeled as such, and is discouraged. NEONATAL MULE DEER FAWN SURVIVAL IN WEST-CENTRAL COLORADO ABSTRACT Declining mule deer (Odocoileus hemionus) populations resulting from apparent low recruitment brought management and political focus on neonatal fawn survival. We captured mule deer fawns on the Uncompahgre Plateau (5,957 km2 ) in west-central Colorado, USA, at a mean age of 3 days (range from newborn to 6 days), and we radiomarked them with mortality-sensing drop-off radiocollars. Two hundred thirty fawns were radiomarked with samples of 50 in 1999, 88 in 2000, and 92 in 2001. Designated neonate survival period was from capture to 14 December. Survival was different among years (X/ = 6.160, P = 0.046) with 3-year mean survival of0.501. Cause-specific mortality ordered from highest to lowest was sick/starve, coyote, unknown, trauma, bear, and feline. Neither all predation combined (coyote, bear, and feline; P = 0.379) nor coyote predation alone (P > 0.989) differed among years. By 31 July, 74% of the sick/starve mortality and 75% of the predation mortality had taken place with 76% of mortality from all sources occurring by this date. Mean fawn weights at capture were different among years (P = 0.044). We also found a difference in hind foot length among years (P = 0.002). Weight and hind foot means were different between 2000 and 200 I (P > 0. 017) with 1999 not different from either 2000 or 2001 (P < 0.017). Mean capture date was 19 June (SD= 4.83 days) and median capture date was 19 June (range= 9 Jun to 6 Jul) with 94.78% of all captures occurring between 13 and 30 June. This implies that most does were bred during their first estrous cycle. Neonatal survival through 14 December did not completely account for observed low f:d ratios. We hypothesized fetus mortality during late pregnancy or mortality of fawns at birth (before they could be detected for capture) as potential causes of poor recruitment. � https://cpw.cvlcollections.org/files/original/1933dd5b6643a23a758d154f2f16e804.pdf 682e51cc96e1eec2b8ad71cc0a683102 PDF Text Text Colorado Division of Wildlife July 2007 – June 2008 WILDLIFE RESEARCH REPORT State of: Cost Center: Work Package: Task No.: Colorado 3430 3001 3 Federal Aid Project No. W-185-R : : : : Division of Wildlife Mammals Research Deer Conservation Pilot Evaluation of Predator-Prey Dynamics On the Uncompahgre Plateau Period Covered: July 1, 2007 - June 30, 2008 Authors: M.W. Alldredge, E.J. Bergman, C.J. Bishop, K.A. Logan, D.J. Freddy Personnel: B. Dunne, V. Yovovich, E. Phillips, M. Schuette All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT In an attempt to address predator-prey dynamics, we initiated a pilot study to evaluate cougar predation relative to prey distribution across the southern half of the Uncompahgre Plateau in southwestern Colorado. As part of ongoing mule deer and cougar research in this area, we estimated cougar kill rates and prey selection by sampling different sized clusters of cougar GPS locations across the landscape. Cluster size ranged from 1 location to >30 locations/cluster. In the vicinity of each sampled cluster, we searched for cougar prey items to determine whether a kill had occurred and to classify prey by species and age. This field effort was primarily focused in areas with extensive historical mule deer population and winter range distribution data. Simultaneously, a pilot effort to collect distribution and movement data of elk over this same geographic area was conducted. As predicted, cougar kill sites were associated with deer and elk distribution. The greatest density of kill sites occurred across mid-upper elevation deer winter range where overlap of wintering elk and deer was greatest. We investigated 462 clusters during this pilot study. Kill probability increased as cluster size increased ( /3ˆ = 0.353, SE = 0.0706). Kill probability exceeded 0.9 with ≥ 10 locations/cluster and approached 1 with ≥ 15 locations/cluster. The probability of a kill was high if a cougar spent >2 days in the same general area, and a kill was essentially certain if a cougar spent >3 days in the same general area. There was some probability of a kill at clusters that comprised only 1 location, indicating that isolated cougar locations may periodically be associated with kills and should not be ruled out when using GPS location data to address cougar prey utilization. Our estimates of kill probability are conservative because the estimates assume detection probability was 1, which is unlikely. Cougars killed adult deer, fawn deer, adult elk, and calf elk in roughly equal proportions. Each prey class comprised 0.22 0.24 of the total kill. Kill composition varied as a function of percent vegetative cover and elevation. Future research should evaluate detection probability, which underlies the interpretation of cougar kill rates. 87 �WILDLIFE RESEARCH REPORT PILOT EVALUATION OF PREDATOR-PREY DYNAMICS ON THE UNCOMPAHGRE PLATEAU MATHEW W. ALLDREDGE, ERIC J. BERGMAN, CHAD J. BISHOP, KENNETH A. LOGAN, AND DAVID J. FREDDY P.N. OBJECTIVE To assess if a sampling based approach to collecting cougar predation data can efficiently result in unbiased data. To make a pilot assessment of how cougar kills are spatially distributed over prey winter range. SEGMENT OBJECTIVES 1. Use and evaluate the efficiency of a GPS collar, GIS and statistical sampling based approach to investigate potential cougar kill sites. 2. Estimate mule deer density on three study areas and extrapolate results onto surrounding mule deer range. 3. Overlay locations of 5 elk, collected via GPS collars, on mule deer winter range boundaries to gain preliminary information as to how much spatial overlap occurs between the species and to determine where cougar kills occur in relation to the mule deer and elk space use. INTRODUCTION Predator prey interactions have always been a topic of interest for wildlife managers and ecologists. However, due to the complexities of studying natural systems, behavioral theories pertaining to the subject are often developed in invertebrate, aquatic or small mammal systems, often under controlled laboratory conditions (Mathews et al. 2006, Schmitz 2006, Werner and Peacor 2006). Similarly, many models are developed within theoretical frameworks (Keeling et al. 2000, Mitchell and Lima 2002). While developing theories under these conditions is almost inherently necessary, their subsequent transition to free ranging systems is not frequent (Ryall and Fahrig 2006). Of the free ranging systems where theories are developed and tested, most deal with avian species (Lima and Bednekoff 1999, Roth et al. 2006), where as application to large mammalian systems is less frequent. Of the mammalian predator prey systems that have been studied, most have been conducted in preservation/park settings that largely exclude human influence (Kunkel and Pletscher 1999, Kunkel et al. 1999, Krebs et al. 2001, Creel and Creel 2002, Mao et al. 2005, Wilmers et al. 2006,). Additionally, due to the small number of large scale studies that have been conducted, the ability of managers to draw inference to separate systems (i.e. different species or different ecosystems) is limited. While this existing body of work is invaluable, extrapolation of theories to large mammalian systems could be limited and basing wildlife management decisions on this information may be tenuous. Due to the value of mule deer, elk and cougars as recreationally hunted species in Colorado, there is much interest in understanding the nature and relationship between the population dynamics of these species. However, resulting from the dearth of information pertaining to the interactions of these 3 species, a vast array of opinions and theories pertaining to their impacts on each other have been propagated. As a management agency, the Colorado Division of Wildlife is responsible for supporting or refuting claims with biological data that were collected in a scientifically unbiased manner. To date, these data are largely unavailable. 88 �Currently, the opportunity to develop a predator prey study exists on the Uncompahgre Plateau in southwestern Colorado. Two large scale research programs, independently studying cougar and mule deer, are underway in the same geographic area. Thus, the initial framework to study a top carnivore, and what are thought to be its primary prey species, is in place. However, to date there is little or no information pertaining to elk distribution or population dynamics in this area. The addition of elk spatial data will allow us to assess the feasibility of developing a full study addressing the influence and interactions of cougars, mule deer, and elk. STUDY AREA This pilot study was conducted on the southern half of the Uncompahgre Plateau in southwestern Colorado, near Montrose, Colorado (Figure 1). The study area was defined by the existing boundary for the ongoing cougar research project with prey populations being monitored only in the eastern half of the cougar study area. METHODS Capture and Handling Methods As part of completed, as well as ongoing mule deer research, approximately 75 adult female mule deer were marked with VHF radio collars in the area of interest (Bishop et al. 2005). Additionally, 25 mule deer fawns were captured and radio-collared within the eastern portion of the study area between late-November and late-December 2006 as part of the ongoing mule deer research (Bergman et al. 2005; capture protocols previously approved by CDOW ACUC). All mule deer were captured with baited dropnets (Ramsey 1968, Schmidt et al. 1978, Bartmann et al. 1992) or via helicopter net-gunning (Barrett et al. 1982, van Reenen 1982). As part of the ongoing cougar research project, 19 cougars (15 female, 4 male) were outfitted with GPS collars that allowed on-demand data download interaction with researchers. Cougars were captured primarily via pursuit by dogs as well as in live traps (Logan 2005, capture protocols previously approved by CDOW ACUC). As part of this pilot study, adult female elk (9) were captured via helicopter net-gunning during late-December/January 2006-07 with 5 adult females fitted with drop-off GPS/VHF collars and 4 adult females fitted with VHF permanent collars. Elk were captured on the eastern portion of the study area, directly overlapping areas including radio collared mule deer and cougar. Sample sizes for elk reflected an estimate of what we believed to be an adequate number of elk to provide an initial estimation of elk spatial use in the study area. Ungulate Survival and Location Monitoring On a daily basis, from December through May, we monitored radioed fawns and adult female deer and elk in order to document live/death status. This allowed us to determine accurately the date of death and estimate the proximate cause of death. For animals not heard from the ground, we conducted weekly flights to assess live/death status. Detailed locations of GPS collared elk became available when self-actuating mechanisms caused the GPS collars to drop-off elk in September 2007. Elk GPS collars collected locations every 30 minutes. Identification of Cougar GPS Location Clusters Characteristics of clusters of GPS locations representing cougar-killed ungulate sites (Anderson and Lindzey 2003, Logan 2005) were used to develop a standard algorithm to group GPS points together, to provide a sound sampling frame from which statistical inference could be made about clusters that are not physically investigated. GPS collars collected locations 4 times/day to reflect time periods when cougars are both active and inactive (00:00, 6:00, 12:00 and 19:00). The clustering routine was designed to identify clusters in five unique selection sets in order to identify clusters containing two or more points, those that contained missing GPS locations, and those that were represented by single points. The clustering algorithm was written in Visual Basic and was 89 �designed to run within ARCGIS (Alldredge and Schuette, CDOW unpubl. data 2006). The widths of the spatial and temporal sampling windows were user specified, in order to meet multiple applications and research needs. This also enabled adjustment of the sampling frames to improve cluster specifications as needed. The initial step was to prepare data files for ARCGIS. The main priority was to number all downloaded GPS lat-long location records consecutively to provide a time stamp that could be used in the program. Failed locations were numbered within the data files to maintain the proper time step (i.e. two locations that were separated by a missing location were time stamped in such a way that the clustering algorithm recognized that a missing location existed between the records). At this point data files were imported to ARCGIS and coordinates converted to UTMs. The initial selection set of clusters (S1) were based on clusters consisting of two or more points within a specified distance and time interval. Working with temporal and spatial variables simultaneously is difficult, so we chose to create an association matrix of the combined variables. The units for time were based on GPS locations so that the time between consecutive downloads was one. Cougar locations are attempted 4 times a day, so that one day consisted of 4 time-steps. The association matrix was then constructed as 1 Aij d max e 1 dij ti tj tmax where Aij was the association in time and space between points i and j, dmax was the maximum distance between two points to be considered a cluster, dij was the distance between points i and j, tmax was the maximum number of time steps between points to be considered in a cluster, and ti and tj were the times for locations i and j. This formula weighted the distance between two locations heavier than the time between two locations. It also caused the association Aij to be negative for any locations that were outside the temporal window (separated by more time-steps than tmax). The association between two locations within the specified time interval was greatest for those locations that were spatially closer together. So, the largest value in the association matrix corresponded to the 2 points that were spatially the closest and within the time interval. Initially, dmax was set at 200 m and tmax was set at 16 time steps [4 DAYS] . The initial cluster was selected by choosing the 2 points with the largest association value from the association matrix. The distance was checked to verify that the points were within the specified maximum distance, dmax, and if so, the centroid of the two points was calculated. An association vector Ac was made by calculating the association among the centroid and all other points using the above formula. If all values in Ac were negative, then no points were within the specified time interval, so no additional points were added to the cluster. Then the greatest association value Acmax was selected from Ac and the distance from the centroid to the point corresponding to Acmax was compared to dmax. If the distance was less than dmax then the point was added to the cluster and a new centroid was calculated using all cluster points and a new vector Ac was constructed using the new centroid. This procedure was repeated until no additional points were added to the cluster because either no points were within the specified time interval or the distance from the centroid to all points was greater than dmax. After each cluster was constructed these points were omitted from the association matrix and a new cluster was started by again selecting the greatest value from the matrix and verifying that the distance between points was less than dmax. Points were again added to this cluster as previously described. This entire procedure was repeated until no 2 locations met the temporal or spatial criteria. 90 �All clusters were given a unique identifier, which was based on the animal identification and the Julian date. This completed the selection set for clusters with two or more locations, which were likely to have a high probability of being a kill site. Additional selection sets were constructed from the remaining points as single location clusters. However, not all locations are equal, so the remaining selection sets were created based on whether points were associated with missing locations and based on distance between consecutive locations. The second selection set (S2) of clusters was created from any 2 points that were within a distance dmiss, and were separated by 1 or more missing locations. The cluster was considered to be the area within the distance dmax of each of the known locations (2 areas make up the cluster, and dmiss was initially set at 500 m). The final 2 cluster selection sets consisted of consecutive points that were within the ranges dmax to d2 (S3) and d2 to d3 (S4). To construct these selection sets, the distance between consecutive points was examined and if the distance was within the range dmax to d2 (500 m) then the initial point was added as a cluster to the set S3, or if the distance was within the range d2 to d3 (1000 m) then the initial point was added as a cluster to the set S4. These single-point clusters were assumed to have radius dmax. Points not used in selection sets S1 through S4 were then used in a final selection set S5. These points represented larger movements between consecutive locations and thus were thought to have low probabilities of being associated with a kill site, although these points could be associated with use of small prey items, or kill sites where a cougar was physically disturbed away from a kill site. These single-point clusters were also assumed to have radius dmax. Sampling of Cougar GPS Location Clusters A primary objective of the pilot study was to determine the probability that a given cluster represented a cougar feeding site. Specifically, to evaluate cougar feeding sites as a function of the cluster association matrix. Using the clustering algorithm described above, we attempt to classify each sampled cluster as a cougar feeding site (1) or not a feeding site (0). We expected a high proportion of S1 clusters to represent cougar feeding sites. Conversely, we expected a moderate proportion of S2 and S3 clusters, and a low proportion of S4 and S5 clusters, to represent cougar feeding sites. A secondary objective of the pilot study was to gather preliminary biological data regarding cougar prey utilization, primarily with respect to deer and elk. The secondary objective was most efficiently accomplished by sampling S1 clusters with greater intensity than other clusters. We therefore structured our sampling approach to allow adequate estimation of the proportion of clusters that were cougar feeding sites for each cluster set, while more intensively sampling S1 clusters than all others. With no previous evidence to indicate similarities among individuals based on sex, age, or parental status, sampling was stratified by each individual cougar. GPS collars were downloaded once a month for each cougar and data were analyzed through the clustering algorithm. Clusters within 2 weeks of the download date were selected for the sampling frame, making the maximum time between the predation event and sampling about 1 month by the time field technicians could get to and assess evidence at each cluster site. Clusters were randomly chosen from each selection set for each individual cougar every month in the following manner: S1 = 2 clusters, S2 = 1 cluster, S3 = 1 cluster, and S4 and S5 = 1 cluster on alternating months. Five clusters were sampled each month for each cougar, for a total of 30 clusters per cougar from 1 November 2006, to 15 July 2008. As time allowed, additional clusters were sampled from the selection sets. Our approach forced constant sampling of each cluster set over time regardless of the frequency of clusters within a given set. This prevented a scenario where nearly all sampled clusters in a given month were from sets, S3, S4 and/or S5 (i.e., low probability of finding feeding sites). Our assessment of prey utilization depended on relatively constant detection of cougar feeding sites over time to avoid bias. 91 �However, for each cluster set, the true proportion of clusters representing feeding sites may possibly change over time corresponding to changes in cougar use of feeding sites. If the GPS download data indicated major changes in set-specific cluster frequencies over the sampling period, we maintained the ability to use a proportional-allocation sampling approach if needed. Assuming a binomial distribution and 0.90 of S1 clusters represented cougar feeding sites, our approach enabled us to estimate the true proportion with a 95% confidence interval of +/ 0.07. Assuming 0.5 of S2 clusters represented cougar feeding sites, we were able to estimate the true proportion with a 95% confidence interval of +/ 0.17. Assuming 0.3 of S3 clusters represented feeding sites, we were be able to estimate the true proportion with a 95% confidence interval of +/ 0.15. Finally, assuming 0.1 of S4 and S5 clusters represented feeding sites, we were able to estimate the true proportion with a 95% confidence interval of +/ 0.10. These precision levels were deemed acceptable for the pilot study, and should facilitate development of an optimal sampling scheme in future years for evaluating cougar prey utilization from GPS cluster-location data. Finally, regarding our secondary objective of collecting preliminary prey use data, we were able to estimate the overall proportion of kill sites represented by deer (or the proportion of kill sites represented by elk) with a 95% confidence interval of +/ 0.05 (Anderson and Lindzey 2003, Logan 2005). We used the following protocol to investigate cougar GPS clusters in the field. For S1 clusters, we investigated each cougar GPS location in the cluster by spiraling out a minimum of 20 m from the GPS waypoint while using the GPS unit as a guide, and visually inspecting overlapping view fields in the area for prey remains. Normally, this was sufficient to detect prey remains and other cougar sign (e.g., tracks, beds, toilets) associated with cougar. If prey remains were not detected within 20 m radius of the cluster waypoints, then we expanded our searches to a minimum of 50 m radius around each waypoint. The 20 m and 50 m radius search areas resulted in overlapping view fields of individual waypoints, and took up to 7 hours to complete, depending upon the number of waypoints, topography, and vegetation type and density associated with a cluster. For S2 through S5 clusters, we went to each cougar GPS location and spiraled out 50 m around each waypoint, while using the GPS unit as a guide. Depending on the number of locations, topography, and vegetation type and density, we spent a minimum of 1 hour and up to 3 hours per cluster to judge whether the cluster was a kill site. Estimating Deer, Elk, and Cougar Distributions We examine locations, movements, and kernel home ranges of mule deer, elk, and cougars for spatial overlap and time synchrony using ArcGIS. Our initial analyses are descriptive and should provide insight into patterns of cougar movements and feeding sites in relation to major ungulate species. Based on past observations, we did not expect deer distributions to fluctuate greatly during the winter. However, we did expect elk distributions to fluctuate depending on weather and time. We anticipated being able to generate correlations between species of prey killed by cougars and the relative presence of prey within cougar home ranges. Cougar GPS Cluster Analysis We estimated the probability of locating cougar prey items (i.e., cougar kills) at GPS location clusters using logistic regression in SAS (PROC LOGISTIC; SAS Institute, Cary, NC). We modeled cougar kills as a function of cluster type (S1, S2,…, S5), cluster size (no. locations/cluster), cougar status (adult female with cubs, adult female without cubs, adult male), and season when cluster was investigated (winter, spring, summer, fall). We then analyzed kill composition using a generalized logits model (i.e., multinomial logistic regression) in SAS (PROC LOGISTIC). For this analysis, we used only clusters where prey items were found (i.e., kills). Kill composition was divided into 5 categories: adult deer, fawn deer, adult elk, calf elk, and other (i.e., porcupine, coyote, turkey, unknown). We modeled kill composition as a function of cluster type, cluster size, cougar status, season when kill occurred, elevation, 92 �and percent vegetative cover. We used Akaike‘s information criterion adjusted for sample size (AICc) to select among candidate models in both modeling analyses (Burnham and Anderson 2002). Hypothesis Testing Our preliminary sampling effort of cougar clusters and ungulate distributions provided estimates of cougar kill rates and proportions of deer and elk killed. As data collection continues, we intend to address whether 1) cougar prey mass is positively related to cougar mass (i.e., male cougars kill larger prey than female cougars), 2) cougars prey on deer and elk in proportion to availability (i.e., no selection for prey species), 3) cougars prey on sex or ages of deer or elk populations in proportion to availability (i.e., no selection for prey age classes), 4) cougars alter their use of prey among seasons of the year (i.e., prey-switch between deer and elk, or between juvenile and adult), and 5) maternal cougar home ranges include the highest available densities of ungulate prey. RESULTS AND DISCUSSION Mule Deer Distribution As expected, over the course of the winter, mule deer movements occurred at too fine of spatial and temporal scales to be detected without more intense repeated sampling. However, as they relate to the pilot study, the data gathered are adequate for making basic summaries. Mule deer density appeared to be highly variable across a gradient of winter range (density estimates ranged between 19 and 109 deer/km2 , Figure 1) (Bergman et al. 2008). Relative to the entire Uncompahgre Plateau, the estimates tended to be high, confirming historical information and further justifying the decision to conduct pilot work in this area. Of particular interest in regards to spatial overlap between cougar kill sites and mule deer winter range, the majority of located kill sites were higher in elevation than the greatest concentrations of mule deer. The exception to this trend occurred on the southern most portion of deer winter range where the majority of kill sites were composed of mule deer. As discussed below, an apparent explanation for this may be linked to elk distribution as this area also appeared to be the area of greatest overlap between mule deer and elk. To improve future efforts, several key steps would need to be taken. Mule deer density estimates are relatively course for the majority of winter range included in this pilot study. With the exception of three polygons, deer density estimates were extrapolated from surrounding areas. Furthermore, estimates of deer density for the 3 areas were not collected during the same year and therefore include annual variation. To accurately reflect the conditions, albeit still at a course level, encountered by cougars as they move across mule deer winter range, density estimates should minimally be collected on all segments of winter range on an annual basis. While fine scale movements of deer (i.e. daily movements within winter range) were not incorporated in this study, such data likely would not be hugely beneficial. Fine scale data would be of greatest interest if the focus of the study were shifted to analyze/describe fine scale hunting behavior of individual cougars. Elk Movement and Distribution Elk GPS collar data confirmed our initial expectations that elk movements during winter months were more dynamic than those of deer. The four elk collared with VHF collars left the study area of interest after 7 months and collecting repeated aerial locations was not deemed worthwhile as they were not in areas with radio marked deer or cougar. However, elk did appear to be highly individualistic in regards to space use and movement during winter months. Two elk appeared to concentrate locations over a relatively large geographic area (>75 km2) during the winter months, but restricted movements to stay within these areas. The other 3 elk appeared to utilize relatively small spatial areas (9-10 km2) for 12 week periods before making slightly longer movements (10+ km) to new concentration areas. Plotting known locations for cougar kill sites on elk spatial data suggested that cougar kill sites had a strong correlation to elk distribution (Fig. 2). Based on the more dynamic nature of elk movement during winter, 93 �future efforts to map elk distributions and densities would be better met by saturating the area of interest with GPS collared elk. Annual density estimates, collected via helicopter, would likely only be valid for a relatively short time period (2-4 weeks) due to elk movement and thus making it difficult to track cougar space use and predation patterns in a realistic prey context. By outfitting a large number of elk with GPS collars, resource selection functions for the elk in the area of interest could be built around habitat and elevation selection patterns. Due the large amount of data collected by GPS collars, resource selection functions could justifiably be built at 2-4 week intervals. Kill Probability Associated with Cougar GPS Clusters We investigated 462 clusters during this pilot study (195 S1 clusters, 33 S2 clusters, 71 S3 clusters, 73 S4 clusters, 90 S5 clusters). The probability of locating cougar kills at GPS location clusters varied as a function of cluster type, cluster size, cougar status, and season (Table 1). As expected, S1 clusters were far more likely to be associated with cougar kills than S2 S5 clusters (Figure 3). The probability of a kill at an S1 cluster was 0.505 (95% CI: 0.435, 0.575), whereas kill probability was ≤ 0.12 at all other cluster types. There was some probability of a kill at S4 and S5 clusters, indicating that isolated cougar locations may periodically be associated with kills and should not be ruled out when using GPS location data to address cougar prey utilization. Kill probability increased as cluster size increased ( ˆ = 0.353, SE = 0.0706). Kill probability exceeded 0.9 with ≥ 10 locations/cluster and approached 1 with ≥ 15 locations/cluster (Figure 4). Thus, the probability of a kill was high if a cougar spent >2 days in the same general area, and a kill was essentially certain if a cougar spent >3 days in the same general area. Models receiving the most weight also provided evidence of interactions between cluster size and cougar status and between cluster size and season. The cluster size × cougar status interaction occurred because smaller cluster sizes were more likely to be associated with kills for female cougars than male cougars (Figure 5). For example, female cougars with ≥ 10 locations/cluster indicated a near-certain kill, whereas male cougars with 10 locations/cluster indicated only 0.571 probability of a kill (95% CI: 0.267, 0.830). Adult males were more likely to spend multiple days in an area without a kill than were adult females. The cluster size × season interaction occurred because larger cluster sizes during summer were less likely to indicate a kill than during other seasons (Figure 6). Perhaps cougars were more likely to remain sedentary without a kill nearby during summer months when energetic demands were lower. This result should be interpreted with caution, however, because we collected less data during summer than during other seasons. f/ Our primary reason for including season in the analysis was to evaluate possible differences in detection probability. We expected kills to be difficult to detect during winter and possibly spring months when carcasses and sign would be periodically covered by snow. However, our results did not support this hypothesis, but instead suggested that kills may have been the most difficult to detect during summer. Kills may be difficult to detect in summer range habitats because of extensive foliage or increases in scavenging by bears and/or coyotes. Regardless, carcass detection probability is a significant issue that underlies our entire analysis. That is, it is difficult to fully interpret our findings above without an adequate understanding of detection probability. For example, our summer results could reflect reduced carcass detection probability during summer, or they could reflect changes in cougar behavior during summer as compared to other months. A key point is that our estimates of kill probability for different cluster types and sizes are minimum estimates because these estimates assume detection probability was 1, which is unlikely. Detection probability should be addressed in subsequent research. Cougar Kill Composition Cougars killed adult deer, fawn deer, adult elk, and calf elk in nearly equal proportions (Figure 7). Each prey class comprised 0.22 0.24 of the total kill. Kill composition varied as a function of percent vegetative cover and elevation (Table 2). Adult elk were more likely to be killed in areas with little cover whereas calf elk, adult deer, and other species were more likely to be taken in habitats with heavier cover 94 �(Figure 8). Adult elk and adult deer were more likely to be killed at lower elevations whereas calf elk and other species were more likely to be killed at higher elevations (Figure 9). Unexpectedly, kill composition did not vary in response to cluster type or cluster size (Figure 10). Kill composition could be biased if S1 clusters, or larger cluster sizes, were associated with larger prey items, because it would suggest that larger prey may be more easily detected. However, given that kills of different sized prey occurred in roughly equal probabilities across all cluster sizes, restricting sampling to larger clusters would not necessarily bias kill composition estimates, at least for ungulates. Efficiency would be gained in the field by sampling larger clusters because they are more likely to be associated with kills. Additional data collection will be necessary to determine whether this preliminary finding is valid. Also, we urge caution interpreting this result because it is not biologically intuitive and would lead to biased kill composition data if proven incorrect. SUMMARY Over the past 2 years we have collected data on elk and deer distributions in conjunction with cougar predation data across the southern half of the Uncompahgre Plateau. Part of this effort included the development and implementation of a sampling based approach to estimate cougar kill rates and prey selection from GPS location data. Based on this effort we were able to randomly sample clusters of cougar GPS locations in relation to cluster type/size, which presumably correlates to prey selection and handling time. Mule deer and elk distributions on winter range were as expected with mule deer utilizing lower elevations and elk utilizing both lower and higher elevations with an area of overlap between the two species across deer winter range. Interestingly, cougar kill sites for mule deer generally occurred at midelevations within the range of overlap for deer and elk. Cougar kill sites for elk occurred at all elevations characteristic of elk distribution. As expected, cougar clusters with a large number of points had a high probability of being associated with a predation event and those with few points had a lower probability, especially single point clusters that are spatially distinct from other points. However, evidence of predation was identified at some of the spatially distinct single point clusters, indicating that these types of clusters are important in accurately describing cougar diet composition and predator/prey interactions. The association between cluster size and the probability of a cougar kill was related to season and cougar sex, with larger clusters being less predictive of a kill during summer and for males. Cougars killed elk and deer in approximately equal proportions and killed fawns/calves in equal proportion to adults for both deer and elk. Other prey items that could be detected at GPS locations comprised less than 10% of cougar diets. LITERATURE CITED Anderson, C. R., Jr., and F. G. Lindzey. 2003. Estimating cougar predation rates from GPS location clusters. Journal of Wildlife Management 67:307-316. Barrett, M. W., J. W. Nolan, and L. D. Roy. 1982. Evaluation of a hand-held net-gun to capture large mammals. Wildlife Society Bulletin 10:108-114. Bartmann, R.M. 1983. Composition and quality of mule deer diets on pinyon-juniper winter range, Colorado. Journal of Range Management 36:534-541. Bartmann, R. M., G. C. White, and L. H. Carpenter. 1992. Compensatory mortality in a Colorado mule deer population. Wildlife Monographs 121:5-39. Bergman, E.J., C.J. Bishop, D.J. Freddy and G.C. White. 2005. Evaluation of winter range habitat treatments on over-winter survival and body condition of mule deer. Wildlife Research Report, July: 23-35. Colorado Division of Wildlife, Fort Collins, USA. 95 �Bergman, E.J., C.J. Bishop, D.J. Freddy and G.C. White. 2008. Evaluation of winter range habitat treatments on over-winter survival and body condition of mule deer. Wildlife Research Report, July: In Press. Colorado Division of Wildlife, Fort Collins, USA. Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel inference: a practical information-theoretic approach. Second Edition. Springer-Verlag, New York, New York, USA. Creel, S. and N.M. Creel. 2002. The African wild dog: behavior, ecology and conservation. Princeton University Press, New Jersey, USA. Keeling, M.J., H.B. Wilson and S.W. Pacala. 2000. Reinterpreting space, time-lags, and functional responses in ecological models. Science 290:1758-1761. Krebs, C.J., S. Boutin and R. Boonstra, eds. 2001. Ecosystem dynamics of the boreal forest: the Kluane project. Oxford University Press, New York, USA. Kunkel, K. E., and D. H. Pletscher. 1999. Species-specific population dynamics of cervids in a multipredator ecosystem. Journal of Wildlife Management 63:1082-1093. Kunkel, K. E., T. K. Ruth, D. H. Pletscher, and M. G. Hornocker. 1999. Winter prey selection by wolves and cougars in and near Glacier National Park, Montana. Journal of Wildlife Management 63:901-910. Lima, S.L. and P.A. Bednekoff. 1999. Back to the basics of antipredatory vigilance: Can nonvigilant animals detect attack? Animal Behaviour 58:537-543. Logan, K.A. 2005. Cougar population structure and vital rates on the Uncompahgre Plateau, Colorado. Wildlife Research Report, July: 105-126. Colorado Division of Wildlife, Fort Collins, USA. Mathews, C.G., J.A. Lesku, S.L. Lima and C.J. Amlaner. 2006. Asynchronous eye closure as an antipredator behavior in the western fence lizard (Sceloporus occidentalis). Ethology 112:286-292. Mao, J.S., M.S. Boyce, D.W. Smith, F.J. Singer, D.J. Vales, J.M. Vore and E.H. Merrill. 2006. Habitat selection by elk before and after wolf reintroduction in Yellowstone National Park. Journal of Wildlife Management 69:1691-1707. Mitchell, W.A., and S.L. Lima. 2002. Predator-prey shell games: large-scale movement and its implications for decision making by prey. Oikos 99:249-259. Ramsey, C. W. 1968. A drop-net deer trap. Journal of Wildlife Management 32:187-190. Roth, T.C., II, S.L. Lima and W.E. Vetter. 2006. Determinants of predation risk in small wintering birds: the hawk's perspective. Behavioral Ecology and Sociobiology 60:195-204. Ryall, K.L. and L. Fahrig. 2006. Response of predators to loss and fragmentation of prey habitat: a review of theory. Ecology 87:1086-1093. Schmidt, R. L., W. H. Rutherford, and F. M. Bodenham. 1978. Colorado bighorn sheep-trapping techniques. Wildlife Society Bulletin 6:159-163. Schmitz, O.J. 2006. Predators have large effects on ecosystem properties by changing plant diversity, not plant biomass. Ecology 87:1432-1437. van Reenen, G. 1982. Field experience in the capture of red deer by helicopter in New Zealand with reference to post-capture sequela and management. Pages 408-421 in L. Nielsen, J. C. Haigh, and M. E. Fowler, editors. Chemical immobilization of North American wildlife. Wisconsin Humane Society, Milwaukee, Wisconsin, USA. Werner, E.E. and S.D. Peacor. 2006. Lethal and nonlethal predator effects on an herbivore guild mediated by system productivity. Ecology 87:347-361. Wilmers, C.C., E. Post, R.O. Peterson and J.A. Vucetich. 2006. Predator disease out-break modulates top-down, bottom-up and climatic effects on herbivore population dynamics. Ecology Letters 9:383-389. Prepared by Mathew W. Alldredge, Eric J. Bergman and Chad J. Bishop, Wildlife Researchers 96 �Table 1. Model selection results, based on Akaike‘s Information Criterion with small sample size correction (AICc), of an analysis evaluating the probability of locating cougar prey items (i.e., cougar kills) at GPS location clusters. We modeled cougar kills as a function of cluster type (type; S1, S2,…, S5), cluster size (size; no. GPS locations/cluster), cougar age and sex status (status), and season when cluster was investigated (season; spring, summer, fall, winter). Model Type size season size×season Type size status season size×status size×season Type size status season size×season Type size Type size season Size status season size×status size×season Size season size×season Type size status season Size status season size×season Size status size×status No. Parameters AICc Delta AICc Model Weight 12 16 14 6 9 12 8 11 10 6 365.61 367.84 369.68 370.75 371.80 373.04 374.80 375.71 378.87 380.07 0.00 2.22 4.07 5.13 6.19 7.42 9.19 10.10 13.25 14.46 0.615 0.202 0.081 0.047 0.028 0.015 0.006 0.004 0.001 0.000 Table 2. Model selection results, based on Akaike‘s Information Criterion with small sample size correction (AICc), of an analysis evaluating cougar kill composition at GPS location clusters. We modeled kill composition as a function of cluster type (type; S1, S2,…, S5), cluster size (size; no. GPS locations/cluster), cougar age and sex status (status), season when kill occurred (season; spring, summer, fall, winter), elevation (elev), and percent vegetative cover (cover). Model Elevation cover Elevation cover status Cover Elevation Size elevation Status Status season size elevation cover Size Season Status season size elevation No. Parameters AICc Delta AICc Model Weight 12 20 8 8 12 12 36 8 16 32 347.81 353.25 356.35 366.44 371.41 385.84 386.30 386.47 389.36 399.07 0.00 5.45 8.54 18.64 23.60 38.04 38.49 38.67 41.55 51.27 0.926 0.061 0.013 0.000 0.000 0.000 0.000 0.000 0.000 0.000 97 �,.. .v ,ow,o;o~ ........ '!!!I/J , -dT tt0mo1E ~,mm •~ LION CLOSURE " <• tOtOM ~ , ~ DrMSITT •• T Figure 1. Location of pilot predator-prey research on the Uncompahgre Plateau, southwest Colorado. Ongoing deer research study areas are reflected by red and blue polygons with hash marks, as well as by solid yellow polygons. The ongoing lion research study area is designated by the large red polygon. · L ·,:";..t:•: {~;(j ~:.{t;\ ,: l ; ':!\~.~ lJl .~_/>:/j ':~.,.• Figure 2. Distribution of cougar kill sites (red circles) in relation to mule deer winter range on the southeast portion of the Uncompahgre Plateau, Colorado. Black polygons represent segments of mule deer winter range where density estimates were either estimated or extrapolated to by surrounding areas on which estimates were measured. Gray lines represent Game Management Unit boundaries as designated by the Colorado Division of Wildlife. 98 �,:,/. • · ~ ,~~~ • ..,•·-i~·· ~ ~~lft .\=;·--~.}:1:-: ·· ::,_.~:, :·{ \,-; . ~~~7.:, ·~~- ' !:1 : i; :u:~·.: ,. ~ }t ~ f~~f- S1 S2 ., ,·,; ....Jil:::~-: l;Ii Figure 3. Distribution of cougar kill sites (red circles) in relation to GPS collar locations for 5 elk (black circles) on the southeast portion of the Uncompahgre Plateau, Colorado. Gray lines represent Game Management Unit boundaries as designated by the Colorado Division of Wildlife. 0.7 0.6 0.5 :-2 .... 0 > 0.4 ~ :.cro .0 0.3 ~ ci. 0.2 0.1 0 S3 S4 S5 Cluster Type Figure 4. Probability of a cougar kill at different types of GPS location clusters (with 95% CIs), Uncompahgre Plateau, Colorado, 2006 2008. Refer to the Methods section for a detailed explanation of cluster types. 99 �1 0.9 0.8 0.7 :.2 .... 0.6 0 > ~ :.cro 0.5 .0 ... 0.4 0 c.. 0.3 0.2 0.1 0 1 6 11 16 21 26 31 36 No. Locations in Cluste r Figure 5. Probability of a cougar kill as a function of the number of locations in a GPS cluster (with 95% CI), Uncompahgre Plateau, Colorado, 2006 2008. ... -- -- l o.s 0.8 0.7 ....~ 0.6 g 0 .5 0 ii ro .,Q ... 0.4 0 Q,. 0.3 - AdFemSingle 0.2 - - Ad f-emlubs 0.1 -- AdMale 0 0 5 10 15 20 No. Cluste r Locations 25 30 35 Figure 6. Probability of a cougar kill at GPS location clusters relative to sex and reproduction status (with 95% CIs), Uncompahgre Plateau, Colorado, 2006 2008. Cougar status was defined as single adult female (AdFemSingle), adult female with cubs (AdFemCubs), or adult male (AdMale). 100 �1 0.9 0.8 0.7 ~ 0.6 0 ~ 0.5 - - - Spri ng :c10 ..c e 0.4 - ~ 0.3 - Winter - - Fal l 0.2 - 0.1 - Summer - -- -- -- 0 0 5 10 15 20 25 30 35 No. Cluster Locations Figure 7. Probability of a cougar kill at GPS location clusters by season (with 95% CIs), Uncompahgre Plateau, Colorado, 2006 2008. 0 .35 0.3 - ,_ 0.2 5 :;2 - iv 0 I- 0 .2 0 C: 0 ·z..... 0.1 5 0 C. -- ,- -- - 0 .1 - 0 .05 - 0 ..... ~ - - -- 0 AdDeer AdElk FwnDeer CalfElk Other Figure 8. Prey composition of cougar kills (with 95% CIs) on the Uncompahgre Plateau, Colorado, 2006 2008. Prey items included adult deer (AdDeer), ≥ 6-month-old fawn deer (FwnDeer), adult elk (AdElk), calf elk (CalfElk), and other species (e.g., porcupine, turkey, coyote). 101 �0.8 0.7 0.6 I -Adult Elk I 0.5 0.4 0.3 0.2 0.1 0 Predicted probabilities 0 20 40 60 80 100 0.6 0 .5 - Calf Elk 0.4 -Adult Dee~ - 0.3 0.2 0. 1 0 0 20 40 60 80 100 60 80 100 0.8 0.7 0.6 -Fawn Deer 0.5 - Other 0.4 0.3 0.2 0.1 0 0 20 40 Percent cover Figure 9. Predicted prey composition of cougar kills as a function of vegetative cover (with 95% CIs), Uncompahgre Plateau, Colorado, 2006 2008. 102 �0.7 - - - - - - - - - - - - - - - - - - - 0.6 - + - - - - - - > , - - - - - - -- - - - - - - -- - - - -Adult Deer 0.5 -+------'<-- - > o r - - - - - -----< Predicted probabilities - Adult Elk 0 + - - - ~ - - ~ - - - - - ~ _ _ _ _ _ : := - - - - - , r - - - -, - - 1700 1900 2100 2300 2500 2700 2900 1 0.9 -Fa wn Deer , , I 0.8 I - 0.7 I Calf Elk I I - -Other 0.6 0.5 0.4 ,,, 0.3 ~ 0.2 0.1 --- --- 0 1700 1900 2100 2300 2500 2700 2900 Elevation Figure 10. Predicted prey composition of cougar kills as a function of elevation (m) (with 95% CIs), Uncompahgre Plateau, Colorado, 2006 2008. 103 �0.8 0.7 - Fawn Deer 0.6 - Ca lf Elk 5 10 - Adult Deer - Adult Elk 0.5 0.4 0.3 Predicted probabilities 0.2 0.1 0 0 15 20 25 30 35 30 35 0.6 0.5 - - Other 0.4 0.3 0.2 0.1 ------- .......... 0 0 5 10 - --- -- - - --15 20 25 --- No. locations/cluster Figure 11. Predicted prey composition of cougar kills as a function of the number of locations comprising a GPS location cluster (with 95% CIs), Uncompahgre Plateau, Colorado, 2006 2008. 104 � https://cpw.cvlcollections.org/files/original/b4987169ef2fa5ba06e7e0af8454e131.pdf eb60656e45a79c8ce076d06579b82a7c PDF Text Text 65 Colorado Division of Wildlife Wildlife Research Report July 2001 and July 2002 JOB PROGRESS REPORT State of_ _ _ _ _ _C~ol~o~ra~d~o_ _ _ __ Mammals Research Program Work Package No. _ _~30~0~1~----- Deer Conservation Task No._ _ _ _ _ _____,_4_ _ _ _ _ __ Effect of Nutrition and Habitat Enhancements on Mule Deer Recruitment and Survival Rates Research and Development Project No._____W~-=15~3~-=R~---Period Covered: July 1, 2001 -June 30, 2002 Authors: C. J. Bishop and G. C. White Personnel: D. L. Baker, T. D. I. Beck, G. Bock, S. K. Carroll, D. Coven, K. Crane, D. J. Freddy, L. Gepfert, R. B. Gill, R. Harthan, M. McLain, E. P. Myers, G. C. Miller, J. Olterman, J. A. Padia, T. M. Pojar, C. M. Solohub, B. E. Watkins, CDOW; L. H. Carpenter, WMI; J. Sazrna, B. Welch, BLM, Montrose, CO. ABSTRACT To further understand the factors that caused deer numbers to decline in western Colorado during the 1990s, we designed and initiated a field experiment to measure deer population parameters in response to nutrition and habitat enhancement treatments. During November 2000-March 2002, we captured and radio-collared 112 mule deer in a treatment unit and 109 mule deer in a paired control unit during winter · on the Uncompahgre Plateau in southwest Colorado. We enhanced the nutrition of deer in the treatment unit by providing a safe, pelleted supplemental feed on a daily basis from December through April each winter. Early winter fawn:doe ratios were measured using helicopter and ground classification surveys the year following treatment delivery to determine whether fawn production and survival increased as a result of enhanced nutrition of adult females. Based on multiple age classification surveys, we concluded that the winter nutrition enhancement treatment did not cause an increase in neonatal production and survival during 2001. However, fawn production and summer-fall survival were atypically good during 2001, and not representative of most years during the past decade when the population declined. We also measured overwinter fawn survival rates in response to the treatment. The simplest model which effectively explained survival (x\ 1 = 51.87, P = 0.440) included treatment (x21 = 9.95, P = 0.002) and early winter fawn mass (x21 = 8.33, P = 0.004). From December 1, 2001, through May 31, 2002, the survival rate of fawns was significantly greater (x 21 = 13.216, P < 0.001) in the treatment unit (0.865, SE = 0.056) than in the control unit (0.510, SE = 0.080); and fawns that survived the winter averaged 2.9 kg heavier than fawns that died (F1 = 6.11, P = 0.016). Early winter fawn mass was not different among treatment and control fawns (F1 = 0.36, P = 0.550), thus the effect of the treatment was not confounded with fawn mass. Simply, heavier fawns in both experimental units had higher survival probabilities. During•winter2001-02, which was a mild to average winter, the nutrition enhancement treatment clearly improved overwinter fawn surviv;il, and thus yearling recruitment. We will continue this portion of the resea.n;:h for 2 more years. The results reported here are preliminary and should be treated as such. ~.,,.,. 7:,,-. . • • . ' .. :,: '-~:: .' • • . • . ·-- • • r~11Hrnm~i111mm1li BDOWD16779 ��67 EFFECT OF NUTRITION AND HABITAT ENHANCEMENTS ON MULE DEER RECRUITMENT AND SURVNAL RATES C. J. Bishop and G C. White P. N. OBJECTIVES 1. To determine experimentally whether enhancing mule deer nutrition during winter and early spring by supplemental feeding increases December fawn:doe ratios and overwinter fawn survival. 2. To determine experimentally to what extent habitat treatments replicate the effect of enhanced nutrition from supplemental feeding. SEGMENT OBJECTIVES 1. Capture and radio-collar target sample of adult female mule deer and 6-month-old fawns. 2. Deliver nutrition enhancement treatment to all deer occupying the treatment area. 3. Measure overwinter adult female and fawn survival rates and early winter fawn:doe ratios in the treatment and control areas. INTRODUCTION Mule deer numbers apparently declined during the 1990's throughout much of the West, and have clearly decreased since the peak population levels documented in the 1940's-60's (Gill et al. 1999, Unsworth et al. 1999). Biologists and sportsmen alike have concerns as to what factors may be responsible for declining population trends. Although previous and current research indicates that multiple interacting factors are responsible, habitat and predation have received the focus of attention. A number of studies have evaluated whether predator control increases deer survival, yet results are highly variable (Connolly 1981, Ballard et al. 2001). Together, predator control studies with adequate rigor indicate that predation effects on mule deer are variable as a result of time-specific and site-specific factors. Studies which have demonstrated deer population responses to predator control treatments have failed to determine whether predation is ultimately more limiting than habitat. Numerous research studies have evaluated mule deer habitat quality, but virtually no studies have documented population responses to habitat improvements. In many areas where declining deer numbers are of concern, predation is common yet habitat quality appears to have declined. The question remains as to whether predation, habitat, or some other factor is more limiting to mule deer in these situations, and whether habitat quality can be improved for the benefit of deer. It may also be that no single factor is any more or less important than another, and that a more comprehensive understanding of multi-factor interactions is paramount. We designed a field experiment to measure deer population responses to nutrition and habitat enhancement treatments, to further understand the causative factors underlying observed deer population dynamics. We are conducting the study on the Uncompahgre Plateau, where several predator species (i.e. coyotes, mountain lions, and bears) are present in abundant numbers. In addition to predation, myriad diseases in combination proximately affect survival of the Uncompahgre deer population (Pojar 2000, B.E. Watkins, unpublished data). Predator numbers have not and will not be manipulated in any manner during the course of the study. All factors have been left constant with the exception of deer nutrition and habitat. Deer nutrition is being enhanced by providing supplemental feed to deer during the winter. If December fawn:doe ratios and overwinter fawn survival improve as a direct result of the nutrition enhancement treatment, then we can presume that deer nutrition is ultimately more limiting than . predation or disease. The second phase of the field experiment will incorporate habitat manipulation �68 treatments, which will consist of prescribed fire or mechanical techniques to set back succession of pinyon-juniper habitat in an effort to improve the vigor and quality of winter habitat for mule deer. Deer population responses will be measured in relation to the habitat manipulations in the same manner as the supplemental feed. Thus, the experiment allows us to determine whether nutritional quality of habitat is ultimately more limiting than other factors in a late-seral pinyon-juniper/sagebrush landscape, and if so, whether habitat can be effectively improved for mule deer. The results will also advance our current understanding of multi-factor interactions, with direct implications for mule deer management. MATERIALS AND METHODS Experimental Approach Experimental Design and Study Area We non-randomly selected four areas on the Uncompahgre Plateau to create 4 experimental units (A-D) (Fig. I). Treatments were randomly assigned to the experimental units. The following criteria were used to select experimental units: I.). Deer densities (~50-80 deer/mi2): areas were selected where deer densities were sufficient to meet sample size requirements within the experimental unit, while simultaneously selecting areas that would require feeding less than ~500 animals during a normal winter 2.) Buffer zones: areas were selected such that experimental units would be separated by several miles of non-treatment area (buffers) to prevent deer from occupying more than one experimental unit 3.) Similarity: areas were selected that comprise relatively similar habitat complexes and deer densities that are representative of the overall area 4.) Elk pqpulations: areas were selected to minimize the number of elk present during normal winters Units A and B are receiving the nutrition enhancement treatment in a cross-over experimental design, and are being used to address P .N. Objective I. Unit A served as the treatment unit, while Unit B served as the control, for the first 2 years of research (2000 - 2002). Beginning November 2002, Unit B will receive the treatment while Unit A will serve as the control. Upon completion of P.N. Objective I, Units C and D will be used to conduct phase 2 of the research, or P.N. Objective 2. Habitat in one unit will be manipulated to set back plant succession (treatment), while habitat in the other unit will remain unchanged (control) throughout the experiment. Year Unite UnitD 2004-05 Control Control 2005-06 .. Control Fire/Mechanical Treat. 2006-07 • ':- /Control 2000-01 2001-02 2002-03 2003-04 2007-08 Control.· · v~-g Response Yr 1 Veg Response Yr 2 2008-09 ,:.( .Control Veg Response Yr 3 2009-10 Control Veg Response Yr 4 Figure 1. Schematic representation of experimental units and associated treatments. The nutrition enhancement cross-over design will encompass 4 years; monitoring in the habitat manipulation experimental unit and paired control area will encompass approximately 6 years. • �69 The 2 experimental units (A and B) receiving the nutrition enhancement treatment are (Figs. 2 and 3): (1) The Colona Tract ofthe Billy Creek State Wildlife Area (2) Bureau of Land Management lands adjacent to Shavano Valley as defined by the following: Within Dry Creek Basin Quadrangle (USGS 7.5 Minute), includes Sections 6 and 7 in T. 48 N.-R. 10 W. and Sections 1, 2, 10, 11, 12, 13, 14, 15 in T. 48 N.-R. 11 W. This area roughly includes 38°25'00" -38°27'30" Latitude and 108°00'00" - 108°04'30" Longitude. , . r-' esa County Gunnison County Figure 2. Location of Colona and Shavano (Units A and B) experimental units in Grune Management Unit 62 on the Uncompahgre Plateau, southwest Colorado. �70 Figure 3. Colona and Shavano experimental units (Units A and B), located in Game Management Unit 62 on the Uncompahgre Plateau, southwest Colorado. Polygons represent the nucleus of each experimental unit, which is where animals have been collared and the nutrition enhancement treatment delivered. �71 Response Variables The primary response variable is fawn:doe ratios measured during December and January following the previous winter's treatments. The fawns counted during early winter age classification were born the summer following the winter treatments, and classified when they were 6 months of age. Thus, we are measuring the effect of enhanced adult doe nutrition during winter on subsequent fawn production and survival. Fawn:doe ratios are currently being measured in Units A and B corresponding to P.N. Objective I. The second response variable is overwinter fawn survival, measured from radio-collared fawns during the winter in direct response to the enhanced winter nutrition treatment. We are also measuring overwinter and annual survival of adult does as a function of enhanced winter nutrition. Sample Size The primary response variable is the mean fawn:doe ratios of the radio-collared does wintering on the experimental unit of interest. We desired to detect an effect size, i.e., an increase in fawn:doe ratios in response to the treatments, in the range of 15 to 20 fawns per 100 does. These values were based on simple population models with overwinter fawn survival of 0.444, adult female survival of 0.853, and December fawn:doe ratios of 66 fawns per 100 does to obtain a stationary population (Unsworth et al. 1999). Based on surveys of DAU D-19 from 1992-96, the standard deviation of the fawn:doe ratio for groups with at least one adult female was 57, with a mean of 41. Using an expected standard deviation of 57, the standard error of the mean fawn:doe ratio for 40 radio-collared does is 57/(40 112) = 9.0, which is the expected standard deviation of measured fawn:doe ratios on each of the experimental units. We assessed power of the proposed experiment using SAS Analyst®. We used a two-sample t-test with a sample size of 4, representing the years of the study where treatment effects will be measured. The power of the design to detect an increase of 2 0 fawns per I 00 does is about O. 87. A sample size of 40 fawns per experimental unit per year provides a power of 0.81 to detect a difference of0.15 in survival between 2 experimental units if survival on the control unit is 0.40. We expected to see an increase in fawn survival (effect size) of approximately O.15, because this was the difference measured in the density reduction experiment conducted by White and Bartmann (1998). Capture Methods Deer were captured using baited drop nets (Ramsey 1968, Schmidt et al. 1978) and helicopter net guns (Barrett et al. 1982, van Reenen 1982). Drop nets were baited with certified weed-free alfalfa hay and apple pulp. Drop nets were used as the principal capture technique during a 3-week capture period; helicopter net-gunning was used at the end of the capture period to secure the remainder of deer needed to meet our target sample sizes. All deer were hobbled and blind-folded after being captured. All deer captured via drop nets were carried away from the net to an adjacent handling site using stretchers. Deer were fitted with leather radio collars equipped with mortality sensors, which cause an increase in pulse rate after remaining motionless for 4 hours. Permanent collars were placed on adult females, while temporary collars were placed on fawns. To make collars temporary, one end of the collar was cut in half and reattached using rubber surgical tubing; fawns shed the collars ~6 months post-capture. A rectangular piece of flexible plastic (Ritchey® neck band material) engraved with a unique identifier was stitched to the side of each collar. The unique identifier consisted of 2 symbols for adult females, and only 1 symbol on 2 different colors of plastic for fawns. The identifiers were necessary to visually identify deer from the ground. This has allowed us to effectively document use of the treatment, measure fawn:doe ratios from the ground, and assess experimental unit population size via mark-resight estimators. We recorded the weight, hind foot length and chest girth of each deer, and collected blood samples from most does and fawns to evaluate disease prevalence. �72 Measurement of Fawn:Doe Ratios and Overwinter Survival Each winter we used the radio-collared does to measure fawn:doe ratios in each experimental unit. The resulting fawn:doe ·ratio is a measurement of the previous year's treatment effect. We measured fawn:doe ratios using 2 techniques: (1) We located the sample ofradio-collared does in each experimental unit from a fixed-wing airplane, and used the set of locations to define boundaries for the experimental unit. Shortly after (i.e. 1-2 days), we used a helicopter to systematically fly the defined unit and classify all deer groups encountered. For each group, we documented whether a radio-collared doe was present. (2) We located each radio-collared doe by radio telemetry from the ground. The group of deer with the collared doe was counted and classified by age and sex. Both methods have been employed to gather as much information as possible to determine whether there was a treatment effect. The "true" value cannot be measured perfectly because of the inherent biases and potential sources of error associated with each technique. Thus, by employing both techniques, we have a greater chance of fully understanding whether the treatment caused an effect. We measured survival by radio-monitoring collared deer to determine fate (live/mortality). We also attempted to determine the cause of each mortality, with a primary goal of distinguishing between predation and non-predation mortality causes. Deer were radio-monitored on a daily basis during the winter, which typically allowed us to arrive at mortality sites within 24 hours. Treatment Delivery Deer nutrition was enhanced in the treatment area by providing a safe, pelleted supplemental feed. The supplemental feed was developed through extensive testing with both captive and wild deer (Baker and Hobbs 1985, Baker et al. 1998), and has been safely used in both applieq research and management projects. Pellets were distributed daily using 4wd pickup trucks and ATVs on primitive roads throughout the experimental unit to provide a food source for the entire deer population in the treatment unit. Each 501b. bag of pellets was carried :S:;200m from the truck/A TV and distributed by hand in approximately 2030 small piles of feed in a linear fashion. Numerous bags were distributed in successive order allowing us to create a line of feed that spanned most of the treatment area, which prevented animals from concentrating in any single location. This feeding technique also prevented dominant animals from restricting access to the food supply because of the large area over which pellets were distributed. We attempted to supply pellets ad Iibitum such that a small residual remained when the next day's ration was provided. Collared deer were closely monitored to ensure that treatment deer remained in the experimental unit and actually consumed the feed, and to make sure that non-treatment deer remained in the control unit, which they did. Treatment deer that did not regularly consume the feed were withdrawn from the sample for purposes of measuring treatment effects. The pelleted ration was commercially produced in the form of 2x 1x0.5-cm wafers (Baker and Hobbs 1985). Feed constituents (i.e. digestibility, protein, gross energy etc.) vastly exceeded those of typical winter range deer diets; exact constituent values are provided by Baker et al. (1998). When provided ad libitum, the feed should have allowed deer to meet or exceed nutritional requirements for growth and maintenance (Ullrey et al. 1967, Verme and Ullrey 1972, Thompson et al. 1973, Smith et al. 1975, Baker et al. 1979, Holter et al. 1979). The basis for feeding such high quality pellets was to ensure that the treatment (enhanced nutrition) was effectively delivered to the deer. Our intent was not to determine the exact level of nutrition necessary to increase fawn recruitment, but rather to determine if nutrition is a limiting factor to recruitment. If nutrition is in fact limiting, we will rely on the habitat manipulation treatment to evaluate what exactly can be done via management to increase fawn survival and recruitment. �73 Habitat Manipulations In order to accomplish P.N. Objective 2, habitat will be manipulated in experimental unit D through collaboration with the Uncompahgre Ecosystem Restoration Project (UP), which comprises personnel from the Division of Wildlife, U.S. Forest Service, Bureau of Land Management, Public Lands ' Partnership, and a variety of other public and private stakeholders. The UP committee is using an experimental landscape approach to manipulate various habitats in a mosaic pattern throughout the Uncompahgre Plateau. We will focus our intensive deer monitoring on one of these habitat manipulations that will be conducted in experimental unit D. This portion of the research has not yet been initiated. A complete description of our planned protocols to accomplish P.N. Objective 2 is provided in the Program Narrative (Bishop and White 2000). Statistical Methods Once data collection is completed for the full study, we will test for differences in fawn:doe ratios between experimental units and years using the following statistical model: YiJk = µ + a.1 + /3k + a/3.ik + e;Uk), where YiJk = fawn:doe ratio for the ith deer group in treatment combinationjk; i = 1, 2, ... , n1k(deer groups);}= 1, 2, 3, 4 experimental units (control, supplemental feed, habitat manipulation); k= 1, 2, 3, 4 (6) years; a~k = interactions among experimental units and years; and e;Uk) = random error associated with YiJk• A similar model will be used to analyze overwinter fawn survival, but a logit-link function will be used in place of the identity link function in the above general linear model. A similar model will also be used to test for differences in fawn weights, except the response variable will be fawn mass, and sex and fate (i.e. lived or died) will be included in the model as independent variables. For this progress report, a preliminary fawn:doe ratio analysis was completed using PROC MIXED in SAS (SAS Institute 1997). We used a reduced model with experimental unit as the lone independent variable, and considered experimental unit as a fixed effect and radio-collared does within an experimental unit as random effects. Survival rates were calculated using a Kaplan-Meier survival analysis (Kaplan and Meier 1958, Pollock et al. 1989), and contrasted among experimental units and sexes using a chi-square analysis. We modeled winter fawn survival with a product multinomial model (Grizzle et al. 1969) using PROC CATMOD in SAS (SAS Institute 1989a). Survival was modeled as a function of experimental unit, sex, and capture mass. We used a general linear model in PROC GLM in SAS (SAS Institute 1989b) to test for differences in fawn mass between experimental units, sexes, and fates (i.e. lived or died). Other results in this report are presented as data summaries incorporating means and standard errors, or in some cases, raw data values. These results are incomplete and preliminary in nature, and should be treated as such. RESULTS AND DISCUSSION Deer Capture During November-December 2000, we captured and radio-collared 73 adult female mule deer: 37 in the treatment unit and 36 in the control unit. Due to budgeting constraints, we were unable to capture and radio-collar fawns. During November-December 2001, we captured and radio-collared an additional 32 adult females to replace mortalities from the previous year and to buffer our sample size, resulting in a total of 45 radio-collared does in each experimental unit. We also captured and radio-collared 80 fawns: 40 in each experimental unit. During February 28 -March 1, 2002, we captured an additional 36 does (18 in each experimental unit) as part of a related research project (Bishop et al. 2002). In total, we radiomonitored 221 mule deer (141 adult does and 80 fawns) during November 2000-June 2002. �74 Treatment Delivery 2000-01 From December 15, 2000, through April 19, 2001, we distributed 88 tons of the pelleted ration. For most of the winter and spring, on average, we distributed 0.85 tons offeed each day throughout 22 feeding sites across the 2.3 mi 2 treatment unit. Deer were fed ad libitum because there was always residual feed remaining the next day during the feeding routine. Each sack was distributed in approximately 20-30 distinct, small piles, resulting in > 1000 small piles of feed throughout the treatment unit. This effort allowed deer to effectively access the feed in small groups, and no aggression was ever observed among deer seeking access to the feed. By distributing the feed in this manner, we were able to avoid the negative aspects associated with large-scale feeding operations. Deer adapted to the pelleted supplement right away and utilized it extensively throughout the winter. We continually monitored deer use of the feed from ground observation points, where we obtained 440 visual observations of radio-collared does consuming the feed. These observations, coupled with daily radio-monitoring and periodic aerial relocations, indicate 32 of the 37 radio-collared treatment does spent the entire winter and spring within the boundaries of the treatment unit and received the supplement on a daily basis. Mark-resight population estimates from March helicopter (489 deer, SE = 62) and ground (494 deer, SE = 81) surveys, coupled with feed consumption, indicate we fed roughly 450 to 500 deer during most of the winter and spring. Feed consumption declined coincident with spring green-up, although deer continued to use the feed through mid-late April, at which point they began migrating to summer range. We also fed approximately 25 to 30 elk, but the elk did not affect deer access to the feed. Deer in the control experimental unit did not receive feed or any other treatment. Based on helicopter mark-resight surveys, the deer density in the treatment unit in December was 120 deer/mi2 (SE= 9), but increased shortly after and was 213 deer/mi2 (SE= 27) in March. Deer densities in the control unit changed little from 83 deer/mi2 (SE= 12) in December to 101 deer/mi 2 (SE= 14) in March. 2001-02 From December 15, 2001, through April 25, 2002, we distributed 194 tons of the supplement throughout the treatment unit. For most of the winter and spring, we distributed 2.0-2.1 tons of feed each day. The dramatic increase in supplement distribution from the previous year occurred because a large number of elk descended into the Uncompahgre Valley during mid-late fall/early winter. Elk arrived in unusually large numbers throughout much of the valley prior to the onset of treatment delivery. Once feeding was initiated, approximately 300-500 elk adapted to the feed and remained in or around the 2.3 mi2 treatment unit throughout most of the winter. Given myriad logistical and budgetary constraints, 2.1 tons was the maximum amount of feed we could routinely deliver on a daily basis. Feed was not delivered ad libitum to all deer and elk in the treatment unit throughout the winter because residual feed was rarely observed during the next day's distribution. However, daily field observations indicate most deer approached ad libitum consumption of the supplement. In contrast to the previous winter, deer were waiting for the daily supplement to arrive each morning. Deer then consumed the supplement immediately after it was distributed. Elk were rarely observed utilizing th·e feed until late morning or afternoon, and elk continued to forage in fields below the treatment unit, whereas deer did not. We observed numerous radio-collared deer consuming the pelleted supplement each day; not all of these observations were recorded because of time constraints with distributing the feed. Given this time limitation, we still recorded 818 observations of radio-collared deer consuming the supplemental feed (497 collared doe observations and 321 collared fawn observations). Most days, > 100 and sometimes 200-300 deer were observed utilizing the pellets during the course of distributing the supplement. These observations rarely included elk; thus, deer-elk competition was minimized because of temporal differences in feeding, and deer clearly had first access to the feed. �75 Fawn:Doe Ratios In December 2000, at the beginning of the study and prior to the first year's treatment delivery, fawn:doe ratios were similar in the 2 experimental units. Pre-treatment fawn:doe ratios were 52.6 fawns: 100 does (SE= 5.3) in the treatment unit, and 51.6 fawns: 100 does (SE= 5 .0) in the control unit. In late December 2001 and early January 2002, following the first year's treatment, we conducted 2 age classification helicopter surveys in the treatment and control units. On 12/23/01, we observed 52.8 fawns:100 does (SE = 6.7) in the treatment unit, and 36.7 fawns:100 does (SE= 3.8) in the control unit. On 1/8/02, we observed 54. 7 fawns: 100 does (SE= 6.6) in the treatment unit, and 50.5 fawns: 100 does (SE= 6.0) in the control unit. During December 2001 -February 2002, we obtained fawn:doe ratio estimates from ground observations of radio-collared deer groups for both treatment and control deer. This survey resulted in 61.2 fawns:100 does (SE= 7.8) in the treatment unit, and 74.5 fawns:100 does (SE= 8.5) in the control unit, although the result was not statistically significant (t74 = 1.16, P = 0.249). The fawn:doe ratio results are conflicting, and clearly do not provide evidence that there was any treatment effect. In short, we conclude that the nutrition enhancement treatment did not cause an increase in neonatal production and survival during 2001. However, our results, in conjunction with a December estimate of 64 fawns: 100 does for the entire Uncompahgre deer population (B.E. Watkins, unpublished), indicate fawn production and survival was good during 2001. The observed fawn:doe ratios coupled with overwinter fawn survival and annual adult survival rates indicate the deer population is growing. Considering the past 1-2 decades, this was an atypically good year for the Uncompahgre deer population. It would appear that whatever set of environmental conditions have led to a declining deer population were not present during 2001 in the same manner as in the past. Our main interest lies in observing the effect of the treatment on the deer population in a year where fawn:doe ratios are lower for the population as a whole, similar to what they have been much of the past 15 years. Our results point out the inherent difficulties and biases associated with precisely measuring fawn:doe ratios, particularly in this research study. Ratios obtained from helicopter surveys were based on 2 shortduration flights over small spatial units. Helicopter surveys were complicated by high deer densities in heavy cover, making both deer detection and fawn:doe classifications a considerable challenge. There is a variety of potential biases that may have affected the helicopter surveys, including differential sightability of does and fawns, and incorrectly classifying yearling bucks as adult does. These biases are likely real considering the higher ratios measured during the ground classifications based on the radiocollared does. Ground fawn:doe ratio observations of radio-collared doe groups were made using spotting scopes and field glasses, where we commonly studied the deer for some time. Incorrect classifications during these surveys were likely minimal. For example, small-antlered yearling bucks (e.g. 3 - 6" spikes) were detected from the ground, whereas they were clearly missed on occasion during helicopter surveys. The ground classifications were also preferable to the helicopter surveys in that we obtained repeated observations. We recorded as many as 5 separate ratio observations per radio-collared doe. Overall, we believe the ground fawn:doe ratio estimates, based on individual radio-collared does, provided less biased measurements. Given the inherent difficulties of measuring fawn:doe ratios, and the lack of a clear indication as to the effectiveness of the treatment, we initiated a second study using vaginal implant transmitters in order to capture and radio collar newborn fawns from the radio-collared treatment and control does (Bishop et al. 2002). This new aspect of the research will allow us to gain better estimates of the treatment effect on subsequent fawn production and survival, evaluate cause-specific mortality of treatment/control neonates, and simultaneously provide a greater understanding as to the mechanisms affecting the deer population. Survival Adult Fema/,es During winter 2000-01 (Dec 1, 2000 - May 31, 2001 ), the adult doe survival rate of deer in the treatment unit (0.968, SE= 0.032) was greater (x21 = 2.649, P = 0.104) than the survival rate of deer in the control �76 unit (0.861, SE = 0.058). However, annual adult doe survival rates (Dec 1, 2000 - Nov 30, 200 I) were similar among the treatment and control deer (Trt: S(t) = 0.839, SE= 0.066; Control: S(t) = 0.833, SE= 0.062; x21 = 0.004, P = 0.94 7). Thus, mortalities of control deer occurred primarily during the winter months, while treatment does died primarily during the summer and fall months. During winter 2001-02 (Dec I, 2001 - May 31, 2002), the adult doe survival rate of deer in the treatment unit (0.942, SE= 0.030) was once again greater (x2 1 = 3.116,P = 0.078) than the survival rate of deer in the control unit (0.848, SE= 0.044). At this preliminary stage in the research, the nutrition enhancement treatment has apparently increased survival of adult females during the winter, but the overall annual survival among treatment and control does has not varied. The annual survival rate of does measured thus far aligns with expected survival based on other studies (Unsworth et al. 1999, B.E. Watkins, unpublished). Fawns During winter 2001-02 (Dec 1, 2001 - May 31, 2002), the survival rate of fawns was significantly greater (X 21 = 13 .216, P < 0.001) in the treatment unit (0.865, SE= 0.056) than in the control unit (0.510, SE= 0.080) (Fig. 4). The simplest model which effectively explained survival (x\, = 51.87, P = 0.440) included treatment (x2 1 = 9.95, P = 0.002) and early winter mass (x\ = 8.33, P = 0.004). Fawns receiving the nutrition enhancement treatment, and heavier fawns, had higher survival probabilities. During winter 2001-02, which was a mild to average winter, the nutrition enhancement treatment clearly improved overwinter fawn survival, and thus yearling recruitment. 1 -. I I - -- • I - - I I - 0.9 -■ - ---- -- - ---- - - ------ -- - 1~ 0.8 l L_ I I 0.6 7_ -- • Colona (Treatment) 0.5 --Shavano (Control) 0.4 12/1 12/16 12/.H. 1/15 1/30 .2/14 3/1 3/1.6 3/31. 4./15 I I 4/30 5/15 5/30 Figure 4. Overwinter fawn survival (Dec 1, 2001 - May 31, 2002) in a nutrition enhancement treatment unit (S(t) = 0.865, SE= 0.056) and a control unit (S(t) = 0.510, SE= 0.080), Uncompahgre Plateau, southwest Colorado. Causes of Mortality Adult Females During winter 2000-01, only one adult doe from the treatment unit died, which was road-killed in early May. During June-August, 2001, an additional 4 treatment deer died: 2 from unknown causes that were �77 not predator-related, 1 from a prolapsed uterus, and 1 unknown. In contrast, 5 adult does from the control area died during winter 2000-01 : 3 from malnutrition, 1 from mountain lion predation, and 1 was roadkill ed. One additional deer from the control area died during August 2001 from an unknown cause that was not predator-related. During winter 2001-02, 3 adult does from the treatment unit died: 1 from secondary causes related to chronic arthritis, 1 from predation, and 1 road-killed. The predation and road kill mortalities occurred in mid-late May after deer had left the treatment unit. In June, 2 more treatment adult does died during the fawning period. One doe died while giving birth, and the second doe died from an unknown cause seemingly related to birthing. This second doe, based on a field necropsy, had considerable fat and seemed otherwise healthy. On March 1, 2002, we measured 2 fetuses in utero (Bishop et al. 2002), yet the doe had only 1 fetus in utero when she died. Given her good condition, it is unlikely the second fetus was reabsorbed, which indicates she had already passed the first fetus and died prior to giving birth to the second. During winter 2001-02, 7 adult does from the control unit died: 4 died from mountain lion predation, 1 from entanglement in a fence, 1 from an unknown, non-predator mortality, and 1 unknown. No additional control does died during the month of June. Fawns Five fawns in the treatment unit died during winter 2001-02: 2 from malnutrition/sickness and 3 from disease. Of the 2 fawn mortalities caused by malnutrition/sickness, 1 was a result of basic malnutrition and occurred on December 31, 2001. The other fawn had a combination of heavy parasite loads, scours, and general poor condition. Each of the 3 fawns that died from disease had adequate fat stores. At least one of these fawns died as a result of pneumonia. In the control unit, 19 fawns died during the winter: 5 from malnutrition, 6 from mountain lion/bobcat predation, 4 from coyote/canine predation, 3 unknown predation mortalities, and 1 unknown. A majority of the fawns killed by predators had virtually no femur marrow fat remaining, indicating the predation was likely compensatory in nature. Fawn Mass During winter 2001-02, the early winter mass of radio-collared fawns varied significantly between sexes (F1 = 15.32, P < 0.001) and fates (F1 = 6.ll, P = 0.016). Males averaged 3.6 kg heavier than females, and fawns that survived the winter averaged 2.9 kg heavier than fawns that died. Early winter mass was not different among experimental units (F1 = 0.36, P = 0.550), thus the effect of the treatment was not confounded with fawn mass. The interaction of experimental unit x sex x fate was also significant (F1 = 5.80, P = 0.019), while all other 2-way interactions were not significant. The 3-way interaction occurred because in the control experimental unit, female fawns that survived were not heavier than female fawns that died (Survived: x = 31.0 kg, SE= 1.77; Died: x = 31.5 kg, SE= 1.03); whereas male fawns that survived were considerably heavier than male fawns that died (Survived: x = 38.0 kg, SE= 0.83; Died: x = 32.7 kg, SE= 1.35). In contrast, in the treatment experimental unit, weight differences were more pronounced between surviving and non-surviving females (Survived: x = 33.1 kg, SE= 1.00; Died: x = 28.2 kg, SE= 2.75) than between surviving and non-surviving males (Survived: x = 35.0 kg, SE= 0.87; Died: x = 34.5 kg, SE = 1.21 ). The importance of early winter fawn mass as a predictor of overwinter survival has been documented previously (White et al. 1987, Bishop 1998, White and Bartmann 1998, Unsworth et al. 1999). �78 LITERATURE CITED Baker, D. L., and N. T. Hobbs. 1985. Emergency feeding of mule deer during winter: tests of a supplemental ration. Journal of Wildlife Management 49:934-942. Baker, D. L., D. E. Johnson, L. H. Carpenter, 0. C. Wallmo, and R. B. Gill. 1979. Energy requirements of mule deer fawns in winter. Journal of Wildlife Management 43:162-169. Baker, D. L., G. W. Stout, and M. W. Miller. 1998. A diet supplement for captive wild ruminants. Journal of Zoo and Wildlife Medicine 29: 150-156. Ballard, W. B, D. Lutz, T. W. Keegan, L. H. Carpenter, and J.C. deVos, Jr. 2001. Deer-predator relationships: a review ofrecent North American studies with emphasis on mule and black-tailed deer. Wildlife Society Bulletin 29:99-115. Barrett, M. W., J. W. Nolan, and L. D. Roy. 1982. Evaluation of a hand-held net-gun to capture large mammals. Wildlife Society Bullet.in I 0: I 08-114. Bishop, C. J. 1998. Mule deer fawn mortality and habitat use, and the nutritional quality of bitterbrush and cheatgrass in southwest Idaho. Thesis, University ofldaho, Moscow, Idaho, USA. Bishop, C. J., D. J. Freddy, and G. C. White. 2002. Effects of enhanced nutrition of adult female mule deer on fetal and neonatal survival rates. Colorado Division of Wildlife, Wildlife Research Report, Federal Aid in Wildlife Restoration Project W-153-R, Progress Report. Fort Collins, CO USA. Bishop, C. J., and G. C. White. 2000. Effects of habitat enrichment on mule deer recruitment and survival rates. Colorado Division of Wildlife, Wildlife Research Report, Federal Aid in Wildlife Restoration Project W-153-R-13, Progress Report. Fort Collins, CO, USA. Connolly, G. E. 1981. Limiting factors and population regulation. Pages 245-285 in 0. C. Wallmo, editor. Mule and black-tailed deer of North America. University of Nebraska Press, Lincoln, Nebraska, USA. Gill, R. B., T. D. I. Beck, C. J. Bishop, D. J. Freddy, N. T. Hobbs, R.H. Kahn, M. W. Miller, T. M. Pojar, and G. C. White. 1999. Declining mule deer populations in Colorado: reasons and responses. A report to the Colorado Legislature. Colorado Division of Wildlife, Denver, Colorado, USA. Grizzle, J. E., C. F. Starmer, and G. G. Koch. 1969. Analysis of categorical data by linear models. Biometrics 25:489-504. Holter, J.B., H. H. Hayes, and S. H. Smith. 1979. Protein requirement of yearling white-tailed deer. Journal of Wildlife Management 43:872-879. Kaplan, E. L., and P. Meier. 1958. Nonparametric estimation from incomplete observations. Journal of the American Statistical Association 53:457-481. Pojar, T. M. 2000. Investigating factors contributing to declining mule deer numbers. Colorado Division of Wildlife, Wildlife Research Report, Federal Aid in Wildlife Restoration Project W-153-R-13, Progress Report. Fort Collins, CO, USA. Pollock, K. H., S. R. Winterstein, C._M. Bunck, and P. D. Curtis. 1989. Survival analysis in telemetry studies: the staggered entry design. Journal of Wildlife Management 53:7-15. Ramsey, C. W. 1968. A drop-net deer trap. Journal of Wildlife Management 32: 187-190. SAS Institute. 1989a. SAS/STAT® user's guide, version 6, fourth edition. Volume I. SAS Institute, Cary, North Carolina, USA. SAS Institute. 1989b. SAS/STAT® user's guide, version 6, fourth edition. Volume 2. SAS Institute, Cary, North Carolina, USA. SAS Institute. 1997. SAS/STAT® Software: Changes and Enhancements through Release 6.12. SAS Institute, Cary, North Carolina, USA. Schmidt, R. L., W. H. Rutherford, and F. M. Bodenham. 1978. Colorado bighorn sheep- trapping techniques. Wildlife Society Bulletin 6: 159-163. Smith, S. H., J.B. Holter, H. H. Hayes, and H. Silver. 1975. Protein requirement of white-tailed deer fawns. Journal of Wildlife Management 39:582-589. Thompson, C. B., J.B. Holter, H. H. Hayes, H. Silver, and W. E. Urban, Jr. 1973. Nutrition ofwhitetailed deer. I. Energy requirements of fawns. Journal of Wildlife Management 37:301-311. �79 Ullrey, D. E., W. G. Youatt, H. E. Johnson, L. D. Fay, and B. L. Bradley. 1967. Protein requirement of white-tailed deer fawns. Journal of Wildlife Management 31 :679-685. Unsworth, J. W., D. F. Pac, G. C. White, and R. M. Bartrnann. 1999. Mule deer survival in Colorado, Idaho, and Montana. Journal of Wildlife Management 63:315-326. van Reenen, G. 1982. Field experience in the capture of red deer by helicopter in New Zealand with reference to post-capture sequela and management. Pages 408-421 in L. Nielsen, J.C. Haigh, and M. E. Fowler, editors. Chemical immobilization of North American wildlife. Wisconsin Humane Society, Milwaukee, Wisconsin, USA. Verme, L. J., and D. E. Ullrey. 1972. Feeding and nutrition of deer. Pages 275-291 in D. C. Church, editor. Digestive physiology and nutrition of ruminants. Volume 3 - Practical Nutrition. D. C. Church, Corvallis, Oregon, USA. White, G. C., and R. M. Bartrnann. 1998. Effect of density reduction on overwinter survival of freeranging mule deer fawns. Journal of Wildlife Management 62:214-225. White, G. C., R. A. Garrott, R. M. Bartmann, L. H. Carpenter, and A. W. Alldredge. 1987. Survival of mule deer in northwest Colorado. Journal of Wildlife Management 51 :852-859. �33 JOB PROGRESS REPORT State of Division of Wildlife - Mammals Research Colorado Work Package No. _ _--=-30-=--0=--'1,....__ _ _ __ Deer Conservation Task No. Effect of Nutrition and Habitat Enhancements on Mule Deer Recruitment and Survival Rates 4 Federal Aid Project -----'-W-'---=l-=-8=--5--=R-=------Period Covered: July 1, 2002 - June 30, 2003 Authors: C. J. Bishop, G C. White, D. J. Freddy, and B. E. Watkins Personnel: D. L. Baker, L. Baeten, S. K. Carroll, D. Coven, K. Crane, M. DelTonto, B. Diamond, B. de Vergie, D. Gallegos, J. Garner, L. Gepfert, R. B. Gill, D. Hale, J. Grigg, R. Harthan, W. J. Lassiter, T. Mathieson, J. McMillan, G. C. Miller, M. W. Miller, J. Nicholson, J. A. Padia, T. M. Pojar, J. E. Risher, C. A. Schroeder, W. G. Sinner, C. M. Solohub, J. Thayer, M. A. Thonhoff, L. Wolfe, CDOW; H. VanCampen, CSU; D. Felix, Olathe Spray Service; T. R. Stephenson, California Fish and Game; L. H. Carpenter, WMI; J. SaZina, B. Welch, BLM. ABSTRACT To further understand the factors that caused deer numbers to decline in western Colorado during the 1990s, we designed and initiated a field experiment to measure deer population parameters in response to a nutrition enhancement treatment. During November 2000 - June 2003, we captured and radio-collared 533 individual mule deer evenly distributed among treatment and control units on the Uncompahgre Plateau in southwest Colorado. This included 216 adult females, 94 of which received vaginal implant transmitters (VITs), 160 6-month old fawns, and 157 newborn fawns born from either treatment or control adult does. We enhanced the nutrition of deer in the treatment unit by providing a safe, pelleted supplemental feed on a daily basis from December through April each winter. Early winter fawn:doe ratios were measured using helicopter and ground classification surveys the year following treatment delivery to determine whether fawn production and survival increased as a result of enhanced nutrition of adult females. We also measured overwinter fawn survival rates in response to the treatment. In 2002 and 2003, we measured pregnancy rates, fetus rates, and body condition of treatment and control adult does during late winter using ultrasonography. We also directly measured fetus survival and neonate survival by using VITs to help locate and radio-collar newborn fawns born from treatment and control does. Estimated percent body fat of adult does during late February and early March of 2002 and 2003 was significantly higher (F1. 90 = 108.21, P < 0.001) for treatment deer (10.4%, SE= 0.48, n = 48) than control deer (4.0%, SE= 0.36, n = 46). Serum thyroid hormone concentrations (measured only in 2003) were higher in treatment does than control does as well (F4• 52 = 32 .59, P < 0.001). Pregnancy and fetus rates were similar among treatment and control does. The pregnancy rate of adult does was 0.95 (SE= 0.036, n = 38) and the fetus rate was 1.80 fetuses/doe (SE= 0.10, n = 36) during 2002. Rates were similar in 2003, where we measured a pregnancy rate of 0.92 (SE= 0.034, n =63) and a fetus rate of 1.74 fetuses/doe (SE = 0.069, n = 50) which included 5 yearlings (the fetus rate excluding yearlings was 1.82 fetuses/doe, SE= 0.066, n = 45). The fetus survival rate with treatment and control fetuses combined was 0.86 (SE= 0.073) during 2002 and 0.97 (SE= 0.024) during 2003. Based on multiple early winter age classification surveys, we concluded that the winter nutrition enhancement treatment did not cause an increase in neonatal production and survival during 2001. Neonate survival data coupled with early �34 winter age classification surveys indicated a marginal treatment effect during 2002. However, fawn production and summer-fall survival was relatively good during 2001 and 2002 for the overall population, and not representative of most years during the past decade when the population declined. During 2003, as of late September, survival of newborn treatment fawns was 0.745 (SE= 0.059) and control fawn survival was 0.614 (SE= 0.073). During 2001-02, the overwinter survival rate of fawns was significantly greater (x 21 = 13.216, P < 0.001) in the treatment unit (S(t) = 0.865, SE= 0.056) than in the control unit (S(t) = 0.510, SE= 0.080). Again in 2002-03, the overwinter survival rate of fawns was significantly greater (x 21 = 5.734, P = 0.017) in the treatment unit (S(t) = 0.900, SE= 0.047) than in the control unit (S(t) = 0.691, SE= 0.074). Because of a cross-over over experimental design, the treatment unit during winter 2001-02 became the control unit during winter 2002-03, and vice versa. Thus, the overwinter survival treatment effect was replicated across each experimental unit. Combining both years of data, the best model of overwinter fawn survival (AICc = 148.63) included the treatment effect (x 21 = 14. 71, P < 0.001), early winter fawn mass (x\ = 16.80, P < 0.001), year (x 21 = 3.53, P = 0.060), and sex (x\ = 1.99, P = 0.158). The AIC model selection analysis emphasized the importance of both the treatment effect as well as early winter mass of fawns, because any models without treatment or fawn mass were very poor. Early winter mass was not different among experimental units (F1 = 0.35, P = 0.558), thus the effect of the treatment was not confounded with fawn mass. We will continue this research for 1.5 more years. The results reported here are preliminary and should be treated as such. �35 EFFECT OF NUTRITION AND HABITAT ENHANCEMENTS ON MULE DEER RECRUITMENT AND SURVIVAL RATES C. J. Bishop, G. C. White, D. J. Freddy, and B. E. Watkins P. N. OBJECTIVES 1. To determine experimentally whether enhancing mule deer nutrition during winter and early spring by supplemental feeding increases fetus survival, neonate survival, early winter fawn:doe ratios or overwinter fawn survival. 2. To determine experimentally to what extent habitat treatments replicate the effect of enhanced nutrition from supplemental feeding. SEGMENT OBJECTIVES 1. Capture and radio-collar a target sample of adult female mule deer and 6 month-old fawns during late November through mid-December in a treatment unit and a control unit. 2. Capture a target sample of adult female mule deer in the treatment unit and the control unit to measure pregnancy rates, fetal rates, and body condition during late February to early March, and fit each adult female deer with a radio collar and vaginal implant transmitter. 3. Deliver the nutrition enhancement treatment to all deer occupying the treatment unit from early December through the end of April. 4. Capture and radio-collar a target sample of newborn fawns from treatment and control radio-collared does during June using the vaginal implant transmitters as a technique to determine the timing and location of birth. 5. Measure fetus survival, neonate survival, early winter fawn:doe ratios, overwinter fawn survival, and annual adult female survival based on radio-collared deer from the treatment and control units. INTRODUCTION Mule deer (Odocoileus hemionus) numbers apparently declined during the 1990's throughout much of the West, and have clearly decreased since the peak population levels documented in the 1940's-60's (Gill et al. 1999, Unsworth et al. 1999). Biologists and sportsmen alike have concerns as to what factors may be responsible for declining population trends. Although previous and current research indicates that multiple interacting factors are responsible, habitat and predation have received the focus of attention. A number of studies have evaluated whether predator control increases deer survival, yet results are highly variable (Connolly 1981, Ballard et al. 2001 ). Together, predator control studies with adequate rigor indicate that predation effects on mule deer are variable as a result of time-specific and site-specific factors. Studies which have demonstrated deer population responses to predator control treatments have failed to determine whether predation is ultimately more limiting than habitat. Numerous research studies have evaluated mule deer habitat quality, but virtually no studies have documented population responses to habitat improvements. In many areas where declining deer numbers are of concern, predation is common yet habitat quality appears to have declined. The question remains as to whether predation, habitat, or some other factor is more limiting to mule deer in these situations, and whether habitat quality �36 can be improved for the benefit of deer. It may also be that no single factor is any more or less important than another, and that a more comprehensive understanding of multi-factor interactions is paramount. We designed a field experiment to measure deer population responses to nutrition enhancement treatments, to further understand the causative factors underlying observed deer population dynamics. We are conducting the study on the Uncompahgre Plateau in southwest Colorado, where several predator species are present in abundant numbers: coyotes (Canis latrans), mountain lions (Fe/is concolor), and bears (Ursus americanus). In addition to predation, myriad diseases in combination proximately affect survival of the Uncompahgre deer population (Pojar 2000, B.E. Watkins, unpublished data). Predator numbers have not and will not be manipulated in any manner during the course of the study. All factors have been left constant with the exception of deer nutrition. Deer nutrition is being enhanced by providing supplemental feed to deer occupying a treatment area during the winter. If December fawn recruitment and/or overwinter fawn survival improve as a direct result of the nutrition enhancement treatment, then we can presume that deer nutrition is ultimately more limiting than predation or disease. The second phase of the field experiment, which has not yet been initiated, will incorporate habitat manipulation treatments. The treatments will consist of prescribed fire or mechanical techniques to set back succession of pinyon-juniper (Pinus edulis-Juniperus osteosperma) habitat in an effort to improve the vigor and quality of winter habitat for mule deer. Deer population responses will be measured in relation to the habitat manipulations in the same manner as the supplemental feed. Thus, the experiment allows us to determine whether nutritional quality of winter range habitat is ultimately more limiting than other factors in a late-seral pinyon-juniper/sagebrush (Artemisia spp.) landscape, and if so, whether habitat can be effectively improved for mule deer. The results will also advance our current understanding of multi-factor interactions, with direct implications for mule deer management. MATERIALS AND METHODS Experimental Approach Experimental Design and Study Area We non-randomly selected two areas within mule deer winter range on the Uncompahgre Plateau to create 2 experimental units (A-B) (Fig. 1). The following criteria were used to select experimental units: 1.) Deer densities (~50-80 deer/mi2): areas were selected where deer densities were sufficient to meet sample size requirements within the experimental unit, while simultaneously selecting areas that would require feeding less than ~500-600 animals during a normal winter 2.) Buffer zones: areas were selected such that experimental units would be separated by several miles of non-treatment area (buffers) to prevent deer from occupying more than one experimental unit 3.) Similarity: areas were selected that comprise relatively similar habitat complexes and deer densities that are representative of the overall area 4.) Elk populations: areas were selected to minimize the number of elk present during normal winters �37 Units A and B are receiving the nutrition enhancement treatment in a cross-over experimental design, and are being used to address P.N. Objective 1. Unit A served as the treatment unit, while Unit B served as the control, for the first 2 winters ofresearch (2000 - 2002). Beginning November 2002, Unit B received the treatment while Unit A served as the control. Upon completion of P.N. Objective 1, two additional winter range experimental units will be used to conduct phase 2 of the research, or P.N. Objective 2. Habitat in one unit will be manipulated to set back plant succession (treatment), while habitat in the other unit will remain unchanged (control) throughout the experiment. Year 2000-01 Unit A Unit B 2001-02 2002-03 2003-04 Figure 1. Schematic representation of experimental units and nutrition enhancement treatment allocation. Units A and B are located in winter range habitat on the Uncompahgre Plateau in southwest Colorado. The nutrition enhancement cross-over design will encompass 4 years. The 2 experimental units (A and B) receiving the nutrition enhancement treatment are (Figs. 2 and 3): (1) Experimental unit A includes the Colona Tract of the Billy Creek State Wildlife Area and adjacent land, located approximately 13 km south of Montrose, CO adjacent to U.S. Hwy 550 South. The experimental unit is located within the Colona USGS 7.5 Minute Quadrangle, and roughly includes the polygon defined by the following Zone 13 UTM coordinates: (1) 254000 E, 4250200 N; (2) 252700 E, 4249400 N; (3) 254700 E, 4245600 N; and (4) 256200 E, 4246600 N. (2) Experimental unit B includes Shavano Valley and adjacent land extending west to the Dry Creek Rim. Shavano Valley is located approximately 13 km west of Montrose, CO. The experimental unit is located within the Dry Creek Basin and Montrose West Quadrangles (USGS 7.5 Minute), and roughly includes the polygon defined by the following Zone 13 UTM coordinates: (1) 238400 E, 4262600 N; (2) 232400 E, 4256700 N; (3) 235000 E, 4253600 N; and (4) 239500 E, 4258200 N. In late April and May, prior to fawning, deer from the winter range experimental units migrate to summer range. The summer range study area is defined by movements of the radio-collared deer, which encompass > 1000 mi2 covering the southern portion of the Uncompahgre Plateau and adjacent San Juan Mountains to the south and east (Fig. 2). The summer range study area extends north to the Dry Creek river drainage on the Uncompaghre Plateau, south to Mineral Creek near Silverton, CO, east to the Big Blue river drainage, and west to the San Miguel River canyon. However, a majority of the radio-collared deer summer on the Uncompahgre Plateau between Dry Creek to the north and Horsefly Peak to the south. Winter range elevations range from 1830 m (6000 ft) in Shavano Valley to 2318 m (7600 ft) adjacent to the Dry Creek Rim above Shavano Valley. Winter range habitat is dominated by pinyon-juniper with interspersed sagebrush adjacent to agricultural fields in the Shavano and Uncompahgre Valleys. Summer range elevations occupied by deer range from 1891 m (6200 ft) in the Uncompahgre Valley to 3538 m (11,600 ft) in Imogene Basin southwest of Ouray, CO. Summer range habitats are dominated by sprucefir (Picea spp.-Abies spp.), aspen (Populus tremuloides), ponderosa pine (Pinus ponderosa), Gambel oak (Quercus gambelii), and to a lesser extent, sagebrush and pinyon-juniper at lower elevations. �38 ... ~-· .... ..., ·--.-~ ./ .......a ·<~~~-..:::,,~<•/ ,~---··. i-~~sa County ~- GRAND JUNCTl6N-:'. ---~-:~ ··t Gunnison County Winter Range Exp. Units Summer Ran2e '·· Montrose, County (·'"--''···········•'•""··· .- .•• .... ) :\ ·•. : ·-..t., Figure 2._ Location of Colona and Shavano (Units A and B) experimental units in Game Management Unit 62 on the Uncompahgre Plateau, southwest Colorado; and location of the summe:r range study area throughout the southern Uncompahgre Plateau and adjacent San Juan Mountains �39 Figure 3. Colona and Shavano experimental units (Units A and B), located in Game Management Unit 62 on the Uncompahgre Plateau, southwest Colorado. Polygons represent the nucleus of each experimental unit, which is where animals have been collared and the nutrition enhancement treatment delivered. �40 Response Variables The response variables are fetal and neonatal survival rates, early winter fawn:doe ratios, and overwinter fawn survival rates. The nutrition enhancement treatment is delivered to deer from December through April, fetus survival is assessed during June, neonate survival is measured from June to December, and fawn:doe ratios are measured during the following December and January (1 year after the treatment was initiated). Overwinter fawn survival is measured from December to June as a direct result of the current winter's treatment. We are measuring these response variables in each experimental unit (treatment and control) to determine whether enhanced winter nutrition of adult does increases subsequent newborn fawn production and survival, and whether enhanced winter nutrition of 6-mo. old fawns directly increases overwinter fawn survival. Ultimately, these measurements provide an assessment of the effect of winter range habitat quality on yearling recruitment, and thus population productivity. We are also measuring overwinter and annual survival of adult does as a function of enhanced winter nutrition. Sample Size Fetus/Neonate Survival: We were primarily interested in survival of newborn fawns from radio-collared does that occupy the 2 winter range experimental units. Fetus survival is also important, but difficult to measure. Fetus rates from a sample of radio-collared does can be measured in winter, but the fate of all fetuses cannot be determined the following June because oflogistical constraints. Fetus survival rates can only be measured from some unpredictable fraction of the radio-collared doe sample, making sample size calculations of limited use. Thus, our sample size calculations were based on quantifying neonate survival, not fetus survival. For neonate survival, a sample size of 40 neonates per experimental unit per year provides power of 0.81 to detect a difference of0.15 in survival between 2 experimental units if survival among control fawns is 0.40. We assumed a control survival rate of 0.40 based on neonate survival rates measured recently for the Uncompahgre deer population (Pojar 2000) in combination with December fawn:doe ratios measured during the late 1980's and 1990's, when the Uncompahgre population declined (B. E. Watkins, unpublished data). Based on Bishop et al. (2002), we determined that 60 radio-collared does (30 treatment and 30 control) equipped with vaginal implant transmitters (VITs) would be necessary to capture a minimum of 80 newborn fawns. We also assumed that some fawns would be captured from other treatment and control radio-collared does not equipped with VITs. The 60 radio-collared does with VITs are also being used to evaluate fetus survival; however, logistical constraints limit the power of fetus survival comparisons among experimental units. Early winter fawn:doe ratios: We desired to detect an effect size, i.e., an increase in fawn:doe ratios in response to the treatments, in the range of 15 to 20 fawns per 100 does. These values were based on simple population models with overwinter fawn survival of 0.444, adult female survival of0.853, and December fawn:doe ratios of 66 fawns per 100 does to obtain a stationary population (Unsworth et al. 1999). Based on surveys of the Uncompahgre deer population during the 1990's, the standard deviation of the fawn:doe ratio for groups with at least one adult female was 57, with a mean of 41. Using an expected standard deviation of 57, the standard error of the mean fawn:doe ratio for 40 radio-collared does is 57/(40 112 ) = 9.0, which is the expected standard deviation of measured fawn:doe ratios on each experimental unit. We assessed power using a two-sample t-test with a sample size of 4, representing the 4 years of the study where fawn:doe ratios are being measured in response to enhanced nutrition. Our power to detect an increase of 20 fawns per 100 does based on classification of 40 radio-collared doe groups in each experimental unit is about 0.87. A sample size of 40 fawns per experimental unit per year provides a power of 0.81 to detect a difference of 0.15 in survival between 2 experimental units if survival on the control unit is 0.40. We expected to �41 see an increase in fawn survival (effect size) of approximately 0.15, because this was the difference measured in the density reduction experiment conducted by White and Bartmann (I 998). Adult and 6-month Old Fawn Capture During November and December, adult does and 6-month old fawns were captured using baited drop nets (Ramsey 1968, Schmidt et al. 1978) and helicopter net guns (Barrett et al. 1982, van Reenen 1982). Drop nets were baited with certified weed-free alfalfa hay and apple pulp. Drop nets were used as the principle capture technique for a 3-4 week capture period; helicopter net-gunning was then used at the end of the drop-net capture to secure the remainder of deer needed to meet our target sample sizes. All deer were hobbled and blind-folded after being captured. Deer captured via drop nets were carried away from the net to an adjacent handling site using stretchers. Deer were fitted with leather radio collars equipped with mortality sensors, which cause an increase in pulse rate after remaining motionless for 4 hours. Permanent collars were placed on adult females, while temporary collars were placed on fawns. To make collars temporary, one end of the collar was cut in half and reattached using rubber surgical tubing; fawns shed the collars ~6 months post-capture. A rectangular piece of flexible plastic (Ritchey® neck band material) engraved with a unique identifier was stitched to the side of each collar. The unique identifier consisted of 2 symbols for adult females, and 1 symbol on 2 different colors of plastic for fawns. The identifiers were necessary to visually identify deer from the ground. This allowed us to effectively document use of the treatment, measure fawn:doe ratios from the ground, and assess experimental unit population size via mark-resight estimators. We recorded the weight, hind foot length and chest girth of each deer, and collected blood samples to evaluate disease prevalence. During late February and early March, an additional 30 adult female deer were captured in each experimental unit by net-gunning. Captured deer were ferried by the helicopter to a central processing location, where deer were carried by stretchers to a tent for handling. For each captured deer, we used ultrasonography to measure pregnancy status, fetal rate, and body condition. Only pregnant does were retained and radio-collared. We then inserted a vaginal implant transmitter (VIT) in each doe as a technique for locating the timing and location of her birth site the following June. We also recorded the weight, hind foot length and chest girth of each deer, and collected blood samples to evaluate disease prevalence. Body Condition and Reproductive Status We estimated body fat of treatment and control adult does during mid-late winter using an Aloka 210 (Aloka, Inc., Wallinford, Conn.) portable ultrasound unit with a 5 l\.11-Iz linear transducer. We measured maximum subcutaneous fat thickness on the rump (MAXFAT) following the methodology of Stephenson et al. (1998, 2002). We also measured thickness of the longissimus dorsi muscle via ultrasound (Cook et al. 2001, Stephenson et al. 2002). A small area of hair was shaved to ensure contact between the transducer and the skin. Vegetable oil was applied to the shaved area for conduction purposes and fat/muscle thickness was measured using electronic calipers. We coupled the ultrasound measurements with body condition scores (BCS) obtained from palpation of the ribs, withers, and rump (Cook 2000). MAXF AT and rump BCS measurements were combined into a condition index used to estimate percent body fat (Cook and Cook 2002): % Fat= -6.6387617 + 7.4271417x - l. l 1579443x2 + 0.07733803x3 where x = rLIVINDEX = (MAXFAT-0.15) + rump BCS (ifMAXFAT < 0.15, then rLIVINDEX = rump BCS). The rLIVINDEX and body fat regression was initially developed and validated for elk by Cook et al. (2001 ), and then modified by incorporating a validation of MAXF AT for mule deer performed by Stephenson et al. (2002). �42 During mid-late winter 2003, we also evaluated differences in serum thyroid hormone concentrations between treatment and control adult does. Specifically, we measured total thyroxine (T4), free T4 (Ff4), total tri-iodothyronine (T3), and free T3 (Ff3) following the methodologies of Watkins et al. (1983, 1991). Blood samples were collected at the time of capture, and serum hormone analyses were performed by the Michigan State University Animal Health Diagnostic Laboratory (East Lansing, Michigan). We compared serum thyroid hormone concentrations between treatment and control adult does, and also compared hormone levels to body fat estimates derived from the ultrasonography. We quantified reproductive status (Stephenson et al. 1995, Pojar 2000) with ultrasound via transabdominal scanning using a 3 MHz linear transducer. We searched for fetuses by scanning a portion of the abdomen that was shaved caudal to the last rib and left of the midline. We systematically searched each uterine horn to identify fetal numbers ranging from Oto 3. Whenever possible, we measured eye diameter of each fetus to approximately estimate fetal age and parturition date. Vaginal Implant Transmitters (VITs) We used VITs manufactured by Advanced Telemetry Systems, Inc. (Isanti, MN). The VIT was 76 mm long, excluding antenna length, and had 2 plastic wings with a width of 57 mm when fully spread apart. The plastic wings were used to retain the transmitter in the vagina until parturition. The VIT weighed 15 grams and contained a 10-28 lithium battery programmed to a 12-hour on/off cycle. The diameter of the transmitter/battery was 14 mm, and was encased in an impermeable, water-proof, electrical resin. The transmitter contained an embedded heat-sensor which dictated the frequency pulse rate. When the heat sensor dropped below 90°F, synonymous with transmitter expulsion from the deer, the pulse rate changed from 40 PPM to 80 PPM. VIT batteries were programmed to be active from 0430 to 1630 hrs prior to daylight savings, and thus were active from 0530 to 1730 hrs after daylight savings and during the fawning period. The VIT was inserted into deer using a vaginoscope (Jorgensen Laboratories, Inc., Loveland, CO) and alligator forceps. The vaginoscope was 6" long with a 5/8" internal diameter and had a machined end (smooth surface) to minimize trauma when inserted into the vagina. A discreet mark was placed on the applicator showing the appropriate distance it should be inserted into the deer. The length of a typical mule deer vaginal tract was obtained by taking measurements from road-killed deer and/or other fresh deer carcasses obtained in the study area. Prior to use in the field, VITs were sterilized using a Chlorhexidine solution, air-dried, and sealed in a 3" x 8" sterilization pouch. Sterilization containers with Chlorhexidine solution were used on site during capture to sterilize the vaginoscope and alligator forceps between each use. A new pair of nitrile surgical gloves was used to handle the vaginoscope and VIT for each deer. To insert a VIT, the plastic wings were folded together and placed into the end of the vaginoscope. We then liberally applied sterile KY Jelly to the scope and inserted it into the deer's vagina to the point where the mark on the applicator was reached. The alligator forceps, which extended through the vaginoscope to hold the VIT, was held firmly in place while the scope was pulled out from the vagina. This procedure pushed the VIT out of the scope into the vagina, and the plastic wings spread apart to hold the transmitter in place. The transmitter antenna was typically flush with the vulva, but on occasion extended up to 1 cm beyond the vulva. The tip of the antenna was encapsulated in a wax bead to protect the deer. Neonate Fawn Capture During June we relocated each of the radio-collared does having a VIT each morning using aerial and ground telemetry. Flights began at 0530 hr and were usually completed by 1000 - 1100 hrs. The early flights were crucial for detecting fast signals because shed VITs could exceed 90 °F by mid-day if shed in the open, which caused them to switch back to a slow ("pre-birth") pulse. When a fast ("postpartum") �43 pulse rate was detected, we located the VIT from the ground to determine whether it was shed at the birth site. If the transmitter was located at the birth site, we identified whether any fawn(s) were stillborn. If the fawn(s) were no longer present at the birth site, or could not be found in the vicinity of the birth site, we located the radio-collared doe and searched for fawns at her location. All personnel involved wore surgical gloves to help minimize human scent when handling fawns. For each doe, we attempted to locate each of her fawns and document whether any fawns were stillborn. We attempted to account for each doe's fetuses in order to quantify in utero fetal survival from February to birth. We placed a dropoff radio-collar on each live fawn; radio collars were constructed with elastic neck-band material to facilitate expansion. Hole-punched, leather tabs extended from the end of the elastic and from the transmitter for attachment purposes. Collars were made temporary by cutting the leather tab extending from the elastic and reattaching the leather with latex tubing, which caused the collars to shed from the animal >6 months post-capture. For each fawn, mass and hind foot length were recorded, and a nasal swab sample was collected to screen for Bovine Viral Diarrhea. We then recorded basic vegetation characteristics of the birth site and promptly exited the site. We also routinely located treatment and control radio-collared does not having VITs and attempted to capture their fawns to help achieve our targeted sample size. Each of these does had been previously captured during the research, and were present on either the treatment or control experimental unit during winter. Measurement of Survival Rates and Fawn:Doe Ratios We measured survival rates by radio-monitoring collared deer from the ground and air to determine fate (live/mortality). We also attempted to determine the cause of each mortality, with a primary goal of distinguishing between predation and non-predation mortality causes. Deer were radio-monitored from the ground on a daily basis throughout the year and from the air on approximately a biweekly basis. We were able to detect signals from nearly all radio-collared deer each day during winter, which typically allowed us to arrive at mortality sites within 24 hours of the mortality event. During summer and migration periods, deer were distributed widely and thus were more difficult to radio-monitor. All radiocollared neonates were checked daily throughout the summer and fall, whereas some adult and yearling deer could not be ground-monitored on a routine basis. In result, we typically located neonate mortalities within 24 hours of death, but some adult deer mortalities were not detected for several days, or on rare occasion, for one or more weeks. Fresh, intact neonate carcasses were collected and submitted to the Colorado Division of Wildlife's Wildlife Health Laboratory or the Colorado State University Diagnostic Laboratory for necropsy and tissue analyses. Fresh, intact adult and 6-month old fawn carcasses were also submitted for laboratory necropsy when feasible. Field necropsies were performed on all other deer mortalities, and when appropriate, tissue samples were collected and submitted for analysis. Each winter we used the radio-collared does to measure fawn:doe ratios in each experimental unit. The resulting fawn:doe ratio is a measurement of the previous year's treatment effect. We measured fawn:doe ratios using 2 techniques: (1) We located the sample of radio-collared does in each experimental unit from a fixed-wing airplane, and used the set of locations to define boundaries for the experimental unit. Shortly after (i.e. 1-2 days), we used a helicopter to systematically fly the defined unit and classify all deer groups encountered. For each group, we documented whether a radio-collared doe was present. (2) We located each radio-collared doe by radio telemetry from the ground. The group of deer with the collared doe was counted and classified by age and sex. Both methods were employed to gather as much information as possible to determine whether there was a treatment effect. The "true" value cannot be measured perfectly because of the inherent biases and potential sources of error associated with each technique. Thus, by employing both techniques, we had a greater chance of fully understanding whether the treatment caused an effect. �44 Treatment Delivery Deer nutrition was enhanced in the treatment area by providing a safe, pelleted supplemental feed. The supplemental feed was developed through extensive testing with both captive and wild deer (Baker and Hobbs 1985, Baker et al. 1998), and has been safely used in both applied research and management projects. Pellets were distributed daily using 4wd pickup trucks and ATVs on primitive roads throughout the experimental unit to provide a food source for the entire deer population in the treatment unit. Each 501b. bag of pellets was carried ::;200m from the truck/ATV and distributed by hand in approximately 2030 small piles of feed in a linear fashion. Numerous bags were distributed in successive order allowing us to create linear lines of feed that spanned most of the treatment area, which prevented animals from concentrating in any single location. This feeding technique also prevented dominant animals from restricting access to the food supply because of the large area over which pellets were distributed. We supplied pellets ad libitum such that a small residual remained when the next day's ration was provided. Collared deer were closely monitored to ensure that treatment deer remained in the experimental unit and actually consumed the feed, and to make sure that non-treatment deer remained in the control unit, which they did. The few treatment adult does that moved away from the treatment unit were withdrawn from the sample for purposes of measuring treatment effects. However, to avoid.any biases, all 6-month old fawns captured in the treatment unit were included in survival analyses regardless of whether they accessed the supplement or not. This was because some fawns died shortly after capture (e.g. 2-3 weeks), before we could document whether they had access to the feed. Also, very few fawns that survived more than 2-3 weeks moved away from the treatment unit. The pelleted ration was commercially produced in the form of 2x l x0.5-cm wafers (Baker and Hobbs 1985). Feed constituents (i.e. digestibility, protein, gross energy etc.) vastly exceeded those of typical winter range deer diets; exact constituent values are provided by Baker et al. (1998). When provided ad libitum, the feed should have allowed deer to meet or exceed nutritional requirements for growth and maintenance (Ullrey et al. 1967, Verme and Ullrey 1972, Thompson et al. 1973, Smith et al. 1975, Baker et al. 1979, Holter et al. 1979). The basis for feeding such high quality pellets was to ensure that the treatment (enhanced nutrition) was effectively delivered to the deer. Our intent was not to determine the exact level of nutrition necessary to increase fawn recruitment, but rather to determine if nutrition is a limiting factor to recruitment. If nutrition is in fact limiting, we will rely on habitat manipulation treatments to evaluate what exactly can be done via management to increase fawn survival and recruitment. Statistical Methods A preliminary fawn:doe ratio analysis was completed using PROC MIXED in SAS (SAS Institute 1997). We used a reduced model with experimental unit as the independent variable; we considered experimental unit as a fixed effect and radio-collared does within an experimental unit as random effects. Survival rates were calculated using a Kaplan-Meier survival analysis (Kaplan and Meier 1958, Pollock et al. 1989), and contrasted among experimental units and sexes using a chi-square analysis. For neonate survival analyses, we used a common entry date because a staggered entry would ha.ve biased survival rates low due to early mortalities that occurred before most of the sample was captured. We modeled overwinter fawn survival with a logistic regression model using PROC LOGISTIC in SAS (SAS Institute 1989a); model selection was performed using Akaike's Information Criterion (AlC) (Burnham and Anderson 1998). Survival was modeled as a function of the nutrition enhancement treatment, sex, year, and capture mass. We used a general linear model in PROC GLM in SAS (SAS Institute 1989b) to test for differences in estimated percent body fat between treatment and control adult does and a multivariate model to test for differences in T4, Ff4, T3, and FT3 thryoid hormones between treatment and control does. We then used PROG REG (SAS Institute 1989b) to evaluate the relationship between estimated �45 percent body fat and serum thyroid hormone concentrations. We analyzed fetus survival directly with a binomial survival rate for the subset of fetuses with known fates. We also indirectly analyzed fetus survival by comparing the February fetus rate with the number oflive newborn fawns/doe observed in June using a change-in-ratio estimator (White et al. 1996). Other results in this report are presented as data summaries incorporating means and standard errors, or in some cases, raw data values. These results are incomplete and preliminary in nature, and should be treated as such. RESULTS AND DISCUSSION Deer Capture During November and December 2000-2002, we captured and radio-collared 122 adult female mule deer evenly distributed among the treatment and control units. We also captured and radio-collared 160 6month old fawns during November and December 2001-2002 (40 fawns/unit/year). Due to budgeting constraints, we were unable to radio-collar 6-month old fawns during 2000. We captured an additional 94 adult females during late February and early March 2002-2003 and equipped them with radio collars and VITs. During June 2002-2003, we captured and radio-collared 157 newborn fawns from radio-collared adult females. Thus, the following results are based upon radio-monitoring of 533 individual mule deer evenly distributed among treatment and control units during November 2000-June 2003. Treatment Delivery 2000-01 From December 15, 2000, through April 19, 2001, we distributed 88 tons of the pelleted ration. For most of the winter and spring, on average, we distributed 0.85 tons of feed each day throughout 22 feeding sites across the 2.3 mi2 treatment unit. Deer were fed ad libitum because there was always residual feed remaining the next day during the feeding routine. Each sack was distributed in approximately 20-30 distinct, small piles, resulting in > 1000 small piles of feed throughout the treatment unit. This effort allowed deer to effectively access the feed in small groups, and no aggression was ever observed among deer seeking access to the feed. By distributing the feed in this manner, we were able to avoid the negative aspects associated with large-scale feeding operations. Deer adapted to the pelleted supplement right away and utilized it extensively throughout the winter. We continually monitored deer use of the feed from ground observation points, where we obtained 440 visual observations of radio-collared does consuming the feed. These observations, coupled with daily radio-monitoring and periodic aerial relocations, indicate 32 of the 37 radio-collared treatment does spent the entire winter and spring within the boundaries of the treatment unit and received the supplement on a daily basis. Mark-resight population estimates from March helicopter (489 deer, SE= 62) and ground (494 deer, SE= 81) surveys, coupled with feed consumption, indicate we fed roughly 450 to 500 deer during most of the winter and spring. Feed consumption declined coincident with spring green-up, although deer continued to use the feed through mid-late April, at which point they began migrating to summer range. We also fed approximately 25 to 30 elk, but the elk did not affect deer access to the feed. Deer in the control experimental unit did not receive feed or any other treatment. Based on helicopter mark-resight surveys,· the deer density in the treatment unit in December was 120 deer/mi2 (SE= 9), but increased shortly after and was 213 deer/mi2 (SE= 27) in March. Deer densities in the control unit changed little from 83 deer/mi2 (SE= 12) in December to 101 deer/mi2 (SE= 14) in March. �46 2001-02 From December 15, 2001, through April 25, 2002, we distributed 194 tons of the supplement throughout the treatment unit. For most of the winter and spring, we distributed 2.0-2.1 tons of feed each day. The dramatic increase in supplement distribution from the previous year occurred because a large number of elk descended into the Uncompahgre Valley during mid-late fall/early winter. Elk arrived in unusually large numbers throughout much of the valley prior to the onset of treatment delivery. Once feeding was initiated, approximately 300-500 elk adapted to the feed and remained in or around the 2.3 mi2 treatment unit throughout most of the winter. Given myriad logistical and budgetary constraints, 2.1 tons was the maximum amount of feed we could routinely deliver on a daily basis. Feed was not delivered ad libitum to all deer and elk in the treatment unit throughout the winter because residual feed was rarely observed during the next day's distribution. However, daily field observations indicated most deer approached ad libitum consumption of the supplement. In contrast to the previous winter, deer were waiting for the daily supplement to arrive each morning. Deer then consumed the supplement immediately after it was distributed. Elk were rarely observed utilizing the feed until late morning or afternoon, and elk continued to forage in fields below the treatment unit, whereas deer did not. We observed numerous radio-collared deer consuming the pelleted supplement each day; not all of these observations were recorded because of time constraints with distributing the feed. Given this time limitation, we still recorded 818 observations of radio-collared deer consuming the supplemental feed (497 collared doe observations and 321 collared fawn observations). Most days,> 100 and sometimes 200-300 deer were observed utilizing the pellets during the course of distributing the supplement. These observations rarely included elk; thus, deer-elk competition was minimized because of temporal differences in feeding, and deer clearly had first access to the feed. 2002-03 Beginning December 2002, we switched the treatment and control units consistent with the cross-over experimental design. From December 15, 2002, through April 30, 2003, we distributed 97 tons of the supplement throughout the new treatment unit, which had served as the control unit the previous 2 years. The supplement was distributed daily throughout 29 sites over a larger area (~ 7 mi2) than the first 2 years of research because of the greater size of the experimental unit and broader distribution of radio-collared deer. Residual feed was always present throughout the winter, thus deer were fed ad libitum. Only small groups of elk periodically accessed the supplement, and did not affect deer access. We obtained 286 observations of radio-collared deer consuming the supplement, which were difficult to obtain because the supplement was spread out over a large area and only a single feed site could be observed at any given moment. We also used daily ground radio-monitoring and periodic aerial relocations to document deer access to the supplement. Body Condition Estimated percent body fat of adult does during late February and early March of 2002 and 2003 was significantly higher for treatment deer than control deer {F1, 90 = 108.21, P < 0.001). Over both years combined, mean predicted body fat was 10.4% (SE= 0.48) for treatment adult does and 4.0% (SE= 0.36) for control does. The interaction of experimental unit x year for predicted body fat was also significant (F1, 9o = 21. 79, P < 0.001). This interaction occurred because the difference in body fat between treatment and control deer was greater during 2003 than during 2002. During 2002, mean predicted body fat was 8.2% (SE= 0.92) for treatment adult does and 5.0% (SE= 0.71) for control does, whereas during 2003, mean predicted body fat was 11.7% (SE= 0.35) for treatment does and 3.4% (SE= 0.35) for control does. The body fat estimates reported here should accurately reflect deer, but may be further refined in the �47 future as additional research provides more data on the relationship between body condition indices and estimated percent body fat. In 2003, serum thyroid hormone concentrations were higher in treatment does than control does (F4. 52 = 32.59, P < 0.001). T4 was the most important thyroid hormone in describing the single canonical variable (1.78*T4 - 0.04*T3 + 0.20*FT4 - 0.27*FT3). Not surprisingly, there was a high partial correlation between T4 and FT4 (r = 0.77, P < 0.001) and between T3 and FT3 (r = 0.73, P < 0.001), which has been documented previously (Watkins et al. 1983). When treated as 4 separateANOVAs, T4 (F1,55= 127.45, P < 0.001), FT4 (F1, 55 = 81.72, P < 0.001), and T3 (F1,5 5== 5.39, P== 0.024) were significantly higher in treatment does than control does, whereas FT3 levels were less different among treatment and control deer (F1. 55 == 2.59, P == 0.113). Given these results, we evaluated the relationship between T4 concentrations and estimated percent body fat (derived form ultrasound and BCS indices) using a simple linear regression model(% Fat== -5.114 + 0.106*T4,? = 0.59, P < 0.001). Similar correlations between T4 and actual percent body fat during mid-late winter have been previously documented for white-tailed deer and elk (Watkins et al. 1991, Cook et al. 2001). Fetus Survival and Pregnancy/Fetus Rates We began measuring fetus survival in 2002 as part of our effort to capture and radio-collar newborn fawns born from radio-collared does. Similar numbers of stillborns were observed between treatment and control does during both 2002 and 2003, so all fetus survival analyses reported here represent pooled estimates. In February-March 2002, 36 of 38 adult does captured were pregnant, thus the pregnancy rate was 0.95 (SE= 0.036). We measured an average of 1.80 fetuses/doe (SE== 0.10, n = 36), which included 1.77 fetuses/doe (SE== 0.14, n = 18) in the treatment unit and 1.83 fetuses/doe (SE== 0.15, n == 18) in the control unit. During June 2002, we determined the fate of all fetuses (live or stillborn) from only 14 of the 36 VIT does, largely because of a high VIT battery failure rate. The survival rate of fetuses (n == 22) from these 14 does was 0.86 (SE= 0.073). We also assessed fetus survival using a change-in-ratio estimator between the fetal rate measured in February-March and the observed number of live fawns/doe postpartum in June. In June 2002, considering all does (n == 43) that we located any fawn from, whether live or stillborn, we observed 1.42 (SE== 0.11) live fawns/doe postpartum. This rate should represent a conservative estimate of live fawns/doe postpartum because we inevitably failed to locate all live fawns from each doe. In other words, this estimate would treat any unaccounted fetuses (from the February measurement) as if they were stillborns. For radio-collared does that did not have VITs, and thus we did not have a winter fetus rate measurement, singletons would infer that either the deer only had 1 fetus, or that the other fetus died. It is likely that some of these singletons had a twin that we did not locate. This equates to a conservative fetus survival rate estimate of 0.79 (SE== 0.18). In February-March 2003, 58 of 63 adult does captured were pregnant, resulting in a pregnancy rate of 0.92 (SE== 0.034). Critical personnel and equipment for measuring fetus rates were not continuously available due to capture delays associated with helicopter mechanical problems. Some of the deer fetus counts were performed by inexperienced observers without optimum ultrasound equipment. VITs worked very well, though, allowing us to determine fetus numbers at parturition for many of the deer. Thus, we determined winter fetus rates by using the greatest fetus count for each individual deer, whether obtained using ultrasound during February-March or by locating newborn fawns and stillborns at birthsites during June. We were unable to determine a fetus count for 8 treatment deer because only pregnancy was established with ultrasound and no birthsite assessments were possible in June. These 8 deer were removed from the fetus rate estimates. Of the 50 deer where a fetus count was obtained, 5 were yearlings (2 treatment yearlings, 3 control yearlings). We measured 1.74 fetuses/doe (SE== 0.069, n = 50) overall including yearlings, and 1.82 fetuses/doe (SE== 0.066, n == 45) excluding yearlings. Fetus rates with yearlings included were 1.77 fetuses/doe (SE== 0.091, n == 22) in the treatment unit and 1.70 �48 fetuses/doe (SE= 0.10, n = 28) in the control unit. During June 2003, we determined the fate of all fetuses (live or stillborn) from 33 of the 58 VIT does; the good success was based on VITs commonly being shed at birthsites. The survival rate of fetuses (n = 58) from these 33 does was 0.97 (SE= 0.024). In June 2003, incorporating all does (n = 71) that we located any fawn from, whether live or stillborn, we observed 1.49 (SE= 0.072) live fawns/doe postpartum. Using the change-in-ratio estimator described above, this results in an overall conservative fetus survival rate estimate of 0.86 (SE= 0.15). Neonatal Survival/Fawn: Doe Ratios 2001 In December 2000, at the beginning of the study and prior to the first year's treatment delivery, fawn:doe ratios were similar in the 2 experimental units. Pre-treatment fawn:doe ratios were 52.6 fawns: 100 does (SE= 5.3) in the treatment unit, and 51.6 fawns: 100 does (SE= 5.0) in the control unit. In late December 2001 and early January 2002, following the first year's treatment, we conducted 2 age classification helicopter surveys in the treatment and control units. On 12/23/01, we observed 52.8 fawns: 100 does (SE = 6.7) in the treatment unit, and 36.7 fawns:100 does (SE= 3.8) in the control unit. On 1/8/02, we observed 54. 7 fawns: 100 does (SE = 6.6) in the treatment unit, and 50.5 fawns: 100 does (SE= 6.0) in the control unit. During December 2001 - February 2002, we obtained fawn:doe ratio estimates from ground observations of radio-collared deer groups for both treatment and control deer. This survey resulted in 61.2 fawns:100 does (SE= 7.8) in the treatment unit, and 74.5 fawns:100 does (SE= 8.5) in the control unit, although the result was not statistically significant (h4 = 1.16, P = 0.249). The fawn:doe ratio results are conflicting, and clearly do not provide evidence that there was any treatment effect. In short, we concluded that the nutrition enhancement treatment did not cause an increase in neonatal production and survival during 2001. However, our results, in conjunction with a December estimate of 64 fawns: 100 does for the entire Uncompahgre deer population (B.E. Watkins, unpublished), indicate fawn production and survival was good during 2001. The observed fawn:doe ratios coupled with overwinter fawn survival and annual adult survival rates indicate the deer population was increasing. Considering the past 1-2 decades, this was an atypically good year for the Uncompahgre deer population. It would appear that whatever set of environmental conditions have led to a declining deer population were not present during 2001 in the same manner as in the past. Our main interest lies in observing the effect of the treatment on the deer population in a year where fawn:doe ratios are lower for the population as a whole, similar to what they have been much of the past 15 years. 2002 During June - December 2002, following the second year's treatment, we measured neonate survival directly using radio-collared fawns; however, sample sizes were based on a technique assessment ofVITs and were relatively small for contrasting treatment and control survival of neonates (Bishop et al. 2002). Treatment fawn survival was 0.613 (SE= 0.115, n = 29) and control fawn survival was 0.511 (SE= 0.108, n = 25). In late December 2002 and early January 2003, we once again conducted 2 age classification helicopter surveys in the treatment and control units. On 12/31/02, we observed 91.9 fawns:100 does (SE= 8.4) in the treatment unit, and 52.2 fawns: 100 does (SE= 6.9) in the control unit. On 1/21/03, we observed 52.6 fawns:100 does (SE= 6.4) in the treatment unit, and 36.8 fawns:100 does (SE= 3.9) in the control unit. The combined helicopter survey data indicated 68.1 fawns: 100 does (SE= 5.6) in the treatment unit and 42.8 fawns: 100 does (SE= 3.5) in the control unit. Oppositely, fawn:doe ratio estimates from ground classifications of doe groups during December 2002 - February 2003 were 47.7 fawns: 100 does (SE= 6.3) in the treatment unit, and 63.4 fawns:100 does (SE= 7.5) in the control unit (t108 = 1.61, P = 0.110). As in 2001, fuwn:doe ratio results were conflicting. Helicopter survey data �49 varied between 2 different flights, but consistently indicated a treatment effect. Ground classification data did not indicate a treatment effect. Also, survival data combined with age ratio data indicate neonate production and survival was reasonably favorable during 2002, and not indicative of the low fawn recruitment observed during the late l 980's and l 990's. Our results from 200 I and 2002 point out the inherent difficulties and biases associated with precisely measuring fawn:doe ratios, particularly in this research study. Ratios obtained from helicopter surveys were based on 2 short-duration flights/unit/year over spatially small units. Helicopter surveys were complicated by high deer densities in heavy cover, ma.king both deer detection and fawn:doe classifications a considerable challenge. There is a variety of potential biases that may have affected the helicopter surveys, including differential sightability of does and fawns, double classification of some deer, and incorrectly classifying yearling bucks with small antlers. Ground fawn:doe ratio observations of radio-collared doe groups were made using spotting scopes and field glasses, where we commonly studied the deer for some time. Incorrect classifications during these surveys were likely minimal. For example, small-antlered yearling bucks (e.g. 3 - 6" spikes) were detected from the ground, whereas they were undoubtedly missed on occasion during helicopter surveys. We also obtained repeated observations for some of the radio-collared doe groups from the ground. The main potential bias affecting ground fawn:doe classifications was how observations were made. Many of the ground classifications in the Shavano Valley experimental unit were made by radio-tracking does during the day. On the other hand, a majority of ground classifications in the Colona experimental unit were based on observing deer groups as they entered openings to feed during the late afternoon. Given the inherent difficulties of measuring fawn:doe ratios in the 2 experimental units, and the lack of a clear indication as to the effectiveness of the treatment, we intensified efforts in 2003 to directly measure survival of neonate fawns born from treatment and control radio-collared does. At the completion of the research, we will test whether enhanced winter nutrition of adult does improved newborn fawn survival based on a three-year model of radio-collared neonate survival data. We will continue to measure early winter fawn:doe ratios, but the data will be used cautiously to make inferences regarding treatment effects. 2003 During June 2003, we captured and radio-collared 103 newborn fawns born from treatment and control radio-collared does (55 treatment fawns, 48 control fawns). The VITs worked well; we captured fawns from 41 of the 54 does fitted with VITs. As of late September 2003, treatment fawn survival was 0.745 (SE= 0.059) and control fawn survival was 0.614 (SE= 0.073). Neonate Mortality Causes During 2002, 11 of the 29 treatment fawns died from the following causes: 3 - coyote predation, 2 - bear predation, 1 - felid predation, 1 - predation where the predator was undetermined, 1 - disease/ malnutrition, 1 - abandonment, 1 - road-kill, and 1 - trauma/injury. Twelve of the 25 control fawns died: 6 - malnutrition/disease, 3 - coyote predation, 1 - felid predation, 1 - bear predation, and 1 predation mortality where the predator was undetermined. Thus, 13% of all radio-collared fawns died from malnutrition, 11 % from coyote predation, 6% from bear predation, 4% from felid predation, 4% from predation (unknown predator), and 6% from miscellaneous causes. Currently (June - September 2003), 14 of the 55 treatment fawns have died from the following causes: 6 - disease/malnutrition/starvation, 4 - coyote predation, 3 - predation (unknown predator), and 1 - felid predation. Over the same time period, 18 of the 48 control fawns have died: 8 - coyote predation, 4 - disease/malnutrition/starvation, 3 felid predation, 1 - bear predation, and 2 - unknown. Thus, as of the end of September during 2003, 12% �50 of all radio-collared fawns have died from coyote predation, 10% from disease/malnutrition/starvation, 4% from felid predation, 3% from predation (unknown predator), 1% from bear predation, and 2% from unknown causes. Overwinter Fawn Survival and Mortality Causes During winter 2001-02 (Dec 1, 2001- May 31, 2002), the survival rate of fawns was significantly greater (x\ = 13.216, P < 0.001) in the treatment unit (S(t) = 0.865, SE= 0.056) than in the control unit (S(t) = 0.510, SE= 0.080). Again in 2002-03 (Dec 1, 2002 -May 31, 2003), the overwinter survival rate of fawns was significantly greater (x 21 = 5.734, P = 0.017) in the treatment unit (S(t) = 0.900, SE= 0.047) than in the control unit (S(t) = 0.691, SE= 0.074) (Fig. 4). The treatment unit during winter 2001-02 became the control unit during winter 2002-03, and vice versa. Thus, the overwinter survival treatment effect was replicated across each experimental unit. Combining both years of data, the best model of overwinter fawn survival (AICc = 148.63) included treatment (x\ = 14.71, P < 0.001), early winter fawn mass (x\ = 16.80, P < 0.001), year (x21 = 3.53, P = 0.060), and sex (x2i = 1.99, P = 0.158). The AIC model selection analysis emphasizes the importance of both the treatment effect as well as early winter mass of fawns, because any models without treatment or fawn mass were very poor (Table 1). Survival of fawns receiving the nutrition enhancement treatment was 0.31 higher than survival of control fawns during two mild to average winters, and surviving fawns averaged 2.9 kg heavier than fawns that died. Early winter mass was not different among experimental units (F1 = 0.35, P = 0.558), thus the effect of the treatment was not confounded with fawn mass. Fawn mass was similar between winters as well (F1 = 0.45, P = 0.502). The importance of early winter fawn mass as a predictor of overwinter survival has been documented previously (White et al. 1987, Bishop 1998, White and Bartmann 1998, Unsworth et al. 1999). In summary, the nutrition enhancement treatment improved overwinter fawn survival and thus yearling recruitment, and heavier fawns in each experimental unit had higher survival probabilities. l. •••••, f'" b"'""·•t,:'-«•,.< _.,.•.,,. ~«<: «<~· .,- ««<❖ L ,_.... ·----· ....__. ·---~ <««• ................ .. - - - - - - - - - - - - - - - - - ..,, 0.9 ....... ............. •-: ····t:.:;·· ............. ·.•_ .............................................................................. =: ..............................,, l ;:,;,,;.;.;.;.;.;.;,;,;.;,;,;-.~~ ll..8 • 1~~==-~ .___ _ •••.... ···················v·:;·· ....................................((.........: ..,...:-::0:o:":·:. 0.7 - • • Treatment (2001'"02) -Control (200J.-.02) - " Tttmtrnent. (2002--03) 0.5 · .................. . ()Ji ~··············· .... wmm= Contrul {l002-.03) +--~-~--~-~-~-~~-~-~-~--~-~-__,., 0.4 .U/l l~/1.(i 12.i'.H l.115 l/JO 2l1.4 3,1. ~Vlo ,v..,1 4/lS 4,30 S/15 :5/..~0 Figure 4. Overwinter fawn survival (Dec 1- May 31) in a nutrition enhancement treatment wut (S(t) = 0.865, SE= 0.056, 2001-02; S(t) = 0.900, SE= 0.047, 2002-03) and a control w1.it (S(t) = 0.510, SE= 0.080, 2001-02; S(t) = 0.691, SE= 0.074, 2002-03), Uncompahgre Plateau, southwest Colorado. �51 # Ii Param eters Model Name -2 Log Likelih od (K) AIC AICc C Treatment + Sex + Year + Mass 143.231 5 153.231 148.626 0 Treatment + Year + Mass Treatment + Sex + Year + Trt*Year + Mass Treatment + Sex + Mass 145.286 4 153.286 149.548 0.92 143.059 6 155.059 149.615 0.99 146.898 4 154.898 151.159 2.53 Treatment + Mass 148.957 3 154.957 152.113 3.49 Sex + Year+ Mass 160.345 4 168.345 164.606 15.98 Treatment 165.845 2 169.845 167.922 19.30 Sex +Year 178.195 3 184.195 181.351 32.73 AIC Table 1. Model selection results for a logistic regression analysis of overwinter mule deer fawn survival in southwest Colorado. Enhanced nutrition (freatment) and early winter fawn mass were the critical predictors of survival. Model selection was performed using Akaike's Information Criterion (AIC). During winter 2001-02, five fawns in the treatment unit died: 2 from malnutrition/sickness and 3 from disease. Of the 2 fawn mortalities caused by malnutrition/sickness, I was a result of basic malnutrition and occurred on December 31, 2001, shortly after the treatment was initiated. The other fawn died early as well and had a combination of heavy parasite loads, scours, and general poor condition. Each of the 3 fawns that died from disease had adequate fat stores. At least one of these fawns died as a result of pneumonia. In the control unit, 19 fawns died during the winter: 5 from malnutrition, 6 from mountain lion/bobcat predation, 4 from coyote/canine predation, 3 unknown predation mortalities, and 1 unknown. A majority of the fawns killed by predators had virtually no femur marrow fat remaining, indicating the predation was likely compensatory in nature. During winter 2002-03, where the initial control unit became the treatment following the cross-over, four fawns died in the treatment unit: 3 from coyote predation and 1 unknown mortality. In the control unit, 12 fawns died during the winter: 4 from coyote predation, 2 from malnutrition, 1 from mountain lion predation, 1 was road-killed, and 4 causes were unknown. As in the previous winter, these fawns had virtually no femur marrow fat remaining, indicating very poor condition. Adult Female Survival and Causes of Mortality During winter 2000-01 (Dec 1, 2000 - May 31, 2001 ), the adult doe survival rate in the treatment unit (S(t) = 0.968, SE= 0.032) was greater (x\ = 2.649, P = 0.104) than the survival rate in the control unit (S(t) = 0.861, SE= 0.058). However, annual adult doe survival rates (Dec 1, 2000 - Nov 30, 2001) were similar among the treatment and control deer (Trt: S(t) = 0.839, SE= 0.066; Control: S(t) = 0.833, SE= 0.062; 1 = 0.004, P = 0.947). We observed a similar result the following year. The 2001-02 overwinter adult doe survival rate in the treatment unit (S(t) = 0.942, SE= 0.030) was greater (x 21 = 3.116, P = 0.078) than survival in the control unit (S(t) = 0.848, SE= 0.044), yet annual adult doe survival was similar among treatment and control deer (Trt: S(t) = 0.824, SE= 0.049; Control: S(t) = 0.818, SE= 0.047; x\ = 0.090, P = 0.764). Thus, mortalities of control deer occurred primarily during the winter months, while treatment does died primarily during the summer and fall months. x2 �52 During winter 2002-03, following the treatment cross-over, overwinter adult doe survival rates were similar among treatment and control deer (Trt: S(t) = 0.945, SE= 0.024; Control: S(t) = 0.924, SE= 0.028; x\ = 0.360, P = 0.549). The main difference from the previous 2 years was that overwinter survival of adult does in the Shavano experimental unit increased in 2002-03 upon receiving the treatment. Current annual adult doe survival rates (Dec 1, 2002 - Oct 7, 2003) are 0.888 (SE= 0.034) for treatment does and 0.835 (SE= 0.039) for control does. The treatment has apparently had a minimal impact on annual adult doe survival, and annual survival rates measured thus far align with expected survival based on other studies (Unsworth et al. 1999, B.E. Watkins, unpublished). During 2000-02, when the Colona experimental unit received the treatment and the Shavano experimental unit was the control, 16 treatment and 16 control does died. The 16 treatment does died from the following categories: 4 - road-killed, 3 - while giving birth, 3 - predation (undetermined predator), 2 non-predation unknown (intact carcasses with no evidence of predation or scavenging), 1 - disease (chronic arthritis), 1 - mountain lion predation, and 2 - unknown. Predation was not a major mortality factor for treatment does, and a majority of mortalities were independent of nutrition (does were in good condition). The 16 control doe mortalities included the following causes: 5 - mountain lion predation, 3 - malnutrition, 2 - non-predation unknown, 1 - road-killed, 1 - bear predation, 1 - injury (fence), 1 legal harvest, and 2 - unknown. Predation and malnutrition were the major mortality causes of control deer. Interestingly, during this 2-year period, we did not document any coyote predation on adult does. Thus far during 2003, with Shavano as the treatment and Colona as the control, there have been 9 treatment doe mortalities: 3 - coyote predation, 3 - disease/infection, 1 - road-killed, and 2 unknown. Two of the coyote mortalities, 2 of the disease mortalities, and the road-kill occurred on adult does in good condition. There have been 14 control doe mortalities thus far in 2003: 3 - coyote predation, 3 malnutrition/disease, 3 - non-predation unknown, 1 - mountain lion predation, 1 - road-kill, and 3 unknown. As we saw during 2000 - 2002, malnutrition and predation were the major mortality factors of control does. LITERATURE CITED Baker, D. L., and N. T. Hobbs. 1985. Emergency feeding of mule deer during winter: tests of a supplemental ration. Journal of Wildlife Management 49:934-942. Baker, D. L., D. E. Johnson, L. H. Carpenter, 0. C. Wallmo, and R. B. Gill. 1979. Energy requirements of mule deer fawns in winter. Journal of Wildlife Management 43:162-169. Baker, D. L., G. W. Stout, and M. W. Miller. 1998. A diet supplement for captive wild ruminants. Journal of Zoo and Wildlife Medicine 29: 150-156. Ballard, W. B, D. Lutz, T. W. Keegan, L. H. Carpenter, and J.C. deVos, Jr. 2001. Deer-predator relationships: a review of recent North American studies with emphasis on mule and black-tailed deer. Wildlife Society Bulletin 29:99-115. Barrett, M. W., J. W. Nolan, and L. D. Roy. 1982. Evaluation of a hand-held net-gun to capture large mammals. Wildlife Society Bulletin 10: 108-114. Bishop, C. J. 1998. Mule deer fawn mortality and habitat use, and the nutritional quality ofbitterbrush and cheatgrass in southwest Idaho. Thesis, University of Idaho, Moscow, Idaho, USA. �53 Bishop, C. J., D. J. Freddy, and G. C. White. 2002. Effects of enhanced nutrition of adult female mule deer on fetal and neonatal survival rates: a pilot study to address feasibility. Colorado Division of Wildlife, Wildlife Research Report, Federal Aid in Wildlife Restoration Project W-153-R, Job Final Report. Fort Collins, Colorado, USA. Burnham, K. P., and D.R. Anderson. 1998. Model selection and inference: a practical informationtheoretic approach. Springer-Verlag, New York, New York, USA. Connolly, G. E. 1981. Limiting factors and population regulation. Pages 245-285 in 0. C. Wallmo, editor. Mule and black-tailed deer of North America. University of Nebraska Press, Lincoln, Nebraska, USA. Cook, J. G., and R. C. Cook. 2002. An informal training guide to condition evaluation in elk and deer. National Council for Air and Stream Improvement, Unpublished Report. Cook, R. C. 2000. Studies of body condition and reproductive physiology in Rocky Mountain elk. Thesis, University ofldaho, Moscow, Idaho, USA. Cook, R. C., J. G. Cook, D. L. Murray, P. Zager, B. K. Johnson, and M. W. Gratson. 2001. Development of predictive models of nutritional condition for Rocky Mountain elk. Journal of Wildlife Management 65:973-987. Gill, R. B., T. D. I. Beck, C. J. Bishop, D. J. Freddy, N. T. Hobbs, R.H. Kahn, M. W. Miller, T. M. Pojar, and G. C. White. 1999. Declining mule deer populations in Colorado: reasons and responses. A report to the Colorado Legislature. Colorado Division of Wildlife, Denver, Colorado, USA. Holter, J. B., H. H. Hayes, and S. H. Smith. 1979. Protein requirement of yearling white-tailed deer. Journal of Wildlife Management 43:872-879. Kaplan, E. L., and P. Meier. 1958. Nonparametric estimation from incomplete observations. Journal of the American Statistical Association 53:457-481. Pojar, T. M. 2000. Investigating factors contributing to declining mule deer numbers. Colorado Division of Wildlife, Wildlife Research Report, Federal Aid in Wildlife Restoration Project W-153-R-13, Progress Report. Fort Collins, CO, USA. Pollock, K. H., S. R. Winterstein, C. M. Bunck, and P. D. Curtis. 1989. Survival analysis in telemetry studies: the staggered entry design. Journal of Wildlife Management 53:7-15. Ramsey, C. W. 1968. A drop-net deer trap. Journal ofWildlife Management 32:187-190. SAS Institute. 1989a. SAS/STAT® user's guide, version 6, fourth edition. Volume 1. SAS Institute, Cary, North Carolina, USA. SAS Institute. 1989b. SAS/STAT® user's guide, version 6, fourth edition. Volume 2. SAS Institute, Cary, North Carolina, USA. SAS Institute. 1997. SAS/STAT® Software: Changes and Enhancements through Release 6.12. SAS Institute, Cary, North Carolina, USA. �54 Schmidt, R. L., W. H. Rutherford, and F. M. Bodenham. 1978. Colorado bighorn sheep-trapping techniques. Wildlife Society Bulletin 6: 159-163. Smith, S. H., J.B. Holter, H. H. Hayes, and H. Silver. 1975. Protein requirement of white-tailed deer fawns. Journal of Wildlife Management 39:582-589. Stephenson, T. R., V. C. Bleich, B. M. Pierce, and G. P. Mulcahy. 2002. Validation of mule deer body composition using in vivo and post-mortem indices of nutritional condition. Wildlife Society Bulletin 30:557-564. Stephenson, T. R., K. J. Hundertmark, C. G. Schwartz, and V. Van Ballenberghe. 1998. Predicting body fat and body mass in moose with ultrasonography. Canadian Journal of Zoology 76:717-722. Stephenson, T. R., J. W. Testa, G. P. Adams, R G. Sasser, C. G. Schwartz, and K. J. Hundertmark. 1995. Diagnosis of pregnancy and twinning in moose by ultrasonography and serum assay. Akes 31:167-172. Thompson, C. B., J.B. Holter, H. H. Hayes, H. Silver, and W. E. Urban, Jr. 1973. Nutrition of whitetailed deer. I. Energy requirements of fawns. Journal of Wildlife Management 37:301-311. Ullrey, D. E., W. G. Youatt, H. E. Johnson, L. D. Fay, and B. L. Bradley. 1967. Protein requirement of white-tailed deer fawns. Journal of Wildlife Management 31 :679-685. Unsworth, J. W., D. F. Pac, G. C. White, and R. M. Bartmann. 1999. Mule deer survival in Colorado, Idaho, and Montana. Journal of Wildlife Management 63:315-326. van Reenen, G. 1982. Field experience in the capture of red deer by helicopter in New Zealand with reference to post-capture sequela and management. Pages 408-421 in L. Nielsen, J. C. Haigh, and M. E. Fowler, editors. Chemical immobilization of North American wildlife. Wisconsin Humane Society, Milwaukee, Wisconsin, USA. Verme, L. J., and D. E. Ullrey. 1972. Feeding and nutrition of deer. Pages 275-291 in D. C. Church, editor. Digestive physiology and nutrition of ruminants. Volume 3 - Practical Nutrition. D. C. Church, Corvallis, Oregon, USA. Watkins, B. E., D. E. Ullrey, R. F. Nachreiner, and S. M. Schmitt. 1983. Effects of supplemental iodine and season on thyroid activity of white-tailed deer. Journal of Wildlife Management 47:45-58. Watkins, B. E., J. H. Witham, D. E. Ullrey, D. J. Watkins, and J.M. Jones. 1991. Body composition and condition evaluation of white-tailed deer fawns. Journal ofWildlife Management 55:39-51. White, G. C., and R. M. Bartmann. 1998. Effect of density reduction on overwinter survival of freeranging mule deer fawns. Journal of Wildlife Management 62:214-225. White, G. C., R. A. Garrott, R. M. Bartmann, L. H. Carpenter, and A. W. Alldredge. 1987. Survival of mule deer in northwest Colorado. Journal of Wildlife Management 51:852-859. White, G. C., A. F. Reeve, F. G. Lindzey, and K. P. Burnham. 1996. Estimation of mule deer winter mortality from age ratios. Journal of Wildlife Management 60:37-44. �Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Project No. Work Package No. Task No. : : : Colorado 3001 : 4 Federal Aid Project: W-185-R Cost Center 3430 Mammals Research Deer Conservation Effect of Nutrition and Habitat Enhancements on Mule Deer Recruitment and Survival Rates : Period Covered: July 1, 2003 - June 30, 2004 Authors: C. J. Bishop, G. C. White, D. J. Freddy, and B. E. Watkins Personnel: D. L. Baker, L. Baeten, T. Banulis, E. J. Bergman, S. K. Carroll, D. Coven, K. Crane, M. DelTonto, B. Diamond, B. deVergie, D. Gallegos, J. Garner, L. Gepfert, R. B. Gill, D. Hale, J. Grigg, H. Halbritter, R. Harthan, M. Johnston, W. J. Lassiter, T. Mathieson, J. McMillan, G. C. Miller, M. W. Miller, J. Nicholson, J. A. Padia, T. M. Pojar, R. Powers, J. E. Risher, C. A. Schroeder, W. G. Sinner, C. M. Solohub, J. Thayer, M. A. Thonhoff, R. Wertsbaugh, L. Wolfe, CDOW; H. VanCampen, CSU; D. Felix, Olathe Spray Service; T. R. Stephenson, California Fish and Game; L. H. Carpenter, WMI; J. Sazma, B. Welch, BLM. All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT To further understand the factors that caused deer numbers to decline in western Colorado during the 1990s, we designed and initiated a field experiment to measure deer population parameters in response to a nutrition enhancement treatment. During November 2000 – June 2004, we captured and radio-collared 810 individual mule deer evenly distributed among treatment and control units on the Uncompahgre Plateau in southwest Colorado. This included 293 adult females, 154 of which received vaginal implant transmitters (VITs), 241 6-month old fawns, and 276 newborn fawns born from either treatment or control adult does. We enhanced the nutrition of deer in the treatment unit by providing a safe, pelleted supplemental feed on a daily basis from December through April each winter. Early winter fawn:doe ratios were measured using helicopter and ground classification surveys the year following treatment delivery to determine whether fawn production and survival increased as a result of enhanced nutrition of adult females. We also measured overwinter fawn survival rates in response to the treatment. During 2002 – 2004, we measured pregnancy rates, fetus rates, and body condition of treatment and control adult does during late winter using ultrasonography. We also directly measured fetus survival and neonate survival by using VITs to help locate and radio-collar newborn fawns born from treatment and control does. Estimated percent body fat of adult does during late February and early March, 2002-04, 21 �was significantly higher (F1, 148 = 153.41, P < 0.001) for treatment deer (9.8%, SE = 0.36, n = 78) than control deer (4.3%, SE = 0.26, n = 76). Serum thyroid hormone concentrations, measured only in 2003 and 2004, were higher in treatment does than control does (F4, 108 = 46.59, P < 0.001). Pregnancy and fetus rates were similar among treatment and control does. The pregnancy rate of adult does was 0.95 (SE = 0.036, n = 38) and the fetus rate was 1.80 fetuses/doe (SE = 0.10, n = 36) during 2002. Rates were similar in 2003, where we measured a pregnancy rate of 0.92 (SE = 0.034, n = 63) and a fetus rate of 1.74 fetuses/doe (SE = 0.069, n = 50) which included 5 yearlings (the fetus rate excluding yearlings was 1.82 fetuses/doe, SE = 0.066, n = 45). In 2004, we measured a pregnancy rate of 0.94 (SE = 0.029, n = 66) and the fetus rate was 1.97 fetuses/doe (SE = 0.053, n = 60), which included 4 yearlings (the fetus rate excluding yearlings was 2.00 fetuses/doe, SE = 0.051, n = 56). The fetus survival rate with treatment and control fetuses combined was 0.86 (SE = 0.073) during 2002, 0.97 (SE = 0.024) during 2003, and 0.90 (SE = 0.040) during 2004. Fetus survival was similar among treatment and control deer during 2002 – 2003, but not 2004, where treatment fetus survival was 1.00 (SE = 0.000, n = 33) and control fetus survival was 0.76 (SE = 0.085, n = 25). Based on multiple early winter age classification surveys, we concluded the winter nutrition enhancement treatment did not cause an increase in neonatal production and survival during 2001. However, fawn production and summer-fall survival was relatively good for the overall population, and not representative of most years during the past decade when the population declined. During June – December, 2002−2003, survival of newborn treatment fawns was 0.620 (SE = 0.067) and control fawn survival was 0.493 (SE = 0.070). Survival data coupled with early winter age classification surveys provided evidence the nutrition enhancement treatment increased December fawn recruitment during 2002 and 2003. During December – June, 2001−2004, the overwinter survival rate of fawns was significantly greater (χ21 = 18.781, P < 0.001) in the treatment unit (S(t) = 0.895, SE = 0.029) than in the control unit (S(t) = 0.655, SE = 0.044). Because of a cross-over experimental design, the treatment unit during winter 2001-02 became the control unit during winters 2002-04, and vice versa. Thus, the treatment effect was replicated across each experimental unit. Combining all years of data, the best model of overwinter fawn survival (AICc = 207.65) included the nutrition enhancement treatment (χ21 = 19.04, P < 0.001), early winter fawn mass (χ21 = 23.27, P < 0.001), and year (χ21 = 6.20, P = 0.045). The AIC model selection analysis emphasized the importance of both the treatment effect as well as early winter mass of fawns, because any models without treatment or fawn mass were poor. Early winter mass of control fawns was slightly higher than that of treatment fawns (F1, 231 = 3.00, P = 0.085); thus the effect of the treatment was not confounded with fawn mass. Data collection will not be completed until January 2005. The results reported here are preliminary and should be treated as such. 22 �JOB PROGRESS REPORT EFFECT OF NUTRITION AND HABITAT ENHANCEMENTS ON MULE DEER RECRUITMENT AND SURVIVAL RATES C. J. Bishop, G. C. White, D. J. Freddy, and B. E. Watkins P. N. OBJECTIVES 1. To determine experimentally whether enhancing mule deer nutrition during winter and early spring by supplemental feeding increases fetus survival, neonate survival, early winter fawn:doe ratios or overwinter fawn survival. 2. To determine experimentally to what extent habitat treatments replicate the effect of enhanced nutrition from supplemental feeding. SEGMENT OBJECTIVES 1. Capture and radio-collar a target sample of adult female mule deer and 6 month-old fawns during late November through mid-December in a treatment unit and a control unit. 2. Capture a target sample of adult female mule deer in the treatment unit and the control unit to measure pregnancy rates, fetal rates, and body condition during late February to early March, and fit each adult female deer with a radio collar and vaginal implant transmitter. 3. Deliver the nutrition enhancement treatment to all deer occupying the treatment unit from early December through the end of April. 4. Capture and radio-collar a target sample of newborn fawns from treatment and control radio-collared does during June using the vaginal implant transmitters as a technique to determine the timing and location of birth. 5. Measure fetus survival, neonate survival, early winter fawn:doe ratios, overwinter fawn survival, and annual adult female survival based on radio-collared deer from the treatment and control units. INTRODUCTION Mule deer (Odocoileus hemionus) numbers apparently declined during the 1990’s throughout much of the West, and have clearly decreased since the peak population levels documented in the 1940’s60’s (Unsworth et al. 1999, Gill et al. 2001). Biologists and sportsmen alike have concerns as to what factors may be responsible for declining population trends. Although previous and current research indicates multiple interacting factors are responsible, habitat and predation have received the focus of attention. A number of studies have evaluated whether predator control increases deer survival, yet results are highly variable (Connolly 1981, Ballard et al. 2001). Together, predator control studies with adequate rigor indicate predation effects on mule deer are variable as a result of time-specific and sitespecific factors. Studies which have demonstrated deer population responses to predator control treatments have failed to determine whether predation is ultimately more limiting than habitat. Numerous research studies have evaluated mule deer habitat quality, but virtually no studies have documented population responses to habitat improvements. In many areas where declining deer numbers are of concern, predation is common yet habitat quality appears to have declined. The question remains as to whether predation, habitat, or some other factor is more limiting to mule deer in these situations, and whether habitat quality can be improved for the benefit of deer. It may also be that no single factor is any more or less important than others, and a more comprehensive understanding of multi-factor interactions is needed. We designed a field experiment to measure deer population responses to nutrition enhancement treatments, to further understand the causative factors underlying observed deer population dynamics. We 23 �are conducting the study on the Uncompahgre Plateau in southwest Colorado, where several predator species are present in abundant numbers: coyotes (Canis latrans), mountain lions (Felis concolor), and bears (Ursus americanus). In addition to predation, myriad diseases in combination proximately affect survival of the Uncompahgre deer population (Pojar and Bowden 2004, B.E. Watkins, unpublished data). Predator numbers have not and will not be manipulated in any manner during the course of the study. All factors have been left constant with the exception of deer nutrition. Deer nutrition is being enhanced by providing supplemental feed to deer occupying a treatment area during the winter. If December fawn recruitment and/or overwinter fawn survival improve as a direct result of the nutrition enhancement treatment, then we can presume that deer nutrition is ultimately more limiting than predation or disease. The second phase of the field experiment, which has not yet been initiated, will incorporate habitat manipulation treatments. The treatments will consist of prescribed fire or mechanical techniques to set back succession of pinyon-juniper (Pinus edulis-Juniperus osteosperma) habitat in an effort to improve the vigor and quality of winter habitat for mule deer. Deer population responses will be measured in relation to the habitat manipulations in the same manner as the supplemental feed. Thus, the experiment will evaluate whether nutritional quality of winter range habitat is ultimately more limiting than other factors in a late-seral pinyon-juniper/sagebrush (Artemisia spp.) landscape, and if so, whether habitat can be effectively improved for mule deer. The results will also advance our current understanding of multifactor interactions, with direct implications for mule deer management. MATERIALS AND METHODS Experimental Approach Experimental Design and Study Area We non-randomly selected two areas within mule deer winter range on the Uncompahgre Plateau to create 2 experimental units (A-B) (Fig. 1). The following criteria were used to select experimental units: 1.) Deer densities (~50-80 deer/mi2): areas were selected where deer densities were sufficient to meet sample size requirements within the experimental unit, while simultaneously selecting areas that would require feeding less than ~500-600 animals during a normal winter 2.) Buffer zones: areas were selected such that experimental units would be separated by several miles of non-treatment area (buffers) to prevent deer from occupying more than one experimental unit 3.) Similarity: areas were selected that comprise relatively similar habitat complexes and deer densities that are representative of the overall area 4.) Elk populations: areas were selected to minimize the number of elk present during normal winters Units A and B are receiving the nutrition enhancement treatment in a cross-over experimental design, and are being used to address P.N. Objective 1. Unit A served as the treatment unit, while Unit B served as the control, for the first 2 winters of research (2000 – 2002). Beginning November 2002, Unit B received the treatment while Unit A served as the control. Upon completion of P.N. Objective 1, two additional winter range experimental units will be used to conduct phase 2 of the research, or P.N. Objective 2. Habitat in one unit will be manipulated to set back plant succession (treatment), while habitat in the other unit will remain unchanged (control) throughout the experiment. 24 �Year 2000-01 Unit A Treatment Unit B Control 2001-02 Control 2002-03 Treatment Control Treatment 2003-04 Control Treatment Figure 1. Schematic representation of experimental units and nutrition enhancement treatment allocation. Units A and B are located in winter range habitat on the Uncompahgre Plateau in southwest Colorado. The nutrition enhancement cross-over design will encompass 4 years. The 2 experimental units (A and B) receiving the nutrition enhancement treatment are (Figs. 2 and 3): (1) Experimental unit A includes the Colona Tract of the Billy Creek State Wildlife Area and adjacent land, located approximately 13 km south of Montrose, CO adjacent to U.S. Hwy 550 South. The experimental unit is located within the Colona USGS 7.5 Minute Quadrangle, and roughly includes the polygon defined by the following Zone 13 UTM coordinates: (1) 254000 E, 4250200 N; (2) 252700 E, 4249400 N; (3) 254700 E, 4245600 N; and (4) 256200 E, 4246600 N. (2) Experimental unit B includes Shavano Valley and adjacent land extending west to the Dry Creek Rim. Shavano Valley is located approximately 13 km west of Montrose, CO. The experimental unit is located within the Dry Creek Basin and Montrose West Quadrangles (USGS 7.5 Minute), and roughly includes the polygon defined by the following Zone 13 UTM coordinates: (1) 238400 E, 4262600 N; (2) 232400 E, 4256700 N; (3) 235000 E, 4253600 N; and (4) 239500 E, 4258200 N. In late April and May, prior to fawning, deer from the winter range experimental units migrate to summer range. The summer range study area is defined by movements of the radio-collared deer, which encompass >1000 mi2 covering the southern portion of the Uncompahgre Plateau and adjacent San Juan Mountains to the south and east (Fig. 2). The summer range study area extends north to the Dry Creek river drainage on the Uncompaghre Plateau, south to Mineral Creek near Silverton, CO, east to the Big Blue river drainage, and west to the San Miguel River canyon. However, a majority of the radio-collared deer summer on the Uncompahgre Plateau between Dry Creek to the north and Highway 62 to the south. Winter range elevations range from 1830 m (6000 ft) in Shavano Valley to 2318 m (7600 ft) adjacent to the Dry Creek Rim above Shavano Valley. Winter range habitat is dominated by pinyonjuniper with interspersed sagebrush adjacent to agricultural fields in the Shavano and Uncompahgre Valleys. Summer range elevations occupied by deer range from 1891 m (6200 ft) in the Uncompahgre Valley to 3538 m (11,600 ft) in Imogene Basin southwest of Ouray, CO. Summer range habitats are dominated by spruce-fir (Picea spp.-Abies spp.), aspen (Populus tremuloides), ponderosa pine (Pinus ponderosa), Gambel oak (Quercus gambelii), and to a lesser extent, sagebrush and pinyon-juniper at lower elevations. 25 �Uncompahgre Plateau Mesa County GRAND JUNCTION Delta County U om nc GMU 62 r hg pa e u ea at Pl Montrose County GMU 61 DELTA Shavano E.U. Gunnison County Winter Range Exp. Units MONTROSE Colona Montrose County E.U. Summer Range Ouray County Sanmiguel County Figure 2. Location of Colona and Shavano (Units A and B) experimental units in Game Management Unit 62 on the Uncompahgre Plateau, southwest Colorado; and location of the summer range study area throughout the southern Uncompahgre Plateau and adjacent San Juan Mountains. 26 �Hwy 550 Uncompahgre Valley Colona Exp. Unit Shavano Valley Shavano Exp. Unit Figure 3. Colona and Shavano experimental units (Units A and B), located in Game Management Unit 62 on the Uncompahgre Plateau, southwest Colorado. 27 �Response Variables The response variables are fetal and neonatal survival rates, early winter fawn:doe ratios, and overwinter fawn survival rates. The nutrition enhancement treatment is delivered to deer from December through April, fetus survival is assessed during June, neonate survival is measured from June to December, and fawn:doe ratios are measured during the following December and January (1 year after the treatment was initiated). Overwinter fawn survival is measured from December to June as a direct result of the current winter’s treatment. We are measuring these response variables in each experimental unit (treatment and control) to determine whether enhanced winter nutrition of adult does increases subsequent newborn fawn production and survival, and whether enhanced winter nutrition of 6-mo. old fawns directly increases overwinter fawn survival. Ultimately, these measurements provide an assessment of the effect of winter range habitat quality on yearling recruitment, and thus population productivity. We are also measuring overwinter and annual survival of adult does as a function of enhanced winter nutrition. Sample Size Fetus/Neonate Survival: We were primarily interested in survival of newborn fawns from radiocollared does that occupy the 2 winter range experimental units. Fetus survival is also important, but difficult to measure. Fetus rates from a sample of radio-collared does can be measured in winter, but the fate of all fetuses cannot be determined the following June because of logistical constraints. Fetus survival rates can only be measured from some unpredictable fraction of the radio-collared doe sample, making sample size calculations of limited use. Thus, our sample size calculations were based on quantifying neonate survival, not fetus survival. For neonate survival, a sample size of 40 neonates per experimental unit per year provides power of 0.81 to detect a difference of 0.15 in survival between 2 experimental units if survival among control fawns is 0.40. We assumed a control survival rate of 0.40 based on neonate survival rates measured recently for the Uncompahgre deer population (Pojar and Bowden 2004) in combination with December fawn:doe ratios measured during the late 1980’s and 1990’s, when the Uncompahgre population declined (B. E. Watkins, unpublished data). Based on Bishop et al. (2002), we determined that 60 radio-collared does (30 treatment and 30 control) equipped with vaginal implant transmitters (VITs) would be necessary to capture a minimum of 80 newborn fawns. We also assumed that some fawns would be captured from other treatment and control radio-collared does not equipped with VITs. The 60 radio-collared does with VITs are also being used to evaluate fetus survival; however, logistical constraints limit the power of fetus survival comparisons among experimental units. Early winter fawn:doe ratios: We desired to detect an effect size, i.e., an increase in fawn:doe ratios in response to the treatments, in the range of 15 to 20 fawns per 100 does. These values were based on simple population models with overwinter fawn survival of 0.444, adult female survival of 0.853, and December fawn:doe ratios of 66 fawns per 100 does to obtain a stationary population (Unsworth et al. 1999). Based on surveys of the Uncompahgre deer population during the 1990’s, the standard deviation of the fawn:doe ratio for groups with at least one adult female was 57, with a mean of 41. Using an expected standard deviation of 57, the standard error of the mean fawn:doe ratio for 40 radio-collared does is 57/(401/2) = 9.0, which is the expected standard deviation of measured fawn:doe ratios on each experimental unit. We assessed power using a two-sample t-test with a sample size of 4, representing the 4 years of the study where fawn:doe ratios are being measured in response to enhanced nutrition. Our power to detect an increase of 20 fawns per 100 does based on classification of 40 radio-collared doe groups in each experimental unit is about 0.87. A sample size of 40 fawns per experimental unit per year provides a power of 0.81 to detect a difference of 0.15 in survival between 2 experimental units if survival on the control unit is 0.40. We expected to see an increase in fawn survival (effect size) of approximately 0.15, because this was the difference measured in the density reduction experiment conducted by White and Bartmann (1998). 28 �Adult and 6-month Old Fawn Capture During November and December, adult does and 6-month old fawns were captured using baited drop nets (Ramsey 1968, Schmidt et al. 1978) and helicopter net guns (Barrett et al. 1982, van Reenen 1982). Drop nets were baited with certified weed-free alfalfa hay and apple pulp. Drop nets were used as the principle capture technique for a 3-4 week capture period; helicopter net-gunning was then used at the end of the drop-net capture to secure the remainder of deer needed to meet our target sample sizes. All deer were hobbled and blind-folded after being captured. Deer captured via drop nets were carried away from the net to an adjacent handling site using stretchers. Deer were fitted with leather radio collars equipped with mortality sensors, which cause an increase in pulse rate after remaining motionless for 4 hours. Permanent collars were placed on adult females, while temporary collars were placed on fawns. To make collars temporary, one end of the collar was cut in half and reattached using rubber surgical tubing; fawns shed the collars ≥6 months post-capture. A rectangular piece of flexible plastic (Ritchey® neck band material) engraved with a unique identifier was stitched to the side of each collar. The unique identifier consisted of 2 symbols for adult females, and 1 symbol on 2 different colors of plastic for fawns. The identifiers were necessary to visually identify deer from the ground. This allowed us to effectively document use of the treatment, measure fawn:doe ratios from the ground, and assess experimental unit population size via mark-resight estimators. We recorded the weight, hind foot length and chest girth of each deer, and collected blood samples to evaluate disease prevalence. During late February and early March, an additional 30 adult female deer were captured in each experimental unit by net-gunning. Captured deer were ferried by the helicopter to a central processing location, where deer were carried by stretchers to a tent for handling. For each captured deer, we used ultrasonography to measure pregnancy status, fetal rate, and body condition. Only pregnant does were retained and radio-collared. We then inserted a vaginal implant transmitter (VIT) in each doe as a technique for locating the timing and location of her birth site the following June. We also recorded the weight, hind foot length and chest girth of each deer, and collected blood samples to evaluate disease prevalence. Body Condition and Reproductive Status We estimated body fat of treatment and control adult does during mid-late winter using an Aloka 210 (Aloka, Inc., Wallinford, Conn.) or SonoVet 2000 (Universal Medical Systems, Bedford Hills, NY) portable ultrasound unit with a 5 MHz linear transducer. We measured maximum subcutaneous fat thickness on the rump (MAXFAT) following the methodology of Stephenson et al. (1998, 2002). We also measured thickness of the longissimus dorsi muscle via ultrasound (Cook et al. 2001, Stephenson et al. 2002). A small area of hair was shaved to ensure contact between the transducer and the skin. Vegetable oil was applied to the shaved area for conduction purposes and fat/muscle thickness was measured using electronic calipers. We coupled the ultrasound measurements with body condition scores (BCS) obtained from palpation of the ribs, withers, and rump (Cook 2000). MAXFAT and rump BCS measurements were combined into a condition index used to estimate percent body fat (Cook and Cook 2002): % Fat = -6.6387617 + 7.4271417x – 1.11579443x2 + 0.07733803x3 where x = rLIVINDEX = (MAXFAT – 0.15) + rump BCS (if MAXFAT < 0.15, then rLIVINDEX = rump BCS). The rLIVINDEX and body fat regression was initially developed and validated for elk by Cook et al. (2001), and then modified by incorporating a validation of MAXFAT for mule deer performed by Stephenson et al. (2002). During mid-late winter, we also evaluated differences in serum thyroid hormone concentrations between treatment and control adult does. Specifically, we measured total thyroxine (T4), free T4 (FT4), total tri-iodothyronine (T3), and free T3 (FT3) following the methodologies of Watkins et al. (1983, 1991). Blood samples were collected at the time of capture, and serum hormone analyses were performed by the Michigan State University Animal Health Diagnostic Laboratory (East Lansing, Michigan). We 29 �compared serum thyroid hormone concentrations between treatment and control adult does, and also compared hormone levels to body fat estimates derived from the ultrasonography. We quantified reproductive status (Stephenson et al. 1995, Andelt et al. 2004) with ultrasound via transabdominal scanning using a 3 MHz linear transducer. We searched for fetuses by scanning a portion of the abdomen that was shaved caudal to the last rib and left of the midline. We systematically searched each uterine horn to identify fetal numbers ranging from 0 to 3. Whenever possible, we measured eye diameter of each fetus to approximately estimate fetal age and parturition date. Vaginal Implant Transmitters (VITs) We used VITs manufactured by Advanced Telemetry Systems, Inc. (Isanti, MN). The VIT was 76 mm long, excluding antenna length, and had 2 plastic wings with a width of 57 mm when fully spread apart. The plastic wings were used to retain the transmitter in the vagina until parturition. The VIT weighed 15 grams and contained a 10-28 lithium battery programmed to a 12-hour on/off cycle. The diameter of the transmitter/battery was 14 mm, and was encased in an impermeable, water-proof, electrical resin. The transmitter contained an embedded heat-sensor which dictated the frequency pulse rate. When the heat sensor dropped below 90°F, synonymous with transmitter expulsion from the deer, the pulse rate changed from 40 PPM to 80 PPM. VIT batteries were programmed to be active from 0430 to 1630 hrs prior to daylight savings, and thus were active from 0530 to 1730 hrs after daylight savings and during the fawning period. The VIT was inserted into deer using a vaginoscope (Jorgensen Laboratories, Inc., Loveland, CO) and alligator forceps. The vaginoscope was 6” long with a 5/8” internal diameter and had a machined end (smooth surface) to minimize trauma when inserted into the vagina. A discreet mark was placed on the applicator showing the appropriate distance it should be inserted into the deer. The length of a typical mule deer vaginal tract was obtained by taking measurements from road-killed deer and/or other fresh deer carcasses obtained in the study area. Prior to use in the field, VITs were sterilized using a Chlorhexidine solution, air-dried, and sealed in a 3” x 8” sterilization pouch. Sterilization containers with Chlorhexidine solution were used on site during capture to sterilize the vaginoscope and alligator forceps between each use. A new pair of nitrile surgical gloves was used to handle the vaginoscope and VIT for each deer. To insert a VIT, the plastic wings were folded together and placed into the end of the vaginoscope. We then liberally applied sterile KY Jelly to the scope and inserted it into the deer’s vagina to the point where the mark on the applicator was reached. The alligator forceps, which extended through the vaginoscope to hold the VIT, was held firmly in place while the scope was pulled out from the vagina. This procedure pushed the VIT out of the scope into the vagina, and the plastic wings spread apart to hold the transmitter in place. The transmitter antenna was typically flush with the vulva, but on occasion extended up to 1 cm beyond the vulva. The tip of the antenna was encapsulated in a wax bead to protect the deer. Neonate Fawn Capture During June we relocated each of the radio-collared does having a VIT each morning using aerial and ground telemetry. Flights began at 0530 hr and were usually completed by 1000 – 1100 hrs. The early flights were crucial for detecting fast signals because shed VITs could exceed 90 °F by mid-day if shed in the open, which caused them to switch back to a slow (“pre-birth”) pulse. When a fast (“postpartum”) pulse rate was detected, we located the VIT from the ground to determine whether it was shed at the birth site. If the transmitter was located at the birth site, we identified whether any fawn(s) were stillborn. If the fawn(s) were no longer present at the birth site, or could not be found in the vicinity of the birth site, we located the radio-collared doe and searched for fawns at her location. All personnel involved wore surgical gloves to help minimize human scent when handling fawns. For each doe, we attempted to locate each of her fawns and document whether any fawns were stillborn. We attempted to account for each doe’s fetuses in order to quantify in utero fetal survival from February to birth. We placed a drop-off radio-collar on each live fawn; radio collars were constructed with elastic neck-band 30 �material to facilitate expansion. Hole-punched, leather tabs extended from the end of the elastic and from the transmitter for attachment purposes. Collars were made temporary by cutting the leather tab extending from the elastic and reattaching the leather with latex tubing, which caused the collars to shed from the animal >6 months post-capture. For each fawn, mass and hind foot length were recorded, and a nasal swab sample was collected to screen for Bovine Viral Diarrhea. We then recorded basic vegetation characteristics of the birth site and promptly exited the site. We also routinely located and attempted to capture fawns from treatment and control radiocollared does not having VITs to help achieve our targeted sample size. Each of these does had been previously captured during the research, and were present on either the treatment or control experimental unit during winter. Measurement of Survival Rates and Fawn:Doe Ratios We measured survival rates by radio-monitoring collared deer from the ground and air to determine fate (live/mortality). We also attempted to determine the cause of each mortality, with a primary goal of distinguishing between predation and non-predation mortality causes. Deer were radiomonitored from the ground on a daily basis throughout the year and from the air on approximately a biweekly basis. We were able to detect signals from nearly all radio-collared deer each day during winter, which typically allowed us to arrive at mortality sites within 24 hours of the mortality event. During summer and migration periods, deer were distributed widely and thus were more difficult to radiomonitor. All radio-collared neonates were checked daily throughout the summer and fall, whereas some adult and yearling deer could not be ground-monitored on a routine basis. In result, we typically located neonate mortalities within 24 hours of death, but some adult deer mortalities were not detected for several days, or on rare occasion, for one or more weeks. Fresh, intact neonate carcasses were collected and submitted to the Colorado Division of Wildlife’s Wildlife Health Laboratory or the Colorado State University Diagnostic Laboratory for necropsy and tissue analyses. Fresh, intact adult and 6-month old fawn carcasses were also submitted for laboratory necropsy when feasible. Field necropsies were performed on all other deer mortalities, and when appropriate, tissue samples were collected and submitted for analysis. Each winter we used the radio-collared does to measure fawn:doe ratios in each experimental unit. The resulting fawn:doe ratio is a measurement of the previous year’s treatment effect. We measured fawn:doe ratios using 2 techniques: (1) We located the sample of radio-collared does in each experimental unit from a fixed-wing airplane, and used the set of locations to define boundaries for the experimental unit. Shortly after (i.e. 1-2 days), we used a helicopter to systematically fly the defined unit and classify all deer groups encountered. For each group, we documented whether a radio-collared doe was present. (2) We located each radio-collared doe by radio telemetry from the ground. The group of deer with the collared doe was counted and classified by age and sex. Both methods were employed to gather as much information as possible to determine whether there was a treatment effect. The “true” value cannot be measured perfectly because of the inherent biases and potential sources of error associated with each technique. Thus, by employing both techniques, we had a greater chance of fully understanding whether the treatment caused an effect. Treatment Delivery Deer nutrition was enhanced in the treatment area by providing a safe, pelleted supplemental feed. The supplemental feed was developed through extensive testing with both captive and wild deer (Baker and Hobbs 1985, Baker et al. 1998), and has been safely used in both applied research and management projects. Pellets were distributed daily using 4wd pickup trucks, ATVs, and snowmobiles on primitive roads throughout the experimental unit to provide a food source for the entire deer population in the treatment unit. Each 50lb. bag of pellets was carried ≤200m from the vehicle and distributed by hand in approximately 20-30 small piles of feed in a linear fashion. Numerous bags were 31 �distributed in successive order allowing us to create linear lines of feed that spanned most of the treatment area, which prevented animals from concentrating in any single location. This feeding technique also prevented dominant animals from restricting access to the food supply because of the large area over which pellets were distributed. We supplied pellets ad libitum where a small residual remained when the next day’s ration was provided. Collared deer were closely monitored to ensure that treatment deer remained in the experimental unit and actually consumed the feed, and to make sure that non-treatment deer remained in the control unit, which they did. The few treatment adult does that moved away from the treatment unit were withdrawn from the sample for purposes of measuring treatment effects. However, to avoid any biases, all 6-month old fawns captured in the treatment unit were included in survival analyses regardless of whether they accessed the supplement or not. This was because some fawns died shortly after capture (e.g. 2-3 weeks), before we could document whether they had access to the feed. Also, very few fawns that survived more than 2-3 weeks moved away from the treatment unit. The pelleted ration was commercially produced in the form of 2×1×0.5-cm wafers (Baker and Hobbs 1985). Feed constituents (i.e. digestibility, protein, gross energy etc.) vastly exceeded those of typical winter range deer diets; exact constituent values are provided by Baker et al. (1998). When provided ad libitum, the feed should have allowed deer to meet or exceed nutritional requirements for growth and maintenance (Ullrey et al. 1967, Verme and Ullrey 1972, Thompson et al. 1973, Smith et al. 1975, Baker et al. 1979, Holter et al. 1979). The basis for feeding such high quality pellets was to ensure that the treatment (enhanced nutrition) was effectively delivered to the deer. Our intent was not to determine the exact level of nutrition necessary to increase fawn recruitment, but rather to determine if nutrition is a limiting factor to recruitment. If nutrition is in fact limiting, we will rely on habitat manipulation treatments to evaluate what exactly can be done via management to increase fawn survival and recruitment. Statistical Methods A preliminary fawn:doe ratio analysis was completed using PROC MIXED in SAS (SAS Institute 1997). We used a reduced model with experimental unit as the independent variable; we considered experimental unit as a fixed effect and radio-collared does within an experimental unit as random effects. Survival rates were calculated using a Kaplan-Meier survival analysis (Kaplan and Meier 1958, Pollock et al. 1989), and contrasted among experimental units and sexes using a chi-square analysis. For neonate survival analyses, we used a common entry date because a staggered entry would have biased survival rates low due to early mortalities that occurred before most of the sample was captured. We modeled overwinter fawn survival with a logistic regression model using PROC LOGISTIC in SAS (SAS Institute 1989a); model selection was performed using Akaike’s Information Criterion (AIC) (Burnham and Anderson 1998). Survival was modeled as a function of the nutrition enhancement treatment, sex, year, and capture mass. We used a general linear model in PROC GLM in SAS (SAS Institute 1989b) to test for differences in estimated percent body fat between treatment and control adult does and a multivariate model to test for differences in T4, FT4, T3, and FT3 thryoid hormones between treatment and control does. We then used PROG REG (SAS Institute 1989b) to evaluate the relationship between estimated percent body fat and serum thyroid hormone concentrations. We analyzed fetus survival directly with a binomial survival rate for the subset of fetuses with known fates. We also indirectly analyzed fetus survival by comparing the February fetus rate with the number of live newborn fawns/doe observed in June using a change-in-ratio estimator (White et al. 1996). Other results in this report are presented as data summaries incorporating means and standard errors, or in some cases, raw data values. These results are incomplete and preliminary, and should be treated as such. 32 �RESULTS AND DISCUSSION Deer Capture During November and December 2000-2003, we captured and radio-collared 139 adult female mule deer evenly distributed among the treatment and control units. We also captured and radio-collared 241 6-month-old fawns during November and December 2001-2003 (40 fawns/unit/year). Due to budgeting constraints, we were unable to radio-collar 6-month old fawns during 2000. We captured an additional 154 adult females during late February and early March 2002-2004 and equipped them with radio collars and VITs. During June 2002-2004, we captured and radio-collared 276 newborn fawns from radio-collared adult females. Thus, the following results are based upon radio-monitoring of 810 individual mule deer evenly distributed among treatment and control units during November 2000-June 2004. Treatment Delivery 2000-01 From December 15, 2000, through April 19, 2001, we distributed 88 tons of the pelleted ration. For most of the winter and spring, on average, we distributed 0.85 tons of feed each day throughout 22 feeding sites across the 2.3 mi2 treatment unit. Deer were fed ad libitum because there was always residual feed remaining the next day during the feeding routine. Each sack was distributed in approximately 20-30 distinct, small piles, resulting in >1000 small piles of feed throughout the treatment unit. This effort allowed deer to effectively access the feed in small groups, and no aggression was ever observed among deer seeking access to the feed. By distributing the feed in this manner, we were able to avoid the negative aspects associated with large-scale feeding operations. Deer adapted to the pelleted supplement right away and utilized it extensively throughout the winter. We continually monitored deer use of the feed from ground observation points, where we obtained 440 visual observations of radiocollared does consuming the feed. These observations, coupled with daily radio-monitoring and periodic aerial relocations, indicate 32 of the 37 radio-collared treatment does spent the entire winter and spring within the boundaries of the treatment unit and received the supplement on a daily basis. Mark-resight population estimates from March helicopter (489 deer, SE = 62) and ground (494 deer, SE = 81) surveys, coupled with feed consumption, indicate we fed roughly 450 to 500 deer during most of the winter and spring. Feed consumption declined coincident with spring green-up, although deer continued to use the feed through mid-late April, at which point they began migrating to summer range. We also fed approximately 25 to 30 elk, but the elk did not affect deer access to the feed. Deer in the control experimental unit did not receive feed or any other treatment. Based on helicopter mark-resight surveys, the deer density in the treatment unit in December was 120 deer/mi2 (SE = 9), but increased shortly after and was 213 deer/mi2 (SE = 27) in March. Deer densities in the control unit changed little from 83 deer/mi2 (SE = 12) in December to 101 deer/mi2 (SE = 14) in March. 2001-02 From December 15, 2001, through April 25, 2002, we distributed 194 tons of the supplement throughout the treatment unit. For most of the winter and spring, we distributed 2.0-2.1 tons of feed each day. The dramatic increase in supplement distribution from the previous year occurred because a large number of elk descended into the Uncompahgre Valley during mid-late fall/early winter. Elk arrived in unusually large numbers throughout much of the valley prior to the onset of treatment delivery. Once feeding was initiated, approximately 300-500 elk adapted to the feed and remained in or around the 2.3 mi2 treatment unit throughout most of the winter. Given myriad logistical and budgetary constraints, 2.1 tons was the maximum amount of feed we could routinely deliver on a daily basis. Feed was not delivered ad libitum to all deer and elk in the treatment unit throughout the winter because residual feed was rarely observed during the next day’s 33 �distribution. However, daily field observations indicated most deer approached ad libitum consumption of the supplement. In contrast to the previous winter, deer were waiting for the daily supplement to arrive each morning. Deer then consumed the supplement immediately after it was distributed. Elk were rarely observed utilizing the feed until late morning or afternoon, and elk continued to forage in fields below the treatment unit, whereas deer did not. We observed numerous radio-collared deer consuming the pelleted supplement each day; not all of these observations were recorded because of time constraints with distributing the feed. Given this time limitation, we still recorded 818 observations of radio-collared deer consuming the supplemental feed (497 collared doe observations and 321 collared fawn observations). Most days, >100 and sometimes 200-300 deer were observed utilizing the pellets during the course of distributing the supplement. These observations rarely included elk; thus, deer-elk competition was minimized because of temporal differences in feeding, and deer clearly had first access to the feed. 2002-03 Beginning December 2002, we switched the treatment and control units consistent with the crossover experimental design. From December 15, 2002, through April 30, 2003, we distributed 97 tons of the supplement throughout the new treatment unit, which had served as the control unit the previous 2 years. The supplement was distributed daily throughout 29 sites over a larger area (~7 mi2) than the first 2 years of research because of the greater size of the experimental unit and broader distribution of radiocollared deer. Residual feed was always present throughout the winter, thus deer were fed ad libitum. Only small groups of elk periodically accessed the supplement, and did not affect deer access. We obtained 286 observations of radio-collared deer consuming the supplement, which were difficult to obtain because the supplement was spread out over a large area and only a single feed site could be observed at any given moment. We also used daily ground radio-monitoring and periodic aerial relocations to document deer access to the supplement. 2003-04 From December 10, 2003, through April 30, 2004, we distributed 197 tons of the supplement throughout the treatment unit. The increase in supplement distribution occurred because of an increase in elk on the upper portion of the experimental unit. However, unlike winter 2001-02, residual feed was present throughout the winter and deer were fed ad libitum. By targeting a portion of the daily feed distribution to elk, we restricted elk to the upper extent of the deer winter range for most of the winter. Thus, elk had a minimal affect on deer access to the supplement. We obtained 413 observations of radiocollared deer consuming the supplement. As before, we also used daily ground radio-monitoring and periodic aerial relocations to document deer access to the supplement. Body Condition Estimated percent body fat of adult does during late February and early March, 2002–2004, was significantly higher for treatment deer than control deer (F1, 148 = 153.41, P < 0.001). Over all years combined, mean predicted body fat was 9.8% (SE = 0.36) for treatment adult does and 4.3% (SE = 0.26) for control does. The interaction of experimental unit × year for predicted body fat was also significant (F2, 148 = 14.39, P < 0.001). This interaction occurred because the difference in body fat between treatment and control deer was greater during 2003 than during 2002 or 2004. Mean predicted body fat was 8.2% (SE = 0.92) for treatment adult does and 5.0% (SE = 0.71) for control does during 2002, and 9.0% (SE = 0.53) for treatment does and 4.7% (SE = 0.36) for control does during 2004. The difference was greater during 2003, where mean predicted body fat was 11.7% (SE = 0.35) for treatment does and 3.4% (SE = 0.35) for control does. The body fat estimates reported here should accurately reflect deer, but may be further refined in the future as additional research provides more data on the relationship between body condition indices and estimated percent body fat. Serum thyroid hormone concentrations, measured during 2003 and 2004, were higher in treatment does than control does (F4, 108 = 46.59, P < 0.001) (Table 1). Hormone concentrations also 34 �varied between years (F4, 108 = 14.21, P < 0.001), but the experimental unit × year interaction was not significant (F4, 108 = 1.68, P = 0.160). Thus, each year thyroid hormone concentrations were higher in treatment does than control does. T4 was the most important thyroid hormone in describing the canonical variable for differences between treatment and control does (1.04*T4 − 0.02*T3 + 0.77*FT4 – 0.17*FT3). As expected, there was a high partial correlation between T4 and FT4 (r = 0.67, P < 0.001) and between T3 and FT3 (r = 0.60, P < 0.001), which has been documented previously (Watkins et al. 1983). When treated as 4 separate ANOVAs, T4 (F1, 111 = 165.97, P < 0.001), FT4 (F1, 111 = 144.37, P < 0.001), T3 (F1, 111 = 13.84, P < 0.001), and FT3 (F1, 111 = 8.26, P = 0.005) were significantly higher in treatment does than control does. Given these results, we evaluated the relationship between T4 concentrations and estimated percent body fat (derived from ultrasound and BCS indices) using a simple linear regression model (% Fat = −3.122 + 0.090*T4, r2 = 0.52, P < 0.001). Similar correlations between T4 and actual percent body fat during mid-late winter have been previously documented for white-tailed deer and elk (Watkins et al. 1991, Cook et al. 2001). Table 1. Total thyroxine (T4) and total tri-iodothyronine (T3) concentrations (nmol/l), and free T4 (FT4) and free T3 (FT3) concentrations (pmol/l), measured during late February in adult female mule deer occupying a nutrition enhancement treatment unit and a control unit on the Uncompahgre Plateau in southwest Colorado, 2003-04. Thyroid Hormone T3 (SE) FT3 (SE) 146.6 (3.53) FT4 (SE) 30.0 (1.27) 1.65 (0.058) 4.10 (0.130) Control 92.3 (3.56) 17.1 (0.65) 1.42 (0.080) 3.71 (0.210) Treatment 131.9 (4.48) 24.8 (1.39) 2.08 (0.075) 4.21 (0.154) Control 90.0 (3.54) 12.5 (0.59) 1.70 (0.104) 3.60 (0.188) Year Exp. Unit T4 (SE) 2003 Treatment 2004 Fetus Survival and Pregnancy/Fetus Rates We began measuring fetus survival in 2002 as part of our effort to capture and radio-collar newborn fawns born from radio-collared does. Similar numbers of stillborns were observed between treatment and control does during both 2002 and 2003, so fetus survival estimates for those years are not differentiated by experimental unit. In February-March 2002, 36 of 38 adult does captured were pregnant, thus the pregnancy rate was 0.95 (SE = 0.036). We measured an average of 1.80 fetuses/doe (SE = 0.10, n = 36), which included 1.77 fetuses/doe (SE = 0.14, n = 18) in the treatment unit and 1.83 fetuses/doe (SE = 0.15, n = 18) in the control unit. During June 2002, we determined the fate of all fetuses (live or stillborn) from only 14 of the 36 VIT does, largely because of a high VIT battery failure rate. The survival rate of fetuses (n = 22) from these 14 does was 0.86 (SE = 0.073). We also assessed fetus survival using a change-in-ratio estimator between the fetal rate measured in February-March and the observed number of live fawns/doe postpartum in June. In June 2002, considering all does (n = 43) that we located any fawn from, whether live or stillborn, we observed 1.42 (SE = 0.11) live fawns/doe postpartum. This rate should represent a conservative estimate of live fawns/doe postpartum because we inevitably failed to locate all live fawns from each doe. In other words, this estimate would treat any unaccounted fetuses (from the February measurement) as if they were stillborns. For radio-collared does that did not have VITs, and thus we did not have a winter fetus rate measurement, singletons would infer that either the deer only had 1 fetus, or that the other fetus died. It is likely that some of these singletons had a twin that we did not locate. This equates to a conservative fetus survival rate estimate of 0.79 (SE = 0.18). 35 �In February-March 2003, 58 of 63 adult does captured were pregnant, resulting in a pregnancy rate of 0.92 (SE = 0.034). Critical personnel and equipment for measuring fetus rates were not continuously available due to capture delays associated with helicopter mechanical problems. Some of the deer fetus counts were performed by inexperienced observers without optimum ultrasound equipment. VITs worked very well, though, allowing us to determine fetus numbers at parturition for many of the deer. Thus, we determined winter fetus rates by using the greatest fetus count for each individual deer, whether obtained using ultrasound during February-March or by locating newborn fawns and stillborns at birthsites during June. We were unable to determine a fetus count for 8 treatment deer because only pregnancy was established with ultrasound and no birthsite assessments were possible in June. These 8 deer were removed from the fetus rate estimates. Of the 50 deer where a fetus count was obtained, 5 were yearlings (2 treatment yearlings, 3 control yearlings). We measured 1.74 fetuses/doe (SE = 0.069, n = 50) overall including yearlings, and 1.82 fetuses/doe (SE = 0.066, n = 45) excluding yearlings. Fetus rates with yearlings included were 1.77 fetuses/doe (SE = 0.091, n = 22) in the treatment unit and 1.70 fetuses/doe (SE = 0.10, n = 28) in the control unit. During June 2003, we determined the fate of all fetuses (live or stillborn) from 33 of the 58 VIT does; the good success was based on VITs commonly being shed at birthsites. The survival rate of fetuses (n = 58) from these 33 does was 0.97 (SE = 0.024). In June 2003, incorporating all does (n = 71) that we located any fawn from, whether live or stillborn, we observed 1.49 (SE = 0.072) live fawns/doe postpartum. Using the change-in-ratio estimator described above, this results in an overall conservative fetus survival rate estimate of 0.86 (SE = 0.15). In February 2004, the overall pregnancy rate was 0.94 (SE = 0.029, n = 66) and the fetus rate was 1.97 fetuses/doe (SE = 0.053, n = 60), which included 4 yearlings. Excluding yearlings, the fetus rate was 2.00 fetuses/doe (SE = 0.051, n = 56). Fetus rates were 1.90 fetuses/doe (SE = 0.074, n = 30) in the treatment unit and 2.03 fetuses/doe (SE = 0.076, n = 30) in the control unit with yearlings included, and 1.93 (SE = 0.069, n = 29) in the treatment unit and 2.07 (SE = 0.074, n = 27) in the control unit with yearlings excluded. We determined the fate of all fetuses (live or stillborn) from 31 of the 60 VIT does. The overall fetus survival rate was 0.90 (SE = 0.040, n = 58). Different from 2002 or 2003, each of these stillborns were from control does. The survival rate of control fetuses was 0.76 (SE = 0.085, n = 25) as compared to a survival rate of 1.00 (n = 33) for treatment fetuses. Using data from all does (n = 82) in which we located any fawn, the conservative change-in-ratio fetus survival estimate was 0.79 (SE = 0.13) overall, 0.88 (SE = 0.17) for treatment deer, and 0.69 (SE = 0.14) for control deer. Neonatal Survival/Fawn:Doe Ratios 2001 In December 2000, at the beginning of the study and prior to the first year’s treatment delivery, fawn:doe ratios were similar in the 2 experimental units. Pre-treatment fawn:doe ratios were 52.6 fawns:100 does (SE = 5.3) in the treatment unit, and 51.6 fawns:100 does (SE = 5.0) in the control unit. In late December 2001 and early January 2002, following the first year’s treatment, we conducted 2 age classification helicopter surveys in the treatment and control units. On 12/23/01, we observed 52.8 fawns:100 does (SE = 6.7) in the treatment unit, and 36.7 fawns:100 does (SE = 3.8) in the control unit. On 1/8/02, we observed 54.7 fawns:100 does (SE = 6.6) in the treatment unit, and 50.5 fawns:100 does (SE = 6.0) in the control unit. During December 2001 – February 2002, we obtained fawn:doe ratio estimates from ground observations of radio-collared deer groups for both treatment and control deer. This survey resulted in 61.2 fawns:100 does (SE = 7.8) in the treatment unit, and 74.5 fawns:100 does (SE = 8.5) in the control unit, although the result was not statistically significant (t74 = 1.16, P = 0.249). The fawn:doe ratio results are conflicting, and clearly do not provide evidence that there was any treatment effect. In short, we concluded that the nutrition enhancement treatment did not cause an increase in neonatal production and survival during 2001. However, our results, in conjunction with a December estimate of 64 fawns:100 does for the entire Uncompahgre deer population (B.E. Watkins, 36 �unpublished), indicate fawn production and survival was good during 2001. The observed fawn:doe ratios coupled with overwinter fawn survival and annual adult survival rates indicate the deer population was increasing. Considering the past 1-2 decades, this was an atypically good year for the Uncompahgre deer population. 2002 During June – December 2002, following the second year’s treatment, we measured neonate survival directly using radio-collared fawns; however, sample sizes were based on a technique assessment of VITs and were relatively small for contrasting treatment and control survival of neonates (Bishop et al. 2002). Treatment fawn survival was 0.613 (SE = 0.115, n = 29) and control fawn survival was 0.511 (SE = 0.108, n = 25). In late December 2002 and early January 2003, we once again conducted 2 age classification helicopter surveys in the treatment and control units. On 12/31/02, we observed 91.9 fawns:100 does (SE = 8.4) in the treatment unit, and 52.2 fawns:100 does (SE = 6.9) in the control unit. On 1/21/03, we observed 52.6 fawns:100 does (SE = 6.4) in the treatment unit, and 36.8 fawns:100 does (SE = 3.9) in the control unit. The combined helicopter survey data indicated 68.1 fawns:100 does (SE = 5.6) in the treatment unit and 42.8 fawns:100 does (SE = 3.5) in the control unit. Oppositely, fawn:doe ratio estimates from ground classifications of doe groups during December 2002 – February 2003 were 47.7 fawns:100 does (SE = 6.3) in the treatment unit, and 63.4 fawns:100 does (SE = 7.5) in the control unit (t108 = 1.61, P = 0.110). As in 2001, fawn:doe ratio results were conflicting. Helicopter survey data varied between 2 different flights, but consistently indicated a treatment effect. Ground classification data did not indicate a treatment effect. Also, survival data combined with age ratio data indicate neonate production and survival was reasonably favorable during 2002, and not indicative of the low fawn recruitment observed during the late 1980’s and 1990’s. 2003 During June 2003, we captured and radio-collared 103 newborn fawns born from treatment and control radio-collared does (55 treatment fawns, 48 control fawns). The VITs worked well; we captured fawns from 41 of the 54 does fitted with VITs. Treatment fawn survival (June – Dec) was 0.624 (SE = 0.082) and control fawn survival was 0.483 (SE = 0.093). Final standard errors were larger than expected because a number of fawns shed collars prematurely when crossing fences during fall migration. Using helicopter surveys, we measured 62.4 fawns:100 does (SE = 5.3) in the treatment unit and 50.0 fawns:100 does (SE = 4.9) in the control unit. Estimates from ground classifications of doe groups were 68.0 fawns:100 does (SE = 7.6) in the treatment unit and 62.1 fawns:100 does (SE = 7.6) in the control unit. Age ratio estimates from the helicopter and the ground were more consistent during 2003 than in past years. Overall, observed fawn:doe ratios were consistent with treatment and control fawn survival rates measured from June to December. 2002-03 Survival rate point estimates were very similar during 2002 and 2003. Combined over both years, treatment survival (S(t) = 0.620, SE = 0.067) was higher (P = 0.189) than control survival (S(t) = 0.493, SE = 0.070). The high censor rate due to shed collars during fall affected the p-value. Neonate survival through July 15, 2002 and 2003, was significantly higher (P = 0.006) for treatment fawns (S(t) = 0.833, SE = 0.041) than control fawns (S(t) = 0.634, SE = 0.057). We are currently measuring 2004 neonate survival rates, which will be necessary for final interpretations as to the effectiveness of the treatment. Our results from 2001 and 2002 emphasize the inherent difficulties and biases associated with precisely measuring fawn:doe ratios, particularly in this research study. Ratios obtained from helicopter surveys were based on 2 short-duration flights/unit/year over spatially small units. Helicopter surveys were complicated by high deer densities in heavy cover, making both deer detection and fawn:doe classifications a considerable challenge. There is a variety of potential biases that may have affected the 37 �helicopter surveys, including differential sightability of does and fawns, double classification of some deer, and incorrectly classifying yearling bucks with small antlers. Ground fawn:doe ratio observations of radio-collared doe groups were made using spotting scopes and field glasses, where we commonly studied the deer for some time. Incorrect classifications during these surveys were likely minimal. For example, small-antlered yearling bucks (e.g. 3 – 6” spikes) were detected from the ground, whereas they were undoubtedly missed on occasion during helicopter surveys. We also obtained repeated observations for some of the radio-collared doe groups from the ground. The main potential bias affecting ground fawn:doe classifications was how observations were made. Many of the ground classifications in the Shavano Valley experimental unit were made by radio-tracking does during the day. On the other hand, a majority of ground classifications in the Colona experimental unit were based on observing deer groups as they entered openings to feed during the late afternoon. Our age ratio results were more consistent during 2003. Deer were not as concentrated during helicopter surveys, and unlike previous years, almost all of the ground classification data for the Colona experimental unit was obtained by radio-tracking does during the day. Given the inherent difficulties of measuring fawn:doe ratios in the 2 experimental units, and the lack of a clear indication as to the effectiveness of the treatment, we will only cautiously use fawn:doe ratios to make inferences regarding treatment effects. At the completion of the research, we will test whether enhanced winter nutrition of adult does improved newborn fawn survival based on a three-year model of the radio-collared neonate survival data. Neonate Mortality Causes During June − December of 2002 and 2003, 32 of 84 treatment fawns died: 8 – coyote predation, 2 – bear predation, 2 – felid predation, 3 – predation where the predator was undetermined, 9 – disease/starvation/ malnutrition, 1 – abandonment, 2 – trauma/injury, 1 – road-kill, 2 – unknown, and 2 – poached. The two poached fawns were censored from analyses evaluating the effect of the treatment. Converted to mortality rates based on the Kaplan-Meier survival analysis, 11.4% of all treatment fawns died from disease/starvation/malnutrition, 10.1% from coyote predation, 3.8% from predation where the predator was undetermined, 2.5% each from bear predation, felid predation, injury/trauma, and unknown causes, and 1.3% each from abandonment and road-kill. Simplified, 18.9% of all treatment fawns died from predation, 11.4% died from disease/starvation/malnutrition, and 7.6% died from other or unknown causes. During June – December of 2002 and 2003, 35 of 72 control fawns died: 12 – coyote predation, 4 – felid predation, 2 – bear predation, 1 – predation where the predator was undetermined, 11 – disease/starvation/ malnutrition, 1 – trauma/injury, and 4 – unknown. Converted to mortality rates based on the Kaplan-Meier survival analysis, 17.4% of all control fawns died from coyote predation, 15.9% died from disease/starvation/malnutrition, 5.8% each from felid predation and unknown causes, 2.9% from bear predation, and 1.4% each from trauma/injury and predation where the predator was undetermined. Simplified, 27.5% of all control fawns died from predation, 15.9% from disease/starvation/malnutrition, and 7.2% from other or unknown causes. In summary, mortality rates due to predation and disease/starvation/malnutrition were lower for treatment fawns than control fawns. Overwinter Fawn Survival and Mortality Causes During winter 2001-02 (Dec 10, 2001 – June 15, 2002), the survival rate of fawns was significantly greater (χ21 = 13.216, P < 0.001) in the treatment unit (S(t) = 0.865, SE = 0.056) than in the control unit (S(t) = 0.510, SE = 0.080). Similarly, in 2002-03 (Dec 10, 2002 – June 15, 2003), the overwinter survival rate of fawns was significantly greater (χ21 = 5.734, P = 0.017) in the treatment unit (S(t) = 0.900, SE = 0.047) than in the control unit (S(t) = 0.691, SE = 0.074). Again in 2003-04 (Dec 10, 2003 – June 15, 2004), the overwinter survival rate of fawns was significantly greater (χ21 = 3.852, P = 0.050) in the treatment unit (S(t) = 0.920, SE = 0.045) than in the control unit (S(t) = 0.756, SE = 0.067). Combining survival data across all 3 winters, treatment fawn survival (S(t) = 0.895, SE = 0.029) was 38 �higher (χ21 = 18.781, P < 0.001) than control fawn survival (S(t) = 0.655, SE = 0.044) (Fig. 4). The treatment unit during winter 2001-02 became the control unit during winters 2002-03 and 2003-04, and vice versa. Thus, the overwinter survival treatment effect was replicated across each experimental unit. Combining all years of data, the best model of overwinter fawn survival (AICc = 207.65) included treatment (χ21 = 19.04, P < 0.001), early winter fawn mass (χ21 = 23.27, P < 0.001), and year (χ21 = 6.20, P = 0.045). The AIC model selection analysis emphasizes the importance of both the treatment effect as well as early winter mass of fawns, because any models without treatment or fawn mass were very poor (Table 2). Survival of fawns receiving the nutrition enhancement treatment was 0.24 higher than survival of control fawns during three mild to average winters, and surviving fawns averaged 3.5 kg heavier than fawns that died. Early winter mass of control fawns was slightly higher than that of treatment fawns (F1, 231 = 3.00, P = 0.085); thus the effect of the treatment was not confounded with fawn mass. Fawn mass was similar between winters as well (F2, 231 = 1.31, P = 0.273). The importance of early winter fawn mass as a predictor of overwinter survival has been documented previously (White et al. 1987, Bishop 1998, White and Bartmann 1998, Unsworth et al. 1999). In summary, the nutrition enhancement treatment improved overwinter fawn survival and thus yearling recruitment, and heavier fawns in each experimental unit had higher survival probabilities. 1 ~ -, - -. -- ~ - 0.9 \ 0.8 7 '- - --- -. - ---- \., 'L...- - 0.7 - ------ - Treatment 0.6 - - -Control 0.5 ' 12/1 ' 12/16 12131 ' 1/15 ' 100 ' 2/14 ' ' ' ' 3/1 3/16 3/31 4115 ' 4/30 ' ' 5/15 51.30 Figure 4. Overwinter fawn survival (Dec 10 – June 15, 2001 – 2004) in a nutrition enhancement treatment unit (S(t) = 0.895, SE = 0.029) and a control unit (S(t) = 0.655, SE = 0.044) on the Uncompahgre Plateau, southwest Colorado. 39 ' 6/14 �Table 2. Model selection results for a logistic regression analysis of overwinter mule deer fawn survival in southwest Colorado, 2001− 2004. Enhanced nutrition (Treatment) and early winter fawn mass were the critical predictors of survival. Model selection was performed using Akaike’s Information Criterion (AIC). # -2 Log ∆ Param Likelih AIC eters c (K) AIC AICc Model Name od Treatment + Year + Mass Treatment + Sex + Year + Mass Treatment + Sex + Year + Trt*Year + Mass Treatment + Sex + Mass 202.254 5 212.254 207.649 0 201.227 6 213.227 207.783 0.13 201.060 8 217.060 210.026 2.38 207.179 4 215.179 211.440 3.79 Treatment + Mass 208.556 3 214.556 211.712 4.06 Sex + Year + Mass 223.598 5 233.598 228.993 21.34 Treatment 235.739 2 239.739 237.816 30.17 Sex + Year 248.878 4 256.878 253.139 45.49 During winters 2001-04, 12 of 115 treatment fawns died: 5 from coyote predation, 3 from disease/illness, 2 from malnutrition, 1 from trauma/injury, and 1 unknown. Each of the 3 fawns that died from disease had adequate fat stores. At least one of these fawns died as a result of pneumonia. Converted to mortality rates based on the Kaplan-Meier survival analysis, 4.3% of all treatment fawns died from coyote predation, 2.6% from disease/illness, 1.7% from malnutrition, 0.9% from trauma/injury, and 0.9% from unknown causes. Simplified, 4.3% of all treatment fawns died from predation, 4.3% from disease/malnutrition, and 1.8% from other or unknown causes. During winters 2001-04, 41 of 120 control fawns died: 13 from coyote predation, 8 from mountain lion predation, 8 from malnutrition, 6 from unknown causes, 3 from predation where the predator was undetermined, 2 were road-killed, and 1 from trauma/injury. Converted to mortality rates based on the Kaplan-Meier survival analysis, 10.9% of all control fawns died from coyote predation, 6.7% from mountain lion predation, 6.7% from malnutrition, 5.0% from unknown causes, 2.5% from predation where the predator was undetermined, 1.7% from road-kill, and 0.8% from trauma/injury. Simplified, 20.1% of all control fawns died from predation, 6.7% from malnutrition, and 7.5% from other or unknown causes. Most fawns killed by predators had little or no femur marrow fat remaining, indicating the predation was likely compensatory in nature. Adult Female Survival and Causes of Mortality During winter 2000-01 (Dec 1, 2000 – May 31, 2001), the adult doe survival rate in the treatment unit (S(t) = 0.968, SE = 0.032) was greater (χ21 = 2.649, P = 0.104) than the survival rate in the control unit (S(t) = 0.861, SE = 0.058). However, annual adult doe survival rates (Dec 1, 2000 – Nov 30, 2001) were similar among the treatment and control deer (Trt: S(t) = 0.839, SE = 0.066; Control: S(t) = 0.833, SE = 0.062; χ21 = 0.004, P = 0.947). We observed a similar result the following year. The 2001-02 overwinter adult doe survival rate in the treatment unit (S(t) = 0.942, SE = 0.030) was greater (χ21 = 3.116, P = 0.078) than survival in the control unit (S(t) = 0.848, SE = 0.044), yet annual adult doe survival was similar among treatment and control deer (Trt: S(t) = 0.824, SE = 0.049; Control: S(t) = 0.818, SE = 0.047; χ21 = 0.090, P = 0.764). Thus, mortalities of control deer occurred primarily during the winter months, while treatment does died primarily during the summer and fall months. 40 �During winter 2002-03, following the treatment cross-over, overwinter adult doe survival rates were similar among treatment and control deer (Trt: S(t) = 0.945, SE = 0.024; Control: S(t) = 0.924, SE = 0.028; χ21 = 0.360, P = 0.549). The main difference from the previous 2 years was that overwinter survival of adult does in the Shavano experimental unit increased in 2002-03 upon receiving the treatment. However, annual adult doe survival rates (Dec 1, 2002 – Nov 30, 2003) were higher (χ21 = 2.016, P = 0.156) for treatment does 0.888 (SE = 0.034) than control does 0.813 (SE = 0.041). The main difference from the previous 2 years was overwinter survival of adult does in the Shavano experimental unit increased in 2002-03 upon receiving the treatment. Summer-fall survival was similar in that Colona adult does had higher mortality rates than Shavano adult does. Thus, in 2002-03, there was no difference between survival rates of treatment and control adult does during winter but there was evidence of higher annual survival of treatment adult does. During winter 2003-04, overwinter adult doe survival rates were higher (χ21 = 3.843, P = 0.050) among treatment does (S(t) = 0.979, SE = 0.014) than control does (S(t) = 0.915, SE = 0.027). Thus far in 2004, annual adult doe survival rates (Dec 1, 2003 – 8/31, 2004) are 0.951 (SE = 0.021) for treatment does and 0.896 (SE = 0.029) for control does. Considering all years, the treatment has improved overwinter adult doe survival but had a relatively minor affect on annual survival. Considering only the past 2 years, there is evidence the treatment has had a positive affect on annual survival. Annual survival rates measured in this study align with expected survival based on other studies (Unsworth et al. 1999, B.E. Watkins, unpublished). During 2000-02, when the Colona experimental unit received the treatment and the Shavano experimental unit was the control, 16 treatment and 16 control does died. The 16 treatment does died from the following categories: 4 – road-killed, 3 – while giving birth, 3 – predation (undetermined predator), 2 – non-predation unknown (intact carcasses with no evidence of predation or scavenging), 1 – disease (chronic arthritis), 1 – mountain lion predation, and 2 – unknown. Predation was not a major mortality factor for treatment does, and a majority of mortalities were independent of nutrition (does were in good condition). The 16 control doe mortalities included the following causes: 5 – mountain lion predation, 3 – malnutrition, 2 – non-predation unknown, 1 – road-killed, 1 – bear predation, 1 – injury (fence), 1 – legal harvest, and 2 – unknown. Predation and malnutrition were the major mortality causes of control deer. Interestingly, during this 2-year period, we did not document any coyote predation on adult does. During Dec 2002 – August 2004, with Shavano as the treatment and Colona as the control, there have been 14 treatment doe mortalities: 5 – disease/infection, 5 unknown causes, 3 – coyote predation, and 1 – road-killed. As we saw during 2000-02, predation was not a major mortality factor for treatment does, and a majority of mortalities were independent of nutrition. There have been 26 control adult doe mortalities during this same time period: 7 – malnutrition/disease, 5 – road-kill, 4 – coyote predation, 4 – unknown causes, 3 – mountain lion predation, and 3 – non-predation unknown. Malnutrition, predation, and road-kill were the major mortality factors of control does during 2002-04. SUMMARY We successfully enhanced nutrition of deer occupying the treatment units. There was no evidence the treatment positively influenced fetus survival until 2004, when virtually all stillborn fetuses were from control adult does. We currently have evidence the treatment caused an increase in neonate survival; however, data collection is incomplete. The treatment caused a significant increase in overwinter fawn survival, which is where the greatest differences occurred between treatment and control deer. Overwinter adult doe survival increased as a result of the treatment, but annual survival was more similar among treatment and control adult does. Results reported here are based on preliminary analyses, and in some cases, incomplete data sets. Final analyses will be conducted once data collection is complete. 41 �LITERATURE CITED Andelt, W. F., T. M. Pojar, and L. W. Johnson. 2004. Long-term trends in mule deer pregnancy and fetal rates in Colorado. Journal of Wildlife Management 68:542-549. Baker, D. L., and N. T. Hobbs. 1985. Emergency feeding of mule deer during winter: tests of a supplemental ration. Journal of Wildlife Management 49:934-942. Baker, D. L., D. E. Johnson, L. H. Carpenter, O. C. Wallmo, and R. B. Gill. 1979. Energy requirements of mule deer fawns in winter. Journal of Wildlife Management 43:162-169. Baker, D. L., G. W. Stout, and M. W. Miller. 1998. A diet supplement for captive wild ruminants. Journal of Zoo and Wildlife Medicine 29:150-156. Ballard, W. B, D. Lutz, T. W. Keegan, L. H. Carpenter, and J. C. deVos, Jr. 2001. Deer-predator relationships: a review of recent North American studies with emphasis on mule and black-tailed deer. Wildlife Society Bulletin 29:99-115. Barrett, M. W., J. W. Nolan, and L. D. Roy. 1982. Evaluation of a hand-held net-gun to capture large mammals. Wildlife Society Bulletin 10:108-114. Bishop, C. J. 1998. Mule deer fawn mortality and habitat use, and the nutritional quality of bitterbrush and cheatgrass in southwest Idaho. Thesis, University of Idaho, Moscow, Idaho, USA. Bishop, C. J., D. J. Freddy, and G. C. White. 2002. Effects of enhanced nutrition of adult female mule deer on fetal and neonatal survival rates: a pilot study to address feasibility. Colorado Division of Wildlife, Wildlife Research Report, Federal Aid in Wildlife Restoration Project W-153-R, Job Final Report. Fort Collins, Colorado, USA. Burnham, K. P., and D. R. Anderson. 1998. Model selection and inference: a practical informationtheoretic approach. Springer-Verlag, New York, New York, USA. Connolly, G. E. 1981. Limiting factors and population regulation. Pages 245-285 in O. C. Wallmo, editor. Mule and black-tailed deer of North America. University of Nebraska Press, Lincoln, Nebraska, USA. Cook, R. C. 2000. Studies of body condition and reproductive physiology in Rocky Mountain elk. Thesis, University of Idaho, Moscow, Idaho, USA. Cook, R. C., and J. G. Cook. 2002. An informal training guide to condition evaluation in elk and deer. National Council for Air and Stream Improvement, Unpublished Report. Cook, R. C., J. G. Cook, D. L. Murray, P. Zager, B. K. Johnson, and M. W. Gratson. 2001. Development of predictive models of nutritional condition for Rocky Mountain elk. Journal of Wildlife Management 65:973-987. Gill, R. B., T. D. I. Beck, C. J. Bishop, D. J. Freddy, N. T. Hobbs, R. H. Kahn, M. W. Miller, T. M. Pojar, and G. C. White. 2001. Declining mule deer populations in Colorado: reasons and responses. Colorado Division of Wildlife Special Report Number 77. Denver, Colorado, USA. Holter, J. B., H. H. Hayes, and S. H. Smith. 1979. Protein requirement of yearling white-tailed deer. Journal of Wildlife Management 43:872-879. Kaplan, E. L., and P. Meier. 1958. Nonparametric estimation from incomplete observations. Journal of the American Statistical Association 53:457-481. Pojar, T. M., and D. C. Bowden. 2004. Neonatal mule deer fawn survival in west-central Colorado. Journal of Wildlife Management 68:550-560. Pollock, K. H., S. R. Winterstein, C. M. Bunck, and P. D. Curtis. 1989. Survival analysis in telemetry studies: the staggered entry design. Journal of Wildlife Management 53:7-15. Ramsey, C. W. 1968. A drop-net deer trap. Journal of Wildlife Management 32:187-190. SAS Institute. 1989a. SAS/STAT® user’s guide, version 6, fourth edition. Volume 1. SAS Institute, Cary, North Carolina, USA. SAS Institute. 1989b. SAS/STAT® user’s guide, version 6, fourth edition. Volume 2. SAS Institute, Cary, North Carolina, USA. SAS Institute. 1997. SAS/STAT® Software: Changes and Enhancements through Release 6.12. SAS Institute, Cary, North Carolina, USA. 42 �Schmidt, R. L., W. H. Rutherford, and F. M. Bodenham. 1978. Colorado bighorn sheep- trapping techniques. Wildlife Society Bulletin 6:159-163. Smith, S. H., J. B. Holter, H. H. Hayes, and H. Silver. 1975. Protein requirement of white-tailed deer fawns. Journal of Wildlife Management 39:582-589. Stephenson, T. R., V. C. Bleich, B. M. Pierce, and G. P. Mulcahy. 2002. Validation of mule deer body composition using in vivo and post-mortem indices of nutritional condition. Wildlife Society Bulletin 30:557-564. Stephenson, T. R., K. J. Hundertmark, C. G. Schwartz, and V. Van Ballenberghe. 1998. Predicting body fat and body mass in moose with ultrasonography. Canadian Journal of Zoology 76:717-722. Stephenson, T. R., J. W. Testa, G. P. Adams, R. G. Sasser, C. G. Schwartz, and K. J. Hundertmark. 1995. Diagnosis of pregnancy and twinning in moose by ultrasonography and serum assay. Alces 31:167-172. Thompson, C. B., J. B. Holter, H. H. Hayes, H. Silver, and W. E. Urban, Jr. 1973. Nutrition of whitetailed deer. I. Energy requirements of fawns. Journal of Wildlife Management 37:301-311. Ullrey, D. E., W. G. Youatt, H. E. Johnson, L. D. Fay, and B. L. Bradley. 1967. Protein requirement of white-tailed deer fawns. Journal of Wildlife Management 31:679-685. Unsworth, J. W., D. F. Pac, G. C. White, and R. M. Bartmann. 1999. Mule deer survival in Colorado, Idaho, and Montana. Journal of Wildlife Management 63:315-326. van Reenen, G. 1982. Field experience in the capture of red deer by helicopter in New Zealand with reference to post-capture sequela and management. Pages 408-421 in L. Nielsen, J. C. Haigh, and M. E. Fowler, editors. Chemical immobilization of North American wildlife. Wisconsin Humane Society, Milwaukee, Wisconsin, USA. Verme, L. J., and D. E. Ullrey. 1972. Feeding and nutrition of deer. Pages 275-291 in D. C. Church, editor. Digestive physiology and nutrition of ruminants. Volume 3 – Practical Nutrition. D. C. Church, Corvallis, Oregon, USA. Watkins, B. E., D. E. Ullrey, R. F. Nachreiner, and S. M. Schmitt. 1983. Effects of supplemental iodine and season on thyroid activity of white-tailed deer. Journal of Wildlife Management 47:45-58. Watkins, B. E., J. H. Witham, D. E. Ullrey, D. J. Watkins, and J. M. Jones. 1991. Body composition and condition evaluation of white-tailed deer fawns. Journal of Wildlife Management 55:39-51. White, G. C., and R. M. Bartmann. 1998. Effect of density reduction on overwinter survival of freeranging mule deer fawns. Journal of Wildlife Management 62:214-225. White, G. C., R. A. Garrott, R. M. Bartmann, L. H. Carpenter, and A. W. Alldredge. 1987. Survival of mule deer in northwest Colorado. Journal of Wildlife Management 51:852-859. White, G. C., A. F. Reeve, F. G. Lindzey, and K. P. Burnham. 1996. Estimation of mule deer winter mortality from age ratios. Journal of Wildlife Management 60:37-44. Prepared by _______________________ Chad J. Bishop, Wildlife Researcher 43 �Colorado Division of Wildlife July 2004 – June 2005 WILDLIFE RESEARCH REPORT State of Cost Center Work Package Task No. Colorado 3430 3001 4 : : : : Federal Aid Project: W-185-R : Division of Wildlife Mammals Research Deer Conservation Effect of Nutrition and Habitat Enhancements on Mule Deer Recruitment and Survival Rate Period Covered: July 1, 2004 - June 30, 2005 Authors: C. J. Bishop, G. C. White, D. J. Freddy, and B. E. Watkins Personnel: D. L. Baker, L. Baeten, T. M. Banulis, E. J. Bergman, S. K. Carroll, M. J. Catanese, D. L. Coven, K. Crane, M. L. DelTonto, B. Diamond, B. deVergie, P. Ehrlich, D. Gallegos, J. Garner, L. Gepfert, R. B. Gill, D. Hale, J. L. Grigg, H. J. Halbritter, R. Harthan, M. D. Johnston, W. J. Lassiter, C. T. Larsen, T. Mathieson, J. W. McMillan, G. C. Miller, M. W. Miller, J. D. Nicholson, J. A. Padia, T. M. Pojar, R. M. Powers, J. E. Risher, C. A. Schroeder, W. G. Sinner, C. M. Solohub, M. H. Swan, K. Taurman, J. A. Thayer, M. A. Thonhoff, C. E. Tucker, R. M. Wertsbaugh, L. L. Wolfe, CDOW; H. VanCampen, CSU; D. Felix, Olathe Spray Service; T. R. Stephenson, California Fish and Game; L. H. Carpenter, WMI; J. Sazma, B. Welch, BLM. Project support received from Federal Aid in Wildlife Restoration, Mule Deer Foundation, and Colorado Habitat Partnership Program. All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT We measured mule deer (Odocoileus hemionus) population parameters in response to a nutrition enhancement treatment to evaluate the relative importance of habitat quality as a limiting factor of mule deer in western Colorado. During November 2000 – June 2004, we captured and radio-collared 810 individual mule deer evenly distributed among treatment and control units on the Uncompahgre Plateau in southwest Colorado. Our sample included 293 adult females, 154 of which received vaginal implant transmitters (VITs), 241 6-month-old fawns, and 276 newborn fawns born from either treatment or control adult does. We enhanced the nutrition of deer in the treatment unit by providing a safe, pelleted supplemental feed on a daily basis from December through April each winter. The treatment unit during winters 2000−01 and 2001−02 became the control unit during winters 2002−03 and 2003−04, and vice versa. Thus, the treatment effect was replicated across each experimental unit. Early winter fawn:doe ratios were measured using helicopter and ground classification surveys the year following treatment delivery to determine whether fawn production and survival increased as a result of enhanced nutrition of adult females. During winters 2001–02 through 2003−04, we measured pregnancy rates, fetus rates, and body condition of treatment and control adult does using ultrasonography. We measured fetus survival and neonate survival by using VITs to help locate and radio-collar newborn fawns born from treatment 37 �and control does. We also measured overwinter fawn survival rates in response to the treatment. Estimated percent body fat of adult does during late February and early March, 2002−04, was higher (F1, 148 = 153.41, P < 0.001) for treatment deer (9.8%, SE = 0.36, n = 78) than control deer (4.3%, SE = 0.26, n = 76). Serum thyroid hormone concentrations, measured only in 2003 and 2004, were higher in treatment does than control does (F4, 108 = 46.59, P < 0.001). Pregnancy and fetus rates were similar among treatment and control does. The pregnancy rate of adult does was 0.95 (SE = 0.036, n = 38) and the fetus rate was 1.80 fetuses/doe (SE = 0.10, n = 36) during 2002. Rates were similar in 2003, where we measured a pregnancy rate of 0.92 (SE = 0.034, n = 63) and a fetus rate of 1.74 fetuses/doe (SE = 0.069, n = 50) which included 5 yearlings (the fetus rate excluding yearlings was 1.82 fetuses/doe, SE = 0.066, n = 45). In 2004, pregnancy rate was 0.94 (SE = 0.029, n = 66) and fetus rate was 1.97 fetuses/doe (SE = 0.053, n = 60), which included 4 yearlings (fetus rate excluding yearlings was 2.00 fetuses/doe, SE = 0.051, n = 56). Based on multiple early winter age classification surveys, we lacked evidence to determine whether the winter nutrition enhancement treatment had any effect on neonatal production and survival during 2001, which provided additional incentive to directly measure fetus and neonate survival. During 2002−2004, fetus-neonate survival from 1 March−15 December was higher (χ21 = 3.846, P = 0.050) for treatment fawns (S(t) = 0.528, SE = 0.027) than control fawns (S(t) = 0.401, SE = 0.025). Survival data coupled with early winter age classification surveys provided evidence the nutrition enhancement treatment increased December fawn recruitment during 2002−2004. During 15 December– 15 June, 2001−2004, the overwinter survival rate of fawns was greater (χ21 = 18.781, P < 0.001) in the treatment unit (S(t) = 0.895, SE = 0.029) than in the control unit (S(t) = 0.655, SE = 0.044). Using a staggered entry survival process with data combined over years, survival of treatment fetuses to 1 year of age (S(t) = 0.458, SE = 0.031) was 0.18 higher (χ21 = 13.20, P < 0.001) than survival of control fetuses to 1 year of age (S(t) = 0.276, SE = 0.026). The finite rate of population increase, λ, based on our measurements of treatment population parameters was 1.20, which would cause the deer population to double in approximately 4 years. The finite rate of increase calculated from control deer was 1.04, indicating a stable or slightly increasing population. The nutrition enhancement treatment therefore had a dramatic effect on deer population performance, indicating habitat quality was ultimately limiting the population. Our results provide a foundation for focusing deer management efforts on improving habitat quality in western Colorado pinyon-juniper (Pinus edulis-Juniperus osteosperma) ecosystems with corresponding research efforts to quantify the effects of habitat manipulations on deer performance. 38 �WILDLIFE RESEARCH REPORT EFFECT OF NUTRITION AND HABITAT ENHANCEMENTS ON MULE DEER RECRUITMENT AND SURVIVAL RATES CHAD J. BISHOP, GARY C. WHITE, DAVID J. FREDDY, AND BRUCE E. WATKINS P. N. OBJECTIVES 1. To determine experimentally whether enhancing mule deer nutrition during winter and early spring via supplementation increases fetus survival, neonate survival, overwinter fawn survival, or ultimately, population productivity. 2. To determine experimentally to what extent habitat treatments replicate the effect of enhanced nutrition from supplemental feeding. SEGMENT OBJECTIVES 1. Radio-monitor and measure survival of the sample of radio-collared mule deer adult does and fawns. 2. Measure early winter fawn:doe ratios using both aerial helicopter surveys and ground classifications of deer groups associated with radio-collared adult does. 3. Summarize and analyze data and publish information in an annual Job Progress Report. 4. Complete a peer-reviewed manuscript for publication in a scientific journal pertaining to the effectiveness of vaginal implant transmitters for capturing mule deer neonates exclusively from radiomarked adult does (Appendix I). INTRODUCTION Mule deer (Odocoileus hemionus) numbers apparently declined during the 1990s throughout much of the West, and have clearly decreased since the peak population levels documented during the 1940s−1960s (Unsworth et al. 1999, Gill et al. 2001). Biologists and sportsmen alike have concerns as to what factors may be responsible for declining population trends. Although previous and current research indicates multiple interacting factors are responsible, habitat and predation have received the focus of attention. A number of studies have evaluated whether predator control increases deer survival, yet results are highly variable (Connolly 1981, Ballard et al. 2001). Together, predator control studies with adequate rigor and statistical power indicate predation effects on mule deer are variable as a result of time-specific and site-specific factors. Studies which have demonstrated deer population responses to predator control treatments have failed to determine whether predation is ultimately more limiting than habitat when considering long term population changes. Numerous research studies have evaluated mule deer habitat quality, but virtually no studies have documented population responses to habitat improvements. In many areas where declining deer numbers are of concern, predation is common yet habitat quality appears to have declined. The question remains as to whether predation, habitat, or some other factor is more limiting to mule deer in these situations, and whether habitat quality can be improved for the benefit of deer. It may also be that no single factor is responsible for observed deer declines, and a more comprehensive understanding of multi-factor interactions is needed. We designed and implemented a field experiment where we measured deer population responses to a nutrition enhancement treatment to further understand the causative factors underlying observed deer population dynamics. We conducted the study on the Uncompahgre Plateau in southwest Colorado, where several predator species were present in abundant numbers: coyotes (Canis latrans), mountain lions (Felis concolor), and bears (Ursus americanus). In addition to predation, myriad diseases in combination have proximately affected survival of the Uncompahgre deer population (Pojar and Bowden 39 �2004, B. E. Watkins, Colorado Division of Wildlife, unpublished data). Predator numbers were not manipulated in any manner during the course of the study. All factors were left constant with the exception of deer nutrition. Deer nutrition was enhanced by providing supplemental feed to deer occupying a treatment area during winter. We measured December fawn recruitment and overwinter fawn survival in response to the treatment to determine whether deer nutrition was ultimately more limiting than predation or disease. The second phase of the research will incorporate habitat manipulation treatments. The treatments will consist of prescribed fire or mechanical techniques to set back succession of pinyon-juniper (Pinus edulis-Juniperus osteosperma) habitat in an effort to improve the vigor and quality of winter habitat for mule deer. Deer population responses will be measured in relation to the habitat manipulations in the same manner as the supplemental feed. Thus, the experiment evaluates whether nutritional quality of winter range habitat is ultimately more limiting than other factors in a lateseral pinyon-juniper and sagebrush (Artemisia spp.) landscape, and if so, whether habitat can be effectively improved for mule deer. The results advance our understanding of multi-factor interactions, with direct implications for mule deer management. STUDY AREA We non-randomly selected two experimental units (A−B) within mule deer winter range on the Uncompahgre Plateau (Figure 1) to facilitate a cross-over experimental design for evaluating the effects of enhanced deer nutrition during winter on annual population performance. We used the following criteria to select experimental units: 1.) Deer densities (≥80 deer/mi2): we selected areas where deer densities were sufficient to meet sample size requirements within the experimental unit, while simultaneously selecting areas that would require feeding no more than 600−800 animals during a normal winter; 2.) Buffer zones: we selected areas such that experimental units would be separated by several miles of non-treatment area (buffers) to prevent deer from occupying more than one experimental unit; 3.) Similarity: we selected areas that comprised relatively similar habitat complexes and deer densities that were representative of the overall area; 4.) Elk populations: we selected areas in an effort to minimize the number of elk present during normal winters. Units A and B received the nutrition enhancement treatment in a cross-over experimental design to address P.N. Objective 1. Unit A served as the treatment unit, while Unit B served as the control, for the first 2 winters of research (2000 – 2002). During winters 2002−03 and 2003−04, Unit B received the treatment while Unit A served as the control. Upon completion of P.N. Objective 1, additional winter range experimental units will be used to conduct phase 2 of the research, or P.N. Objective 2. Habitat in treatment units will be manipulated to set back plant succession, while habitat in control units will remain unchanged throughout the experiment. Experimental units A and B were defined as follows (Figures 2 and 3): (1) Experimental unit A included the Colona Tract of the Billy Creek State Wildlife Area and adjacent land, located approximately 13 km south of Montrose, CO adjacent to U.S. Hwy 550 South. The experimental unit was located within the Colona USGS 7.5 Minute Quadrangle, and roughly included the polygon defined by the following Zone 13 UTM coordinates: (1) 254000 E, 4250200 N; (2) 252700 E, 4249400 N; (3) 254700 E, 4245600 N; and (4) 256200 E, 4246600 N. (2) Experimental unit B included Shavano Valley and adjacent land extending west to the Dry Creek Rim. Shavano Valley is located approximately 13 km west of Montrose, CO. The experimental unit was located within the Dry Creek Basin and Montrose West Quadrangles (USGS 7.5 Minute), and 40 �roughly included the polygon defined by the following Zone 13 UTM coordinates: (1) 238400 E, 4262600 N; (2) 232400 E, 4256700 N; (3) 235000 E, 4253600 N; and (4) 239500 E, 4258200 N. In late April and May, prior to fawning, deer from the winter range experimental units migrated to summer range. We defined the summer range study area by movements of the radio-collared deer captured on winter range; summer range encompassed >1000 mi2 covering the southern portion of the Uncompahgre Plateau and adjacent San Juan Mountains (Figure 2). The summer range study area extended north to the Dry Creek river drainage on the Uncompaghre Plateau, south to Mineral Creek near Silverton, CO, east to the Big Blue River drainage, and west to the San Miguel River canyon. However, a majority of the radio-collared deer summered on the Uncompahgre Plateau between Dry Creek to the north and Highway 62 to the south. Winter range elevations ranged from 1830 m (6000 ft) in Shavano Valley to 2318 m (7600 ft) adjacent to the Dry Creek Rim above Shavano Valley. Winter range habitat was dominated by pinyonjuniper with interspersed sagebrush adjacent to agricultural fields in the Shavano and Uncompahgre Valleys. Summer range elevations occupied by deer ranged from 1891 m (6200 ft) in the Uncompahgre Valley to 3538 m (11,600 ft) in Imogene Basin southwest of Ouray, CO. Summer range habitats were dominated by spruce-subalpine fir (Picea spp.-Abies lasiocarpa), aspen (Populus tremuloides), sagebrush, ponderosa pine (Pinus ponderosa), Gambel oak (Quercus gambelii), and to a lesser extent, pinyon-juniper at lower elevations. METHODS Response Variables We measured fetal and neonatal survival rates, early winter fawn:doe ratios, and overwinter fawn survival rates of deer occupying the treatment and control units. We delivered the nutrition enhancement treatment to deer from December through April, assessed fetus survival during June, measured neonate survival from June to December, and fawn:doe ratios during December−February (1 year after the treatment was initiated). We measured overwinter fawn survival from December to June in direct response to the current winter’s treatment. Our measurements determined whether enhanced winter nutrition of adult does increased subsequent newborn fawn production and survival, and whether enhanced winter nutrition of 6−12-month old fawns increased overwinter fawn survival. Ultimately, these measurements provided an assessment of the effect of winter range habitat quality on yearling recruitment, and thus population productivity. We also measured overwinter and annual survival of adult does as a function of enhanced winter nutrition. Sample Size Fetus/Neonate Survival: Fetus and neonate sample sizes were not independent of one another because each resulted from the sample size of radio-collared does. We therefore needed a target sample size of either fetuses or neonates to generate our adult doe sample size. We based our sample size calculations on quantifying neonate survival because it was our highest priority and we could generate reliable estimates. Target fetus sample sizes were difficult to estimate because of uncertainty identifying fetus fates. That is, many fetuses measured in utero during winter were not accounted for as live or dead at parturition. Fetus survival rates could only be measured from some unpredictable fraction of the radiocollared doe sample, making sample size calculations of limited use. For neonate survival, a sample size of 40 neonates per experimental unit per year would provide power of 0.81 to detect a difference of 0.15 in survival between the 2 experimental units if survival among control fawns was 0.40. We assumed a control survival rate of 0.40 based on previous neonate survival rates measured on the Uncompahgre (Pojar and Bowden 2004) in combination with December fawn:doe ratios measured during the late 1980s and 1990s, when the Uncompahgre population declined (B. E. Watkins, Colorado Division of Wildlife, 41 �unpublished data). We considered 40 neonates per experimental unit per year a minimum sample size because we ideally wanted to detect a difference in neonate survival of <0.15 between experimental units. Based on Bishop et al. (2002), we determined that 60 radio-collared does (30 treatment and 30 control) equipped with vaginal implant transmitters (VITs) would be necessary to capture a minimum of 80 newborn fawns. We also assumed that some fawns would be captured from other treatment and control radio-collared does not equipped with VITs. The 60 radio-collared does with VITs were also used to evaluate fetus survival; however, logistical constraints limited the power of fetus survival comparisons among experimental units. Early winter fawn:doe ratios: We desired to detect an effect size, i.e., an increase in fawn:doe ratios in response to the treatments, in the range of 15 to 20 fawns per 100 does. These values were based on population models with overwinter fawn survival of 0.444, adult female survival of 0.853, and December fawn:doe ratios of 66 fawns per 100 does to obtain a stationary population (Unsworth et al. 1999). Based on surveys of the Uncompahgre deer population during the 1990s, the standard deviation of the fawns:100 does ratio for groups with at least one adult female was 57, with a mean of 41. Using an expected standard deviation of 57, the standard error of the mean fawns:100 does ratio for 40 radiocollared does is 57/(401/2) = 9.0, which is the expected standard deviation of measured fawns:100 does ratios on each experimental unit. We assessed power using a two-sample t-test with a sample size of 4, representing the 4 years of the study where fawn:doe ratios were measured in response to enhanced nutrition. Our power to detect an increase of 20 fawns per 100 does based on classification of 40 radiocollared doe groups in each experimental unit was about 0.87. Overwinter fawn survival: Our sample size of 40 fawns per experimental unit per year provided a power of 0.81 to detect a difference of 0.15 in survival between the 2 experimental units assuming a control survival rate of 0.40. We expected to see an increase in fawn survival (effect size) of approximately 0.15, because this was the difference measured in the density reduction experiment conducted by White and Bartmann (1998). We assumed a control survival rate of 0.40 based on longterm data from Colorado, Idaho, and Montana (Unsworth et al. 1999). However, recent data from 5 deer populations in Colorado indicates overwinter fawn survival has commonly been ≥70% during the past 6 years (Colorado Division of Wildlife, unpublished data). Adult and 6-month Old Fawn Capture We captured adult does and 6-month-old fawns during November and December using baited drop nets (Ramsey 1968, Schmidt et al. 1978) and helicopter net guns (Barrett et al. 1982, van Reenen 1982). We baited drop nets with certified weed-free alfalfa hay and apple pulp. We used drop nets as the principle capture technique for a 3−4 week capture period; we then used helicopter net-gunning at the end of the drop-net capture to secure the remainder of deer needed to meet our target sample sizes. All deer were hobbled and blind-folded after being captured. We used stretchers to carry deer away from the net when using drop nets. Deer were fitted with nylon-belting radio collars equipped with mortality sensors; pulse rate increased after remaining motionless for 4 hours. We placed permanent collars on adult females and temporary collars on fawns. To make collars temporary, we cut one end of the collar in half and reattached the two ends using rubber surgical tubing; fawns shed the collars ≥6 months post-capture. We stitched a rectangular piece of flexible plastic (Ritchey® neck band material) engraved with a unique identifier to the side of each collar. The unique identifier consisted of 2 symbols for adult females, and 1 symbol on 2 different colors of plastic for fawns. We used the identifiers to visually identify deer from the ground, which allowed us to effectively document use of the treatment, measure fawn:doe ratios, and assess experimental unit population size via mark-resight estimators. We recorded mass (kg), hind foot length (cm), and chest girth (cm) of each deer, and collected blood samples to evaluate disease prevalence. 42 �During late February and early March, we captured an additional 30 adult female deer in each experimental unit by net-gunning. Captured deer were ferried by the helicopter to a central processing location, where deer were carried by stretchers to a tent for handling. We used ultrasonography to measure pregnancy status, fetal rate, and body condition of each captured deer. We retained and radiocollared pregnant does only. We then inserted a vaginal implant transmitter (VIT) in each doe as a technique for locating the timing and location of her birth site the following June. We also recorded the weight (kg), hind foot length (cm), and chest girth (cm) of each deer, and collected blood samples to evaluate disease prevalence. Body Condition and Reproductive Status We estimated body fat of treatment and control adult does during mid-late winter using an Aloka 210 (Aloka, Inc., Wallinford, Conn.) or SonoVet 2000 (Universal Medical Systems, Bedford Hills, NY) portable ultrasound unit with a 5 MHz linear transducer. We measured maximum subcutaneous fat thickness on the rump (MAXFAT) following the methodology of Stephenson et al. (1998, 2002). We also measured thickness of the longissimus dorsi muscle via ultrasound (Cook et al. 2001, Stephenson et al. 2002). A small area of hair was shaved to ensure contact between the transducer and the skin. Lubricant was applied to the shaved area for conduction purposes and fat and muscle thickness were measured using electronic calipers. We coupled the ultrasound measurements with body condition scores (BCS) obtained from palpation of the ribs, withers, and rump (Cook 2000). MAXFAT and rump BCS measurements were combined into a condition index used to estimate percent body fat (Cook and Cook 2002): % Fat = -6.6387617 + 7.4271417x – 1.11579443x2 + 0.07733803x3 where x = rLIVINDEX = (MAXFAT – 0.15) + rump BCS (if MAXFAT < 0.15, then rLIVINDEX = rump BCS). The rLIVINDEX and body fat regression was initially developed and validated for elk by Cook et al. (2001), and then modified by incorporating a validation of MAXFAT for mule deer performed by Stephenson et al. (2002). We also evaluated differences in serum thyroid hormone concentrations between treatment and control adult does during mid-late winter. Specifically, we measured total thyroxine (T4), free T4 (FT4), total tri-iodothyronine (T3), and free T3 (FT3) following the methodologies of Watkins et al. (1983, 1991). Blood samples were collected at the time of capture, and serum hormone analyses were performed by the Michigan State University Animal Health Diagnostic Laboratory (East Lansing, Michigan). We compared serum thyroid hormone concentrations between treatment and control adult does, and also compared hormone levels to body fat estimates derived from the ultrasonography. We quantified reproductive status (Stephenson et al. 1995, Andelt et al. 2004) with ultrasound via transabdominal scanning using a 3 MHz linear transducer. We searched for fetuses by scanning a portion of the abdomen that was shaved caudal to the last rib and left of the midline. We systematically searched each uterine horn to identify fetal numbers ranging from 0 to 3. Whenever possible, we measured eye diameter of each fetus to approximately estimate fetal age and parturition date. Vaginal Implant Transmitters (VITs) We used VITs manufactured by Advanced Telemetry Systems, Inc. (Isanti, MN). The VIT was 76 mm long, excluding antenna length, and had 2 silicone wings with a width of 57 mm when fully spread apart. The silicone wings were used to retain the transmitter in the vagina until parturition. The VIT weighed 15 grams and contained a 10−28 lithium battery programmed to a 12-hour on/off cycle. The diameter of the transmitter (excluding wings) was 14 mm, and was encased in an impermeable, water-proof, electrical resin. The transmitter contained an embedded heat-sensor which dictated the frequency pulse rate. When the heat sensor dropped below 90°F, synonymous with transmitter expulsion from the deer, the pulse rate changed from 40 PPM to 80 PPM. VIT batteries were programmed to be active from 0430 to 1630 hrs prior to daylight savings, and thus were active from 0530 to 1730 hrs after daylight savings and during the fawning period. We inserted VITs into deer using a vaginoscope 43 �(Jorgensen Laboratories, Inc., Loveland, CO) and alligator forceps. The vaginoscope was 6” long with a 5/8” internal diameter and had a machined end (smooth surface) to minimize trauma when inserted into the vagina. A discreet mark was placed on the applicator showing approximate insertion distance. We obtained the length of a typical mule deer vaginal tract by taking measurements from road-killed deer and other fresh deer carcasses obtained in the study area. Prior to use in the field, VITs were sterilized using chlorhexidine, air-dried, and sealed in a 3” × 8” sterilization pouch. We used sterilization containers with diluted chlorhexidine on site during capture to sterilize the vaginoscope and alligator forceps between each use. We used a new pair of nitrile surgical gloves to handle the vaginoscope and VIT for each deer. To insert a VIT, the plastic wings were folded together and placed into the end of the vaginoscope. We liberally applied sterile KY Jelly® to the scope and inserted it into the vaginal canal until the tip of the VIT antenna was approximately flush with the vulva. We used the alligator forceps, which extended through the vaginoscope, to firmly hold the VIT in place while the scope was pulled out from the vagina. The VIT silicone wings spread apart upon removal of the scope to hold the transmitter in place. The transmitter antenna was typically flush with the vulva, but on occasion extended up to 1 cm beyond the vulva. The tip of the antenna was encapsulated in a wax bead to protect the deer. All capture and handling procedures, including VIT techniques, were approved by the Colorado Division of Wildlife’s Animal Care and Use Committee (project protocols 11−2000 and 1−2002). Neonate Fawn Capture All radio-collared adult does were relocated from the air during late May to identify likely fawning areas. During each morning of June we checked VIT signal status by aerially relocating radiocollared does having VITs. Implant radio-signals could not be easily monitored from the ground because of weak signal strength and a large study area. Flights began at 0530 hours and were usually completed by 1000–1100 hours. Early flights were necessary to detect fast signals because temperature sensors of VITs expelled in open habitats and subject to sunlight often exceeded 90°F by mid-day, which caused VITs to switch back to a slow (i.e. prepartum) pulse. When a fast (i.e. postpartum) pulse rate was detected, we ground-tracked both the VIT and radio-collar frequencies simultaneously because the shed VIT and adult doe were typically in close proximity to one another. We attempted to observe behavior of the collared doe, establish whether the VIT was shed at a birth site, and search for fawns in the vicinity of the doe and expelled VIT. In cases where the doe had moved away from the VIT (e.g. >200 m), we located the VIT to determine whether shedding occurred at a birth site and whether any stillborn fawn(s) were present, and subsequently located the collared doe to search for fawns at her location. We attempted to account for each doe’s fetuses as live or stillborn fawns in order to quantify in utero fetus survival from February to birth. All personnel wore surgical gloves when handling fawns to help minimize human scent. We placed a drop-off radio-collar on each live fawn; radio collars were constructed with elastic neck-band material to facilitate expansion. Hole-punched, vinyl-belting tabs extended from the end of the elastic and from the transmitter for attachment purposes. We made collars temporary by cutting the vinyl tab extending from the elastic and reattaching the belting with latex tubing, which caused the collars to shed from the animal >6 months post-capture. Some collars were shed prematurely (i.e. 4−5 months postcapture) in association with fences during fall migration. For each fawn, mass (kg) and hind foot length (cm) were recorded, and a nasal swab sample was collected to screen for Bovine Viral Diarrhea. We then recorded basic vegetation characteristics of the birth site and promptly exited the site. We ground-relocated most of the radio-collared does not receiving VITs approximately every other day during June in an attempt to capture additional fawns from treatment and control does. We did the same for any VIT doe whose implant failed because of premature expulsion or battery failure. We relied on doe behavior and searches in the vicinity of the collared does to locate fawns. We worked in 44 �pairs and partitioned the study area into segments, whereby each 2-person team was responsible for one segment. We used 3−4 teams during 2002 and 5−6 teams during 2003 and 2004. Measurement of Survival Rates and Fawn:Doe Ratios We measured survival rates by radio-monitoring collared deer from the ground and air to determine fate (i.e. lived or died). We also attempted to determine the cause of each mortality, with a primary goal of distinguishing between predation and non-predation mortality causes. We radiomonitored deer from the ground on a daily basis year-round and from the air on approximately a biweekly basis. We detected signals from nearly all radio-collared deer each day during winter, which typically allowed us to arrive at mortality sites within 24 hours of the mortality event. During summer and migration periods, deer were distributed widely and thus were more difficult to radio-monitor. All radiocollared neonates were checked daily throughout the summer and fall, whereas some adult and yearling deer could not be ground-monitored on a routine basis. In result, we typically located neonate mortalities within 24 hours of death, but some adult deer mortalities were not detected for several days, or on rare occasion, for one or more weeks. Fresh, intact neonate carcasses were collected and submitted to the Colorado Division of Wildlife’s Wildlife Health Laboratory or the Colorado State University Diagnostic Laboratory for necropsy and tissue analyses. Fresh, intact adult and 6-month-old fawn carcasses were also submitted for laboratory necropsy when feasible. Field necropsies were performed on all other deer mortalities, and when appropriate, tissue samples were collected and submitted for analysis. Each winter we used the radio-collared does to measure fawn:doe ratios in each experimental unit. The resulting fawn:doe ratio was a measurement of the previous year’s treatment effect. We measured fawn:doe ratios using 2 techniques: (1) We located the sample of radio-collared does in each experimental unit from a fixed-wing airplane, and used the set of locations to define boundaries for the experimental unit. Shortly after (i.e. 1−2 days), we used a helicopter to systematically fly the defined unit and classify all deer groups encountered. For each group, we documented whether a radio-collared doe was present. (2) We located each radio-collared doe by radio telemetry from the ground. The group of deer with the collared doe was counted and classified by age and sex. Both methods were employed to gather as much information as possible to determine whether there was a treatment effect. The “true” value cannot be measured perfectly because of the inherent biases and potential sources of error associated with each technique. Thus, by employing both techniques, we had a greater chance of fully understanding whether the treatment caused an effect. Treatment Delivery We enhanced deer nutrition in the treatment experimental unit by providing a safe, pelleted supplemental feed. The supplemental feed was developed through extensive testing with both captive and wild deer (Baker and Hobbs 1985, Baker et al. 1998), and has been safely used in both applied research and management projects. We distributed pellets daily using 4wd pickup trucks, ATVs, and snowmobiles on primitive roads throughout the experimental unit to provide a food source for the entire deer population in the treatment unit. We carried each 50 lb. bag of pellets ≤200 m from the vehicle and distributed it by hand in approximately 20−30 small piles of feed in a linear fashion. We distributed numerous bags in successive order to create straight lines of feed that spanned most of the treatment area, which prevented animal concentrations. Our feeding technique also prevented dominant animals from restricting access to the food supply because of the large area over which pellets were distributed. We attempted to supply pellets ad libitum such that residual pellets remained when the next day’s ration was provided. We closely monitored collared deer to ensure that treatment deer remained in the experimental unit and actually consumed the feed, and to make sure that non-treatment deer remained in the control unit, which they did. The few treatment adult does that distinctly moved away from the treatment unit were withdrawn from the sample for purposes of measuring treatment effects. However, to avoid any biases, all 6-month-old fawns captured in the treatment unit were included in survival analyses regardless 45 �of whether they accessed the supplement or not. Some fawns died shortly after capture (i.e. 2−3 weeks), before we could document whether they had access to the feed. Censoring these individuals would have biased treatment survival high relative to control survival. Also, very few fawns that survived more than 2−3 weeks moved away from the treatment unit. The pelleted ration was commercially produced in the form of 2×1×0.5-cm wafers (Baker and Hobbs 1985). Feed quality (e.g. digestible energy, protein) vastly exceeded those of typical winter range deer diets; exact constituent values are provided by Baker et al. (1998). When provided ad libitum, the feed should have allowed deer to meet or exceed nutritional requirements for growth and maintenance (Ullrey et al. 1967, Verme and Ullrey 1972, Thompson et al. 1973, Smith et al. 1975, Baker et al. 1979, Holter et al. 1979). The basis for feeding such high quality pellets was to ensure that the treatment (enhanced nutrition) was effectively delivered to the deer. Our intent was not to determine the exact level of nutrition necessary to increase fawn recruitment, but rather to determine if nutrition was a significant limiting factor to recruitment. We will rely on habitat manipulation treatments to evaluate what exactly can be done via management to increase fawn survival and recruitment if nutrition is deemed a critical limiting factor. Statistical Methods We estimated deer numbers in each experimental unit during the first year of research using helicopter and ground mark-resight surveys. We used the joint hypergeometric maximum likelihood estimator for helicopter surveys and the Bowden estimator for ground surveys, and we analyzed data in Program NOREMARK (Neal et al. 1993, Bowden 1993, White 1996). We used a general linear model in PROC GLM in SAS (SAS Institute 1989) to test for differences in estimated percent body fat between treatment and control adult does and a multivariate model to test for differences in T4, FT4, T3, and FT3 thryoid hormones between treatment and control does. We used PROG REG (SAS Institute 1989) to evaluate the relationship between estimated percent body fat and serum thyroid hormone concentrations. We entered all fawn:doe ratios from helicopter surveys into the CDOW Deer, Elk, and Antelope Management (DEAMAN) database (G. C. White, Colorado State University, software) and computed standard errors based on groups (Bowden et al. 1984). We analyzed fawn:doe ratios from ground surveys using PROC MIXED in SAS (SAS Institute 1997). We used a reduced model with experimental unit as the independent variable; we considered experimental unit as a fixed effect and radio-collared does within an experimental unit as random effects. We analyzed fetus survival with a binomial survival rate from the subset of does where all fetuses had known fates. We also indirectly analyzed fetus survival by comparing the February fetus rate with the number of live newborn fawns/doe observed in June using a change-in-ratio estimator (White et al. 1996). We estimated neonate and overwinter fawn survival and adult doe survival using a Kaplan-Meier survival analysis (Kaplan and Meier 1958, Pollock et al. 1989), and we contrasted survival among experimental units using chi-square analyses. We used a common entry date for analyzing neonate survival because staggered entry would have biased survival rates low due to early mortalities that occurred before most of the sample was captured. We analyzed continuous fetus-neonate-overwinter fawn survival from March of one year to June of the following year using a staggered-entry Kaplan-Meier survival analysis (Pollock et al. 1989). All neonates were entered into the survival analysis on a common date rather than the exact date of capture for the same reason mentioned above. We computed the finite rate of increase, λ, for treatment and control deer by constructing a deterministic age-structured population model using measured pregnancy and fetus rates, fetus survival, neonate survival, overwinter fawn survival, and annual adult doe survival. Results are based on preliminary analyses and should be treated as such. Other results are presented as data summaries incorporating means and standard errors, or in some cases, raw data values. 46 �RESULTS AND DISCUSSION Deer Capture During November and December 2000−2003, we captured and radio-collared 139 adult female mule deer evenly distributed among the treatment and control units. We also captured and radio-collared 241 6-month-old fawns during November and December 2001−2003 (40 fawns/unit/year). Due to budgeting constraints, we were unable to radio-collar 6-month old fawns during 2000. We captured an additional 154 adult females during late February and early March 2002−2004 and equipped them with radio collars and VITs. During June 2002−2004, we captured and radio-collared 276 newborn fawns from radio-collared adult females. Thus, the following results are based upon radio-monitoring of 810 individual mule deer evenly distributed among treatment and control units during November 2000−June 2004. Treatment Delivery 2000−01: We distributed 88 tons of supplemental pellets from December 15, 2000, through April 19, 2001. We distributed an average of 0.85 tons of feed each day throughout 22 feeding sites across the 2.3 mi2 treatment unit during most of the winter and spring. Deer were fed ad libitum because there was always residual feed remaining the next day during the feeding routine. We distributed each sack in approximately 20−30 distinct, small piles, resulting in >1000 small piles of feed throughout the treatment unit. Deer were able to effectively access the feed in small groups, and no aggression was ever observed among deer seeking access to the feed. Deer adapted to the pelleted supplement immediately and utilized it extensively throughout the winter. We continually monitored deer use of the feed from ground observation points, where we obtained 440 visual observations of radio-collared does consuming the feed. These observations, coupled with daily radio-monitoring and periodic aerial relocations, indicated 32 of the 37 radio-collared treatment does spent the entire winter and spring within the boundaries of the treatment unit and received the supplement on a daily basis. Mark-resight population estimates from March helicopter (489 deer, SE = 62) and ground (494 deer, SE = 81) surveys, coupled with feed consumption, indicated we fed roughly 450 to 500 deer during most of the winter and spring. Feed consumption declined coincident with spring green-up, although deer continued to use the feed through mid-late April, at which point they began migrating to summer range. We also fed approximately 25 to 30 elk, but the elk did not affect deer access to the feed. Deer in the control experimental unit did not receive feed or any other treatment. Based on helicopter mark-resight surveys, the deer density in the treatment unit in December was 120 deer/mi2 (SE = 9), but increased shortly after and was 213 deer/mi2 (SE = 27) in March. Deer densities in the control unit changed little from 83 deer/mi2 (SE = 12) in December to 101 deer/mi2 (SE = 14) in March. 2001−02: We distributed 194 tons of the supplement throughout the treatment unit from December 15, 2001, through April 25, 2002. We distributed 2.0−2.1 tons of feed each day for most of the winter and spring. The large increase in supplement distribution from the previous year occurred because a large number of elk descended into the Uncompahgre Valley during late fall. Elk arrived in unusually large numbers throughout much of the valley prior to the onset of treatment delivery. Once feeding was initiated, approximately 300−500 elk adapted to the feed and remained in or around the treatment unit throughout most of the winter. We could not deliver >2.1 tons of pellets per day given myriad logistical and budgetary constraints. Feed was not delivered ad libitum to all deer and elk in the treatment unit throughout the winter because residual feed was rarely observed during the next day’s distribution. However, daily field observations indicated most deer approached ad libitum consumption of the supplement. In contrast to the previous winter, deer were waiting for the daily supplement to arrive each morning. Deer then consumed the supplement immediately after it was distributed. Elk were rarely observed utilizing the 47 �feed until late morning or afternoon, and elk continued to forage in fields below the treatment unit, whereas deer did not. We observed numerous radio-collared deer consuming pellets each day; not all of these observations were recorded because of time constraints with distributing the feed. Given this time limitation, we still recorded 818 observations of radio-collared deer consuming the supplemental feed (497 collared doe observations and 321 collared fawn observations). We observed 100−300 deer utilizing the pellets most days during the course of distributing the supplement. These observations rarely included elk; thus, direct deer-elk competition was minimized because of temporal differences in feeding, and deer had first access to the feed. 2002−03: We switched the treatment and control units consistent with the cross-over experimental design in December 2002. We distributed 97 tons of supplement from December 15, 2002 through April 30, 2003 across the new treatment unit, which had been the control unit the previous 2 years. The supplement was distributed daily throughout 29 sites over a larger area (~7 mi2) than the first 2 years of research because of the greater size of the experimental unit and broader distribution of radiocollared deer. Residual feed was always present throughout the winter, thus deer were fed ad libitum. Only small groups of elk periodically accessed the supplement, and did not affect deer access. We obtained 286 observations of radio-collared deer consuming the supplement, which were difficult to obtain because the supplement was spread out over a large area and only a single feed site could be observed at any given moment. We also used daily ground radio-monitoring and periodic aerial relocations to document deer access to the supplement. 2003−04: We distributed 197 tons of pellets throughout the treatment unit from December 10, 2003, through April 30, 2004. The increase in supplement distribution occurred because elk numbers increased on the upper portion of the experimental unit. However, unlike winter 2001−02, residual feed was present throughout the winter and deer were fed ad libitum. We restricted elk to the upper extent of the deer winter range for most of the winter by allocating a portion of the daily feed distribution exclusively to elk. Thus, elk had a minimal affect on deer access to the supplement. We obtained 413 observations of radio-collared deer consuming the supplement. As before, we also used daily ground radio-monitoring and periodic aerial relocations to document deer access to the supplement. Body Condition Estimated percent body fat of adult does during late February and early March, 2002–2004, was higher for treatment deer than control deer (F1, 148 = 153.41, P < 0.001). Over all years combined, mean predicted body fat was 9.8% (SE = 0.36) for treatment adult does and 4.3% (SE = 0.26) for control does. The interaction of experimental unit × year for predicted body fat was also significant (F2, 148 = 14.39, P < 0.001). This interaction occurred because the difference in body fat between treatment and control deer was greater during 2003 than during 2002 or 2004. Mean predicted body fat was 8.2% (SE = 0.92) for treatment adult does and 5.0% (SE = 0.71) for control does during 2002, and 9.0% (SE = 0.53) for treatment does and 4.7% (SE = 0.36) for control does during 2004. The difference was greater during 2003, where mean predicted body fat was 11.7% (SE = 0.35) for treatment does and 3.4% (SE = 0.35) for control does. The body fat estimates reported here should accurately reflect deer, but may be further refined in the future as additional research provides more data on the relationship between body condition indices and estimated percent body fat. Serum thyroid hormone concentrations, measured during 2003 and 2004, were higher in treatment does than control does (F4, 108 = 46.59, P < 0.001) (Table 1). Hormone concentrations also varied between years (F4, 108 = 14.21, P < 0.001), but the experimental unit × year interaction was not significant (F4, 108 = 1.68, P = 0.160). Thus, each year thyroid hormone concentrations were higher in treatment does than control does. T4 was the most important thyroid hormone in describing the canonical variable for differences between treatment and control does (1.04*T4 − 0.02*T3 + 0.77*FT4 – 48 �0.17*FT3). As expected, there was a high partial correlation between T4 and FT4 (r = 0.67, P < 0.001) and between T3 and FT3 (r = 0.60, P < 0.001), which has been documented previously (Watkins et al. 1983). When treated as 4 separate ANOVAs, T4 (F1, 111 = 165.97, P < 0.001), FT4 (F1, 111 = 144.37, P < 0.001), T3 (F1, 111 = 13.84, P < 0.001), and FT3 (F1, 111 = 8.26, P = 0.005) were significantly higher in treatment does than control does. Given these results, we evaluated the relationship between T4 concentrations and estimated percent body fat (derived from ultrasound and BCS indices) using a simple linear regression model (% Fat = −3.122 + 0.090*T4, r2 = 0.52, P < 0.001). Similar correlations between T4 and actual percent body fat during mid-late winter have been previously documented for white-tailed deer and elk (Watkins et al. 1991, Cook et al. 2001). Pregnancy and Fetus Rates 2002: Adult doe pregnancy rate was 0.95 (SE = 0.037, n = 38) in February−March 2002. We measured an average of 1.80 fetuses/doe (SE = 0.10, n = 36), which included 1.77 fetuses/doe (SE = 0.14, n = 18) in the treatment unit and 1.83 fetuses/doe (SE = 0.15, n = 18) in the control unit. 2003: Adult doe pregnancy rate was 0.92 (SE = 0.034, n = 63) in February−March 2003. Critical personnel and equipment for measuring fetus rates were not continuously available due to capture delays associated with helicopter mechanical problems. Some deer fetus counts were performed by inexperienced observers without optimum ultrasound equipment. VITs worked very well, though, allowing us to determine fetus numbers at parturition for many of the deer. Thus, we determined winter fetus rates by using the greatest fetus count for each individual deer, whether obtained using ultrasound during February−March or by locating newborn fawns and stillborns at birth sites during June. We were unable to determine a fetus count for 8 treatment deer because only pregnancy was established with ultrasound and no birth site assessments were possible in June. These 8 deer were removed from the fetus rate estimates. Of the 50 deer where a fetus count was obtained, 5 were yearlings (2 treatment yearlings, 3 control yearlings). We measured 1.74 fetuses/doe (SE = 0.069, n = 50) overall including yearlings, and 1.82 fetuses/doe (SE = 0.066, n = 45) excluding yearlings. Fetus rates with yearlings included were 1.77 fetuses/doe (SE = 0.091, n = 22) in the treatment unit and 1.70 fetuses/doe (SE = 0.10, n = 28) in the control unit. 2004: In February 2004, adult doe pregnancy rate was 0.94 (SE = 0.029, n = 66) and the fetus rate was 1.97 fetuses/doe (SE = 0.053, n = 60), which included 4 yearlings. Excluding yearlings, the fetus rate was 2.00 fetuses/doe (SE = 0.051, n = 56). Fetus rates were 1.90 fetuses/doe (SE = 0.074, n = 30) in the treatment unit and 2.03 fetuses/doe (SE = 0.076, n = 30) in the control unit with yearlings included, and 1.93 (SE = 0.069, n = 29) in the treatment unit and 2.07 (SE = 0.074, n = 27) in the control unit with yearlings excluded. Pregnancy and fetus rates during our study equaled or exceeded other measured rates recorded in Colorado (Andelt et al. 2004), indicating moderate to high innate productivity potential for both treatment and control does. Our data also indicate that adequate numbers of bucks were available to breed does during the years of our study. Fetus and Neonate Survival/Fawn:Doe Ratios 2000: Fawn:doe ratios were similar in the 2 experimental units in December 2000, prior to the first year’s treatment delivery. Pre-treatment fawn:doe ratios were 52.6 fawns:100 does (SE = 5.3) in the Colona experimental unit and 51.6 fawns:100 does (SE = 5.0) in the Shavano experimental unit. 2001: We conducted 2 age classification helicopter surveys in the treatment and control units in late December 2001 and early January 2002, following the first year’s treatment. On 23 December 2001, we observed 52.8 fawns:100 does (SE = 6.7) in the treatment unit and 36.7 fawns:100 does (SE = 3.8) in 49 �the control unit. On 8 January 2002, we observed 54.7 fawns:100 does (SE = 6.6) in the treatment unit and 50.5 fawns:100 does (SE = 6.0) in the control unit. During December 2001 – February 2002, we obtained fawn:doe ratio estimates from ground observations of radio-collared deer groups for both treatment and control deer. This survey resulted in 61.2 fawns:100 does (SE = 7.8) in the treatment unit and 74.5 fawns:100 does (SE = 8.5) in the control unit, although the result was not statistically significant (t74 = 1.16, P = 0.249). Our fawn:doe ratio results were conflicting and did not provide evidence that there was any treatment effect. We could not make any sound conclusions based on the data, although we generally concluded the nutrition enhancement treatment did not cause a substantial increase in neonatal production and survival during 2001. These data provided the incentive to incorporate direct measurements of fetus and neonate survival into our research. 2002: We measured fetus and neonate survival directly during March – December, 2002, following the second year’s treatment; however, sample sizes were based on a technique assessment of VITs and were relatively small for contrasting survival rates among treatment and control fetuses and neonates (Bishop et al. 2002). During June 2002, we determined the fate of all fetuses (live or stillborn) from only 14 of 36 VIT does, largely because of a high VIT battery failure rate. Numbers of stillborns were similar among treatment and control deer, so we did not differentiate by experimental unit. The survival rate of fetuses (n = 22) from the 14 does was 0.86 (SE = 0.073). We also assessed fetus survival using a change-in-ratio estimator between the fetal rate measured in February−March and the observed number of live fawns/doe postpartum in June. In June 2002, considering all does (n = 43) that we located any fawn from, whether live or stillborn, we observed 1.42 (SE = 0.11) live fawns/doe postpartum. This rate should represent a conservative estimate of live fawns/doe postpartum because we inevitably failed to locate all live fawns from each doe. In other words, this estimate would treat any unaccounted fetuses (from the February measurement) as if they were stillborns. For radio-collared does that did not have VITs, and thus we did not have a winter fetus rate measurement, singletons would infer that either the deer only had 1 fetus, or that the other fetus died. It is likely that some of these singletons had a twin that we did not locate. This equates to a conservative fetus survival rate estimate of 0.79 (SE = 0.18). Treatment fawn survival (Jun – Dec) was 0.613 (SE = 0.115, n = 29) and control fawn survival was 0.511 (SE = 0.108, n = 25). In late December 2002 and early January 2003, we once again conducted 2 age classification helicopter surveys in the treatment and control units. On 31 December 2002, we observed 91.9 fawns:100 does (SE = 8.4) in the treatment unit and 52.2 fawns:100 does (SE = 6.9) in the control unit. On 21 January 2003, we observed 52.6 fawns:100 does (SE = 6.4) in the treatment unit and 36.8 fawns:100 does (SE = 3.9) in the control unit. The combined helicopter survey data indicated 68.1 fawns:100 does (SE = 5.6) in the treatment unit and 42.8 fawns:100 does (SE = 3.5) in the control unit. Conversely, fawn:doe ratio estimates from ground classifications of doe groups during December 2002 – February 2003 were 47.7 fawns:100 does (SE = 6.3) in the treatment unit and 63.4 fawns:100 does (SE = 7.5) in the control unit (t108 = 1.61, P = 0.110). As in 2001, fawn:doe ratio results were conflicting. Helicopter survey data varied between 2 different flights, but consistently indicated a treatment effect. Ground classification data did not indicate a treatment effect. 2003: During June 2003, we determined the fate of all fetuses (live or stillborn) from 33 of 58 VIT does; we had better success because VITs commonly shed at birth sites. The survival rate of fetuses (n = 58) from these 33 does was 0.97 (SE = 0.024). In June 2003, incorporating all does (n = 71) from which we located any fawn, whether live or stillborn, we observed 1.49 (SE = 0.072) live fawns/doe postpartum. Using the change-in-ratio estimator described above, this results in an overall conservative fetus survival rate estimate of 0.86 (SE = 0.15). As in 2002, fetus survival was similar among treatment and control deer and not analyzed separately. During June 2003, we captured and radio-collared 103 newborn fawns born from treatment and control radio-collared does (55 treatment fawns, 48 control fawns). The VITs worked well; we captured 50 �fawns from 41 of the 58 does fitted with VITs. Treatment fawn survival (Jun – Dec) was 0.624 (SE = 0.082) and control fawn survival was 0.483 (SE = 0.093). Final standard errors were larger than expected because a number of fawns shed collars prematurely when crossing fences during fall migration. Using helicopter surveys, we measured 62.4 fawns:100 does (SE = 5.3) in the treatment unit and 50.0 fawns:100 does (SE = 4.9) in the control unit. Estimates from ground classifications of doe groups were 68.0 fawns:100 does (SE = 7.6) in the treatment unit and 62.1 fawns:100 does (SE = 7.6) in the control unit. Age ratio estimates from the helicopter and the ground were more consistent during 2003 than in past years. Overall, observed fawn:doe ratios were consistent with treatment and control fawn survival rates measured from June to December. 2004: We determined the fate of all fetuses from 31 of 60 VIT does. The overall fetus survival rate was 0.90 (SE = 0.040, n = 58). Different from 2002 or 2003, all stillborns were from control does. The survival rate of control fetuses was 0.76 (SE = 0.085, n = 25) as compared to a survival rate of 1.00 (n = 33) for treatment fetuses. Using data from all does (n = 82) in which we located any fawn, the conservative change-in-ratio fetus survival estimate was 0.79 (SE = 0.13) overall, 0.88 (SE = 0.17) for treatment deer, and 0.69 (SE = 0.14) for control deer. We captured and radio-collared 119 newborn fawns born from treatment and control radiocollared does during June 2004 (68 treatment fawns, 51 control fawns). Vaginal implants worked well again, and we had a large sample of non-VIT radio-collared does that we could relocate to opportunistically capture additional treatment and control fawns. Treatment fawn survival (Jun – Dec) was 0.438 (SE = 0.068) and control fawn survival was 0.414 (SE = 0.092). As in 2003, final standard errors were larger than expected because fawns shed collars prematurely during fall migration. Although neonate survival rates were similar among treatment and control fawns, fewer control fawns survived to December because of lower fetus survival. The proportion of fetuses measured in March that were born alive and survived to December during 2004 (i.e. fetus-neonate survival) was 0.438 (SE = 0.068) for treatments and 0.304 (0.073) for controls. Similar to 2002 and 2003, we observed higher December fawn recruitment among treatment deer based on measured survival rates. The difference during 2004 was that stillborn deaths factored in as a larger mortality factor among control deer than during 2002 or 2003. We measured 64.6 fawns:100 does (SE = 5.8) in the treatment unit and 52.7 fawns:100 does (SE = 5.1) in the control unit during helicopter surveys in 2004. Our ground classification estimates were 78.5 fawns:100 does (SE = 6.6) in the treatment unit and 68.7 fawns:100 does (SE = 5.1) in the control unit. Similar to 2003, observed fawn:doe ratios were consistent with treatment and control survival rates. 2002−2004 Fetus-Neonate Survival Summary: Fetus-neonate survival combined over all years of study (1 Mar–15 Dec, 2002–2004) was higher (χ21 = 3.089, P = 0.079) for treatment deer (S(t) = 0.519, SE = 0.048) than for control deer (S(t) = 0.409, SE = 0.052). The high censor rate from shed collars during fall reduced power of the analysis and therefore increased standard errors and the resulting Pvalue. However, at roughly the same time neonate radio-collars were being shed, we captured new samples of fawns for measuring overwinter fawn survival. When fawns captured during November and early December were incorporated into the analysis via staggered entry, fetus-neonate treatment survival (S(t) = 0.528, SE = 0.027) and control survival (S(t) = 0.401, SE = 0.025) rates had tighter standard errors, which reduced the p-value associated with the survival rate comparison (χ21 = 3.846, P = 0.050). The nutrition enhancement treatment had a positive effect on fetus and neonate survival through about the first month postpartum, at which point the treatment stopped having an effect (Figure 4). Fetus-neonate survival through 15 July, 2002–2004, was much higher (χ21 = 6.013, P = 0.014) for treatment fawns (S(t) = 0.746, SE = 0.035) than control fawns (S(t) = 0.583, SE = 0.043). In summary, enhanced nutrition of adult does during winter and early spring caused higher survival of fetuses and fawns, resulting in higher December fawn recruitment (Figure 4). 51 �2001−2004 Fawn:Doe Ratio Summary: Our results from 2001 and 2002 emphasize the inherent difficulties and biases associated with precisely measuring fawn:doe ratios, particularly in this research study. Ratios obtained from helicopter surveys were based on 2 short-duration flights/unit/year over spatially small units. Helicopter surveys were complicated by high deer densities in heavy cover, making both deer detection and fawn:doe classifications a considerable challenge. There were a variety of potential biases that may have affected the helicopter surveys, including differential sightability of does and fawns, double classification of some deer, and incorrect classification of yearling bucks with small antlers. Ground fawn:doe ratio observations of radio-collared doe groups were made using spotting scopes and field glasses, where we commonly studied the deer for some time. Incorrect classifications during these surveys were likely minimal. For example, small-antlered yearling bucks (e.g. 3 – 6” spikes) were detected from the ground, whereas they were undoubtedly missed on occasion during helicopter surveys. We also obtained repeated observations for some of the radio-collared doe groups from the ground. The main potential bias affecting ground fawn:doe classifications was how observations were made. Many of the ground classifications in the Shavano Valley experimental unit were made by radiotracking does during the day. On the other hand, a majority of ground classifications in the Colona experimental unit were based on observing deer groups as they entered openings to feed during the late afternoon. Our age ratio results were more consistent with survival data during 2003 and 2004. Deer were not as concentrated during helicopter surveys, and unlike previous years, a majority of the ground classification data for the Colona experimental unit was obtained by radio-tracking does during the day rather than sitting and waiting for deer to emerge from pinyon-juniper hillsides to feed on sagebrush-grass benches. We relied primarily on fetus-neonate survival data to make inferences regarding treatment effects because of the inherent difficulties measuring fawn:doe ratios in the 2 experimental units. However, we plan to compare observed helicopter and ground fawn:doe ratios with predicted ratios based on fetusneonate survival data as a technique assessment of fawn:doe ratio measurements. This analysis will be incorporated into the job completion report. Neonate Mortality Causes 2002−2003: During June − December of 2002 and 2003, 37 treatment fetuses-neonates died: 3 – stillborn, 8 – coyote predation, 2 – bear predation, 2 – felid predation, 3 – predation where the predator was undetermined, 11 – disease-starvation-malnutrition, 1 – abandonment, 3 – trauma-injury, 2 – unknown, and 2 – poached. The two poached fawns were censored from analyses evaluating the effect of the treatment. Converted to mortality rates based on the Kaplan-Meier survival analysis, 11.4% of all treatment fawns died from disease-starvation-malnutrition, 8.3% from coyote predation, 7.7% were stillborn, 3.1% died each from injury-trauma and from predation where the predator was undetermined, 2.1% each from bear predation, felid predation, and unknown causes, and 1.0% from abandonment. Simplified, 15.6% of all treatment fawns died from predation, 11.4% died from disease-starvationmalnutrition, 7.7% were stillborn, and 6.2% died from other or unknown causes. During June – December of 2002 and 2003, 38 control fetuses-neonates died: 2 – stillborn, 12 – coyote predation, 4 – felid predation, 2 – bear predation, 1 – predation where the predator was undetermined, 12 – diseasestarvation-malnutrition, 1 – trauma-injury, and 4 – unknown. Converted to mortality rates based on the Kaplan-Meier survival analysis, 16.0% of all control fawns died from disease-starvation-malnutrition, 16.0% died from coyote predation, 5.3% each from felid predation and unknown causes, 4.9% were stillborn, 2.7% from bear predation, and 1.3% each from trauma-injury and predation where the predator was undetermined. Simplified, 25.3% of all control fawns died from predation, 16.0% from diseasestarvation-malnutrition, 6.7% from other or unknown causes, and 4.9% were stillborn. In summary, mortality rates due to predation and disease-starvation-malnutrition were lower for treatment fawns than control fawns. 52 �2004: During June – December, 2004, 36 treatment neonates died: 0 – stillborn, 13 – coyote or dog predation, 7 – bear predation, 3 – felid predation, 5 – predation where the predator was undetermined, 2 – disease-starvation-malnutrition, 1 – trauma-injury, and 5 – unknown. Converted to mortality rates based on the Kaplan-Meier survival analysis, 20.3% of all treatment fawns died from canid predation, 10.9% died from bear predation, 7.8% each from unknown causes and from predation where the predator was undetermined, 4.7% from felid predation, 3.1% from disease-starvation-malnutrition, and 1.6% from injury-trauma. Simplified, 43.7% of all treatment fawns died from predation, 9.4% died from other or unknown causes, and 3.1% died from disease-starvation-malnutrition. During June – December, 2004, 32 control fetuses-neonates died: 6 – stillborn, 5 – coyote predation, 4 – bear predation, 1 – felid predation, 2 – predation where the predator was undetermined, 4 – disease-starvation-malnutrition, 5 – injurytrauma, and 5 – unknown. We actually observed 9 stillborns from control does with fetus counts, although only 6 were associated with does in which all fetuses were accounted at parturition. Thus, we used only 6 of the stillborns in our estimate of fetus survival, and therefore stillborn mortality. Converted to mortality rates based on the Kaplan-Meier survival analysis, 24.0% of all control neonates were stillborn, 8.8% died each from coyote predation, injury-trauma, and unknown causes, 7.0% each from bear predation and disease-starvation-malnutrition, 3.5% from predation where the predator was undetermined, and 1.8% from felid predation. Simplified, 24.0% of all control fawns were stillborn, 21.0% died from predation, 17.5% died from other or unknown causes, and 7.0% died from diseasestarvation-malnutrition. Mortality causes were much different during 2004 than either 2002 or 2003. Predation rates were high on treatment fawns while stillborn mortality rates were high among control fawns. Several specific observations during 2004 are worthy of note. Three of the treatment fawn mortalities attributed to coyotes or dogs occurred amongst large herds of sheep which had been released to pasture immediately prior to the mortality events. Bear predation was higher among all fawns during 2004, although 3 of the 7 treatment bear mortalities involved triplets that were killed simultaneously by a bear 1−2 days after the fawns were born. Six treatment fawns captured in the same drainage tributary were killed within a 1-mi2 area; the drainage was in a portion of the study area where no control fawns were captured. However, a single animal did not kill each of the fawns because the mortalities encompassed coyote, felid, and bear predation. Finally, we observed more accidental deaths than typical among control fawns. One control fawn drowned in a river, another fell, one became lodged in a water-filled mudhole, and both a control fawn and a treatment fawn died from injuries sustained while stuck in a woven wire fence. 2002−2004 Summary: Combining all years of data, the survival and cause-specific mortality rates of treatment fawns were: 52.8% survived, 27.2% died from predation (i.e. 13.3% canid, 5.7% bear, 3.2% felid, 5.1% undetermined), 8.2% died from disease-starvation-malnutrition, 4.2% were stillborn, and 7.6% died from other or unknown causes. Survival and cause-specific mortality rates of control fawns were: 40.1% survived, 24.3% died from predation (i.e. 12.9% canid, 4.6% bear, 3.8% felid, 3.0% undetermined), 12.1% died from disease-starvation-malnutrition, 12.1% were stillborn, and 11.4% died from other or unknown causes. The relatively high predation rate of treatment fawns was largely explained by 2004 data alone. As a general summary, control fawns suffered higher rates of disease, illness, malnutrition, and stillborn mortality (i.e. non-predator related mortalities) than did treatment fawns, which explains why survival was higher among treatment fawns (Figure 5). Overwinter Fawn Survival and Mortality Causes During winter 2001−02 (10 Dec 2001–15 Jun 2002), the survival rate of fawns was higher (χ21 = 13.216, P < 0.001) in the treatment unit (S(t) = 0.865, SE = 0.056) than in the control unit (S(t) = 0.510, SE = 0.080). Similarly, in 2002−03 (10 Dec 2002–15 June 2003), the overwinter survival rate of fawns was higher (χ21 = 5.734, P = 0.017) in the treatment unit (S(t) = 0.900, SE = 0.047) than in the control unit (S(t) = 0.691, SE = 0.074). Again in 2003−04 (10 Dec 2003–15 June 2004), the overwinter survival 53 �rate of fawns was higher (χ21 = 3.852, P = 0.050) in the treatment unit (S(t) = 0.920, SE = 0.045) than in the control unit (S(t) = 0.756, SE = 0.067). Combining survival data across all 3 winters, treatment fawn survival (S(t) = 0.895, SE = 0.029) was 0.24 higher (χ21 = 18.781, P < 0.001) than control fawn survival (S(t) = 0.655, SE = 0.044) (Figure 6). The treatment unit during winter 2001−02 became the control unit during winters 2002−03 and 2003−04, and vice versa. Thus, the overwinter survival treatment effect was replicated across each experimental unit. Fawn survival also varied as a function of early winter fawn mass (χ21 = 21.19, P < 0.001). Surviving fawns averaged 3.5 kg heavier than fawns that died. The importance of early winter fawn mass as a predictor of overwinter survival has been documented previously (White et al. 1987, Bishop 1998, White and Bartmann 1998, Unsworth et al. 1999). Early winter mass of treatment fawns ( x = 34.2 kg, SE = 0.418) was similar to control fawns ( x = 34.4, SE = 0.423); thus the effect of the treatment was not confounded with pre-treatment fawn mass. It follows that fawns born from treatment does did not arrive to winter heavier than fawns born from control does, which was not necessarily surprising considering the treatment primarily effected neonate survival through about 1 month postpartum. In summary, the nutrition enhancement treatment improved overwinter fawn survival, and heavier fawns in each experimental unit had higher survival probabilities. During winters 2001−04, 12 of 115 treatment fawns died: 5 from coyote predation, 3 from disease/illness, 2 from malnutrition, 1 from trauma-injury, and 1 unknown. Each of the 3 fawns that died from disease had adequate fat stores. At least one of these fawns died as a result of pneumonia. Converted to mortality rates based on the Kaplan-Meier survival analysis, 4.3% of all treatment fawns died from coyote predation, 2.6% from disease-illness, 1.7% from malnutrition, 0.9% from trauma-injury, and 0.9% from unknown causes. Simplified, 4.3% of all treatment fawns died from predation, 4.3% from disease-malnutrition, and 1.8% from other or unknown causes (Figure 7). During winters 2001−04, 41 of 120 control fawns died: 13 from coyote predation, 8 from mountain lion predation, 8 from malnutrition, 6 from unknown causes, 3 from predation where the predator was undetermined, 2 were road-killed, and 1 from trauma-injury. Converted to mortality rates based on the Kaplan-Meier survival analysis, 10.9% of all control fawns died from coyote predation, 6.7% from mountain lion predation, 6.7% from malnutrition, 5.0% from unknown causes, 2.5% from predation where the predator was undetermined, 1.7% from road-kill, and 0.8% from trauma-injury. Simplified, 20.1% of all control fawns died from predation, 6.7% from malnutrition, and 7.5% from other or unknown causes (Figure 7). Most fawns killed by predators had little or no femur marrow fat remaining, indicating the predation was likely compensatory in nature. Fetus-Neonate-Overwinter Fawn Survival We combined the preceding survival data into a single analysis to express the effect of the treatment across all stages of fawn production and survival. Using a staggered entry survival process with data combined over years, we estimated fawn survival from the fetus stage until one year of age, when fawns were recruited to the yearling (adult) age class (Figure 8). Survival of treatment fetuses to the yearling age class (S(t) = 0.458, SE = 0.031) was 0.18 higher (χ21 = 13.20, P < 0.001) than survival of control fetuses to the yearling age class (S(t) = 0.276, SE = 0.026). Adult Female Survival and Causes of Mortality During winter 2000−01 (1 Dec 2000–31 May 2001), the adult doe survival rate in the treatment unit (S(t) = 0.968, SE = 0.032) was greater (χ21 = 2.649, P = 0.104) than the survival rate in the control unit (S(t) = 0.861, SE = 0.058). However, annual adult doe survival rates (1 Dec 2000–30 Nov 2001) were similar among treatment and control deer (Trt: S(t) = 0.839, SE = 0.066; Control: S(t) = 0.833, SE = 0.062; χ21 = 0.004, P = 0.947). We observed a similar result the following year. The 2001−02 overwinter adult doe survival rate in the treatment unit (S(t) = 0.942, SE = 0.030) was greater (χ21 = 3.116, P = 0.078) than survival in the control unit (S(t) = 0.848, SE = 0.044), yet annual adult doe survival was similar among treatment and control deer (Trt: S(t) = 0.824, SE = 0.049; Control: S(t) = 0.818, SE = 54 �0.047; χ21 = 0.090, P = 0.764). Thus, mortalities of control deer occurred primarily during the winter months, while treatment does died primarily during the summer and fall months. During winter 2002−03, following the treatment cross-over, overwinter adult doe survival rates were similar among treatment and control deer (Trt: S(t) = 0.945, SE = 0.024; Control: S(t) = 0.924, SE = 0.028; χ21 = 0.360, P = 0.549). However, annual adult doe survival rates (1 Dec 2002–30 Nov 2003) were higher (χ21 = 2.016, P = 0.156) for treatment does (S(t) = 0.888, SE = 0.034) than control does (S(t) = 0.813, SE = 0.041). The main difference from the previous 2 years was that overwinter survival of adult does in the Shavano experimental unit increased in 2002−03 upon receiving the treatment. Summer-fall survival was similar in that Colona adult does had higher mortality rates than Shavano adult does. Thus, in 2002−03, there was no difference between survival rates of treatment and control adult does during winter but there was evidence of higher annual survival of treatment adult does. During winter 2003−04, overwinter adult doe survival rates were higher (χ21 = 3.843, P = 0.050) among treatment does (S(t) = 0.979, SE = 0.014) than control does (S(t) = 0.915, SE = 0.027). The annual adult doe survival rate (1 Dec 2003–30 Nov 2004) was 0.895 (SE = 0.030) for treatment does and 0.832 (SE = 0.036) for control does, which was marginally different (χ21 = 1.562, P = 0.211). Considering all years, the treatment improved overwinter adult doe survival but had a relatively minor affect on annual survival. Considering only the past 2 years, the treatment had a positive affect on annual survival. Annual survival rates measured in this study align reasonably well with expected survival based on other studies (Unsworth et al. 1999, Bishop et al. 2005, B. E. Watkins, Colorado Division of Wildlife, unpublished data). During 2000−02, when the Colona experimental unit received the treatment and the Shavano experimental unit was the control, 16 treatment and 16 control does died. The 16 treatment does died from the following categories: 4 – road-killed, 3 – while giving birth, 3 – predation (undetermined predator), 2 – non-predation unknown (intact carcasses with no evidence of predation or scavenging), 1 – disease (chronic arthritis), 1 – mountain lion predation, and 2 – unknown. Predation was not a major mortality factor for treatment does, and a majority of mortalities were independent of nutrition (does were in good condition). The 16 control doe mortalities included the following causes: 5 – mountain lion predation, 3 – malnutrition, 2 – non-predation unknown, 1 – road-killed, 1 – bear predation, 1 – fence injury, 1 – legal harvest, and 2 – unknown. Predation and malnutrition were the major mortality causes of control deer. Interestingly, during this 2-year period, we did not document any coyote predation on adult does. During 2002–04, with Shavano as the treatment and Colona as the control, there were 20 treatment doe mortalities: 6 – disease/infection, 3 – coyote predation, 1 – road-killed, 1 – broken jaw which led to starvation, 1 – fence injury, 1 poached, and 7 unknown causes. As we saw during 2000-02, predation was not a major mortality factor for treatment does, and a majority of mortalities were independent of nutrition. We observed 33 control adult doe mortalities during the same time period: 8 – road-kill, 7 – malnutrition-disease, 5 – coyote predation, 3 – mountain lion predation, 3 – non-predation unknown, 1 – bear predation, 1 – predation where the predator was undetermined, and 5 – unknown causes. Road kill, malnutrition-disease, and predation were the major mortality factors of control does during 2002−04. Road kill was a significant mortality factor of Colona adult does but not Shavano adult does, which partially explains why we failed to see a treatment effect during 2000−02 but did see one during 2002−04. If road-killed deer were censored, greater evidence would exist for a treatment effect during 2000−02 while there would be less evidence of a treatment effect during 2002−04. However, road-kill had minimal effect on the overall 4-year interpretation of the treatment effect on adult doe survival 55 �because of the cross-over design. Ignoring road kill, treatment does tended to die of causes unrelated to nutrition whereas control does were more susceptible to malnutrition and predation. Population Growth Rate The finite rate of population increase, λ, based on our measurements of treatment population parameters was 1.20 (Table 2), which would cause the deer population to double in approximately 4 years. The finite rate of increase calculated from control deer was 1.04 (Table 2), indicating a stable or slightly increasing population. The nutrition enhancement treatment therefore had a dramatic effect on deer population performance, indicating habitat quality was ultimately limiting the population. SUMMARY We successfully enhanced nutrition of deer occupying the treatment units based on our body fat estimates of treatment and control does. Pregnancy and fetus rates were similar among treatment and control does. The treatment caused an increase in both fetus-neonate survival and overwinter fawn survival, resulting in higher yearling recruitment. Overwinter adult doe survival increased as a result of the treatment, but annual survival was more similar among treatment and control adult does. Combining all parameter estimates into a deterministic population model, the treatment population indicated an exceptionally high rate of increase (λ = 1.20) while the control population (λ = 1.04) was indicative of the overall Uncompahgre deer population during 2000−2004. The nutrition enhancement treatment was artificial in the sense that we applied it only to test whether habitat quality was ultimately more limiting than predation or other factors. Our results to do not provide support for managing deer populations with nutrition supplements because our treatment delivery approach could not be applied to a large number of animals over a large area. Rather, our results provide a foundation for focusing deer management efforts on improving habitat quality in western Colorado pinyon-juniper ecosystems with corresponding research efforts to quantify the effects of habitat manipulations on deer. LITERATURE CITED ANDELT, W. F., T. M. POJAR, AND L. W. JOHNSON. 2004. Long-term trends in mule deer pregnancy and fetal rates in Colorado. Journal of Wildlife Management 68:542−549. BAKER, D. L., AND N. T. HOBBS. 1985. Emergency feeding of mule deer during winter: tests of a supplemental ration. Journal of Wildlife Management 49:934−942. _____, D. E. JOHNSON, L. H. CARPENTER, O.C. WALLMO, AND R. B. GILL. 1979. Energy requirements of mule deer fawns in winter. Journal of Wildlife Management 43:162−169. _____, G. W. STOUT, AND M. W. MILLER. 1998. A diet supplement for captive wild ruminants. Journal of Zoo and Wildlife Medicine 29:150−156. BALLARD, W. B., D. LUTZ, T. W. KEEGAN, L. H. CARPENTER, AND J. C. DEVOS, JR. 2001. Deer-predator relationships: a review of recent North American studies with emphasis on mule and black-tailed deer. Wildlife Society Bulletin 29:99−115. BARRETT, M. W., J. W. NOLAN, AND L. D. ROY. 1982. Evaluation of a hand-held net-gun to capture large mammals. Wildlife Society Bulletin 10:108−114. BISHOP, C. J. 1998. Mule deer fawn mortality and habitat use, and the nutritional quality of bitterbrush and cheatgrass in southwest Idaho. Thesis, University of Idaho, Moscow, USA. _____, D. J. FREDDY, AND G. C. WHITE. 2002. Effects of enhanced nutrition of adult female mule deer on fetal and neonatal survival rates: a pilot study to address feasibility. Colorado Division of Wildlife, Wildlife Research Report, Federal Aid in Wildlife Restoration Project W−153−R, Job Final Report. Fort Collins, USA. 56 �_____, J. W. UNSWORTH, AND E. O. GARTON. 2005. Mule deer survival among adjacent populations in southwest Idaho. Journal of Wildlife Management 69:311−321. BOWDEN, D. C. 1993. A simple technique for estimating population size. Department of Statistics, Colorado State University, Fort Collins, USA. _____, A. E. ANDERSON, AND D. E. MEDIN. 1984. Sampling plans for mule deer sex and age ratios. Journal of Wildlife Management 48:500−509. CONNOLLY, G. E. 1981. Limiting factors and population regulation. Pages 245−285 in O. C. Wallmo, editor. Mule and black-tailed deer of North America. University of Nebraska Press, Lincoln, USA. COOK, R. C. 2000. Studies of body condition and reproductive physiology in Rocky Mountain elk. Thesis, University of Idaho, Moscow, USA. _____, AND J. G. COOK. 2002. An informal training guide to condition evaluation in elk and deer. National Council for Air and Stream Improvement, Unpublished Report. _____, _____, D. L. MURRAY, P. ZAGER, B. K. JOHNSON, AND M. W. GRATSON. 2001. Development of predictive models of nutritional condition for Rocky Mountain elk. Journal of Wildlife Management 65:973−987. GILL, R. B., T. D. I. BECK, C. J. BISHOP, D. J. FREDDY, N. T. HOBBS, R. H. KAHN, M. W. MILLER, T. M. POJAR, AND G. C. WHITE. 2001. Declining mule deer populations in Colorado: reasons and responses. Colorado Division of Wildlife Special Report Number 77. Denver, USA. HOLTER, J. B., H. H. HAYES, AND S. H. SMITH. 1979. Protein requirement of yearling white-tailed deer. Journal of Wildlife Management 43:872−879. KAPLAN, E. L., AND P. MEIER. 1958. Nonparametric estimation from incomplete observations. Journal of the American Statistical Association 53:457−481. NEAL, A. K., G. C. WHITE, R. B. GILL, D. F. REED, AND J. H. OLTERMAN. 1993. Evaluation of markresight model assumptions for estimating mountain sheep numbers. Journal of Wildlife Management 57:436−450. POJAR, T. M., AND D. C. BOWDEN. 2004. Neonatal mule deer fawn survival in west-central Colorado. Journal of Wildlife Management 68:550−560. POLLOCK, K. H., S. R. WINTERSTEIN, C. M. BUNCK, AND P. D. CURTIS. 1989. Survival analysis in telemetry studies: the staggered entry design. Journal of Wildlife Management 53:7−15. RAMSEY, C. W. 1968. A drop-net deer trap. Journal of Wildlife Management 32:187−190. SAS INSTITUTE. 1989. SAS/STAT® user’s guide, version 6, fourth edition. Volume 2. SAS Institute, Cary, North Carolina, USA. _____. 1997. SAS/STAT® Software: Changes and Enhancements through Release 6.12. SAS Institute, Cary, North Carolina, USA. SCHMIDT, R. L., W. H. RUTHERFORD, AND F. M. BODENHAM. 1978. Colorado bighorn sheep-trapping techniques. Wildlife Society Bulletin 6:159−163. SMITH, S. H., J. B. HOLTER, H. H. HAYES, AND H. SILVER. 1975. Protein requirement of white-tailed deer fawns. Journal of Wildlife Management 39:582−589. STEPHENSON, T. R., V. C. BLEICH, B. M. PIERCE, AND G. P. MULCAHY. 2002. Validation of mule deer body composition using in vivo and post-mortem indices of nutritional condition. Wildlife Society Bulletin 30:557−564. _____, K. J. HUNDERTMARK, C. G. SCHWARTZ, AND V. VAN BALLENBERGHE. 1998. Predicting body fat and body mass in moose with ultrasonography. Canadian Journal of Zoology 76:717−722. _____, J. W. TESTA, G. P. ADAMS, R. G. SASSER, C. G. SCHWARTZ, AND K. J. HUNDERTMARK. 1995. Diagnosis of pregnancy and twinning in moose by ultrasonography and serum assay. Alces 31:167−172. THOMPSON, C. B., J. B. HOLTER, H. H. HAYES, H. SILVER, AND W. E. URBAN, JR. 1973. Nutrition of white-tailed deer. I. Energy requirements of fawns. Journal of Wildlife Management 37:301−311. 57 �ULLREY, D. E., W. G. YOUATT, H. E. JOHNSON, L. D. FAY, AND B. L. BRADLEY. 1967. Protein requirement of white-tailed deer fawns. Journal of Wildlife Management 31:679−685. UNSWORTH, J. W., D. F. PAC, G. C. WHITE, AND R. M. BARTMANN. 1999. Mule deer survival in Colorado, Idaho, and Montana. Journal of Wildlife Management 63:315−326. VAN REENEN, G. 1982. Field experience in the capture of red deer by helicopter in New Zealand with reference to post-capture sequela and management. Pages 408−421 in L. Nielsen, J. C. Haigh, and M. E. Fowler, editors. Chemical immobilization of North American wildlife. Wisconsin Humane Society, Milwaukee, USA. VERME, L. J., AND D. E. ULLREY. 1972. Feeding and nutrition of deer. Pages 275−291 in D. C. Church, editor. Digestive physiology and nutrition of ruminants. Volume 3 – Practical Nutrition. D. C. Church, Corvallis, Oregon, USA. WATKINS, B. E., D. E. ULLREY, R. F. NACHREINER, AND S. M. SCHMITT. 1983. Effects of supplemental iodine and season on thyroid activity of white-tailed deer. Journal of Wildlife Management 47:45−58. _____, J. H. WITHAM, D. E. ULLREY, D. J. WATKINS, AND J. M. JONES. 1991. Body composition and condition evaluation of white-tailed deer fawns. Journal of Wildlife Management 55:39−51. WHITE, G. C. 1996. NOREMARK: Population estimation from mark-resighting surveys. Wildlife Society Bulletin 24:50−52. _____, AND R. M. BARTMANN. 1998. Effect of density reduction on overwinter survival of free-ranging mule deer fawns. Journal of Wildlife Management 62:214−225. _____, R. A. GARROTT, R. M. BARTMANN, L. H. CARPENTER, AND A. W. ALLDREDGE. 1987. Survival of mule deer in northwest Colorado. Journal of Wildlife Management 51:852−859. _____, A. F. REEVE, F. G. LINDZEY, AND K. P. BURNHAM. 1996. Estimation of mule deer winter mortality from age ratios. Journal of Wildlife Management 60:37−44. Prepared by _______________________ Chad J. Bishop, Wildlife Researcher 58 �Table 1. Total thyroxine (T4) and total tri-iodothyronine (T3) concentrations (nmol/l), and free T4 (FT4) and free T3 (FT3) concentrations (pmol/l), measured during late February in adult female mule deer occupying a nutrition enhancement treatment unit and a control unit on the Uncompahgre Plateau in southwest Colorado, 2003−04. Thyroid Hormone T3 (SE) FT3 (SE) 146.6 (3.53) FT4 (SE) 30.0 (1.27) 1.65 (0.058) 4.10 (0.130) Control 92.3 (3.56) 17.1 (0.65) 1.42 (0.080) 3.71 (0.210) Treatment 131.9 (4.48) 24.8 (1.39) 2.08 (0.075) 4.21 (0.154) Control 90.0 (3.54) 12.5 (0.59) 1.70 (0.104) 3.60 (0.188) Year Exp. Unit T4 (SE) 2003 Treatment 2004 Table 2. Population parameter estimates and population finite rate of increase, λ, for treatment deer that received a nutrition enhancement and control deer that accessed existing habitat only, southwest Colorado, 2002−04. Population Parameter Treatment Control 0.937 0.937 Adult doe fetus rate 1.84 1.84 Fetus survival to birth 0.958 0.879 Neonate survival to December 0.551 0.456 Overwinter fawn survival to June 0.895 0.655 Annual adult doe survival 0.860 0.824 Finite Rate of Increase, λ 1.20 1.04 Adult doe pregnancy ratea a a We used overall estimates of pregnancy and fetus rates because we did not detect meaningful differences between treatment and control deer. Year Unit A Unit B 2000-01 Treatment Control 2001-02 Treatment Control 2002-03 Control Treatment 2003-04 Control Treatment Figure 1. Schematic representation of experimental units and nutrition enhancement treatment allocation. Units A and B were located in winter range habitat on the Uncompahgre Plateau in southwest Colorado. The nutrition enhancement cross-over design encompassed 4 years. 59 �Uncompahgre Plateau Mesa County GRAND JUNCTION Delta County m co Un GMU 62 r hg pa e u ea at Pl Montrose County GMU 61 Sanmiguel County Gunnison County DELTA Shavano E.U. Winter Range MONTROSE Colona Montrose County E.U. Summ er Ouray County Figure 2. Location of Colona and Shavano (Units A and B) experimental units on the Uncompahgre Plateau, southwest Colorado; and location of the summer range study area encompassing the southern Uncompahgre Plateau and adjacent San Juan Mountains. 60 �Hwy 550 Uncompahgre Valley Colona Exp. Unit Shavano Valley Shavano Exp. Unit - 6 Miles - - Figure 3. Colona and Shavano experimental units (Units A and B), located in Game Management Unit 62 on the Uncompahgre Plateau, southwest Colorado. 61 �1 \ 0.9 Treatment 0.8 Control 0.7 0.6 -- - 0.5 0.4 0.3 3/1 3/31 4/30 5/30 6/29 7/29 8/28 9/27 10/27 11/26 Figure 4. Survival (1 Mar –15 Dec, 2002–2004) of mule deer fetuses-neonates born from adult does receiving enhanced nutrition during winter (Treatment, S(t) = 0.528, SE = 0.027) and from adult does accessing existing winter habitat only (Control, S(t) = 0.401, SE = 0.025), southwest Colorado. 0.6 Treatment Control 0.5 0.4 0.3 0.2 0.1 0 Survived Predation Illness/Malnutrition Stillborn Other/ Unknow n Figure 5. Survival and cause-specific mortality rates (1 Mar –15 Dec, 2002–2004) of mule deer fetusesneonates born from adult does receiving enhanced nutrition during winter (Treatment) and from adult does accessing existing winter habitat only (Control), southwest Colorado. 62 �- ..... 1 0 .9 0 .8 0 .7 Treatment 0 .6 Control 0 .5 12/1 12/16 12/31 1/15 1/30 2/14 3/1 3/16 3/31 4/15 4/30 5/15 5/30 6/14 Figure 6. Overwinter fawn survival (10 Dec –15 Jun, 2001–2004) in a nutrition enhancement treatment unit (S(t) = 0.895, SE = 0.029) and a control unit (S(t) = 0.655, SE = 0.044) on the Uncompahgre Plateau, southwest Colorado. 0.9 Treatment Control 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Survived Predation Illness/Malnutrition Other/ Unknow n Figure 7. Overwinter fawn survival and cause-specific mortality rates (10 Dec–15 Jun, 2001–2004) in a nutrition enhancement treatment unit and a control unit on the Uncompahgre Plateau, southwest Colorado. 63 �1 Treatment 0.9 Control 0.8 0.7 0.6 0.5 0.4 0.3 0.2 3/1 4/15 5/30 7/14 8/28 10/12 11/26 1/10 2/24 4/10 5/25 Figure 8. Fawn survival from fetus stage (March) to 1 year of age (June of the following year) for deer receiving enhanced nutrition during winter (Treatment, S(t) = 0.458, SE = 0.031) and deer accessing existing winter habitat only (Control, S(t) = 0.276, SE = 0.026), southwest Colorado, 2002−2004. 64 �APPENDIX I We submitted the following manuscript (referenced here by Abstract) to the Journal of Wildlife Management during summer 2005. USING VAGINAL TRANSMITTERS TO CAPTURE NEONATES FROM MARKED MULE DEER CHAD J. BISHOP, DAVID J. FREDDY, GARY C. WHITE, BRUCE E. WATKINS, THOMAS R. STEPHENSON, AND LISA L. WOLFE ABSTRACT Measuring reproductive success of previously-marked, adult female ungulates enables study of certain complex ecological factors limiting populations. We evaluated the effectiveness of using vaginal implant transmitters (VITs, n = 154) in mule deer (Odocoileus hemionus) combined with repeated relocations of other radio-collared deer for capturing effective samples of neonates (e.g. >100/year) from free-ranging, marked females. We also evaluated the effectiveness of VITs, when used in conjunction with in utero fetus counts, for obtaining direct estimates of fetus survival. During 2003 and 2004, when VIT batteries were placed on a 12-hour duty cycle to lower failure rates, the proportion of VITs that shed ≤3 days prepartum or during parturition was 0.623 (SE = 0.0456), and the proportion shed during parturition was 0.447 (SE = 0.0468). Our neonate capture success rate was 0.880 (SE = 0.0359) from does with VITs shed ≤3 days prepartum or during parturition and 0.307 (SE = 0.0235) from radiocollared does without VITs or whose implants failed to function properly. Combining techniques we captured 275 neonates and 21 stillborns during 2002−2004. We accounted for all fetuses at birth (i.e. live or stillborn) from 78 of the 147 does (0.531, SE = 0.0413) with winter fetus counts, which was heavily dependent on VIT retention success. Deer that shed VITs prepartum were larger and older than deer that retained implants to parturition, indicating a need to develop variable-sized VITs which may be individually fitted to deer in the field. We demonstrated that direct estimates of fetus and neonate survival may be obtained from previously-marked female mule deer in free-ranging populations, thus expanding opportunities for conducting field experiments. Resulting neonate survival estimates lacked bias that is typically associated with other neonate capture techniques. However, current vaginal implant failure rates and overall expense limit applicability of the technique to well-funded studies with adequate personnel. 65 �Colorado Division of Wildlife July 2005 − June 2006 WILDLIFE RESEARCH REPORT State of Cost Center Work Package Task No. Colorado 3430 3001 4 Federal Aid Project: W-185-R : Division of Wildlife : Mammals Research : Deer Conservation : Effect of Nutrition and Habitat Enhancements on Mule Deer Recruitment and Survival Rates : Period Covered: July 1, 2005 − June 30, 2006 Authors: C. J. Bishop, G. C. White, D. J. Freddy, and B. E. Watkins All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT We measured mule deer (Odocoileus hemionus) population parameters in response to a nutrition enhancement treatment to evaluate the relative importance of habitat quality as a limiting factor of mule deer in western Colorado during November 2000 – January 2005. We conducted preliminary data analyses upon completion of field work. We found strong evidence that enhanced nutrition increased fawn recruitment to the yearling age class. During 2002−2004, fetus-neonate survival from 1 March−15 December was higher (χ21 = 3.846, P = 0.050) for treatment fawns (S(t) = 0.528, SE = 0.027) than control fawns (S(t) = 0.401, SE = 0.025). During 15 December–15 June, 2001−2004, the overwinter survival rate of fawns was greater (χ21 = 18.781, P < 0.001) in the treatment unit (S(t) = 0.895, SE = 0.029) than in the control unit (S(t) = 0.655, SE = 0.044). Using a staggered entry survival process with data combined over years, survival of treatment fetuses to 1 year of age (S(t) = 0.458, SE = 0.031) was 0.18 higher (χ21 = 13.20, P < 0.001) than survival of control fetuses to 1 year of age (S(t) = 0.276, SE = 0.026). The finite rate of population increase, λ, was 1.20 for treatment deer and 1.04 for control deer. Our preliminary results provided a foundation for focusing deer management efforts on improving habitat quality in western Colorado pinyon-juniper (Pinus edulis-Juniperus osteosperma) ecosystems with corresponding research efforts to quantify the effects of habitat manipulations on deer performance. During the past year, we monitored post-treatment adult doe survival, identified a set of publications to be completed for submission to scientific journals, initiated final data analyses corresponding to the set of publications, and worked on or completed several manuscripts. A manuscript on the effectiveness of vaginal implant transmitters was accepted for publication in the Journal of Wildlife Management, and a manuscript documenting malignant catarrhal fever in the deer population was submitted to the Journal of Wildlife Diseases. The lead investigator also wrote a portion of a book chapter regarding the effects of excessive herbivory on mule deer populations. 59 �WILDLIFE RESEARCH REPORT EFFECT OF NUTRITION AND HABITAT ENHANCEMENTS ON MULE DEER RECRUITMENT AND SURVIVAL RATES CHAD J. BISHOP, GARY C. WHITE, DAVID J. FREDDY, AND BRUCE E. WATKINS P. N. OBJECTIVE To determine experimentally whether enhancing mule deer nutrition during winter and early spring via supplementation increases fetus survival, neonate survival, overwinter fawn survival, or ultimately, population productivity. SEGMENT OBJECTIVES 1. Radio-monitor and measure post-treatment survival of the sample of radio-collared mule deer adult does. 2. Identify a set of publications to be generated from the research. 3. Initiate final data analyses to support preparation of manuscripts. 4. Prepare manuscripts for submission to scientific journals for publication. INTRODUCTION Mule deer (Odocoileus hemionus) numbers apparently declined during the 1990s throughout much of the West, and have clearly decreased since the peak population levels documented during the 1940s−1960s (Unsworth et al. 1999, Gill et al. 2001). Biologists and sportsmen alike have concerns as to what factors may be responsible for declining population trends. Although previous and current research indicates multiple interacting factors are responsible, habitat and predation have typically received the focus of attention. A number of studies have evaluated whether predator control increases deer survival, yet results are highly variable (Connolly 1981, Ballard et al. 2001). Together, predator control studies with adequate rigor and statistical power indicate predation effects on mule deer are variable as a result of time-specific and site-specific factors. Studies which have demonstrated deer population responses to predator control treatments have failed to determine whether predation is ultimately more limiting than habitat when considering long term population changes. Numerous research studies have evaluated mule deer habitat quality, but virtually no studies have documented population responses to habitat improvements. In many areas where declining deer numbers are of concern, predation is common yet habitat quality appears to have declined. The question remains as to whether predation, habitat, or some other factor is more limiting to mule deer in these situations, and whether habitat quality can be improved for the benefit of deer. It may also be that no single factor is responsible for observed deer declines, and a more comprehensive understanding of multi-factor interactions is needed. We designed and implemented a field experiment where we measured deer population responses to a nutrition enhancement treatment to further understand the causative factors underlying observed deer population dynamics. We conducted the study on the Uncompahgre Plateau in southwest Colorado, where several predator species were present in abundant numbers: coyotes (Canis latrans), mountain lions (Felis concolor), and bears (Ursus americanus). In addition to predation, myriad diseases in combination have proximately affected survival of the Uncompahgre deer population (Pojar and Bowden 2004, B. E. Watkins, Colorado Division of Wildlife, unpublished data). Predator numbers were not manipulated in any manner during the course of the study. All factors were left constant with the exception of deer nutrition. Deer nutrition was enhanced by providing supplemental feed to deer occupying a treatment area during winter. We measured December fawn recruitment and overwinter fawn 60 �survival in response to the treatment to determine whether deer nutrition was ultimately more limiting than predation or disease. A second phase of research was initiated in 2005 to quantify deer population parameters in response to manipulations of pinyon-juniper (Pinus edulis-Juniperus osteosperma) habitat (Bergman et al. 2006). The objective of this research is to determine whether habitat can be effectively improved for mule deer by introducing disturbance into late-seral pinyon-juniper stands. STUDY AREA We non-randomly selected two experimental units (A−B) within mule deer winter range on the Uncompahgre Plateau (Figure 1) to facilitate a cross-over experimental design for evaluating the effects of enhanced deer nutrition during winter on annual population performance. Unit A received a nutrition enhancement treatment during the first 2 winters of research (2000 – 2002) while Unit B served as a control unit. During winters 2002−03 and 2003−04, Unit B received the treatment while Unit A served as the control. In late April and May, prior to fawning, deer from the winter range experimental units migrated to summer range. We defined the summer range study area by movements of the radio-collared deer captured on winter range; summer range encompassed >1000 mi2 covering the southern portion of the Uncompahgre Plateau and adjacent San Juan Mountains (Figure 2). Winter range elevations ranged from 1830 m (6000 ft) in Shavano Valley to 2318 m (7600 ft) adjacent to the Dry Creek Rim above Shavano Valley. Winter range habitat was dominated by pinyon-juniper with interspersed sagebrush adjacent to agricultural fields in the Shavano and Uncompahgre Valleys. Summer range elevations occupied by deer ranged from 1891 m (6200 ft) in the Uncompahgre Valley to 3538 m (11,600 ft) in Imogene Basin southwest of Ouray, CO. Summer range habitats were dominated by spruce-subalpine fir (Picea spp.-Abies lasiocarpa), aspen (Populus tremuloides), sagebrush, ponderosa pine (Pinus ponderosa), Gambel oak (Quercus gambelii), and to a lesser extent, pinyon-juniper at lower elevations. Bishop et al. (2005) provide a detailed study area description. METHODS Refer to Bishop et al. (2005) for field methodology employed during 2000−2005. During fiscal year 2005-06, we continued to monitor radio-collared adult female deer occupying the two experimental units. Our primary research efforts were focused on data analysis and the preparation of manuscripts for publication in scientific journals. The lead investigator completed additional coursework in mathematical statistics, data analysis, and animal nutrition. We submitted or intend to submit the following manuscripts for publication: 1. Effect of enhanced nutrition on the population performance of free-ranging mule deer. Journal of Wildlife Management. a. A separate publication may be submitted to Science focused on the documentation of coyote predation as a compensatory mortality factor during winter. 2. Using vaginal implant transmitters to aid in capture of neonates from marked mule deer. Journal of Wildlife Management. 3. Evaluation of overdispersion in survival analyses of neonate mule deer associated with sibling fawns. Journal of Wildlife Management. 4. Evaluation of serum thyroid hormone levels as an indicator of body condition during late winter. Journal of Wildlife Management. 5. Evaluation of mule deer age and sex ratios as a response variable in field research. Journal of Wildlife Management. 6. Bovine viral diarrhea isolation and seroprevalence in a free-ranging mule deer (Odocoileus hemionus) population in southwest Colorado. Journal of Wildlife Diseases. 7. Malignant catarrhal fever associated with ovine herpesvirus-2 in free-ranging mule deer (Odocoileus hemionus) in Colorado. Journal of Wildlife Diseases. 61 �8. Spatial patterns in mortality causes of neonatal mule deer across a land use gradient in southwest Colorado. (This could go to several different journals or be published as an internal CDOW publication.) 9. Disease assessment in a Colorado mule deer population following a decline. Journal of Wildlife Diseases (or internal CDOW publication). RESULTS AND DISCUSSION A comprehensive presentation and discussion of preliminary results was provided by Bishop et al. (2005). These results have not changed and therefore we do not repeat them here. The following manuscript was accepted for publication (Appendix I): Bishop, C. J., D. J. Freddy, G. C. White, B. E. Watkins, T. R. Stephenson, and L. L. Wolfe. Using vaginal implant transmitters to aid in capture of neonates from marked mule deer. Journal of Wildlife Management. The following manuscript was submitted for publication (Appendix II): Schultheiss, P. C., H. Van Campen, T. R. Spraker, C. J. Bishop, L. L. Wolfe, and B. Podell. Malignant catarrhal fever associated with ovine herpesvirus-2 in free-ranging mule deer (Odocoileus hemionus) in Colorado. Journal of Wildlife Diseases. The following book chapter was completed and currently undergoing external peer review: Watkins, B. E., C. J. Bishop, E. J. Bergman, A. Bronson, B. Hale, D. W. Lutz, B. F. Wakeling, and L. C. Carpenter. Habitat guidelines for mule deer: Colorado Plateau Ecoregion. Mule Deer Working Group, Western Association of Fish and Wildlife Agencies. The lead investigator completed the following courses to assist with data analysis and manuscript preparation: mathematical statistics (2), population dynamics, population analysis, wildlife nutrition, and animal metabolism. A data bootstrap analysis in SAS was initiated to quantify the degree of overdispersion in our neonate survival data. Overdispersion represents extra-binomial variation in sample data arising from violations of independence. Functionally, undetected overdispersion will result in overly precise variance estimates, and ultimately, incorrect inference. Our neonate samples were subject to independence violations because all captured siblings were routinely radio-collared and treated as independent sample units. A known fates analysis will be conducted using Program MARK to quantify the effect of the nutrition enhancement treatment on various stages of fawn survival while simultaneously accounting for temporal variation and individual heterogeneity (i.e., fawn weight and hind foot length). Once these analyses are completed, we will write and submit manuscripts accordingly. The remaining manuscripts will then be handled in order of priority. Our anticipated timeline is detailed below. Draft manuscripts completed in FY 06-07: 1. Effect of enhanced nutrition on the population performance of free-ranging mule deer. Journal of Wildlife Management. 2. Evaluation of overdispersion in survival analyses of neonate mule deer associated with sibling fawns. Journal of Wildlife Management. 3. Evaluation of serum thyroid hormone levels as an indicator of body condition during late winter. Journal of Wildlife Management. Draft manuscripts completed in FY 07-08: 62 �1. Evaluation of mule deer age and sex ratios as a response variable in field research. Journal of Wildlife Management. 2. Bovine viral diarrhea isolation and seroprevalence in a free-ranging mule deer (Odocoileus hemionus) population in southwest Colorado. Journal of Wildlife Diseases. 3. Spatial patterns in mortality causes of neonatal mule deer across a land use gradient in southwest Colorado. The final proposed manuscript related to disease assessment will be completed as time allows. SUMMARY Enhanced winter nutrition of free-ranging deer caused an increase in both fetus-neonate survival and overwinter fawn survival, resulting in higher yearling recruitment (Bishop et al. 2005). Overwinter adult doe survival increased as a result of the treatment, but annual survival was more similar among treatment and control adult does. Combining all parameter estimates into a deterministic population model, the treatment population indicated an exceptionally high rate of increase (λ = 1.20) while the control population (λ = 1.04) was indicative of the overall Uncompahgre deer population during 2000−2004. The nutrition enhancement treatment was artificial in the sense that we applied it only to test whether habitat quality was ultimately more limiting than predation or other factors. Our results to do not provide support for managing deer populations with nutrition supplements because our treatment delivery approach could not be applied to a large number of animals over a large area. Rather, our results provide a foundation for focusing deer management efforts on improving habitat quality in western Colorado pinyon-juniper ecosystems with corresponding research efforts to quantify the effects of habitat manipulations on deer. We are presently in the process of conducting final data analyses and preparing and submitting manuscripts for publication in scientific journals. LITERATURE CITED BALLARD, W. B., D. LUTZ, T. W. KEEGAN, L. H. CARPENTER, AND J. C. DEVOS, JR. 2001. Deer-predator relationships: a review of recent North American studies with emphasis on mule and black-tailed deer. Wildlife Society Bulletin 29:99−115. BERGMAN, E. J., C. J. BISHOP, D. J. FREDDY, AND G. C. WHITE. 2006. Evaluation of winter range habitat treatments on over-winter survival and body condition of mule deer. Federal Aid in Wildlife Restoration Project W−185−R, Job Progress Report. Wildlife Research Report, Colorado Division of Wildlife, Fort Collins, USA. BISHOP, C. J., G. C. WHITE, D. J. FREDDY, AND B. E. WATKINS. 2005. Effect of nutrition and habitat enhancements on mule deer recruitment and survival rates. Federal Aid in Wildlife Restoration Project W−185−R, Job Progress Report. Wildlife Research Report, Colorado Division of Wildlife, Fort Collins, USA. CONNOLLY, G. E. 1981. Limiting factors and population regulation. Pages 245−285 in O. C. Wallmo, editor. Mule and black-tailed deer of North America. University of Nebraska Press, Lincoln, USA. GILL, R. B., T. D. I. BECK, C. J. BISHOP, D. J. FREDDY, N. T. HOBBS, R. H. KAHN, M. W. MILLER, T. M. POJAR, AND G. C. WHITE. 2001. Declining mule deer populations in Colorado: reasons and responses. Colorado Division of Wildlife Special Report Number 77. Denver, USA. POJAR, T. M., AND D. C. BOWDEN. 2004. Neonatal mule deer fawn survival in west-central Colorado. Journal of Wildlife Management 68:550−560. UNSWORTH, J. W., D. F. PAC, G. C. WHITE, AND R. M. BARTMANN. 1999. Mule deer survival in Colorado, Idaho, and Montana. Journal of Wildlife Management 63:315−326. Prepared by _______________________ Chad J. Bishop, Wildlife Researcher 63 �Year Unit A Unit B 2000-01 Treatment Control 2001-02 Treatment Control 2002-03 Control Treatment 2003-04 Control Treatment Figure 1. Schematic representation of experimental units and nutrition enhancement treatment allocation. Units A and B were located in winter range habitat on the Uncompahgre Plateau in southwest Colorado. The nutrition enhancement cross-over design encompassed 4 years. Uncompahgre Plateau Mesa County GRAND JUNCTION Delta County m co Un GMU 62 r hg pa e u ea at Pl Montrose County GMU 61 Sanmiguel County Gunnison County DELTA Shavano E. U. Winter Range Exp. Units MONTROSE Colona Montrose County Summer Range E.U. Ouray County Figure 2. Location of Colona and Shavano (Units A and B) experimental units on the Uncompahgre Plateau, southwest Colorado; and location of the summer range study area encompassing the southern Uncompahgre Plateau and adjacent San Juan Mountains. 64 �APPENDIX I The following manuscript (referenced here by Abstract) was accepted for publication by the Journal of Wildlife Management. USING VAGINAL IMPLANT TRANSMITTERS TO AID IN CAPTURE OF NEONATES FROM MARKED MULE DEER CHAD J. BISHOP, DAVID J. FREDDY, GARY C. WHITE, BRUCE E. WATKINS, THOMAS R. STEPHENSON, AND LISA L. WOLFE ABSTRACT Measuring reproductive success of previously marked, adult female ungulates enables study of certain ecological factors that limit populations. We evaluated the feasibility and efficiency of capturing large samples (i.e., >80/year) of neonate mule deer (Odocoileus hemionus) exclusively from free-ranging, marked adult does using vaginal implant transmitters (VITs, n = 154) and repeated locations of radiocollared does without VITs. We also evaluated the effectiveness of VITs, when used in conjunction with in utero fetal counts, for obtaining direct estimates of fetal survival. During 2003 and 2004, after VIT batteries were placed on a 12-hour duty cycle to lower electronic failure rates, the proportion of VITs that shed ≤3 days prepartum or during parturition was 0.623 (SE = 0.0456), and the proportion shed only during parturition was 0.447 (SE = 0.0468). Our neonate capture success rate was 0.880 (SE = 0.0359) from does with VITs shed ≤3 days prepartum or during parturition and 0.307 (SE = 0.0235) from radiocollared does without VITs or whose implants failed to function properly. Using a combination of techniques, we captured 275 neonates and found 21 stillborns during 2002−2004. We accounted for all fetuses at birth (i.e., live or stillborn) from 78 of the 147 does (0.531, SE = 0.0413) having winter fetal counts, and this rate was heavily dependent on VIT retention success. Deer that shed VITs prepartum were larger than deer that retained VITs to parturition, indicating a need to develop variable-sized VITs that may be fitted individually to deer in the field. We demonstrated that direct estimates of fetal and neonatal survival may be obtained from previously marked female mule deer in free-ranging populations, thus expanding opportunities for conducting field experiments. Survival estimates using VITs lacked bias that is typically associated with other neonate capture techniques. However, current vaginal implant failure rates, and overall expense, limit broad applicability of the technique. 65 �APPENDIX II The following manuscript (referenced here by Abstract) was submitted to the Journal of Wildlife Diseases. MALIGNANT CATARRHAL FEVER ASSOCIATED WITH OVINE HERPESVIRUS-2 IN FREERANGING MULE DEER (Odocoileus hemionus) IN COLORADO PATRICIA C. SCHULTHEISS, HANA VAN CAMPEN, TERRY R. SPRAKER, CHAD J. BISHOP, LISA L. WOLFE, AND BRENDAN PODELL ABSTRACT Malignant catarrhal fever (MCF) was diagnosed in 4 free-ranging mule deer (Odocoileus hemionus) in January and February of 2003. Diagnosis was based on typical histologic lesions of lymphocytic vasculitis and PCR identification of ovine herpesvirus-2 (OHV-2) viral genetic sequences in formalin fixed tissues. The animals were from the Uncompahgre Plateau of southwestern Colorado. Deer from these herds occasionally resided in close proximity to domestic sheep (Ovis aries), the reservoir host of OHV-2, in agricultural valleys adjacent to their winter range. These cases indicate that fatal OHV-2 associated MCF can occur in free-ranging mule deer exposed to domestic sheep that overlap their range. 66 �Colorado Division of Wildlife July 2006 − June 2007 WILDLIFE RESEARCH REPORT State of Cost Center Work Package Task No. Colorado 3430 3001 4 Federal Aid Project: W-185-R : Division of Wildlife : Mammals Research : Deer Conservation : Effect of Nutrition and Habitat Enhancements on Mule Deer Recruitment and Survival Rates : Period Covered: July 1, 2006 − June 30, 2007 Authors: C. J. Bishop, G. C. White, D. J. Freddy, and B. E. Watkins All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT We measured mule deer (Odocoileus hemionus) population parameters in response to a nutrition enhancement treatment to evaluate the relative importance of habitat quality as a limiting factor of mule deer in western Colorado during November 2000 – January 2005. The nutrition enhancement treatment increased survival of fetuses to the yearling age class by 0.14−0.20 depending on year and fawn sex, although 95% confidence intervals slightly overlapped 0. The nutrition treatment also had a positive effect on annual adult doe survival. Survival of does receiving the treatment (Ŝ = 0.879, SE = 0.0206) was higher than survival of control does (Ŝ = 0.833, SE = 0.0253). Our estimate of the population rate of change, λ̂ , was 1.15−1.17 for treatment deer and 1.02−1.06 for control deer, with some overlap in 95% confidence intervals. The treatment caused λ̂ to increase by 0.139 (95% CI: −0.0152, 0.2941) during 2001−02, 0.113 (95% CI: −0.0009, 0.2279) during 2002−03, and 0.145 (95% CI: 0.0176, 0.2723) during 2003−04.Our results provide a foundation for focusing deer management efforts on improving habitat quality in western Colorado pinyon-juniper (Pinus edulis-Juniperus osteosperma) ecosystems with corresponding research efforts to quantify the effects of habitat manipulations on deer performance. During the past year, we had 2 papers published in peer-reviewed journals, we completed final data analyses, and we prepared 3 other manuscripts for publication. The published manuscripts included: 1) a manuscript on the effectiveness of vaginal implant transmitters (Journal of Wildlife Management 71(3):945−954), and 2) a manuscript documenting malignant catarrhal fever in the Uncompahgre deer population (Journal of Wildlife Diseases 43(3):533−537). We also completed a publication on mule deer habitat guidelines for the Colorado Plateau ecoregion, which will be published by the Western Association of Fish and Wildlife Agencies. The estimated publication date is January 2008. 59 �WILDLIFE RESEARCH REPORT EFFECT OF NUTRITION AND HABITAT ENHANCEMENTS ON MULE DEER RECRUITMENT AND SURVIVAL RATES CHAD J. BISHOP, GARY C. WHITE, DAVID J. FREDDY, AND BRUCE E. WATKINS P. N. OBJECTIVE To determine experimentally whether enhancing mule deer nutrition during winter and early spring via supplementation increases fetal survival, neonatal survival, overwinter fawn survival, or ultimately, population productivity. SEGMENT OBJECTIVES 1. Complete final data analyses to support preparation of manuscripts. 2. Prepare manuscripts for submission to scientific journals for publication. 3. Complete dissertation as part of PhD requirements at Colorado State University INTRODUCTION Mule deer (Odocoileus hemionus) numbers apparently declined during the 1990s throughout much of the West, and have clearly decreased since the peak population levels documented during the 1940s−1960s (Unsworth et al. 1999, Gill et al. 2001). Biologists and sportsmen alike have concerns as to what factors may be responsible for declining population trends. Although previous and current research indicates multiple interacting factors are responsible, habitat and predation have typically received the focus of attention. A number of studies have evaluated whether predator control increases deer survival, yet results are highly variable (Connolly 1981, Ballard et al. 2001). Together, predator control studies with adequate rigor and statistical power indicate predation effects on mule deer are variable as a result of time-specific and site-specific factors. Studies which have demonstrated deer population responses to predator control treatments have failed to determine whether predation is ultimately more limiting than habitat when considering long term population changes. Numerous research studies have evaluated mule deer habitat quality, but virtually no studies have documented population responses to habitat improvements. In many areas where declining deer numbers are of concern, predation is common yet habitat quality appears to have declined. The question remains as to whether predation, habitat, or some other factor is more limiting to mule deer in these situations, and whether habitat quality can be improved for the benefit of deer. It may also be that no single factor is responsible for observed deer declines, and a more comprehensive understanding of multi-factor interactions is needed. We designed and implemented a field experiment where we measured deer population responses to a nutrition enhancement treatment to further understand the causative factors underlying observed deer population dynamics. We conducted the study on the Uncompahgre Plateau in southwest Colorado, where several predator species were present in abundant numbers: coyotes (Canis latrans), mountain lions (Felis concolor), and bears (Ursus americanus). In addition to predation, myriad diseases in combination have proximately affected survival of the Uncompahgre deer population (Pojar and Bowden 2004, B. E. Watkins, Colorado Division of Wildlife, unpublished data). Predator numbers were not manipulated in any manner during the course of the study. All factors were left constant with the exception of deer nutrition. Deer nutrition was enhanced by providing supplemental feed to deer occupying a treatment area during winter. We measured December fawn recruitment and overwinter fawn survival in response to the treatment to determine whether deer nutrition was ultimately more limiting than predation or disease. A second phase of research was initiated in 2005 to quantify deer population 60 �parameters in response to manipulations of pinyon-juniper (Pinus edulis-Juniperus osteosperma) habitat (Bergman et al. 2006). The objective of this research is to determine whether habitat can be effectively improved for mule deer by introducing disturbance into late-seral pinyon-juniper stands. STUDY AREA We non-randomly selected two experimental units (A−B) within mule deer winter range on the Uncompahgre Plateau (Figure 1) to facilitate a cross-over experimental design for evaluating the effects of enhanced deer nutrition during winter on annual population performance. Unit A received a nutrition enhancement treatment during the first 2 winters of research (2000 – 2002) while Unit B served as a control unit. During winters 2002−03 and 2003−04, Unit B received the treatment while Unit A served as the control. In late April and May, prior to fawning, deer from the winter range experimental units migrated to summer range. We defined the summer range study area by movements of the radio-collared deer captured on winter range; summer range encompassed >1000 mi2 covering the southern portion of the Uncompahgre Plateau and adjacent San Juan Mountains (Figure 2). Winter range elevations ranged from 1830 m (6000 ft) in Shavano Valley to 2318 m (7600 ft) adjacent to the Dry Creek Rim above Shavano Valley. Winter range habitat was dominated by pinyon-juniper with interspersed sagebrush adjacent to agricultural fields in the Shavano and Uncompahgre Valleys. Summer range elevations occupied by deer ranged from 1891 m (6200 ft) in the Uncompahgre Valley to 3538 m (11,600 ft) in Imogene Basin southwest of Ouray, CO. Summer range habitats were dominated by spruce-subalpine fir (Picea spp.-Abies lasiocarpa), aspen (Populus tremuloides), sagebrush, ponderosa pine (Pinus ponderosa), Gambel oak (Quercus gambelii), and to a lesser extent, pinyon-juniper at lower elevations. Bishop et al. (2005) provide a detailed study area description. METHODS Refer to Bishop et al. (2005) for field methodology employed during 2000−2005. During fiscal year 2006-07, we had 2 papers published in peer reviewed wildlife journals, which were the result of work completed in the previous year. Our primary research efforts were focused on data analysis and the preparation of manuscripts for publication in scientific journals. We spent much of the year evaluating dependence among deer siblings with respect to fetal and neonatal survival analyses. Essentially all statistical analyses are based on an assumption that sample units are independent. In survival analyses, the assumption pertains to independence of fates. That is, we must assume that the death or survival of one sample unit is not related to the fate of another. Predation and maternal condition are both good examples of mechanisms that could cause sibling neonates to lack independent fates. If a coyote or bear kills twin fawns because both were together, clearly those mortality events were not independent. Similarly, a lack of independence would occur if twin fawns each die of starvation because their dam is in poor condition. Data are considered overdispersed when the independence assumption is violated. Overdispersion does not generally affect point estimates, but rather causes variances to be underestimated. We estimated overdispersion in fetal and neonatal survival datasets and incorporated a data bootstrap procedure into Program MARK (White and Burnham 1999), making it easier for others to conduct similar analyses. The procedure in MARK can be generalized to any situation where multiple individuals are marked from the same litter, clutch, pair, trap site, etc. Once we completed the overdispersion analysis, we spent the remainder of the year conducting final data analyses that quantified the effect of enhanced nutrition on population performance. We then prepared several manuscripts that incorporated the various analyses we conducted. The principal investigator also completed a draft of his PhD dissertation. 61 �RESULTS AND DISCUSSION A comprehensive presentation and discussion of preliminary results was provided by Bishop et al. (2005). These results have not changed and therefore we do not repeat them here. The final results are contained in peer-reviewed manuscripts that have either already been published or will be submitted for publication in 2007 or early 2008. The following manuscripts were published in 2007 (abstracts are provided in Appendix I): Bishop, C. J., D. J. Freddy, G. C. White, B. E. Watkins, T. R. Stephenson, and L. L. Wolfe. 2007. Using vaginal implant transmitters to aid in capture of mule deer neonates. Journal of Wildlife Management 71:945−954. . Schultheiss, P. C., H. Van Campen, T. R. Spraker, C. J. Bishop, L. L. Wolfe, and B. Podell. 2007. Malignant catarrhal fever associated with ovine herpesvirus-2 in free-ranging mule deer in Colorado. Journal of Wildlife Diseases 43:533−537. The following book chapter was completed should be published by January 2008: Watkins, B. E., C. J. Bishop, E. J. Bergman, A. Bronson, B. Hale, B. F. Wakeling, L. H. Carpenter., and D. W. Lutz. 2007. Habitat guidelines for mule deer: Colorado Plateau shrubland and forest ecoregion. Mule Deer Working Group, Western Association of Fish and Wildlife Agencies. The following draft manuscripts were prepared in 2007 and will be submitted for publication in 2007 or early 2008 (abstracts are provided in Appendix II): Bishop, C. J., G. C. White, and P. M. Lukacs. In review. Evaluating dependence among mule deer siblings in fetal and neonatal survival analyses. Journal of Wildlife Management. Bishop, C. J., G. C. White, D. J. Freddy, B. E. Watkins, and T. R. Stephenson. In review. Effect of enhanced nutrition on mule deer population performance. Journal of Wildlife Management OR Wildlife Monographs. Bishop, C. J., B. E. Watkins, L. L. Wolfe, D. J. Freddy, and G. C. White. In review. Evaluating mule deer body condition during late winter using serum thyroid hormone concentrations. Journal of Wildlife Management. The following draft dissertation was prepared in 2007 and submitted for review at Colorado State University (abstract is provided in Appendix III): Bishop, C. J. In review. Effect of enhanced nutrition during winter on the Uncompahgre Plateau mule deer population. Dissertation, Colorado State University, Fort Collins, USA. We intend to pursue several additional manuscripts as time allows, listed below in order of priority. 1. Evaluating dependence of fates among mule deer siblings in Colorado, Idaho, and Montana. Journal of Wildlife Management. 2. Bovine viral diarrhea isolation and seroprevalence in a free-ranging mule deer (Odocoileus hemionus) population in southwest Colorado. Journal of Wildlife Diseases. 3. Spatial patterns in mortality causes of neonatal mule deer across a land use gradient in southwest Colorado. Journal of Wildlife Management. 62 �4. Evaluation of mule deer age and sex ratios as a response variable in field research. Journal of Wildlife Management. SUMMARY Enhanced winter nutrition of free-ranging deer caused an increase in both fetus-neonate survival and overwinter fawn survival, resulting in higher yearling recruitment. Overwinter adult doe survival increased as a result of the treatment, and therefore annual survival was higher among treatment than control adult does. Combining all parameter estimates into a deterministic population model, the treatment population indicated an exceptionally high rate of increase while the control population was stable and indicative of the overall Uncompahgre deer population during 2000−2004. The nutrition enhancement treatment was artificial in the sense that we applied it only to test whether habitat quality was ultimately more limiting than predation or other factors. Our results to do not provide support for managing deer populations with nutrition supplements because our treatment delivery approach could not be applied to a large number of animals over a large area. Rather, our results provide a foundation for focusing deer management efforts on improving habitat quality in western Colorado pinyon-juniper ecosystems with corresponding research efforts to quantify the effects of habitat manipulations on deer. We are presently in the process of conducting final data analyses and preparing and submitting manuscripts for publication in scientific journals. LITERATURE CITED BALLARD, W. B., D. LUTZ, T. W. KEEGAN, L. H. CARPENTER, AND J. C. DEVOS, JR. 2001. Deer-predator relationships: a review of recent North American studies with emphasis on mule and black-tailed deer. Wildlife Society Bulletin 29:99−115. BERGMAN, E. J., C. J. BISHOP, D. J. FREDDY, AND G. C. WHITE. 2006. Evaluation of winter range habitat treatments on over-winter survival and body condition of mule deer. Federal Aid in Wildlife Restoration Project W−185−R, Job Progress Report. Wildlife Research Report, Colorado Division of Wildlife, Fort Collins, USA. BISHOP, C. J., G. C. WHITE, D. J. FREDDY, AND B. E. WATKINS. 2005. Effect of nutrition and habitat enhancements on mule deer recruitment and survival rates. Federal Aid in Wildlife Restoration Project W−185−R, Job Progress Report. Wildlife Research Report, Colorado Division of Wildlife, Fort Collins, USA. CONNOLLY, G. E. 1981. Limiting factors and population regulation. Pages 245−285 in O. C. Wallmo, editor. Mule and black-tailed deer of North America. University of Nebraska Press, Lincoln, USA. GILL, R. B., T. D. I. BECK, C. J. BISHOP, D. J. FREDDY, N. T. HOBBS, R. H. KAHN, M. W. MILLER, T. M. POJAR, AND G. C. WHITE. 2001. Declining mule deer populations in Colorado: reasons and responses. Colorado Division of Wildlife Special Report Number 77. Denver, USA. POJAR, T. M., AND D. C. BOWDEN. 2004. Neonatal mule deer fawn survival in west-central Colorado. Journal of Wildlife Management 68:550−560. UNSWORTH, J. W., D. F. PAC, G. C. WHITE, AND R. M. BARTMANN. 1999. Mule deer survival in Colorado, Idaho, and Montana. Journal of Wildlife Management 63:315−326. Prepared by _______________________ Chad J. Bishop, Wildlife Researcher 63 �Year Unit A Unit B 2000-01 Treatment Control 2001-02 Treatment Control 2002-03 Control Treatment 2003-04 Control Treatment Figure 1. Schematic representation of experimental units and nutrition enhancement treatment allocation. Units A and B were located in winter range habitat on the Uncompahgre Plateau in southwest Colorado. The nutrition enhancement cross-over design encompassed 4 years. Uncompahgre Plateau Mesa County GRAND JUNCTION Delta County m co Un GMU 62 r hg pa e u ea at Pl Montrose County GMU 61 Sanmiguel County Gunnison County DELTA Shavano E.U. Winter Range Exp. Units MONTROSE Colona Montrose County Summer Range E.U. Ouray County Figure 2. Location of Colona and Shavano (Units A and B) experimental units on the Uncompahgre Plateau, southwest Colorado; and location of the summer range study area encompassing the southern Uncompahgre Plateau and adjacent San Juan Mountains. 64 �APPENDIX I The following manuscript (referenced here by Abstract) was published in the Journal of Wildlife Management in 2007. USING VAGINAL IMPLANT TRANSMITTERS TO AID IN CAPTURE OF MULE DEER NEONATES CHAD J. BISHOP, DAVID J. FREDDY, GARY C. WHITE, BRUCE E. WATKINS, THOMAS R. STEPHENSON, AND LISA L. WOLFE ABSTRACT Estimating survival of the offspring of marked female ungulates has proven difficult in freeranging populations yet could improve our understanding of factors that limit populations. We evaluated the feasibility and efficiency of capturing large samples (i.e., >80/year) of neonate mule deer (Odocoileus hemionus) exclusively from free-ranging, marked adult does using vaginal implant transmitters (VITs, n = 154) and repeated locations of radio-collared does without VITs. We also evaluated the effectiveness of VITs, when used in conjunction with in utero fetal counts, for obtaining direct estimates of fetal survival. During 2003 and 2004, after we placed VIT batteries on a 12-hour duty cycle to lower electronic failure rates, the proportion that shed ≤3 days prepartum or during parturition was 0.623 (SE = 0.0456), and the proportion of VITs shed only during parturition was 0.447 (SE = 0.0468). Our neonate capture success rate was 0.880 (SE = 0.0359) from does with VITs shed ≤3 days prepartum or during parturition and 0.307 (SE = 0.0235) from radio-collared does without VITs or whose implants failed to function properly. Using a combination of techniques, we captured 275 neonates and found 21 stillborns during 2002−2004. We accounted for all fetuses at birth (i.e., live or stillborn) from 78 of the 147 does (0.531, SE = 0.0413) having winter fetal counts, and this rate was heavily dependent on VIT retention success. Deer that shed VITs prepartum were larger than deer that retained VITs to parturition, indicating a need to develop variable-sized VITs that may be fitted individually to deer in the field. We demonstrated that direct estimates of fetal and neonatal survival may be obtained from previously marked female mule deer in free-ranging populations, thus expanding opportunities for conducting field experiments. Survival estimates using VITs lacked bias that is typically associated with other neonate capture techniques. However, current vaginal implant failure rates, and overall expense, limit broad applicability of the technique. Citation: Bishop, C. J., D. J. Freddy, G. C. White, B. E. Watkins, T. R. Stephenson, and L. L. Wolfe. 2007. Using vaginal implant transmitters to aid in capture of mule deer neonates. Journal of Wildlife Management 71:945−954. 65 �The following manuscript (referenced here by Abstract) was published in the Journal of Wildlife Diseases in 2007: MALIGNANT CATARRHAL FEVER ASSOCIATED WITH OVINE HERPESVIRUS-2 IN FREERANGING MULE DEER IN COLORADO PATRICIA C. SCHULTHEISS, HANA VAN CAMPEN, TERRY R. SPRAKER, CHAD J. BISHOP, LISA L. WOLFE, AND BRENDAN PODELL ABSTRACT Malignant catarrhal fever (MCF) was diagnosed in 4 free-ranging mule deer (Odocoileus hemionus) in January and February of 2003. Diagnosis was based on typical histologic lesions of lymphocytic vasculitis and PCR identification of ovine herpesvirus-2 (OHV-2) viral genetic sequences in formalin fixed tissues. The animals were from the Uncompahgre Plateau of southwestern Colorado. Deer from these herds occasionally resided in close proximity to domestic sheep (Ovis aries), the reservoir host of OHV-2, in agricultural valleys adjacent to their winter range. These cases indicate that fatal OHV-2 associated MCF can occur in free-ranging mule deer exposed to domestic sheep that overlap their range. Citation: Schultheiss, P. C., H. Van Campen, T. R. Spraker, C. J. Bishop, L. L. Wolfe, and B. Podell. 2007. Malignant catarrhal fever associated with ovine herpesvirus-2 in free-ranging mule deer in Colorado. Journal of Wildlife Diseases 43:533−537. 66 �APPENDIX II The following draft manuscripts (referenced here by Abstract) were prepared in 2007 and will be submitted to the Journal of Wildlife Management. EVALUATING DEPENDENCE AMONG MULE DEER SIBLINGS IN FETAL AND NEONATAL SURVIVAL ANALYSES CHAD J. BISHOP, GARY C. WHITE, AND PAUL M. LUKACS ABSTRACT The assumption of independent sample units is potentially violated in deer (Odocoileus spp.) fetal and neonatal survival analyses where twin and triplet siblings comprise a high proportion of the sample. Violation of the independence assumption causes sample data to be overdispersed relative to a binomial model, and therefore requires a variance inflation factor, c, to obtain appropriate estimates of sampling variances. We evaluated overdispersion in fetal and neonatal mule deer (O. hemionus) datasets where more than half of the sample units were comprised of siblings. We developed a likelihood function for estimating fetal survival when the fates of some fetuses are unknown, and we used several variations of the binomial model to estimate neonatal survival. We compared theoretical variance estimates obtained from these analyses with empirical variance estimates obtained from data bootstrap analyses to estimate the overdisperion parameter, c. Our estimates of c for fetal survival ranged from 0.678 to 1.118, which provided virtually no evidence of overdispersion. For neonatal survival, 3 different models indicated that ĉ ranged from 1.1 to 1.4 and averaged 1.24−1.26, providing evidence of limited overdispersion (i.e., limited sibling dependence). Our results indicate that fates of sibling mule deer fetuses and neonates may often be independent even though they have the same dam. Predation tends to act independently on sibling neonates because of dam-neonate behavioral adaptations, although we observed several cases of a predator(s) killing siblings. The effect of maternal characteristics on sibling fate dependence is less straightforward and may vary by circumstance. We recommend that future neonatal survival studies incorporate additional sampling intensity to accommodate modest overdispersion (i.e., ĉ = 1.25), which would facilitate a corresponding ĉ adjustment in a model selection analysis using quasi-likelihood without a reduction in power. 67 �EFFECT OF ENHANCED NUTRITION ON MULE DEER POPULATION PERFORMANCE CHAD J. BISHOP, GARY C. WHITE, DAVID J. FREDDY, BRUCE E. WATKINS, AND THOMAS R. STEPHENSON ABSTRACT Concerns over declining mule deer (Odocoileus hemionus) populations during the 1990s prompted research efforts to identify and understand key limiting factors of deer. Similar to past deer declines, a top priority of state wildlife agencies was to evaluate the relative importance of habitat and predation. We therefore evaluated the effect of enhanced nutrition of deer during winter and spring on fecundity and survival rates using a field experiment involving free-ranging mule deer on the Uncompahgre Plateau in southwest Colorado. The nutrition enhancement treatment represented an instantaneous increase in nutritional carrying capacity of a pinyon (Pinus edulis) and Utah juniper (Juniperus osteosperma) winter range, and was intended to simulate optimum habitat quality. Prior studies on the Uncompahgre Plateau indicated predation and disease were the most common proximate causes of deer mortality. By manipulating nutrition and leaving predation as it was, we determined whether habitat quality was ultimately a critical limiting factor of the deer population. We measured fetal, neonatal, and overwinter fawn survival, and annual adult doe survival, which we then used to estimate population rate of change as a function of enhanced nutrition. Pregnancy and fetal rates were high for all deer, regardless of the nutrition treatment. Fetal and neonatal survival rates were higher among deer that received the nutrition enhancement treatment than deer that served as experimental controls. Overwinter fawn survival was considerably higher among treatment deer than control deer. Overwinter survival increased by 0.16−0.31 depending on year and fawn sex, and none of the 95% confidence intervals associated with the effect overlapped 0. The nutrition enhancement treatment increased survival of fetuses to the yearling age class by 0.14−0.20 depending on year and fawn sex, although 95% confidence intervals slightly overlapped 0. The nutrition treatment also had a positive effect on annual adult doe survival. Survival of does receiving the treatment (Ŝ = 0.879, SE = 0.0206) was higher than survival of control does (Ŝ = 0.833, SE = 0.0253). Our estimate of the population rate of change, λ̂ , was 1.15−1.17 for treatment deer and 1.02−1.06 for control deer, with some overlap in 95% confidence intervals. The treatment caused λ̂ to increase by 0.139 (95% CI: −0.0152, 0.2941) during 2001−02, 0.113 (95% CI: −0.0009, 0.2279) during 2002−03, and 0.145 (95% CI: 0.0176, 0.2723) during 2003−04. We documented density dependence in the Uncompahgre deer population because survival of fawns and does increased considerably in response to enhanced nutrition. We found strong evidence that coyote (Canis latrans) predation of ≥6 month old fawns and adult does was compensatory. Our results demonstrate that observed coyote predation is not useful for evaluating whether coyotes are negatively impacting a deer population. We also found evidence that mountain lion (Puma concolor) predation was compensatory. Disease was not compensatory among adult does. We found that winter range habitat quality was a limiting factor of the Uncompahgre Plateau deer population. We recommend the implementation and evaluation of habitat treatments designed to set back succession and increase productivity of late-seral pinyon-juniper habitats that presently dominate the landscape because of the absence of fire. 68 �EVALUATING MULE DEER BODY CONDITION DURING LATE WINTER USING SERUM THYROID HORMONE CONCENTRATIONS CHAD J. BISHOP, BRUCE E. WATKINS, LISA L. WOLFE, DAVID J. FREDDY, AND GARY C. WHITE ABSTRACT Body condition of ungulates is ultimately a determinant of fecundity and survival rates. Researchers have recently developed ultrasonography and body condition scoring techniques that allow reliable estimation of body fat in several ungulate species, but the approach is not feasible to employ in all circumstances, particularly in routine population monitoring programs. There remains a need for a reliable blood chemistry index that could be used to assess relative condition of different deer populations or groups. We evaluated the relationship between estimated body fat of free-ranging mule deer and serum concentrations of total thyroxine (T4), total triiodothyronine (T3), free T4 (FT4), and free T3 (FT3), during late February−early March in southwest Colorado. Deer body fat varied widely because we imposed a nutrition treatment on one-half of our sample. Concentrations of T4 and FT4 were 48.23 nmol/l (SE = 3.88) and 12.69 pmol/l (SE = 1.12) higher, respectively, in deer that received the nutrition treatment than deer that did not receive the treatment. Our optimal model of estimated body fat included T4, T42, FT4 and deer chest girth (%Fât = –4.8015 – 0.0946×T4 + 0.000603×T42 + 0.1474×FT4 + 0.1426×chest girth, r2 = 0.609). Ultrasound and body condition scoring should be used to estimate body fat whenever possible. However, in cases where only a blood sample can be obtained, we documented the potential utility of T4 and FT4 during late winter for evaluating relative body condition of mule deer. 69 �APPENDIX III The following draft dissertation (referenced here by Abstract) was prepared in 2007 and will be submitted to Colorado State University. EFFECT OF ENHANCED NUTRITION DURING WINTER ON THE UNCOMPAHGRE PLATEAU MULE DEER POPULATION CHAD J. BISHOP ABSTRACT Mule deer (Odocoileus hemionus) populations declined across much of the West during the 1990s, prompting state wildlife agencies to pursue explorations of mule deer limiting factors. The greatest concern of agencies and sportsmen was whether declining habitat quality or predation was ultimately responsible for the observed declines. In Colorado, the Uncompahgre Plateau mule deer population received the most attention, having substantially declined from the 1980s through the late 1990s. Biologists hypothesized that poor winter range habitat quality was the primary cause of the observed decline. In contrast, many of the Colorado Division of Wildlife’s (CDOW) external constituents hypothesized that high rates of predation were keeping the mule deer herd below nutritional carrying capacity. The habitat quality hypothesis indicated CDOW should pursue habitat improvements in the pinyon (Pinus edulis) and juniper (Juniperus osteosperma) winter range, whereas the predator hypothesis suggested CDOW should pursue efforts to reduce predator populations, particularly coyote (Canis latrans) populations. The competing hypotheses represented very different paradigms of population limitation. I therefore evaluated the effect of enhanced nutrition during winter on the Uncompahgre deer population as a way to evaluate the importance of habitat quality versus that of predation. I conducted a field study incorporating a crossover experimental design to quantify the effect of enhanced nutrition on fetal, neonatal, overwinter fawn, and annual adult doe survival rates. I captured and radio-collared samples of deer in 2 experimental units (EUs) on winter range. I delivered the nutrition treatment to deer occupying one EU (treatment) and did not administer the treatment to deer in the other EU (control). Established field techniques were not sufficient to allow me to quantify the effect of the treatment on fetal and neonatal survival. I therefore pursued an exploration of vaginal implant transmitters as a mechanism to capture necessary samples of newborn fawns on summer range exclusively from radio-collared does that occupied the winter range EUs (Chapter 1). This effort allowed me to estimate fetal and neonatal survival as a function of the treatment. In broad terms, I demonstrated that direct estimates of fetal and neonatal survival may be obtained from previously marked female mule deer in free-ranging populations, thus expanding opportunities for conducting field experiments. I encountered additional challenges with estimation of fetal and neonatal survival. First, I was unable to determine the fate of all fetuses that I documented in utero. I therefore developed a likelihood function for estimating fetal survival when the fates of some fetuses are unknown (Chapter 2). Second, a majority of my fetal and neonatal samples were comprised of siblings, indicating my data were potentially overdispersed. Overdispersion causes sample variances to be underestimated and requires a variance inflation factor, c. To estimate c, I compared theoretical variance estimates with empirical variance estimates obtained from data bootstrap analyses (Chapter 2). I found little evidence of overdispersion in my fetal survival data, and I found modest overdispersion in my neonatal sample data (ĉ = 1.25). Although some overdispersion was detected, my result indicated that fates of sibling mule deer neonates may often be independent even though they have the same dam and use the environment similarly. I discuss reasons for this in Chapter 2. 70 �After resolving issues with fetal and neonatal survival estimation, I quantified the effect of the nutrition enhancement treatment on fetal, neonatal, overwinter fawn, and annual adult doe survival (Chapter 3). I then used these parameter estimates, along with estimated fecundity rates, in an agestructured, deterministic population model to estimate the effect of the treatment on the population rate of change, λ. The treatment caused λ̂ to increase by an average of 0.133 (SD = 0.0168) during the 3 years of my study. I documented density dependence in the Uncompahgre deer population because survival of fawns and does increased considerably in response to enhanced nutrition. I found strong evidence that coyote predation of ≥6-month-old fawns and adult does was compensatory. Finally, I found that winter range habitat quality was a limiting factor of the Uncompahgre Plateau deer population. I completed my principal study objectives in the first 3 chapters of the dissertation. However, my research afforded the opportunity to evaluate the utility of serum thyroid hormones in mule deer as an index to body condition (Chapter 4). Concentrations of total thyroxine (T4) and free T4 (FT4) were substantially higher in treatment deer than control deer. I also found that serum thyroid hormones were highly correlated with estimated body fat in mule deer during late winter. Concentrations of T4 and FT4 could be useful for evaluating relative condition of different deer groups or populations, and for roughly estimating body fat of individual animals during late winter. n summary, I demonstrated that winter range habitat quality was ultimately limiting the Uncompahgre mule deer population. Observed predation was primarily compensatory, particularly of ≥6month-old fawns and adult does. My findings indicate that CDOW should implement and evaluate habitat treatments in late-seral pinyon-juniper habitat as a means to increase habitat productivity for mule deer. My findings provide no support for predator reduction programs. 71 �Colorado Division of Wildlife July 2007 June 2008 WILDLIFE RESEARCH REPORT State of: Cost Center: Work Package: Task No.: Colorado 3430 3001 4 Federal Aid Project No. W-185-R Period Covered: July 1, 2007 : : : : : Division of Wildlife Mammals Research Deer Conservation Effect of Nutrition and Habitat Enhancements On Mule Deer Recruitment and Survival Rates June 30, 2008 Authors: C. J. Bishop, G. C. White, D. J. Freddy, and B. E. Watkins All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT We measured mule deer (Odocoileus hemionus) population parameters in response to a nutrition enhancement treatment to evaluate the relative importance of habitat quality as a limiting factor of mule deer in western Colorado during November 2000 – January 2005. The nutrition enhancement treatment increased survival of fetuses to the yearling age class by 0.14 0.20 depending on year and fawn sex; 95% confidence intervals slightly overlapped 0. Averaged across sexes and years, survival of treatment fetuses to the yearling age class was 0.447 (SE = 0.0519), whereas survival of control fetuses to the yearling age class was 0.271 (SE = 0.0418). The treatment caused fetal to yearling survival to increase by 0.177 (SE = 0.0818, 95% CI: 0.0163, 0.3370). The nutrition treatment also had a positive effect on annual adult female survival. Survival of adult females receiving the treatment (Ŝ = 0.879, SE = 0.0206) was higher than survival of control adult females (Ŝ = 0.833, SE = 0.0253). Our estimate of the population rate of change, ˆ , was 1.165 (SE = 0.0358) for treatment deer and 1.033 (SE = 0.0380) for control deer. The r nutrition treatment caused ˆ to increase by 0.133 (SE = 0.0428). We documented food limitation in the Uncompahgre deer population because survival of fawns and adult females increased considerably in response to enhanced nutrition. Our results provide a foundation for focusing deer management efforts on improving habitat quality in western Colorado pinyon-juniper (Pinus edulis-Juniperus osteosperma) ecosystems with corresponding research efforts to quantify the effects of habitat manipulations on deer performance. During 2007 08, we published one paper from this research in the Journal of Wildlife Management (JWM 72(5):1085 1093), we had another paper accepted for publication in Journal of Wildlife Management, and we had one paper accepted for publication in Wildlife Monographs pending suitable revision. The lead principal investigator published a Dissertation to complete requirements for a Ph.D. at Colorado State University. We previously published a manuscript on the effectiveness of vaginal implant transmitters (VITs) for capturing newborn fawns from specific adult females (Bishop et al. 2007). As a follow-up to this component of our research, we worked with Advanced Telemetry Systems (ATS, Isanti, MN) to develop a VIT with modified retention wings. The modified VIT will be ready for fieldtesting in 2009. r 39 �WILDLIFE RESEARCH REPORT EFFECT OF NUTRITION AND HABITAT ENHANCEMENTS ON MULE DEER RECRUITMENT AND SURVIVAL RATES CHAD J. BISHOP, GARY C. WHITE, DAVID J. FREDDY, AND BRUCE E. WATKINS P. N. OBJECTIVE To determine experimentally whether enhancing mule deer nutrition during winter and early spring via supplementation increases fetal survival, neonatal survival, overwinter fawn survival, or ultimately, population productivity. SEGMENT OBJECTIVES 1. Publish manuscripts in peer-reviewed scientific journals. 2. Publish dissertation as part of Ph.D. requirements at Colorado State University INTRODUCTION Mule deer (Odocoileus hemionus) numbers apparently declined during the 1990s throughout much of the West, and have clearly decreased since the peak population levels documented during the 1940s- 1960s (Unsworth et al. 1999, Gill et al. 2001). Biologists and sportsmen alike have concerns as to what factors may be responsible for declining population trends. Although previous and current research indicates multiple interacting factors are responsible, habitat and predation have typically received the focus of attention. A number of studies have evaluated whether predator control increases deer survival, yet results are highly variable (Connolly 1981, Ballard et al. 2001). Together, predator control studies with adequate rigor and statistical power indicate predation effects on mule deer are variable as a result of time-specific and site-specific factors. Studies which have demonstrated deer population responses to predator control treatments have failed to determine whether predation is ultimately more limiting than habitat when considering long term population changes. Numerous research studies have evaluated mule deer habitat quality, but virtually no studies have documented population responses to habitat improvements. In many areas where declining deer numbers are of concern, predation is common yet habitat quality appears to have declined. The question remains as to whether predation, habitat, or some other factor is more limiting to mule deer in these situations, and whether habitat quality can be improved for the benefit of deer. It may also be that no single factor is responsible for observed deer declines, and a more comprehensive understanding of multi-factor interactions is needed. We designed and implemented a field experiment where we measured deer population responses to a nutrition enhancement treatment to further understand the causative factors underlying observed deer population dynamics. We conducted the study on the Uncompahgre Plateau in southwest Colorado, where several predator species were present in abundant numbers: coyotes (Canis latrans), mountain lions (Felis concolor), and bears (Ursus americanus). In addition to predation, myriad diseases in combination have proximately affected survival of the Uncompahgre deer population (Pojar and Bowden 2004, B. E. Watkins, Colorado Division of Wildlife, unpublished data). Predator numbers were not manipulated in any manner during the course of the study. All factors were left constant with the exception of deer nutrition. Deer nutrition was enhanced by providing supplemental feed to deer occupying a treatment area during winter. We measured December fawn recruitment and overwinter fawn survival in response to the treatment to determine whether deer nutrition was ultimately more limiting 40 �than predation or disease. A second phase of research was initiated in 2005 to quantify deer population parameters in response to manipulations of pinyon-juniper (Pinus edulis-Juniperus osteosperma) habitat (Bergman et al. 2007). The objective of this research is to determine whether habitat can be effectively improved for mule deer by introducing disturbance into late-seral pinyon-juniper stands. STUDY AREA We non-randomly selected two experimental units (A B) within mule deer winter range on the Uncompahgre Plateau (Figure 1) to facilitate a cross-over experimental design for evaluating the effects of enhanced deer nutrition during winter on annual population performance. Unit A received a nutrition enhancement treatment during the first 2 winters of research (2000 – 2002) while Unit B served as a control unit. During winters 2002 03 and 2003 04, Unit B received the treatment while Unit A served as the control. In late April and May, prior to fawning, deer from the winter range experimental units migrated to summer range. We defined the summer range study area by movements of the radio-collared deer captured on winter range; summer range encompassed >1000 mi2 covering the southern portion of the Uncompahgre Plateau and adjacent San Juan Mountains (Figure 2). Winter range elevations ranged from 1830 m (6000 ft) in Shavano Valley to 2318 m (7600 ft) adjacent to the Dry Creek Rim above Shavano Valley. Winter range habitat was dominated by pinyon-juniper with interspersed sagebrush adjacent to agricultural fields in the Shavano and Uncompahgre Valleys. Summer range elevations occupied by deer ranged from 1891 m (6200 ft) in the Uncompahgre Valley to 3538 m (11,600 ft) in Imogene Basin southwest of Ouray, CO. Summer range habitats were dominated by spruce-subalpine fir (Picea spp.-Abies lasiocarpa), aspen (Populus tremuloides), sagebrush, ponderosa pine (Pinus ponderosa), Gambel oak (Quercus gambelii), and to a lesser extent, pinyon-juniper at lower elevations. Bishop et al. (2005) provide a detailed study area description. METHODS Refer to Bishop et al. (2005) or Bishop (2007) for field methodology employed during 2000 2005. During fiscal year 2007 08, we had 1 paper published and 2 papers accepted for publication in peer-reviewed scientific journals. Thus, our primary research efforts were focused on preparation of manuscripts for publication. We completed and published a paper in Journal of Wildlife Management focused on mule deer sibling dependence in context of fetal and neonatal survival analyses. In this paper, we also presented a likelihood function for estimating fetal survival when the fates of some fetuses are unknown. We spent much of the year preparing and submitting a manuscript to Wildlife Monographs. This particular publication documents the effect of enhanced nutrition on all aspects of mule deer productivity, survival, and population rate of change. Finally, we prepared and submitted a manuscript documenting the utility of serum thyroid hormone concentrations for evaluating mule deer body condition in late winter with this manuscript accepted for publication following two substantive revisions. The principal investigator also published his Ph.D. dissertation. A component of this project was an evaluation of vaginal implant transmitters (VITs) as a tool for locating neonatal mule deer fawns from targeted adult females (Bishop et al. 2007). To build on this research, we worked with Advanced Telemetry Systems (ATS, Isanti, MN) to develop a VIT with modified retention wings during 2007 08. We intend to evaluate the modified VIT in conjunction with ongoing mule deer energy development research in northwest Colorado. RESULTS AND DISCUSSION A comprehensive presentation and discussion of all results from this study is provided by Bishop (2007) and is not repeated here. These results and conclusions are being systematically published in peer- 41 �reviewed journals. The following manuscripts were published in 2007 and 2008 (abstracts are provided in Appendix I): Bishop, C. J. 2007. Effect of enhanced nutrition during winter on the Uncompahgre Plateau mule deer population. Dissertation, Colorado State University, Fort Collins, USA. Bishop, C. J., D. J. Freddy, G. C. White, B. E. Watkins, T. R. Stephenson, and L. L. Wolfe. 2007. Using vaginal implant transmitters to aid in capture of mule deer neonates. Journal of Wildlife Management 71:945 954. Bishop, C. J., G. C. White, and P. M. Lukacs. 2008. Evaluating dependence among mule deer siblings in fetal and neonatal survival analyses. Journal of Wildlife Management 72:1085 1093. Schultheiss, P. C., H. Van Campen, T. R. Spraker, C. J. Bishop, L. L. Wolfe, and B. Podell. 2007. Malignant catarrhal fever associated with ovine herpesvirus-2 in free-ranging mule deer in Colorado. Journal of Wildlife Diseases 43:533 537. The following manuscripts were accepted for publication in 2008 and will most likely be published in 2009 (abstracts are provided in Appendix II): Bishop, C. J., B. E. Watkins, L. L. Wolfe, D. J. Freddy, and G. C. White. 2009. Evaluating mule deer body condition using serum thyroid hormone concentrations. Journal of Wildlife Management: In press. Bishop, C. J., G. C. White, D. J. Freddy, B. E. Watkins, and T. R. Stephenson. 2009. Effect of enhanced nutrition on mule deer population rate of change. Wildlife Monographs: in review. (Manuscript has been tentatively accepted pending suitable revision). We intend to pursue several additional manuscripts as time allows, listed below in order of priority. 1. Evaluating dependence of fates among mule deer siblings in Colorado, Idaho, and Montana. Journal of Wildlife Management. 2. Bovine viral diarrhea isolation and seroprevalence in a free-ranging mule deer (Odocoileus hemionus) population in southwest Colorado. Journal of Wildlife Diseases. 3. Spatial patterns in mortality causes of neonatal mule deer across a land use gradient in southwest Colorado. Journal of Wildlife Management. 4. Evaluation of mule deer age and sex ratios as a response variable in field research. Journal of Wildlife Management. SUMMARY Enhanced winter nutrition of free-ranging deer caused an increase in both fetus-neonate survival and overwinter fawn survival, resulting in higher yearling recruitment. Overwinter adult female survival increased as a result of the nutrition treatment, and therefore annual survival was higher among treatment than control adult females. Combining all parameter estimates into a deterministic population model, the treatment population indicated an exceptionally high rate of increase while the control population was stable and indicative of the overall Uncompahgre deer population during 2000 2004. The nutrition enhancement treatment was artificial in the sense that we applied it only to test whether habitat quality was ultimately more limiting than predation or other factors. Our results to do not provide support for managing deer populations with nutrition supplements because our treatment delivery approach could not 42 �be applied to a large number of animals over a large area. Rather, our results provide a foundation for focusing deer management efforts on improving habitat quality in western Colorado pinyon-juniper ecosystems with corresponding research efforts to quantify the effects of habitat manipulations on deer. LITERATURE CITED Ballard, W. B., D. Lutz, T. W. Keegan, L. H. Carpenter, and J. C. DeVos, Jr. 2001. Deer-predator relationships: a review of recent North American studies with emphasis on mule and black-tailed deer. Wildlife Society Bulletin 29:99 115. Bergman, E. J., C. J. Bishop, D. J. Freddy, and G. C. White. 2007. Evaluation of winter range habitat treatments on over-winter survival and body condition of mule deer. Wildlife Research Report July: 73-96. Colorado Division of Wildlife, Fort Collins, USA. Bishop, C. J. 2007. Effect of enhanced nutrition during winter on the Uncompahgre Plateau mule deer population. Dissertation, Colorado State University, Fort Collins, USA. Bishop, C. J., D. J. Freddy, G. C. White, B. E. Watkins, T. R. Stephenson, and L. L. Wolfe. 2007. Using vaginal implant transmitters to aid in capture of mule deer neonates. Journal of Wildlife Management 71:945 954. Bishop, C. J., G. C. White, D. J. Freddy, and B. E. Watkins. 2005. Effect of nutrition and habitat enhancements on mule deer recruitment and survival rates. Wildlife Research Report July: 3766. Colorado Division of Wildlife, Fort Collins, USA. Connolly, G. E. 1981. Limiting factors and population regulation. Pages 245 285 in O. C. Wallmo, editor. Mule and black-tailed deer of North America. University of Nebraska Press, Lincoln, USA. Gill, R. B., T. D. I. Beck, C. J. Bishop, D. J. Freddy, N. T. Hobbs, R. H. Kahn, M. W. Miller, T. M. Pojar, And G. C. White. 2001. Declining mule deer populations in Colorado: reasons and responses. Colorado Division of Wildlife Special Report Number 77. Fort Collins, USA. Pojar, T. M., And D. C. Bowden. 2004. Neonatal mule deer fawn survival in west-central Colorado. Journal of Wildlife Management 68:550 560. Unsworth, J. W., D. F. Pac, G. C. White, And R. M. Bartmann. 1999. Mule deer survival in Colorado, Idaho, and Montana. Journal of Wildlife Management 63:315 326. Prepared by _______________________ Chad J. Bishop, Wildlife Researcher 43 �Year Unit A Unit B 2000-01 Treatment Control 2001-02 Treatment Control 2002-03 Control Treatment 2003-04 Control Treatment Figure 1. Schematic representation of experimental units and nutrition enhancement treatment allocation. Units A and B were located in winter range habitat on the Uncompahgre Plateau in southwest Colorado. The nutrition enhancement cross-over design encompassed 4 years. Uncompahgre Plateau Mesa County GRAND JUNCTION Delta County U nc GMU 62 Gunnison County DELTA om Winter Range Exp. Units pa hg re ea at Pl GMU 61 u Montrose County Sanmiguel County Shavano E.U. MONTROSE Colona Montrose County Summer Range E.U. Ouray County Figure 2. Location of Colona and Shavano (Units A and B) experimental units on the Uncompahgre Plateau, southwest Colorado; and location of the summer range study area encompassing the southern Uncompahgre Plateau and adjacent San Juan Mountains. 44 �APPENDIX I – PUBLISHED PROJECT PAPERS The following Colorado State University dissertation (referenced here by Abstract) was published in 2007. EFFECT OF ENHANCED NUTRITION DURING WINTER ON THE UNCOMPAHGRE PLATEAU MULE DEER POPULATION CHAD J. BISHOP ABSTRACT Mule deer (Odocoileus hemionus) populations declined across much of the West during the 1990s, prompting state wildlife agencies to explore mule deer limiting factors. The greatest concern of agencies and sportsmen was whether declining habitat quality, predation, or both were responsible for the observed declines. In Colorado, the Uncompahgre Plateau mule deer population received the most attention because of a steep population decline from the 1980s through the late 1990s. Biologists hypothesized that poor quality of the pinyon (Pinus edulis) and juniper (Juniperus osteosperma) winter range was the primary cause of the observed decline. In contrast, many of the Colorado Division of Wildlife‘s (CDOW) constituents hypothesized that high predation rates were keeping the mule deer herd below nutritional carrying capacity. These hypotheses represented very different paradigms of population limitation. Perhaps more importantly, the competing views suggested that CDOW should pursue one of two very different management strategies: 1) implement habitat improvements in the pinyon-juniper winter range, or 2) implement efforts to reduce predator populations, particularly coyote (Canis latrans) populations. Information was needed to guide the decision process. I therefore evaluated the effect of enhanced nutrition during winter on the Uncompahgre deer population as a way to evaluate the importance of habitat quality versus that of predation. I conducted a field study incorporating a crossover experimental design to quantify the effect of enhanced nutrition on fetal, neonatal, overwinter fawn, and annual adult doe survival rates. I captured and radio-collared samples of deer in 2 experimental units (EUs) on winter range. I delivered the nutrition treatment to deer occupying one EU (treatment) and did not administer the treatment to deer in the other EU (control). Established field techniques were not sufficient to allow me to quantify the effect of the treatment on fetal and neonatal survival. I therefore pursued an exploration of vaginal implant transmitters as a mechanism to capture necessary samples of newborn fawns on summer range exclusively from radio-collared does that occupied the winter range EUs (Chapter 1). This effort allowed me to estimate fetal and neonatal survival as a function of the treatment. In broad terms, I demonstrated that direct estimates of fetal and neonatal survival may be obtained from previously marked female mule deer in free-ranging populations, thus expanding opportunities for conducting field experiments. I encountered additional challenges with estimation of fetal and neonatal survival. First, I was unable to determine the fate of all fetuses that I documented in utero. I therefore developed a likelihood function for estimating fetal survival when the fates of some fetuses are unknown (Chapter 2). Second, a majority of my fetal and neonatal samples were comprised of siblings, indicating my data were potentially overdispersed. Overdispersion causes sample variances to be underestimated and requires a variance inflation factor, c. To estimate c, I compared theoretical variance estimates with empirical variance estimates obtained from bootstrap analyses of the data (Chapter 2). I found little evidence of overdispersion in my fetal survival data, and I found modest overdispersion in my neonatal sample data (ĉ = 1.25). Although some overdispersion was detected, my results indicated that fates of sibling mule deer neonates may often be independent even though they have the same dam and use the environment similarly. I discuss reasons for this in Chapter 2. 45 �After resolving issues with fetal and neonatal survival estimation, I quantified the effect of the nutrition enhancement treatment on fetal, neonatal, overwinter fawn, and annual adult doe survival (Chapter 3). I then used these parameter estimates, along with estimated fecundity rates, in an agestructured, deterministic population model to estimate the effect of the treatment on the population rate of r change, ˆ . The treatment caused ˆ to increase by an average of 0.133 (SD = 0.0168) during the 3 years of my study. I documented density dependence in the Uncompahgre deer population because survival of fawns and does increased considerably in response to enhanced nutrition. I found strong evidence that coyote predation of ≥6-month-old fawns and adult does was compensatory. Finally, I found that winter range habitat quality was a limiting factor of the Uncompahgre Plateau deer population. r I completed my principal study objectives in the first 3 chapters of the dissertation. However, my research afforded the opportunity to evaluate the utility of serum thyroid hormones in mule deer as an index to body condition (Chapter 4). Concentrations of total thyroxine (T4) and free T4 (FT4) were substantially higher in treatment deer than control deer. I also found that serum thyroid hormones were highly correlated with estimated body fat in mule deer during late winter. Concentrations of T4 and FT4 could be useful for evaluating relative condition of different deer groups or populations, and for roughly estimating body fat of individual animals during late winter. In summary, I demonstrated that winter range habitat quality was ultimately limiting the Uncompahgre mule deer population. Observed predation was primarily compensatory, particularly of ≥6month-old fawns and adult does. My findings indicate that CDOW should evaluate habitat treatments in late-seral pinyon-juniper habitat as a means to increase habitat productivity for mule deer. Citation: Bishop, C. J. 2007. Effect of enhanced nutrition during winter on the Uncompahgre Plateau mule deer population. Dissertation, Colorado State University, Fort Collins, USA. 46 �The following manuscript (referenced here by Abstract) was published in the Journal of Wildlife Management in 2007. USING VAGINAL IMPLANT TRANSMITTERS TO AID IN CAPTURE OF MULE DEER NEONATES CHAD J. BISHOP, DAVID J. FREDDY, GARY C. WHITE, BRUCE E. WATKINS, THOMAS R. STEPHENSON, AND LISA L. WOLFE ABSTRACT Estimating survival of the offspring of marked female ungulates has proven difficult in freeranging populations yet could improve our understanding of factors that limit populations. We evaluated the feasibility and efficiency of capturing large samples (i.e., >80/year) of neonate mule deer (Odocoileus hemionus) exclusively from free-ranging, marked adult does using vaginal implant transmitters (VITs, n = 154) and repeated locations of radio-collared does without VITs. We also evaluated the effectiveness of VITs, when used in conjunction with in utero fetal counts, for obtaining direct estimates of fetal survival. During 2003 and 2004, after we placed VIT batteries on a 12-hour duty cycle to lower electronic failure rates, the proportion that shed 3 days prepartum or during parturition was 0.623 (SE = 0.0456), and the proportion of VITs shed only during parturition was 0.447 (SE = 0.0468). Our neonate capture success rate was 0.880 (SE = 0.0359) from does with VITs shed 3 days prepartum or during parturition and 0.307 (SE = 0.0235) from radio-collared does without VITs or whose implants failed to function properly. Using a combination of techniques, we captured 275 neonates and found 21 stillborns during 2002 2004. We accounted for all fetuses at birth (i.e., live or stillborn) from 78 of the 147 does (0.531, SE = 0.0413) having winter fetal counts, and this rate was heavily dependent on VIT retention success. Deer that shed VITs prepartum were larger than deer that retained VITs to parturition, indicating a need to develop variable-sized VITs that may be fitted individually to deer in the field. We demonstrated that direct estimates of fetal and neonatal survival may be obtained from previously marked female mule deer in free-ranging populations, thus expanding opportunities for conducting field experiments. Survival estimates using VITs lacked bias that is typically associated with other neonate capture techniques. However, current vaginal implant failure rates, and overall expense, limit broad applicability of the technique. Citation: Bishop, C. J., D. J. Freddy, G. C. White, B. E. Watkins, T. R. Stephenson, and L. L. Wolfe. 2007. Using vaginal implant transmitters to aid in capture of mule deer neonates. Journal of Wildlife Management 71:945 954. 47 �The following manuscript (referenced here by Abstract) was published in the Journal of Wildlife Management in 2008. EVALUATING DEPENDENCE AMONG MULE DEER SIBLINGS IN FETAL AND NEONATAL SURVIVAL ANALYSES CHAD J. BISHOP, GARY C. WHITE, AND PAUL M. LUKACS ABSTRACT The assumption of independent sample units is potentially violated in survival analyses where siblings comprise a high proportion of the sample. Violation of the independence assumption causes sample data to be overdispersed relative to a binomial model, which leads to underestimates of sampling variances. A variance inflation factor, c, is therefore required to obtain appropriate estimates of variances. We evaluated overdispersion in fetal and neonatal mule deer (Odocoileus hemionus) datasets where more than half of the sample units were comprised of siblings. We developed a likelihood function for estimating fetal survival when the fates of some fetuses are unknown, and we used several variations of the binomial model to estimate neonatal survival. We compared theoretical variance estimates obtained from these analyses with empirical variance estimates obtained from data bootstrap analyses to estimate the overdisperion parameter, c. Our estimates of c for fetal survival ranged from 0.678 to 1.118, which indicate little to no evidence of overdispersion. For neonatal survival, 3 different models indicated that ĉ ranged from 1.1 to 1.4 and averaged 1.24 1.26, providing evidence of limited overdispersion (i.e., limited sibling dependence). Our results indicate that fates of sibling mule deer fetuses and neonates may often be independent even though they have the same dam. Predation tends to act independently on sibling neonates because of dam-neonate behavioral adaptations. The effect of maternal characteristics on sibling fate dependence is less straightforward and may vary by circumstance. We recommend that future neonatal survival studies incorporate additional sampling intensity to accommodate modest overdispersion (i.e., ĉ = 1.25), which would facilitate a corresponding ĉ adjustment in a model selection analysis using quasi-likelihood without a reduction in power. Our computational approach could be used to evaluate sample unit dependence in other studies where fates of individually marked siblings are monitored. Citation: Bishop, C. J., G. C. White, and P. M. Lukacs. 2008. Evaluating dependence among mule deer siblings in fetal and neonatal survival analyses. Journal of Wildlife Management 72:1085 1093. 48 �The following manuscript (referenced here by Abstract) was published in the Journal of Wildlife Diseases in 2007: MALIGNANT CATARRHAL FEVER ASSOCIATED WITH OVINE HERPESVIRUS-2 IN FREE-RANGING MULE DEER IN COLORADO PATRICIA C. SCHULTHEISS, HANA VAN CAMPEN, TERRY R. SPRAKER, CHAD J. BISHOP, LISA L. WOLFE, AND BRENDAN PODELL ABSTRACT Malignant catarrhal fever (MCF) was diagnosed in 4 free-ranging mule deer (Odocoileus hemionus) in January and February of 2003. Diagnosis was based on typical histologic lesions of lymphocytic vasculitis and PCR identification of ovine herpesvirus-2 (OHV-2) viral genetic sequences in formalin fixed tissues. The animals were from the Uncompahgre Plateau of southwestern Colorado. Deer from these herds occasionally resided in close proximity to domestic sheep (Ovis aries), the reservoir host of OHV-2, in agricultural valleys adjacent to their winter range. These cases indicate that fatal OHV-2 associated MCF can occur in free-ranging mule deer exposed to domestic sheep that overlap their range. Citation: Schultheiss, P. C., H. Van Campen, T. R. Spraker, C. J. Bishop, L. L. Wolfe, and B. Podell. 2007. Malignant catarrhal fever associated with ovine herpesvirus-2 in free-ranging mule deer in Colorado. Journal of Wildlife Diseases 43:533 537. 49 �APPENDIX II PROJECT PAPERS ACCEPTED FOR PUBLICATION The following manuscript (referenced here by Abstract) was accepted for publication by the Journal of Wildlife Management during 2008 but has not yet been published. EVALUATING MULE DEER BODY CONDITION USING SERUM THYROID HORMONE CONCENTRATIONS CHAD J. BISHOP, BRUCE E. WATKINS, LISA L. WOLFE, D. J. FREDDY, AND GARY C. WHITE ABSTRACT Body condition of ungulates is a determinant of fecundity and survival rates. Ultrasonography and body condition scoring techniques allow reliable estimation of body fat but may not be feasible to employ in some circumstances. A reliable blood chemistry index for assessing relative condition of different ungulate populations or groups would be useful in ongoing population monitoring programs. We provided a nutrition supplement (treatment) to a group of free-ranging mule deer (Odocoileus hemionus) during 2 consecutive winters in southwest Colorado. In late February each year, we evaluated whether percent body fat and serum concentrations of total thyroxine (T4), total triiodothyronine (T3), free T4 (FT4), and free T3 (FT3) were higher among treatment deer than an adjacent group of deer that did not receive the treatment (control). As a corroborative analysis, we modeled body fat as a function of thyroid hormone concentrations and morphometric variables. Estimated body fat of treatment deer averaged 12.3% (SE = 0.327), whereas estimated body fat of control deer averaged 7.0% (SE = 0.333), during the 2 winters of study. Concentrations of T4 and FT4 averaged 48.07 nmol/l (SE = 3.80) and 12.61 pmol/l (SE = 1.04) higher, respectively, in treatment deer than control deer. Our optimal model of estimated body fat included T4, T42, FT4, and deer chest girth (%Fât = –4.8015 – 0.0946X T4 + 0.000603X T42 + 0.1474 X FT4 + 0.1426 X chest girth, R2 = 0.609). Serum thyroid hormones effectively discerned treatment deer from control deer and were related to estimated body fat. Ultrasound and body condition scoring should be used to estimate body fat whenever possible. However, in cases where only a blood sample can be obtained, we documented potential utility of T4 and FT4 during late winter for evaluating relative body condition of mule deer. 50 �The following manuscript (referenced here by Abstract) was tentatively accepted for publication by Wildlife Monographs during 2008 and is still in the revision stage. EFFECT OF ENHANCED NUTRITION ON MULE DEER POPULATION RATE OF CHANGE CHAD J. BISHOP, GARY C. WHITE, DAVID J. FREDDY, BRUCE E. WATKINS, AND THOMAS R. STEPHENSON ABSTRACT Concerns over declining mule deer (Odocoileus hemionus) populations during the 1990s prompted research efforts to identify and understand key limiting factors of deer. Similar to past deer decline incidents, a top priority of state wildlife agencies was to evaluate the relative importance of habitat and predation. We therefore evaluated the effect of enhanced nutrition of deer during winter and spring on fecundity and survival rates using a life table response experiment involving free-ranging mule deer on the Uncompahgre Plateau in southwest Colorado. The nutrition enhancement treatment represented an instantaneous increase in nutritional carrying capacity of a pinyon (Pinus edulis) and Utah juniper (Juniperus osteosperma) winter range, and was intended to simulate optimum habitat quality. Prior studies on the Uncompahgre Plateau indicated predation and disease were the most common proximate causes of deer mortality. By manipulating nutrition and leaving natural predation unaltered, we determined whether habitat quality was ultimately a critical factor limiting the deer population. We measured fetal, neonatal, and overwinter fawn survival, and annual adult female survival, which we then used to estimate population rate of change as a function of enhanced nutrition. Pregnancy and fetal rates were high for all deer, regardless of the nutrition treatment. Fetal and neonatal survival rates were higher among deer that received the nutrition enhancement treatment than deer that served as experimental controls. Overwinter fawn survival increased for treatment deer by 0.16 0.31 depending on year and fawn sex, and none of the 95% confidence intervals associated with the effect overlapped 0. Nutrition enhancement increased survival of fetuses to the yearling age class by 0.14 0.20 depending on year and fawn sex, although 95% confidence intervals slightly overlapped 0. Annual survival of adult females receiving the treatment (Ŝ = 0.879, SE = 0.0206) was higher than survival of control adult females (Ŝ = 0.833, SE = 0.0253). Our estimate of the population rate of change, ˆ , was 1.165 (SE = 0.0358) for r treatment deer and 1.033 (SE = 0.0380) for control deer. The nutrition treatment caused ˆ to increase by 0.133 (SE = 0.0428). We documented density dependence in the Uncompahgre deer population because survival of fawns and adult females increased considerably in response to enhanced nutrition. We found strong evidence that coyote (Canis latrans) predation of ≥6-month-old fawns and adult females was compensatory. Our results demonstrate that observed coyote predation, by itself, is not useful for evaluating whether coyotes are negatively impacting a deer population. We also found evidence that mountain lion (Puma concolor) predation was compensatory. Disease mortality was not compensatory among adult females. We found that winter range habitat quality was a limiting factor of the Uncompahgre Plateau mule deer population. Therefore, we recommend evaluating habitat treatments for deer that are designed to set-back succession and increase productivity of late-seral pinyon-juniper habitats that presently dominate the winter range. r 51 � https://cpw.cvlcollections.org/files/original/310a46686431af346629c8bed1afe33b.pdf b28821a288fc91e16e6fbe72992dbdb2 PDF Text Text 81 ! f Colorado Division of Wildlife Wildlife Research Report July 2001 and July 2002 FINAL REPORT I l. State of_______C-=-=-o=lo=r=ad=o=-------- Division of Wildlife - Mammals Research Work Package No . _ _--=3=.0-=-0.::....1_ _ _ _ __ Deer Conservation Task No. _ _ _ _ _ _ _4.c..__ _ _ _ __ Pilot Study- Use of Ultrasound and Vaginal Implants Federal Aid Project_ _W~-1~8~5~-R~----- Research and Development Period Covered: July 1, 2001-June 30, 2002 Authors: C. J. Bishop, D. J. Freddy, and G. C. White, Ph.D. i l. Personnel: D. L. Baker, T. Baker, R. Bavin, T. D. I. Beck, S. K. Carroll, D. Coven, K. Crane, M. Del Tonto, L. Gepfert, J. Grigg, M. McLain, G. C. Miller, M. W. Miller, J. Olterman, J. A. Padia, T. M. Pojar, J. Risher, C. M. Solohub, M. Thonhoff, B. E. Watkins, L. Wolfe, CDOW; T. R. Stephenson, California Dept. of Fish and Game; R. C. Cook, National Council for Air and Stream Improvement. ABSTRACT Field research of mule deer could be greatly enhanced if newborn fawns could be captured from specific adult females from which data has already been collected. We evaluated the logistical feasibility and effectiveness of using vaginal implant transmitters (VITs) to determine the location and timing of birth of specific, radio-collared adult female mule deer. The VITs were manufactured by Advanced Telemetry Systems, Inc. (Isanti, MN). We placed VITs in 36 adult female deer on February 28 and March 1, 2002. At this time, we recorded data such as fetal rate, body condition, body mass, and serology from a blood sample. In June 2002, we intensively radio-monitored the VITs to determine when they were expelled from the deer. When a VIT was shed, we immediately located it to determine whether a birth site was present and to locate/capture neonates. The proportion of VITs that were expelled at or near the birth site with the transmitter functioning correctly was 0.33 (SD= 0.083). Of 36 VIT trials, 3 were censored, 7 were shed prematurely, 15 had battery failures, and 11 were successful (9 of which led to birth sites and subsequent fawn captures). In spite of the high VIT failure rate, we captured and radio-collared a total of 54 fawns from 38 adult does during June 11 - July 1, 2002. Twenty-six fawns were captured from 17 of the 36 VIT does and 28 fawns were captured from 21 radio-collared does that did not have VITs. Contrary to our own expectations, we successfully captured fawns from radio-collared does by relocating the does on a routine basis. However, this technique was inefficient and required a total capture effort of 1700 man-hours (212 man-days) during a 22-day period, or roughly 4 man-days/fawn. The VIT battery failures were clearly the main problem we experienced, which can be corrected. The amount of time and effort saved by the 11 successful VITs justifies their use, particularly with continued refinement of the VIT design. BDOW016782 ��83 EFFECTS OF ENHANCED WINTER NUTRITION OF ADULT FEMALE MULE DEER ON FETAL AND NEONATAL SURVIVAL RATES: A PILOT STUDY TO ADDRESS FEASIBILITY C. J. Bishop, D. J. Freddy, and G. C. White PROJECT OBJECTIVE 1. To evaluate the feasibility of utilizing ultrasound techniques and vaginal implant transmitters in adult female mule deer to measure stillborn fetus mortality and to locate and capture specific neonate fawns. SEGMENT OBJECTIVES 1. • Prepare a ·Program Narrative for a I -year pilot study. 2. Conduct the I-year pilot study to evaluate logistical feasibility of field techniques, collect data necessary for subsequent sample size calculations, and to obtain preliminary biological data. 3. Prepare a Job Final Report for the I-year pilot study. INTRODUCTION Background The Colorado Division of Wildlife initiated 2 studies on the Uncompahgre Plateau in response to chronically low December fawn:doe ratios throughout the 1990's and an overall decline in total deer numbers. In 1997, an ongoing survival study was initiated to quantify overwinter fawn survival and annual adult survival rates, and to identify mortality agents (B.E. Watkins, Colorado Division of Wildlife, unpublished data). In 1999, another study began to quantify pregnancy/fetal rates and to measure survival rates and cause-specific mortality of neonate fawns (Pojar and Andelt 1999, Pojar 2000). These studies have provided several significant findings to date. First, overwi~ter survival rates of fawns, and annual survival rates of does and bucks, are above average when compared to measurements obtained elsewhere (Unsworth et al. 1999). Second, adult doe pregnancy rates (93%) and fetal rates (1.7 fawns/doe) in February are normal (Pojar and Andelt 1999). Third, summer fawn survival has been relatively low overall, with malnutrition/sickness and predation being the primary causes of mortality (Pojar 2000). Based on these findings, in utero fetus survival/summer fawn surviyal is clearly the limiting factor to population growth on the Plateau. Summer fawn survival has been measured directly, while the extent of in utero fetus mortality from February to birth has been back-calculated utilizing expected versus observed December fawn:doe ratios based on the observed summer fawn survival. Although this fetus mortality has not been measured directly, there is considerable evidence that some portion of viable fetuses in February are not surviving to birth. Given the magnitude ofmalnutrition/skkness observed in newborn fawns, the question of prepartum adult doe nutrition is paramount. Summer range habitat quality on the Uncompahgre Plateau is seemingly good, and arguably better than many other deer summer ranges throughout the intermountain West. However, lo'Yer transitional and winter range habitat quality appears to be limited in terms of forage diversity and quality. We understand the inherent limits in nutritional quality of winter range forage, but hypothesize that winter range habitat quality on the Plateau may not be meeting the minimum nutritional requirements of pregnant adult does. �84 In 2000, we initiated a research study on the Uncompahgre Plateau to evaluate the effects of enhanced nutrition of mule deer during winter on fawn production and survival (Bishop and White 2000, Bishop and White 2002). The objectives of this research are twofold: 1) to determine experimentally whether enhancing the nutrition of deer during winter and early spring by supplemental feeding increases overwinter fawn survival and/or December fawn:doe ratios the following December; and 2) to determine experimentally to what extent habitat treatments replicate the effect of enhanced nutrition from supplemental feeding. We are addressing these objectives by radio-collaring adult does and 6-month old fawns in a treatment experimental unit and a control experimental unit. The current phase of the experiment uses supplemental feed as a nutrition enhancement treatment, while the second phase will use habitat manipulations (e.g. prescribed fire, mechanical) as the treatment. The main focus of this research is to determine whether a decline in winter range habitat quality has been a causative factor of poor December fawn recruitment on the Uncompahgre Plateau during the past decade. More specifically, we are determining whether m~trition enhancements and/or habitat treatments on winter range cause an increase in fawn:doe ratios the following December. Our primary response variable is December fawn:doe ratios measured from radio-collared does in the treatment and control units. December fawn:doe ratios represent a combined approximation of fawn production (# fetuses produced and successfully brought to term) and survival (% of newborn fawns surviving from birth to December). Fawn:doe ratios are influenced by the number of yearling females in the population, the proportion of small yearling bucks that may be misidentified as does, doe harvests etc. Irrespective of these inherent biases, low December fawn:doe ratios typically indicate either poor fawn production, low summer fawn survival, or a combination of both. To improve our current study design evaluating the importance of habitat/nutrition, we initiated a I-year pilot study in an attempt to obtain separate, direct measurements of in utero fetus survival and summer fawn survival for treatment and control does. These direct measurements, if obtainable, would be preferable to December fawn:doe ratios, and would provide a better understanding of the effect, if any, of the nutrition enhancement treatment on the deer population. The purpose of the pilot study was to evaluate the logistical feasibility of field techniques necessary to accomplish the research, and to collect data for subsequent sample size calculations assuming the research progressed into a full-scale study. In order to directly measure in utero fetus survival and summer fawn survival of radio-collared does from the treatment and control areas, we needed to record winter fetal rates of the collared does, and then locate and capture the collared does' fawns the following June. Winter fetal rates can be measured using established ultrasound techniques. However, there are no established techniques for locating and capturing a large sample of newborn fawns from specific, individual does. Previous attempts to capture neonates from radio-collared does in forest-shrub habitats have been largely unsuccessful (M.A. Hurley, Idaho Dept. of Fish and Game, pers. comm.; T. M. Pojar, Colorado Division of Wildlife, pers. comm.). There are 2 major problems: 1) there is no effective way to determine when any given doe will give birth to her fawn(s); and 2) it is often very difficult to find a fawn simply based on locating the doe, because the fawns are often bedded in heavy cover some distance from the doe (e.g. 50-100 yards). To overcome these problems, we conducted a I-year research study to evaluate the use of vaginal implant transmitters in adult does as a technique to determine both the timing and location of birthing. Vaginal Implant Transmitters For some time, radio-transmitter implants in the vaginas of deer have been considered as a technique for. locating and capturing newborn fawns from radio-collared does immediately following parturition. Early attempts to employ this technique were largely unsuccessful in terms of both effectiveness and animal �85 welfare concerns (Garrott and Bartmann 1984, Giessman and Dalton 1984, Nelson 1984). This early technique used sutures to partially close the vulva in order to retain the transmitter in the vagina. More recently, Bowman and Jacobsen ( 1998) developed and employed a modified vaginal implant transmitter (VIT) for white-tailed deer, which met better success. This transmitter had plastic wings to retain the transmitter in the vagina until parturition; thus, no sutures were used. They found no indications that animals were negatively impacted by the newly designed VIT; however, retention rate of implants to parturition was only 75%, and sample sizes were small. Within the last 2 years, several studies have been initiated using (modified) VITs to study white-tailed deer (M. Carstenson and G. D. Delguidice, University of Minnesota, pers. comm.), black-tailed deer (N. Pamplin and D. Jackson, Oregon State University and Oregon Dept. offish and Wildlife, pers. comm.), and elk (J. Noyes and B. Johnson, Oregon Dept. offish and Wildlife, pers. comm.; J. Vore, Montana Fish, Wildlife, and Parks, pers. comm.) These ongoing studies have found greater success with VITs in terms ofretention to parturition, and have not documented any detrimental effects to the animals. Given the success at finding birth sites and fawns, these studies do not indicate that vaginal implants cause major problems with in utero fetus survival or birthing. MATERIALS AND METHODS Experimental Design and Study Area This research was conducted in conjunction with our ongoing research study evaluating the effects of enhanced winter nutrition on overwinter mule deer fawn survival and early winter fawn:doe ratios. The research is being conducted in 2 experimental units on winter range on the Uncompahgre Plateau. The 2 units are receiving a nutrition enhancement treatment in a cross-over experimental design. Unit A served as the treatment unit, and Unit B served as the control, for the first 2 years of research (2000 - 2002). Beginning November 2002, Unit B will receive the treatment while Unit A will serve as the control. Bishop and White (2002) provide a complete description of the experimental design and study area. The 2 experimental units (A and B) receiving the nutrition enhancement treatment are (Fig. 1): (1) The Colona Tract of the Billy Creek State Wildlife Area (2) Bureau of Land Management lands adjacent to Shavano Valley as defined by the following: Within Dry Creek Basin Quadrangle (USGS 7.5 Minute), includes Sections 6 and 7 in T. 48 N.-R. 10 W. and Sections 1, 2, 10, 11, 12, 13, 14, 15 in T. 48 N.-R. 11 W. This area roughly includes 38°25'00" 38°27'30" Latitude and 108°00'00" - 108°04'30" Longitude. In late April and May, prior to fawning, deer from the winter range experimental units migrate to summer range. The summer range study area encompasses 800 mi2 covering the southern portion of the Uncompahgre Plateau and adjacent San Juan Mountains to the south and east (Fig. 1). Winter range elevations range from 1830 m (6000 ft) in Shavano Valley to 2318 m (7600 ft) adjacent to the Dry Creek Rim above Shavano Valley. Winter range habitat is dominated by pinyon-juniper with interspersed sagebrush adjacent to agricultural fields in the Shavano and Uncompahgre Valleys. Summer range elevations occupied by deer range from 1891 m (6200 ft) in the Uncompahgre Valley to 3538 m (11,600 ft) in Imogene Basin southwest of Ouray, CO. Summer range habitats are dominated by sprucefir, aspen, ponderosa pine, Gamb~l oak, and to a lesser extent, sagebrush and pinyon-juniper at lower elevations. �86 Sample Size The main objective of the pilot study was to assess "success/failure" ofVITs from an equipment functionality and field logistics standpoint. Our first definition of success/failure was the proportion of vaginal implant transmitters that were expelled at the birth site with the heat sensor functioning correctly (transmitter success rate). We set our sample size based on an initial assumption that 0.90 of the transmitters would function correctly, which necessitated 36 radio-collared does to estimate the transmitter success rate with a 95% confidence interval of+/- 0.10. Our second definition of success/failure was the proportion of successful fawn captures that occurred from the sample of 36 does (fawn capture success rate). We defined a successful fawn capture as locating and capturing at least 1 fawn from a doe equipped with a vaginal implant. Prior to the study, we assumed a 0.80 rate of finding at least 1 fawn/radio-collared doe. With 36 radio-collared does, this would allow us to measure a fawn capture success rate with a 95% confidence interval of+/- 0.13. This level of precision was sufficient for us to evaluate the overall success/failure of vaginal implant transmitters as a field technique for locat_ing and capturing newborn fawns from specific radio-collared does. Captur.e and Handling Techniques On February 28 and March 1, 2002, we captured a total of36 adult female deer utilizing helicopter net guns (Barrett et al. 1982, van Reenen 1982). Eighteen deer were captured in the nutrition enhancement treatment unit, and 18 in the control unit. Captured deer were ferried by the helicopter to a central processing location. Most deer captured in each experimental unit were chemically immobilized using a combination of ketamine and xylazine to facilitate the ultrasound and VIT insertion procedures. Ketamine and xylazine were mixed in a 5: 1 ratio (200:40 mg/ml), and administered intravenously at a dosage rate of approximately 1.5-2.0 ml/45 kg animal body mass. Immediately prior to release, drug effects were (partially) reversed with an intravenous injection of yohimbine at a rate of-12 mg/45 kg animal body mass. Each deer was aged based on tooth replacement and wear, and only deer ~-5 years old were retained. For each captured deer, we used ultrasonography to measure pregnancy status, fetal rate, and body condition. If the doe was pregnant, she was then radio-collared using a fixed length, permanent collar. Each radio collar had Ritchey® neck band material stitched to the left side with a unique identifier engraved on it for visual identification purposes. We then inserted the VIT and released the deer. We performed the ultrasound and VIT insertion procedures in a 10 x 12 ft wall frame tent located at the processing site to avoid problems associated with weather conditions and helicopter rotor wash. We also recorded the weight of each deer, recorded a body condition score, and collected a blood sample for serology tests. Ultrasonography We estimated body fat using an Aloka 210 (Aloka, Inc., Wallinford, Conn.) portable ultrasound unit with a 5 MHz linear transducer. Maximum subcutaneous fat thickness on the rump was measured immediately cranial to the cranial process of the tuber ischium. Proper orientation was assured by scanning along a line between the spine, at its closest point to the tuber coxae (hip bone), and the caudal process of the tuber ischium (pin bone). A small area of hair was shaved to ensure contact between the transducer and the skin. A conducting gel was applied to the shaved area and fat thickness was measured using electronic calipers. We quantified reproductive status using a 3 MHz linear transducer. To permit transabdominal scanning, a portion of the abdomen was shaved caudal to the last rib and -left of the mid line, and gel was applied. Both uterine horns were systematically scanned to identify fetal numbers ranging from Oto 3. Upon identification of a fetus, we measured, whenever possible, eye diameter, crown-rump length, biparietal �87 40 0 40 80 Miles Figure I. Map showing the locations of the Colona and Shavano experimental units (Units A and B) on winter range, and the location of deer summer range, on the Uncompahgre Plateau, southwest Colorado. The summer range study area was defined by the summer distribution of deer that were captured on the winter range experimental units. �88 distance, and skull length. In most cases, we only obtained an eye diameter measurement. Morphometric measurements were collected to estimate fetal age and parturition date. Vaginal Implant Transmitters (VITs) The VITs we used were manufactured by Advanced Telemetry Systems, Inc. (Isanti, MN). The VIT was 76 mm long, excluding antenna length, and had 2 plastic wings with a width of 57 mm when fully spread apart. The plastic wings were used to retain the transmitter in the vagina until parturition. The VIT weighed 15 grams and contained a 10-28 lithium battery. The diameter of the transmitter/battery was 14 mm, and was encased in an impermeable, water-proof, electrical resin. The transmitter contained an embedded heat-sensor which dictated the frequency pulse rate. When the heat sensor dropped below 86°F, synonymous with transmitter expulsion from the deer, the pulse rate changed from 40 PPM to 80 PPM. The VIT was inserted into deer using a vaginoscope (Jorgensen Laboratories, Inc., Loveland, CO) and alligator forceps. The vaginoscope was 6" long with a 5/8" internal diameter and had a machined end (smooth surface) to minimize trauma when inserted into the vagina. A discreet mark was placed on the applicator showing the appropriate distance it should be inserted into the deer. The length of a typical mule deer vaginal tract was obtained by taking measurements from road-killed deer and/or other fresh deer carcasses obtained in the study area. Prior to use in the field, VITs were sterilized using a Chlorhexidine solution, air-dried, and sealed in a 3" x 8" sterilization pouch. Sterilization containers with Chlorhexidine solution were used on site during capture to sterilize the vaginoscope and alligator forceps between each use. A new pair of nitrile surgical gloves was used to handle the vaginoscope and VIT for each deer. To insert a VIT, the plastic wings were folded together and placed into the end of the vaginoscope. We then liberally applied sterile KY Jelly to the scope and inserted it into the deer's vagina to the point where the mark on the applicator was reached. The a11igator forceps, which extended through the vaginoscope to hold the VIT, was held firmly in place while the scope was pulled out from the vagina. This procedure pushed the VIT out of the scope into the vagina, and the plastic wings spread apart to hold the transmitter in place. The transmitter antenna was typically flush with the vulva, but on occasion extended up to 1 cm beyond the vulva. The tip of the antenna was encapsulated is a wax bead to protect the deer. Adult Doe Monitoring From March through May, we regularly monitored the radioed does as part of our current research experimental design, which included daily monitoring for live/death status (Bishop and White 2002). We also used aerial telemetry to relocate each of the does every couple of weeks during the remainder of the winter. Fetus Survival and Neonate Capture During June 10-30, 2002, we relocated each of the radio-collared does having a VIT each morning using a fixed wing aircraft. Flights began at 6:00 AM and were usually completed by 10:00-11:00 AM. The early flights were crucial for detecting fast signals because shed VITs were often warmer than 86 °F by mid-day, which caused them to switch back to a slow ("pre-birth") pulse. When a fast ("postpartum") pulse rate was detected, we located the VIT from the ground to determine whether it was shed at the birth site. If the transmitter was located at the birth site, we identified whether any fawn(s) were stillborn. If the fawn(s) were no longer present at the birth site, or could not be found in the vicinity of the birth site, we located the radio-collared doe and searched for fawns at her location. All personnel involved wore surgical gloves to help minimize human scent when handling fawns. For each doe, we attempted to document whether any fawns were stillborn, locate each of her fawns, radio-collar and weigh the fawns, record basic vegetation characteristics of the birth site, and promptly exit the site. We attempted to �89 account for each doe's fetuses in order to evaluate the efficacy of using this technique to quantify in utero fetal survival from February to birth. We then radio-monitored each of the radio-collared fawns on a daily basis to measure survival rates of treatment and control fawns and to assess cause-specific mortality. We also periodically located other radio-collared does that did not have VITs and attempted to capture their fawns to help achieve our targeted sample size. Each of these does were part of the nutrition enhancement research, and were present on either the treatment or control experimental unit during winter. RESULTS AND DISCUSSION VIT Effectiveness The proportion ofVITs that were expelled at or near the birth site with the transmitter functioning correctly was 0.33 (SD= 0.083). Of36 VIT trials, 3 were censored, 7 were shed prematurely, 15 had battery failures, and 11 were successful. Censors: Two adult does died in May well before fawning and were still carrying the VITs. One doe was never relocated after leaving winter range. These 3 deer were censored because there was no test of whether the VIT functioned correctly or not. Premature Sheds: Three VITs were shed in May or early June well before fawning (May 18, May 19, and June 6). The other 4 VITs were shed during the fawning period, but at least 1-2 days before the respective does gave birth. Battery Failures: With only 1 exception, all battery failures occurred just before or during the fawning period. This was the glaring problem with the VIT success rate. The battery we had hoped to use had a warranty life of 116 days and a capacity of 232 days. Unfortunately, this battery was discontinued by Advanced Telemetry Systems (ATS) just before our research study began due to poor results. The battery we subsequently used in the VITs had a warranty life of only 94 days, and a capacity of 188 days. We needed our batteries to last 120 days for this research. ATS bad found good results with this shorter life battery, and recommended its use because it typically lasted well beyond the warranty battery life. We knew this was a risk at the onset of the research, but had confidence the batteries would last the necessary 120 days based on ATS recommendations. Had the batteries not failed, we likely would have bad a 60-70% success rate. Successes: We had 11 transmitters function correctly. One transmitter was still in the doe and functioning at the end of our capture period. Of the remaining 10 VITs, 9 allowed us to efficiently locate and capture fawns, typically at the birth site, and account for each of the given doe's fetuses measured in February/March. We located 14 fawns, one of which was a stillborn, from these 9 VITs. Our fawn capture success rate for the 33 available does was 0.61 (SD= 0.086), meaning we captured at least 1 fawn from 61 % of the VIT radio-collared does. In 3 instances, we opportunistically located a VIT with a failed battery by radio-tracking the doe and searching for her fawns. In total, we located 30 fawns from VIT does, 3 of which were stillborn, and 1 we weren't able to radio-collar. Fawn Capture We captured and radio-collared a total of 54 fawns from 38 adult does (1.42 fawns/doe) during June 11July 1, 2002 (Fig. 2). We found 4 stillborns at birthsites, and 2 suspected stillborns at or near birthsites which could have been early neonate mortalities. We captured 30 fawns from treatment does and 24 from control does. Twenty-six fawns were captured from 17 of the VIT does (1.53 fawnsNIT doe) and 28 fawns were captured from 21 radio-collared does that did not have VITs (1.33 fawns/non-VIT doe). We documented a total of 62 live fawns from the 38 does (1.63 fawns/doe), although we only captured 54 fawns because 1 fawn from a set of twins escaped on multiple occasions. �90 Capture Effort The 15 VIT battery failures caused considerable problems during the fawn capture. As it turned out, VITs helped us capture only 13 fawns and locate 1 stillborn fawn. The other 41 fawns and 5 stillborns were captured/located by routinely radio-tracking collared does and searching for fawns at their locations. This required an intensive field effort. We worked approximately 1700 man-hours (212 man-days) during a 22-day period to capture the 54 fawns, or roughly 4 man-days/fawn. Our objective pre-fawning was to capture 55 fawns; thus, even with the VIT failures, we were able to capture the necessary fawns from radio-collared does. The field effort would have been considerably less had we not had the battery failures. 8 7 "O "'::I - 6 i... 5 0. C;I u 4 -"' ,.. ~ ~ 3 ~ ::it: 2 1 0 I I 6110/02 6113102 II 6116102 6119102 I 6122/02 6125/02 I I 6128/02 7/1102 Figure 2. Number of newborn fawns captured by day during June 11-July 1, 2002. All fawns were captured from radio-collared does throughout the southern portion of the Uncompahgre Plateau and adjacent San Juan Mountains in southwestern Colorado. Fetus Survival In February-March 2002, we measured an average of 1.80 fetuses/doe (SE= 0.14, n = 36), which included 1.77 fetuses/doe (SE= 0.14, n = 18) in the treatment unit and 1.83 fetuses/doe (SE= 0.15, n = 18) in the control unit. In June 2002, considering all does that we located any fawn from, whether live or stillborn, we observed 1.42 (SE= 0.11, n = 43) live fawns/doe postpartum. This rate includes the 6 stillborns, and should represent a conservative estimate of live fawns/doe postpartum because we inevitably failed to locate all live fawns from each doe. In other words, this estimate would treat any unaccounted fetuses (from the February measurement) as if they were stillborns. For does that did not have VITs, and thus we did not have a winter fetus rate measurement, singletons would infer that either the deer only had I fetus, or that the other fetus died. It is likely that many of these singletons had a twin that we did not locate. This equates to a conservative fetus survival rate estimate of 0.79 (SE= 0.063). We accounted for all the fetuses of 14 VIT does in June 2002, 8 of which were the direct result of the VIT �91 functioning correctly. Of these 14 does, the fetus survival rate was 0.86 (SD= .096). This data point lacks precision and may potentially be biased because we did not account for an adequate number of fetuses due to the VIT failures. Of 10 VITs that functioned correctly and were shed during fawning, we accounted for all recorded fetuses in 8 of the deer. One of the other 2 VITs that functioned correctly allowed us to account for 2 of 3 fetuses measured in February. Clearly, to gain a more reliable estimate of fetus survival, a high percentage of the VITs must function correctly so that more birth sites can be located, and more fetuses can be accounted for. Neonate Survival We accomplished our neonate capture objectives even with the VIT failures; it simply required a greater effort. If this technique were incorporated into the nutrition enhancement experiment at full scale, we would need to capture approximately 80 newborn fawns (40 each from treatment and control does). Assuming we can purchase implants with longer-lived batteries, we could feasibly capture 80 fawns by increasing our sample size ofVIT does. CONCLUSIONS The VITs were largely successful except for the 15 battery failures. The battery problem can easily be corrected by working with Advanced Telemetry Systems to locate and utilize a reliable battery with a longer life. Such batteries exist, and have been used routinely in small mammal and avian radio transmitters. The 7 premature sheds were expected to some extent, and not a major concern. The VITs that did not fail were highly useful for determining the location and timing of birth, which is of critical importance for capturing fawns from individual, radio-collared does. We found the use ofVITs to be a successful field technique given our objectives, assuming a reliable, longer-lived battery can be incorporated into the current design. Literature Cited Barrett, M. W., J. W. Nolan, and L. D. Roy. 1982. Evaluation of a hand-held net-gun to capture large mammals. Wildlife Society Bulletin 10: 108-114. Bishop, C. J., and G. C. White. 2000. Effects of habitat enrichment on mule deer recruitment and survival rates. Colorado Division of Wildlife, Wildlife Research Report, Federal Aid in Wildlife Restoration Project W-153-R-13, Progress Report. Fort Collins, CO, USA. Bishop, C. J., and G. C.. White. 2002. Effects of nutrition and habitat enhancements on mule deer recruitment and survival rates. Colorado Division of Wildlife, Wildlife Research Report, Federal Aid in Wildlife Restoration Project W-153-R, Progress Report. Fort Collins, CO, USA. Bowman, J. L., and H. A. Jacobson. 1998. An improved vaginal-implant transmitter for locating whitetailed deer birth sites and fawns. Wildlife Society Bulletin 26:295-298. Garrott, R. A, and R. M. Bartmann. 1984. Evaluation of vaginal implants for mule deer. Journal of Wildlife Management 48:646-648. Giessman, N. F., and C. J. Dalton. 1984. White-tailed deer fawn mortality in the southeastern Missouri Ozarks. Missouri Department of Conservation, Jefferson City, Pittman-Robertson Project W-13R-35. �92 Nelson, T. A. 1984. Production and survival of white-tailed deer fawns on Crab Orchard National Wildlife Refuge. Thesis, Southern Illinois University, Carbondale, IL, USA. Pojar, T. M. 2000. Investigating factors contributing to declining mule deer numbers. Colorado Division of Wildlife, Wildlife Research Report, Federal Aid in Wildlife Restoration Project W-153-R-13, Progress Report. Fort Collins, CO, USA. Pojar, T. M., and W. F. Andelt. 1999. Investigating factors contributing to declining mule deer numbers. Colorado Division of Wildlife, Wildlife Research Report, Federal Aid in Wildlife Restoration Project W-153-R-12, Progress Report. Fort Collins, CO, USA. Unsworth, J. W., D. F. Pac, G. C. White, and R. M. Bartmann. 1999. Mule deer survival in Colorado, Idaho, and Montana. Journal of Wildlife Management 63:315-326. van Reenen, G. 1982. Field experience in the capture of red deer by helicopter in New Zealand with reference to post-capture sequela and management. Pages 408-421 in L. Nielsen, J.C. Haigh, and M. E. Fowler, editors. Chemical immobilization of North American wildlife. Wisconsin Humane Society, Milwaukee, Wisconsin, USA. � https://cpw.cvlcollections.org/files/original/821afeab84e170a005e542df8312032e.pdf 8ab60fcf9e2995d808e819786f0f1e27 PDF Text Text 93 Colorado Division of Wildlife Wildlife Research Report July 2001 and July 2002 PROGRESS REPORT State of_ _ _ _ _ _--=C-=o=lo=r-=a=do-"-------- Division of Wildlife - Mammals Research Work Package_ _ _ _~30~0~1~------ Deer Conservation Task No. 5 -------~------- Federal Aid Project No. Work Package No. Work Package No. Work Package No. Work Package No. Work Package No. Improved Population Modeling DEAMAN System Administration W-185-R Research and Development and the following non-Federal Aid projects 0661 Grouse Conservation 0664 Prairie Dog Conservation 0670 Lynx Conservation 0850 Peregrine Falcon Recovery 3006 Other Small Game Conservation Period Covered: July 1, 2001 - June 30, 2002 Author: G. C. White Personnel: C. Bishop, G. Miller, T. E. Remington, D. J. Freddy, T. M. Shenk, L. Stevens, J. Craig, R. Kahn, D. C. Bowden, F. Pusateri, J. Dennis, M. M. Conner, D. Walsh, B. Lubow. ABSTRACT Progress towards the objectives of this job include: Consulting assistance to CDOW on harvest surveys, terrestrial inventory systems, and population modeling procedures was provided. Estimates of spring and fall turkey, spring snow goose, sharptailed and sage grouse, chukars, ptarmigan, Abert's squirrels, and general small game harvest were computed from survey data, and programs and harvest estimates provided to CDOW via email and CD ROM. Computer code written in SAS to compute these estimates and display results graphically was also provided. Computer code was also written in SAS to estimate the compliance rate of Colorado small game license holders with the Harvest Information Program. The DEAMAN software package for the storage, summary, and analysis of big game population and harvest data was revised further as a Windows 95/98/NT/2000/ME/XP program. The capability to incorporate data on radio-collared animals to estimate survival with the Kaplan-Meier estimator and display movement data was added, and distributed to terrestrial biologist via the WWW at http://www.cnr.colostate.edu/~gwhite/deaman. A 3-day workshop was conducted with CDOW Terrestial Biologists in the use ofDEAMAN and population modeling procedures, mainly to instruct personnel on the use of spreadsheet models for ungulate population dynamics. In addition, numerous questions were answered via meetings with biologists, and via email. A paper, coauthored with Bruce Lubow, was published in the Journal of Wildlife Management on past efforts to develop a realistic mule deer population model based on data collected with current CDOW procedures. Data from the Piceance Basin were used to illustrate the modeling technique. The full '· BDOW016783 �94 citation is: White, G. C., and B. Lubow. 2002. Fitting spreadsheet population models to multiple sources of observed data. Journal of Wildlife Management 66:300-309. A paper on use of population viability analyses applicable to animals monitored with mark-encounter data was published with T. M. Shenk and A. B. Franklin. The full citation is: White, G. C., A. B. Franklin, and T. M. Shenk. 2002. Estimating parameters of PVA models from data on marked animals. Pages 169-190 in S. R. Beissinger and D. R. McCullough, editors. Population Viability Analysis. University of Chicago Press, Chicago, Illinois, USA. A paper on analysis of radio-tracking data for estimation of survival pertinent to monitoring the reintroduced lynx population in Colorado was published with T. M. Shenk: White, G. C., and T. M. Shenk. 2001. Population estimation with radio-marked animals. Pages 329-350 in J. J. Millspaugh and J.M. Marzluff, editors. Design and Analysis of Wildlife Radiotelemetry Studies. Academic Press, San Diego, California, USA A paper on the estimation of population size from correlated sampling unit estimates of the variable of interest was submitted to the Journal of Wildlife Management. The methodology developed in this paper is proposed for use in a joint Colorado/Utah survey of the colony area of white-tailed and Gunnison prairie dogs in western Colorado and eastern Utah. The full citation is: Bowden, D. C., G. C. White, A. B. Franklin, and J. L. Ganey. 2003. Estimating population size with correlated sampling unit estimates. Journal of Wildlife Management 67:1-10. A paper on the estimation of survival of Greater Sage-grouse in North Park, Colorado, was submitted to the Journal of Wildlife Management: Zablan, M.A., C. E. Braun, and G. C. White. 2003. Estimation of northern sage-grouse survival in North Park, Colorado. Journal of Wildlife Management 67:144154. Assistance in the analysis of candidate systems to estimate deer abundance in GMU IO was provided. A research study to examine the impact of nutrition on the decline of mule deer fecundity during the last 20 years was continued. I have provided input on estimation of the number of deer on the feed sites, and developed an estimator of fawn survival rates based on radio-collared does and fall and spring fawn:doe ratios. Data were collected and analyzed on spatial distribution, movement of radio-collared animals, and population sizes related to estimating the spread and impacts of chronic wasting disease in deer populations. A report summarizing these findings was provided to CDOW personnel involved with the study. A graduate research project by Dan Walsh to evaluate utility of Iek counts of Greater Sage-grouse in Middle Park is ongoing. Mark-resight methods are being used to estimate lek attendance and population size. Preliminary results of this work were reported at the Sage-grouse Working Group Meeting on Population Analysis held June 24-25, 2002, in Torrey, Utah. Computer programs to assist researchers with field data collection of feeding rates of tame Greater Sagegrouse were written for the HP Jornada Pocket Computer. In addition, a program for computing triangulation Iocations:ofradioed animals was:wrltte·n for the HP Jorilada Pocket Ccimputer for use in the field to evaluate interactively and graphically the quality of triangulated locations. A model of the Colorado peregrine falcon population was constructed from estimates of survival and reproduction derived from banded birds (1973-2001) and monitored nests (1989-2001). Data were supplied by Jerry Craig. Survival estimates for 0-1, 1-2, and 2+ year old birds were 0.544, 0.670, and 0.800, respectively, with standard errors of 0.0765, 0.091, and 0.0544. Average young produced per pair was 1.660 (SE= 0.0443), but there was considerable variation across years (min =·I.388, SE= 0.1548, in 1995; max= 2.122, SE= 0.1393, in 2000). Based on a population model constructed from these estimates, the annual rate of population increase was 1.028957 for females first reproducing at 3 years of age, and 1.080316 for females first reproducing at 2 years of age. Given these high rates of population increase, some take by falconers could be accommodated by the Colorado peregrine population. An analysis to estimate the effort required to estimate the percent of eastern Colorado inhabited by blacktailed prairie dogs was completed and results provided to CDOW personnel involved with the effort. For the first 5 strata surveyed, I have computed estimates of prairie dog colony areas to assist the CDOW personnel conducting the survey. Results to date suggest the survey is working well. �95 CONSULTING SERVICES FOR MARK-RECAPTURE ANALYSES G. C. White PROJECT OBJECTIVES Assess the status of Colorado peregrine falcon population based on parameters estimated from banded birds and monitored nests. SEGMENT OBJECTIVES 1. Develop a model of the Colorado peregrine falcon population based on survival and reproduction estimates derived from banded birds (1973-2001) and monitored nests (1989-2001). 2. Using this model, determine the impact oflimited take by falconers on Colorado peregrine populations. RESULTS AND DISCUSSION Methods Survival Rate Estimation A total of938 peregrines were banded as nestlings during the interval 1974-2000. From these, 11 live resightings and 53 dead recoveries were obtained. Survival was estimated with Program MARK (White and Burnham 1999) using the joint live and dead recoveries model of Burnham (1993). Program MARK uses the Seber (1970) parameterization for dead recoveries in the Burnham model, so parameters for survival (S), probability that a band from a dead bird is recovered (r), and the probability of a live bird being resighted and the band read (p). The fidelity parameter (F) was fixed to 1 because of the sparseness of the data. Reproduction A total of 142 nesting sites were monitored for the number of young fledged starting in 1973 through 2001, although the number of sites increased with year as the breeding population increased in Colorado. Only data from 1989 through 2001 were used in the analysis presented here because I wanted at least 30 nests per year to estimate the effects modeled. Mean number of young fledged per site was estimated by year. Estimates of the variance components by site and year was est_imated with PROC MIXED of SAS ·(Littell et al. 1996). Structures considered for the variance of the repeated observations within sites were first order autoregressive, first order heterogeneous autoregressive, compound symmetry, heterogeneous compound symmetry, exponential local effects, also known as dispersion effects, in a log-linear variance model, and a null structure with zero covariances and constant variances across years, also commonly known as the variance components structure (Littell et al. 1996). Random effects considered in the models were year and site effects. A fixed effect for even and odd years was also included in the models because there is a defined sequence of a year of high reproductive output, followed by a year of low reproductive output, followed by another year of high reproductive output, etc. The even/odd year fixed effect models this oscillating reproductive output. Selection among models for both survival and reproduction parameter estimation was performed with the AICc criterion recommended by Burnham and Anderson (1998). �96 Population Model A model of the female segment of the Colorado peregrine population based on the estimates of survival and reproduction was constructed in an Excel spreadsheet, and also as a SAS program. The model included 4 age classes (No, N 1, N 2, and N 3), even though the survival estimates used in the model had survival the same for all birds 3+ years old. The parameters in the deterministic model were number of fledglings per reproducing female (F), proportion of fledglings that are female (assumed to be SR= 0.5), survival for 4 age classes (So, S 1, S2, and S3), and the proportion of females that breed on their second birthday (B2). In addition, a parameter for the proportion of fledglings removed by falconers was included in the model to allow the estimation of the effects of human take (1). The difference equations for the transition from year t to year t+ 1 in model were: N1.1+1=No., So, No.1+1=(B2 N2.1+1+ N3.1+1J F SR (1-T). A stochastic model was developed from the above deterministic model by replacing the mean fledglings per female with a value randomly drawn from the observed values for the years 1989 through 2001. The VLOOKUP function of Excel was used to randomly select one of the 13 observed values for each year in the model. That is, the fledgling rate was sampled with replacement from the observed values. This stochastic reproduction model was also extended to incorporate demographic stochasticity in the survival process. Instead of multiplying the population segment by a fixed survival rate to obtain the number of survivors, the process is treated as a binomial process, with the number of survivors drawn from a binomial distribution with the appropriate survival rate. However, given the size of the peregrine population in Colorado, demographic stochasticity had little or no effect on the model predictions. Results Survival Rate Estimation The encounters_ of_:rparked birds were spar!,~, and thus preclude complex models involving both time and age effects. The a priori list of models· co'risidered (Table 1) included models to evaluate age differences in survival for birds during their first 4 years of life, and time-specific effects in survival, live resighting rate, and dead recovery rate. The minimum AICc models all included time-specific variation in the band recovery rate, but did not suggest time-specific variation was required to model the live resighting rates. The minimum AICc model was {S(a2) p(.) r(t)}, with the second best model {S(a3) p(.) r(t)} only 0.388 units above the minimum. I chose to use estimates from the 3-age model because this model is more realistic biologically, and the small difference in AICc values does not suggest that the 2age class model is much better than the 3-age class model. Parameter estimates for the 3-age class model (Table 2) are reasonable in that survival increases with age, but all have large standard errors. Reproduction Estimation Average number of fledglings produced per nest during the interval 1989-2001 was 1.66059 with SE 0.044296 (Table 3). However, as shown in Table 3, there is considerable variation from year to year in the number of fledglings per breeding pair. �97 Model selection results (Table 4) for estimation of the variance components for site and year suggest that a first order autoregressive variance structure is required. Based on the minimum AI Cc model with the AR(l) variance structure, the estimate of the variance component for year was 0.01609, for site was 0.08438, with a residual variance component of 1.6069. The autocorrelation coefficient between consecutive years within a site as 0.1109. Thus, the variance components due to year and site were relatively minor compared to the residual variance in young fledged per breeding pair. The even/odd year effect was estimated as 0.2482 with a SE of0.1065 (P < 0.0403). Thus, the magnitude of the every-other-year oscillation is estimated to be 0.2482 young per nest. Population Model The deterministic population model provided an estimate of A = 1.028957 with T= 0 and F= 0 for the survival parameter values in Table 2 and fledglings per reproducing female of 1.66059, estimated as the mean fledgling rate for the 1989-2001 interval. Thus, the model predicts that the Colorado peregrine population is increasing 2.9% per year, even with no 2-year old birds reproducing. The effect ofreproduction from 2-year birds is shown in Figure 1, and suggests that if all of the 2year old birds breed, A = 1.080316. Results from the model with stochastic reproduction provided estimates of A consistent with the deterministic model, e.g., for 10,000 simulations, A = 1.0290287 with SE= 0.000631839, giving a 95% confidence interval of 1.0277902 to 1.0302673 that encompasses the value estimated from the deterministic model. The large standard errors of the parameter estimates used to build the population model (Tables 2 and 3) suggest that the estimate of A obtained will also have a large standard error. I used Monte Carlo simulation techniques to draw values of each of the survival and reproduction parameters from a normal distribution with mean equal to the parameter estimate and standard deviation equal to the standard error of the parameter estimate. With a single set of parameters so obtained, 10,000 values of A were averaged to obtain a mean for that parameter set. This process was repeated for 1000 parameter sets to obtain a SD of A of 0.0612896. This value can be interpreted as a SE of the estimated A that accounts for the SE of the input parameters. Average standard errors from the 10,000 simulations for the 1000 parameter sets was 0.0015142, suggesting that the preceding SD is not affected by the small amount of variation associated with each of the 1000 estimates. Discussion The best AI Cc model for estimation of survival required 28 parameters to estimate time-specific band recovery rates, r(t). A more parsimonious model would provide estimates of survival with better precision. One approach to obtaining a more parsimonious model would be to include a covariate that models the variation in r(t). The estimated rate of increase from the population projection model is relatively high for a wildlife population. However, the number of nest sites monitored each year (Table 3) provides confirmation of the high estimates of A. An exponential model regression, log.,(No. Nest Monitored)= /3 0 + /3 1 Year, was used to estimate the rate of increase of numbers of nest monitored. From this regression, jJ 1 = 0.07883 with SE= 0.00755 (P < 0.0001), giving an estimate of the annual rate of increase of the population (A) of exp(0.07883) = 1.082. This value exceeds the maximum value predicted from the population model even with 100% of the 2-year old birds breeding. Although the �98 estimate obtained from the number of nests monitored suffers from confounding with monitoring effort and effort to find new nests, the results still suggest that the Colorado peregrine population has had a high annual rate of increase, and that the predictions from the population model described here are consistent with another estimate of A . The high annual rate of increase suggests that moderate take can be accommodated by the Colorado peregrine population without affecting the population growth rate. Based on the population model presented with no 2-year old birds breeding, A = l with 17.5% of the fledged young taken. With 50% of 2-year old birds breeding, A = l with 26. 7% of the fledged young taken, and with 100% of 2year old birds breeding, A = l with 34.0% of the fledged young taken. Summary A model of the Colorado peregrine falcon population was constructed from estimates of survival and reproduction derived from banded birds (1973-2001) and monitored nests (1989-2001). Survival estimates for 0-1, 1-2, and 2+ year old birds were 0.544, 0.670, and 0.800, respectively, with standard errors of 0.0765, 0.091, and 0.0544. Average young produced per pair was 1.660 (SE= 0.0443), but there was considerable variation across years (min= 1.388, SE= 0.1548, in 1995; max= 2.122, SE= 0.1393, in 2000). Based on a population model constructed from these estimates, the annual rate of population increase was 1.028957 for females first reproducing at 3 years of age, and 1.080316 for females first reproducing at 2 years of age. Given these high rates of population increase, some take could be accommodated by the Colorado peregrine population. Literature Cited Burnham, K. P. 1993. A theory for combined analysis of ring recovery and recapture data. Pages 199-213 in J.-D. Lebreton and P. M. North, editors. Marked individuals in the study of bird population. Birkhauser Verlag, Basel, Switzerland. Burnham, K. P., and D.R. Anderson. 1998. Model selection and inference: a practical information-theoretic approach. Springer-Verlag, New York, New York, USA. 353 pp. Littell, R. C., G. A. Milliken, W.W. Stroup, and R. D. Wolfinger. 1996. SAS® System for Mixed Models. SAS Institute Inc., Cary, NC, USA. 633pp. Seber, G. A. F. 1970. Estimating time-specific survival and reporting rates for adult birds from band returns. Biometrika 57:313-318. White, G. C., and K. P. Burnham. 1999. Program MARK: survival estimation from populations of marked animals. Bird Study 46 Supplement:120-138. �99 Table 1. Set of a priori models considered in Program MARK (White and Burnham 1999) for estimation of survival of peregrines banded as nestlings with the joint live-dead model of Burnham (1993). Model names include survival (S), live resighting probability (p) and probability that the band from a dead bird is recovered (r). For survival, models with time-specific survival (t), constant survival(.), and 2 (a2), 3 (a3), and 4 (a4) age classes were considered. Constant and time-specific models for both p and r were also considered. The fidelity parameter (F) was fixed to 1 in all models. Model AICc Delta AICc AICc Weights Model Likelihood Num. Par Deviance 31 161.38 {S(a2) p(.) r(t)} 710.841 0 0.46193 {S(a3) p(.) r(t)} 711.229 0.388 0.38047 0.8237 32 159.626 {S(a4) p(.) r(t)} 713.332 2.491 0.13294 0.2878 33 159.583 {S(.) p(.) r(t)} 716.701 5.86 0.02467 0.0534 30 169.377 {S(a3) p(.) r(.)} 742.565 31.724 0 0 5 247.202 {S(a2) p(.) r(.)} 742.803 31.962 0 0 4 249.462 {S(t)p(.) r(.)} 744.811 33.97 0 0 30 197.487 {S(.) p(.) r(.)} 753.754 42.913 0 0 3 262.429 {S(a4)p(.) r(.)} 886.175 175.334 0 0 6 388.786 �100 Table 2. Parameter estimates from the 3-age class model {S( a3) p(.) r(t)} of peregrine falcons in Colorado, estimated from birds banded 1974-2000. Parameter Estimate F (fixed to l) SE LCI UCI 0 Sage 0-1 0.543995 0.076538 0.394531 0.685934 Sage 1-2 0.669762 0.098121 0.459491 0.828723 Sage 2-3+ 0.800291 0.054382 0.672873 0.886454 p 0.005662 0.002265 0.002581 0.012374 r 1974 0 lE-07 -3E-07 3E-07 r 1975 0 5E-07 -l.IE-06 l.IE-06 r 1976 0 3E-07 -7E-07 7E-07 r 1977 0.42494 0.275041 0.07526 0.870289 r 1978 0.781612 0.215533 0.231516 0.977021 r 1979 0.20686 0.101744 0.071797 0.467918 r 1980 0.190922 0.093857 0.066924 0.437053 r 1981 0 0 -lE-07 lE-07 r 1982 0 0 0 0 r 1983 0.074126 0.052037 0.017792 0.261367 r 1984 0.035033 0.034906 0.004775 0.215516 r 1985 0 0 0 0 r 1986 0.042528 0.042355 0.00575 0.254375 r 1987 0 0 0 0 r 1988 0.042219 0.029772 0.010304 0.157271 r 1989 0.016702 0.016695 0.002311 0.11077 r 1990 0.019441 0.01942 0.002685 0.127406 r 1991 0.021935 0.021894 0.003025 0.142181 r 1992 0.142769 0.053181 0.066349 0.280741 r 1993 0.01603 0.01601 0.002223 0.106429 r 1994 0.055134 0.031693 0.017401 0.161262 r 1995 0.083715 0.041537 0.030642 0.208902 r 1996 0.034717 0.024458 0.008529 0.130709 r 1997 0.060956 0.034981 0.019218 0.176984 r 1998 0 0 0 0 r 1999 0.040953 0.041348 0.005395 0.251596 r2000 0.052539 0.053344 0.006742 0.311771 r2001 0.069547 0.07131 0.008547 0.39323 �101 Table 3. Summary of number of ~oung fledged Eer nest for Colorado Eeregrines, 1989-2001. Year Number of Nests Fledglings/Nest SD SE 1989 33 1.90909 1.35471 0.235825 1990 42 1.47619 1.15269 0.177864 1991 52 1.65385 1.16963 0.162198 1992 54 1.62963 1.29289 0.17594 1993 55 1.65455 1.36379 0.183893 1994 64 1.625 1.26617 0.158271 1995 67 1.38806 1.26677 0.154761 1996 83 1. 71084 1.29285 0.141909 1997 71 1.42254 1.26109 0.149664 1998 81 1.95062 1.27379 0.141532 1999 88 1.59091 1.33594 0.142412 2000 98 2.12245 1.37927 0.139327 2001 90 1.35556 1.36003 0.14336 Mean 878 1.66059 1.31254 0.044296 Table 4. Model selection results for estimation of variance components of year and site with the MIXED procedure of SAS (Littell et al. 1996). Variance Structure Fixed Effects Random Effects AICc Delta AICc AR(l) Even/Odd Years Year Site 2950.4 0 AR(l) Even/Odd Years 2951.8 1.4 AR(l) Year Site 2952.3 1.9 AR(l) Year 2953.7 3.3 Default Even/Odd Years Year Site 2954.6 4.2 cs Even/Odd Years Year Site 2954.6 4.2 Default Year Site 2956.5 6.1 cs cs Year 2956.5 6.1 Year Site 2958.5 8.1 AR(l) Site 2961.3 10.9 2962.8 12.4 AR(l) ARH(l) Even/Odd Years Year Site 29702 19.8 EXP(YEAR) Even/Odd Years Year Site 2973.6 23.2 �102 - 1.09 ~ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ~ _g 1.08 .c ~ 1.07 : 1.06 ra fu 1.05 C: o 1.04 s ~ 1.03 1.02 - + - - - - - - - - . - - - - - - - , - - - - - - ~ - - - - - ~ - - - - ~ - - - - - - - ! 0.6 0 0.2 0.4 0.8 1 1.2 Proportion of 2-year olds Breeding Figure I. Predicted effect from the population model of the proportion of2-year old females breeding on the rate of increase in the population ( A ). � https://cpw.cvlcollections.org/files/original/692941f935a0f48eb05b20d1ed7ee5e6.pdf b9a962c77578f5367ea5688da1291ac8 PDF Text Text 103 Colorado Division of Wildlife Wildlife Research Report July 2001 and July 2002 JOB PROGRESS REPORT State of _ _ _ _ _ ____:C:::.;o""l""'o""'ra~d::..,o"------- Mammals Research Work Package -------=3:....:0:...::0"""1_ _ _ _ _ __ Deer Conservation Study No. _ _ _ _ _ _____::;9_ _ _ _ _ __ Evaluation ofGnRH-PAP as a Long-term Fertility Control Agent in Female Mule Deer (Odocoileus hemionus hemionus) Federal Aid Project No. W-153-R-13 Research and Development Period Covered: July 1, 2001 -June 30, 2002 Author: Dan L. Baker, Ph.D. Personnel: M.A. Wild, T. M. Nett, T. Davis, E. Jones, B. Hochmoth ABSTRACT We evaluated the effects of GnRH-PAP on mule deer pregnancy rates, duration of suppression of luteinizing hormone and progesterone secretion, blood chemistry and hematology, and reproductive behavior during 1 November to 30 December, 2001. Twenty-two adult female mule deer were assigned to one of 3 experimental groups. Nine female mule deer were treated with GnRH-PAP and 9 females served as untreated controls. The dose of GnRH-PAP used in this experiment did not lower pregnancy rates in female mule deer. Treated and control females tested positive for pregnancy specific protein B on all sampling dates and all delivered healthy fawns in July. At 30 days posttreatment, luteinizing hormone and progesterone were not different (P > 0.58) in treated and control mule deer. Reproductive behaviors of GnRH-PAP treated females were not different from controls. We conclude that the dose of GnRH-PAP administered in this experiment was ineffective in suppressing reproduction in female mule deer. ,i·.. • : ( ,.:;._ ... ' �105 EVALUATION OF GnRH-PAP AS A LONG-TERM FERTILITY CONTROL AGENT IN FEMALE MULE DEER (ODOCOILEUS HEM/ONUS HEM/ONUS) Dan L. Baker P. N. OBJECTIVES 1. Develop a practical and acceptable technology for long-term control of fertility in female mule deer. 2. Demonstrate the feasibility of controlling mule deer population growth in a field application. SEGMENT OBJECTIVES 1. Evaluate the effectiveness of GnRH-PAP in preventing pregnancy in captive female mule deer. 2. Evaluate the duration of GnRH-PAP suppression of LH and progesterone secretion in female mule deer. 3. Assess the behavioral and physiological side-effects (if any) of GnRH-PAP in captive female mule deer. INTRODUCTION Controlling the growth of animal populations is fundamental to maintaining proper balance between wildlife and the habitats they occupy. This is particularly true for wild ungulates. Overabundant ungulates can cause serious degradation of plant communities, and preventing such damage requires controlling their numbers. Hunting has traditionally been used to control ungulate populations but there are increasingly more situations where hunting is infeasible. Such areas include urban areas where safety of people and property may be threatened, or national parks and refuges where populations are managed primarily for non-consumptive uses like wildlife viewing and photography or on military installations and industrial parks because of concerns for security. In these situations, alternatives to hunting or culling as a means of controlling ungulate numbers are needed. Fertility control offers a potenti~I alternative for controlling the growth of overabundant ungulate populations when traditional methods are infeasible or unacceptable (Kirkpatrick and Turner, 1985; Bamford, 1990; Garrot, 1995). However, current technology does not provide a means for controlling populations that is practical, economical and without undesirable side-effects (reviewed by Fagerstone et al., 2001). For most free-ranging wild ungulate populations, permanent sterilization has been proposed as the most efficacious approach to population management (Hone 1992, Garrot 1995, Hobbs et al. 2000). A promising new non-steroidal, non-immunological approach to permanent infertility involves analogs of gonadotropin-releasing hormone (GnRH). GnRH is a molecule produced in the hypothalamus of the brain. It directs specific cells in the anterior pituitary gland to synthesize and secrete two important reproductive hormones; follicle stimulating hormone (FSH) and luteinizing hormone (LH). These latter two hormones, known as gonadotropins, control the proper functioning of the ovaries in the female and the testes in the male. �106 Analogs of GnRH have the potential to permanently inhibit reproduction. By coupling a superactive analog of GnRH to a cytotoxin, it should be possible to specifically target that toxin to LH and FSHsecreting cells in the anterior pituitary gland (Collier and Kaplan 1984, Pastan et al. 1986). Therefore, a single GnRH-toxin conjugate has the potential to induce sterility in both sexes and numerous species of animals. There are many natural cytotoxins available for conjugating to GnRH. Toxins are composed of two subunits, a toxic subunit and a binding subunit. In order to target the toxin to specific cell types (rather than all cells) within the body, the binding subunit of the toxin can be removed and replaced by a molecule that will bind to only one cell type, in our case an analog of GnRH. This will target the toxin to gonadotropin secreting cells in the anterior pituitary gland. This approach has several potential advantages over other methods of contraception. These include: 1) a single treatment should permanently sterilize an animal; 2) the same treatment should be effective in both males and females and in different vertebrate species; 3) GnRH-toxin conjugate is a protein and is metabolized from the body within a few days of treatment, therefore it poses no threat to non-target species; 4) the small volume required for contraception facilitates microencapsulation and administration by syringe dart or biodegradable bullet. Proposed Research To our knowledge, only limited investigations have been conducted with GnRH-PAP in wild ungulates (Nett et al. 2001). In order to provide an estimate of the dose ofGnRH-PAP conjugate required for contraception, it is essential that the potency of GnRH analog be determined in each species and at different phases of the reproductive cycle. We addressed this question in a series of GnRH challenge trials with captive mule deer at the Foothills Wildlife Research Facility in Fort Collins, Colorado (Baker et al. 1996). In these experiments, we determined the most effective dose of GnRH analog in female mule deer during the breeding season to be 1 µg/50 kg BW. This is the minimum dose of GnRH analog that will illicit maximum LH secretion. The next questions that needed to be answered were "how effective is GnRH-PAP in preventing pregnancy, how durable are its effects over time, and are there unacceptable side effects? The objective of this experiment is to begin to address these questions. Specifically, our objectives were: 1) to evaluate the effectiveness ofGnRH-PAP in preventing pregnancy in mule deer; 2) to evaluate the duration of GnRH-PAP suppression ofLH and progesterone secretion; 3) to assess the behavioral and physiological side-effects (if any) of GnRH-P AP treatment. MATERIALS AND METHODS Reproductive Biology Mule deer (Odocoileus hemionus hemionus) are polytocous, multiovular, spontaneous ovulators that exhibit highly seasonal patterns of reproduction that are controlled by photoperiod regimens. The onset of the breeding season occurs during decreasing daily photoperiods of autumn and is preceded by a period of deep an estrous in summer (Plotka et al.1977). The first ovulation of the breeding season is usually preceded by one or more silent ovulations associated with the formation of short-lived corpora lutea that serve to synchronize the first overt estrus within a herd (Thomas and McT.Cowan 1975). In temperate North America, the majority of conceptions occur in late November, but recurrent estrous �cycles of 24 -28 days are possible through March if females fail to conceive (Knox et al. 1988). In early spring, coincidental with increasing day length, reproductive cycles cease and females remain anestrous until October. For pregnant females, parturition generally occurs in late May or early June, after a gestation period of about 200 days (Anderson and Medin 1966). Most females produce one fawn when they are two years old, and one or two annually thereafter (Cowan 1956). Experimental Design We evaluated the effects of GnRH-PAP on mule deer pregnancy rates, duration of suppression of LH and progesterone secretion, blood chemistry and hematology, and reproductive behavior during 1 November to 25 December 2002. Twenty-two adult female and 2 adult male mule deer were used in this experiment. Females were assigned to one of 3 experimental groups based on their tractability for handling and blood sampling. Nine female mule deer (Group A) were treated with GnRH-PAP and 9 females (Group B) served as untreated controls and were used to compare pregnancy rates, blood chemistry and hematology, and reproductive behavior to those of treated animals. Immediately following the pretreatment GnRH challenge trial, and before catheters were removed, 9 females in Group A were administered an optimum dose ofGnRH-PAP (lµg/50kgBW) N. Females in Groups A and B were maintained together with 2 adult male mule deer in a 2 hectare pasture. The remaining 4 females (Group C) served as untreated, non-pregnant controls and were placed in a separate pasture (0.5 ha) without direct contact with male deer. We compared LH and progesterone secretion of these females to those treated with GnRH-PAP (Group A). Sample size requirements were based on the variances observed in these measurements from previous studies with captive mule deer and expected effectiveness of treatment (Baker et al. 1996, Nett et al. 2001). Experimental Animals In order to meet sample size requirements calculated for this experiment (see pages 5-7), 5, adult, freeranging female mule deer were captured from urban areas of the front range of Colorado and transported to FWRF. We attempted to capture the most human-habituated deer as possible in order to minimize any stress related to captivity. All deer were captured and handled under the supervision of a veterinarian using one of several previously approved methods (Conner and Miller, CDOW ACUC 12-1999 & addenda); however thiafentanil oxalate (0.1 mg/kg) was substituted for carfentanil citrate as a primary tranquilization drug. Once tranquilized, does were blindfolded, condition and vital signs checked, eartagged, collared, vaccinated with 7-way clostridial v~ccine, treated with Ivermectin (0.1 mg/kg) and long-acting penicillin. Each doe was then placed in a closed vehicle (covered horse trailer), and sedation reversed and blindfold removed. After release at FWRF, deer were observed daily for any signs of post-capture injuries. Response Measurements Hormonal evaluation. Prior to application of GnRH-PAP, we measured the LH response of each female in Groups A and C to a challenge dose of GnRH analog. Results from this trial provided a pretreatment baseline for comparison to future posttreatment LH responses. This and succeeding LH challenge trials were conducted as follows: On day 1 of the trial, deer were moved from 2 ha pastures to individual isolation pens, sedated (4-6 ml, 2:1 ketamine (200 mg/ml):xylazine hydrochloride (100 mg/ml, IM), and fitted nonsurgically with indwelling jugular catheters. Animals were reversed with yohimbine (0.125 mg/kg, N). On day 2, we administered GnRH analog (1µ /50 kg BW) through the cannula and �108 collect blood samples (5 ml) at 0, 60, 120,180,240,300,360, and 480 minutes postinjection. Following the last blood collection, catheters were removed and each animal given an antibiotic (ceftiofur, (1 mg/kg, IV). Animals were then returned to 2 ha pastures. Serum was.stored at -20 °C until analyzed for LH (Niswender et al. 1969). The duration of contraceptive effectiveness was assessed by conducting similar GnRH challenge trials each month from November, 2001 to December 2003. Analysis. Responsiveness of the pituitary to GnRH challenge was assessed in three ways: 1) maximum LH (ng/ml) response achieved postinjection minus baseline, 2) time required to reach maximum LH, and 3) total amount ofLH secreted (ng/ml/min). Pregnancy rates. We assessed contraceptive effectiveness by determining the pregnancy rates of treated (Group A) and control (Group B) deer. A single blood sample (10 ml) was taken via jugular venipuncture from each animal for pregnancy-specific protein B (PSPB) analysis approximately 60, 90, and 220 days post-conception (Willard et al. 1998). Animal handling and blood collections for PSPB followed methods previously described for hormonal assessment and were collected in conjunction with these measurements. Neonates born to any experimental animal were incorporated into the resident FWRF mule deer herd. General health. Limited knowledge of the effects of GnRH-P AP on nutrition, body weight dynamics, blood chemistry and general health of mule deer have been reported in a previous study at this facility (Nett et al. 2001). However, since a different toxin conjugate is being tested in this experiment, we evaluated these potential side-effects here as well. We assessed physiological side-effects of GnRH by comparing serum chemistry, hematology, and body weight dynamics of treated (Group A) and untreated, non-pregnant mule deer (Group C). Blood collections and body weight measurements were made in conjunction with GnRH challenge trials. Blood samples for hematology and serum chemistry analysis were collected prior to treatment and at 90 days posttreatment then submitted for analysis to Colorado State University, Veterinary Teaching Hospital, Clinical Pathology Laboratory, Fort Collins, Colorado, USA. Serum chemistry profiles were obtained using a Hatachi 917 autoanalyzer (Roche/Boehringer Mannheim, Indianapolis, Indiana, USA) for the following parameters: glucose, creatinine, phosphorus, calcium, magnesium, total protein, albumin, globulin, albumin/globulin ratio, bilirubin, creatinine kinase, aspartate aminotransferase, gamma-glutamyltransferase, sorbitol dehydrogenase, sodium, potassium, chloride, and biocarbonate. Values for the following hematological parameters were obtained using an ADVIA 120 autoanalyzer (Bayer Corporation, Tarrytown, New York, USA): nucleated cells, neutrophils, lymphocytes, monocytes, eosinophils, plasma protein, erythrocyte, hemoglobin, packed cell volume, mean corpuscular volume, mean corpuscular hemoglobin concentration, platelets, and fibrinogen. Reproductive behavior. The effectiveness of GnRH-PAP as a, fertility control agent is dependent upon permanent suppression of ovulation and steroidogenesis. Thus, we tested 2 hypotheses relative to the effects of leuprolide on reproductive behavior of mule deer: (1) because GnRH-PAP is expected to suppress gonadotrophin secretion and ovulation, we predicted that sexual interactions during the breeding season (Nov 1 - Dec 20) would be reduced in treated females (Group A) compared to untreated controls (Group B), and (2) once untreated females become pregnant, reproductive behaviors would cease and sexual interactions would be similar between untreated and treated females during the postbreeding season (Jan 10 - Mar 31 ). �109 To test these hypotheses, we examined the effects of GnRH-PAP on reproductive interactions of male and female deer during 2 time periods; breeding season (defined as the period November 1- December 20, 2001) and postbreeding season (defined as the period January 10 - March 27, 2002). On November 1, 2001, female deer in Group A were treated with GnRH-P AP and released with untreated controls (Group B) into 2 ha paddocks. Four days later (November 5), we placed 2 adult male mule deer with these groups and initiate behavioral observations. All females were individually identified with color/numeric-coded neck collars. Animals selected as treatments and controls were unknown to observers. Behavioral measurements were made from a distance of 50-350 m from an elevated tower (10 m) using binoculars and a spotting scope during the day, and a spotlight and night vision scope at night. We recorded selected behaviors using a lap-top computer with a behavioral software program. We used focal animal sampling procedures to sample reproductive behaviors of all experimental animals • over a 24-hour period (Lehner, 1996). Previous studies (Baker et al. 2000) have shown that mule deer are most active in morning (0500-0800), late day (1400-1700) and night (2000-2400). Thus, time-of-day sampling periods was randomly assigned each week using a randomized block design. Each sampling period consisted of at least two hours of continuous observations. Sample size estimation for pregnancy rate and behavior measurements. Sample size calculations were based on results of previous investigations (Baker 2001 ). We used male pre-copulatory behavioral rates to estimate sample size because these rates were much higher than other rates (Table 1). Table 1. Behavior rates for mule deer (Baker 2001 ). Behavior Category Treatment Group Copulatory Control Leuprolide Male Pre-Copulatory Control Leuprolide SQ Female Pre-Copulatory Control Leuprolide SQ General Breeding Control Leuprolide SQ Mean SE (# behaviors/day) 0.20 0.29 0.17 0.24 8.28 22.18 1.77 2.50 0.22 1.35 0.34 0.47 2.00 2.96 0.41 0.58 For control females, we directly bootstrapped a given sample size from the male pre-copulatory rates exhibited towards control does from the Table 1. That is, for a sample size of 6, we randomly selected 6 does (with replacement) for the given observation period, from the 8 control does. For treated does, we followed the same procedure except that we multiplied the response by the effect size. For example, for a +50% effect size, we multiplied the control response by 1.5. We then ran sample sizes for increased behavioral rates because this would be the most problematic to the animal. However, the power should be almost the same for a -50% effect size. We considered the higher behavior rates because increased behavior and corresponding energy output would be the most critical to the animal. This approach captured the day to day variability in behavior rates, because we bootstrapped for each observation period, but it assumes that the variability in behavior is the same for the control and treatment does. Next, we ran the procedure for 52 and 104 observation periods. We assumed 104 observation periods would be acceptable for 2 reasons. First, our proposed collection schedule was: �IIO a. November - 4 weeks x 3 observation periods/day x 5 days/week= 60 obs periods b. December - 4 weeks x 2 observation periods/day x 5 days/week= 40 obs periods Total = I 00 observation periods. Since power was not nearly as sensitive to the number of observation periods as to the number of animals used in the experiment; we decided that ifwe were somewhat below the 104 observations used in sample size simulations, it would not change the power meaningfully. Power results were based on the number of times an effect was detected during 100 simulations. Results from male precopulatory behavioral rates indicated that a sample size of 14 does (7 control and 7 treatment) should provide power of>90% to detect a 50% effect size. Table 2. Sample size calculations for a given effect size using male precopulatory behavior rates. Effec tsize Sample Size* Number of Observation Periods 25% 12 104 16 104 8 104 10 104 12 104 12 52 40% 9 104 50% 7 104 8 104 8 52 30% a. Power 0.05 0.10 0.05 0.10 0.05 0.10 0.05 0.10 0.05 0.10 0.05 0.10 0.05 0.10 0.05 0.10 0.05 0.10 0.05 0.10 57% 74% 81% 91% 68% 70% 72% 83% 90% 95% 68% 81% 91% 97% 94% 96% 97% 100% 74% 88% _ _ 6_· ·- •••• * Sample size for each group, e.g. 12 means 12 treatment and 12 control does. The daily male precopulatory rate for the control group was 8.3 behaviors/day. For a 50% effect size, we could detect a difference between the control mean of 8.3 behaviors/day and a treatment mean of <4.2 and> 12.5 behaviors per day. For a 40% difference in behavior rates, we could detect a difference between the control mean of 8.3 behaviors/day and a treatment mean of <5 and> 11.6 behaviors per day. After estimating the number of animals needed to detect behavioral differences, we calculated power to detect differences in pregnancy rates. If we have 7 control and 7 treatment animals, and 1 treatment animal gets pregnant, and 1 control animal does not get pregnant, we had >95% power to detect this difference. Basically, to detect a difference in pregnancy rates in the case where 1 treatment animal gets pregnant, and 1 control animal does not get pregnant, we could have as low as 5 treatment and 5 control animals and still have >90% power to detect a difference. Thus, the sample size needed to detect �111 differences in behavior rates was sufficient to detect differences in pregnancy rates. Our calculations indicated that the optimum sample size for these measurements to be 14 animals (7 control : 7 treated), however, there is a high probability of losing at least one or two animals per year to non-treatment related mortality. Therefore, in order to insure meaningful measurements over the 2-year investigation, we increased our sample size to 18 animals (9 control: 9 treated). Statistical Analysis We analyzed for differences among hormone levels using least squares analysis of variance for general linear models (SAS Institute 1993). Responses to treatments were analyzed with one-way analysis of variance for a randomized complete block design with repeated measures structure. Treatment effects were tested using the animal-within-treatment variance as the error term. Time was treated as a withinsubject effect using a multivariate approach to repeated measures (Morrison, 1976). A "protected" least significant difference test (Milliken and Johnson 1984) was used to separate means when the overall Ftest indicated significant treatment effects (P < 0.05). We tested specific reproductive behavior hypotheses that mean behavior rate was not different between treatment and control groups for both the breeding and postbreeding seasons using an ANOVA model with a repeated measures structure. Similar to the hormonal analysis, time was be treated as a within subject effect using multivariate approach to repeated measures (Morrison, 1976). To test for treatment effects, we accounted for time-of-day, date effects and their interactions. PROC GENMOD (SAS Institute 1993) was used to estimate and test for differences in mean behavior rate by treatment, time-ofday, and date. Means and standard errors were estimated using least squares, and hypothesis tests were based on type III generalized estimating equations that account for correlation in repeated measurements. RESULTS AND DISCUSSION Pregnancy rates. This dose of GnRH-PAP did not lower pregnancy rates in female mule. Treated and control females tested positive for PSPB on all sampling dates and all delivered healthy fawns in July August, 2002. All fawns were incorporated into the existing captive mule deer herd at the FWRF. Hormonal evaluation. GnRR-PAP did not cause a significant reduction in LH in treated female mule deer. Peak serum LH concentrations for treated animals, after 30 days posttreatment, averaged 8.6 ± 1.97 ng/ml and 9.7 ± 3.0 ng/ml for controls. Based on these responses, GnRH challenge trials were terminated at+ 30 days posttreatment (4 Dec 2001 ). Reproductive behavior. We observed male to male dominance interac~ions immediately following their release into the pastures with treated and untreated females. Within 10 days, one male established dominance. Thereafter, the subdominant male retreated to remote locations within the pastures and rarely interacted with females or the dominant male for the remainder of the experiment. Contrary to our hypothesis, sexual interactions were not different (P > 0.65) during the breeding and post-breeding seasons between GnRH-treated females and untreated controls for any of the breeding behavior categories. We observed almost no sexual interactions between the dominant male and treated or untreated females during the postbreeding season. We conclude :from these measurements that the dose ofGnRH-PAP administered in this experiment was ineffective in suppressing reproduction in female mule deer. Future experiments should investigate the effects of higher levels of GnRH-PAP on reproductive performance in this species. �112 LITERATURE CITED Anderson, A. E., and D. E. Medin. 1966. The breeding season in migratory mule deer. Outdoor Facts, Number 60, Colorado Division of Wildlife. Baker, D. L. 2001. Technical support for deer population management at the Rocky Mountain Arsenal National Wildlife Refuge, Denver, Colorado. Pages 215-232 in Wildlife Research Report, Mammals Research, Federal Aid Projects, Job Progress Report, Project W-153-R-4, SP 1, Jl. Colorado Division of Wildlife, Fort Collins, Colorado, USA. _ _ _. 1996. Regulation of mule deer population growth by fertility control: laboratory, field, and simulation experiments. Pages 81-87 in Wildlife Research Report, Mammals Research, Federal Aid Projects, Job Progress Report, Project W-153-R-4, SP 1, Jl. Colorado Division of Wildlife, Fort Collins, Colorado, USA. Bomford, M. 1990. A role for fertility control in wildlife management? Department of Primary Industries and Energy Bureau of Rural Resources Bulletin No 7 Australian Government Publishing Service Canberra Australia.. Collier, R. J., and D. A. Kaplan. 1984. Immunotoxins. Scientific American 251 :64. Cowan, I. McT. 1956. Life and times of the coast black-tailed deer. Pages 56-78 in The Deer ofNorth America, ed. W. P. Taylor. Wildlife Management Institute, Washington, D. C. Fagerstone K. A., M. Coffey, P. Curtis, R. Dolbeer, G. Killian, L.A. Miller, and L. Wilmot. 2001. Wildlife contraception. Wildlife Society Technical Review. Proceedings of the Wildlife Society 8th Annual Conference, Reno, USA. Garrot, R. A. 1995. Effective management of free-ranging ungulate populations using contraception Wildlife Society Bulletin 23 :445-452. Hone, J. 1992. Rate of increase and fertility control. Journal of Applied Ecology 29:695-698. Hobbs, N. T., D. C. Bowden, and D. L. Baker. 2000. Effects of fertility control on populations of . ungulates: general, stage-structured models. Journal of Wildlife Management 64: 473-491. Knox, W. M., K. V. Miller, and R. L. Marchinton. 1988. Recurrent estrous cycles in white-tailed deer. Journal of Mammalogy 69:384-386. Lehner, P. N. 1996. Handbook of Ethological Methods. Second Edition. Cambridge University Press, . Cambridge, UK. Milliken, G. A., and D. E. Johnson. 1984. Analysis ofMessy Data. Volume I. Designed Experiments. Lifetime Leaming Publications, Belmont,California, USA Morrison, D. F. 1976. Multivariate Statistical Methods. McGraw-Hill Book Co., New York, USA Pastan, I., M. C. Willingham, and D. J. FitzGerald. 1986. Immunotoxins. Cell 47:641-648. Plotka, E. D., U.S. Seal, G. C. Schmoller, P. D. Karns, and K. D. Keenlyne. 1997. Reproductive steroids in the white-tailed deer ( Odocoileus virginianus borealis). l. Seasonal changes in the female. Biology ofReproductiorr 16:340-3_43. ·.' • . Thomas, D. C., and I. McT. Cowan. 1975. The pattern ofreproduction in female Colum.bian black-tailed deer, Odocoileus hemionus columbianus. Journal of Reproduction and Fertility 44:261-272. Nett, T. M., D. L. Baker, and M.A. Wild. 2001. Evaluation ofGnRH-PAP as a chemosterilant in captive mule deer (Odocoileus hemionus hemionus). Proceedings of the 5th International Symposium on Fertility Control in Wildlife, Kruger National ·Park, South Africa, August 19-22. Niswender G.D., L. E. Reichert, Jr., A. R. Midgley, and A. V. Nalbandov. 1969. Radioimmunoassay for bovine and ovine luteinizing hormone. Endocrinology 84: 1166-1173. Willard, S. T., R. G. Sasser, J. T. Jaques, D.R. White, D. A. Neuendorff, and R. D. Randel. 1998. Early pregnancy detection and hormonal characterization of embryonic-fetal mortality in fallow deer (Dama dama). Theriogenology 49:861-869. � https://cpw.cvlcollections.org/files/original/27f8d0d9a75721e44f08512ea663637a.pdf 5daa8aea2df6d3d0287e0e4d503044c1 PDF Text Text 113 Colorado Division of Wildlife Wildlife Research Report July 2001 and July 2002 PROGRESS REPORT State of ------~C~o~lo~r~a=do-'------Work Package No. Division of Wildlife- Mammals Research 3001 Deer Conservation T~k ________~l~0'------ Chronic W~ting Dise~e in Mule Deer Federal Aid Project ___W~-1~8~5~-=R'----- Monitoring & Management Period Covered: July 1, 2000 through June 30, 2001 Author: Michael W. Miller, D.V.M. Personnel: T. R. Davis, L. L. Wolfe, T. H. Baker, K. T. Larsen, E. S. Williams Interim Report - Preliminary Results This work continues, and precise analysis ofdata has yet to be accomplished. Manipulation or interpretation of these data beyond that contained in this report should be labeled as such and is discouraged. ABSTRACT We continued conducting research on various ~pects of chronic w~ting dise~e (CWD) epidemiology and management. Results of original research, ~ well ~ two review articles, were published or accepted for publication during this segment, and citations are included in the body of the report. In addition to published studies, we completed a study of CWD pathogenesis in mule deer. Seven of 10 orally inoculated deer that survived > 12 mo postinoculation (Pl) developed clinical CWD. Five of the seven deer that showed clinical signs either died or were euthanized in end-stage clinical CWD 20-26 mo PI; the other two were euthanized showing mild or marked clinical signs 20 mo PI according to the established sampling schedule. B~ed on observations of the seven deer that developed clinical CWD, earliest signs were first noticed in individuals about 14.5 to 19 mo PI (mean ± SE = 17.3 ± 0. 7 mo PI). Early clinical signs were both subtle and inconsistent. As clinical dise~e progressed, behavioral changes and loss of body condition became more pronounced and more consistent. Ptyalism (drooling), polydypsia (excessive water consumption), and polyuria (excessive urination), widely regarded~ "classic" signs of CWD, occurred relatively late in clinical courses and were not seen in all c~es. Among the five deer that lived long enough to develop terminal CWD, clinical courses ranged from about 3.5 to 9.5 mo (mean± SE = 5.7 ± 1.2 mo); the shortest clinical course (about 3.5 mo) w~ complicated by acute ~piration pneumonia Immunohistochemistry and histopathology results are pending . .\ I I I. ��115 INTRODUCTION We continued conducting research on various aspects of chronic wasting disease (CWD) epidemiology and management. METHODS Epidemiology & Management Two review articles on CWD epidemiology were accepted for publication during this segment in the Journal of Wildlife Management and in the Revue Scientifique et Technique Office international des Epizooties. Results of three earlier studies on CWD epidemiology were published during this segment in the Journal of Wildlife Diseases and the Journal of Wildlife Management. Pathogenesis & Diagnosis Results of two original studies on CWD diagnosis in mule deer were accepted for publication in the Journal of Wildlife Management and in The Veterinary Record. We completed a study of CWD pathogenesis in mule deer. The methods for this project were included in previous progress reports. We also initiated preliminary field work to develop and evaluate reliable methods for collecting tonsil biopsies from live mule deer for use as a diagnostic and management tool. No results are available for reporting in this reporting period. RESULTS AND DISCUSSION Epidemiology & Management Two review articles on CWD epidemiology were published during this segment: Williams, E. S., and M. W. Miller. 2002. Chronic wasting disease in deer and elk in North America. Revue Scientifique et Technique Office international des Epizooties 21:305-316. Williams, E. S., M. W. Miller, T. J. Kreeger, R.H. Kahn, and E.T. Thorne. 2002. Chronic wasting disease of deer and elk: A review with recommendations for management. Journal of Wildlife Management 66:551-563. Results of three studies on CWD epidemiology were published during this segment: Conner, M. M., C. W. McCarty, and M. W. Miller. 2000. Detection of bias in harvest-based estimates of chronic wasting disease prevalence in mule deer. Journal of Wildlife Diseases 36:691-699. Gross, J.E., and M. W. Miller. 2001. Chronic wasting disease in mule deer: disease dynamics and control. Journal of Wildlife Management 65:205-215. Miller, M. W., E. S. Williams, C. W. McCarty, T. R. Spraker, T. J. Kreeger, C. T. Larsen, and E. T. Thorne. 2000. Epizootiology of chronic wasting disease in free-ranging cervids in Colorado and Wyoming. Journal of Wildlife Diseases 36:676-690. Pathogenesis & Diagnosis Nineteen of 20 mule deer orally inoculated with 5 g brain homogenate from CWD-infected mule deer survived ~ 3 months postinoculation (Pl) and were examined as described in the original study plan; �116 one fawn died < I day PI of capture-related complications, and was not evaluated here. Of the 19 . remaining deer, one died from a cervical fracture at 3 mo PI and 12 others were euthanized at 3, 6, 9, 12, 16, or 20 mo PI according to the study schedule; the other 6 were allowed to survive to terminal stages of CWD or study termination. Seven of 10 orally inoculated deer that survived > 12 mo postinoculation (PI) developed clinical CWD; four of these were males and three were females. Of the three that did not show at least early clinical signs, two were euthanized 16 mo PI but the third appeared clinically normal when euthanized 26 mo PL Five of the seven deer that showed clinical signs either died or were euthanized in end-stage clinical CWD 20-26 mo PI; the other two were euthanized showing mild or marked clinical signs 20 mo PI according to the established sampling schedule. Based on observations of the seven deer that developed clinical CWD, earliest signs (dullness in eyes, diminished alertness, misdirected behaviors, piloerection) were first noticed in individuals about 14.5 to 19 mo PI (mean ± SE = 17.3 ± 0. 7 mo PI). Early on, clinical signs were both subtle and inconsistent. As clinical disease progressed, behavioral changes (e.g., blank staring, uncharacteristic or subdued responses to aversive stimuli, lowered head or other unusual postures, ataxia, inefficient foraging activity) and loss of body condition became more pronounced and more consistent. Ptyalism, polydypsia, and polyuria occurred relatively late in clinical courses, and were not seen in all cases. Among the five deer that lived long enough to develop terminal CWD, clinical courses ranged from about 3.5 to 9.5 mo (mean ±SE= 5.7 ± 1.2 mo); the shortest clinical course (about 3.5 mo) was complicated by acute aspiration pneumonia. Immunohistochemistry and histopathology results are pending. Results of two original studies on CWD diagnosis in mule deer were accepted for publication. One of these (Wolfe et al. 2001) represents the first report of a method for detecting CWD infection in live animals. These publications were: Miller, M. W., and E. S. Williams. 2002. Detecting PrPcwn in mule deer by immunohistochemistry of lymphoid tissues. Veterinary Record 151:610-612. Wolfe, L. L., M. M. Conner, T. H. Baker, V. J. Dreitz, K. P. Burnham, E. S. Williams, N. T. Hobbs, and M. W. Miller. 2002. Evaluation of antemortem sampling to estimate chronic wasting disease prevalence in free-ranging mule deer. Journal of Wildlife Management 66:564-573. � https://cpw.cvlcollections.org/files/original/7242157943a6bfd2a6fc01c60a031612.pdf 37f743c448a3d91a449982ea362e42c9 PDF Text Text 117 Colorado Division of Wildlife Wildlife Research Report July 2001 and July 2002 JOB PROGRESS REPORT State of _ _ _ _ _ _ _c-·o-l~o~ra~d_o_ _ _ __ Mammals Research - Terrestrial Section Work Package No. ---~3"""0""-0.e....1_ _ _ __ Task No. _ _ _ _ _ _ _____:A...:.__ _ _ __ Deer Conservation Deer Aerial Survey Populatim1 Estimation Rangely· Deer Data Analysis Unit D~6, GMU 10 Project No. _ _ _ _ _W"---'-1=5....3"""'-....R--=1~4_ _ __ Period Covered: July 1, 2000-June 30, 2001 Author: D. J. Freddy Personnel: V. Graham, W. deVergie, J. Ellenberger, C. Wagner, P. Schnurr, R. Kahn, R. Velarde, G. Miller, T. Wygant, F. Pusateri ofCDOW, Dr. G. White and M. Kneeland of Colorado State University, J. Unsworth, Idaho Department of Fish and Game, consultants V. Howard, Jr. and T. Bickle, Colorado Mule Deer Association, Colorado Bowhunters Association, Dynamic Aviation, and New Air Aviation. ABSTRACT Sportsmen expressed concerns about the credibility of Colorado's survey sampling methodology to estimate numbers of mule deer (Odocoileus hemionus) in specific populations. We therefore conducted an aerial survey in Colorado Deer Analysis Unit D-6 which was an area of concern to sportsmen. We used helicopters from 28 February to 5 March 2001 to count mule deer on randomly selected quadrats 0.25-mi 2 or l .00-mi2 in size distributed within 11 strata encompassing 364 mi 2 of deer winter range composed of sagebrush (Artemisia tridentata) and pinyon-juniper (Pinus edulis-Juniperous osteosperma) habitats. From these counts, we estimated population size using standard stratified random sample estimators and the Idaho mule deer sightability model. Stratified population estimate was 6,782 ± 2,497 (90% CI) deer. Counts corrected for sightability increased the estimate to 11,052 ± 3,503 (90% CI) deer. Both aerial survey estimates buttressed population estimates of 7,000 to 7,300 deer derived from computer models and were substantially greater than sportsmen's estimate of 1,750 deer. Cost of this validation exercise exceeded 50,000 $US. We interpreted this exercise as a forerunner of the public's interest in challenging agency integrity or methods used to estimate status of ungulate populations. We caution agencies to use tested methodology that can withstand dispassionate public scrutiny. Copies of this report containing the original colored versions of the figures are available for review from the Research Center Library, Colorado Division of Wildlife, 317 West Prospect Road, Fort Collins, CO, 80526, USA. All information in this report is preliminary and subject to further evaluation. l~f il~i BD0WD176 □·2 ��119 PROJECT SUMMARY REPORT DEER AERIAL SURVEY POPULATION ESTIMATION RANGELY DEER DATA ANALYSIS UNIT D-6, GAME MANAGEMENT UNIT 10 COLORADO DIVISION OF WILDLIFE PRESENTED APRIL 19, 2001 DOCUMENTPREPAREDTO/NFORM COLORADO DIVISION OF WILDLIFE COLORADO WILDLIFE COMMISSION COLORADO MULE DEER ASSOC/AT/ON COLORADO BOWHUNTERS ASSOC/AT/ON Includes Draft Manuscript Subject to Future Editorial Review© REPORT PREPARED BY DAVID J. FREDDY WILDLIFE RESEARCHER, MAMMALS RESEARCH COLORADO DIVISION OF WILDLIFE, 317 WEST PROSPECT STREET, FORT COLLINS, CO 80526 DEER AERIAL SURVEY POPULATION ESTIMATION RANGELY DEER DATA ANALYSIS UNIT D-6, GAME MANAGEMENT UNIT 10, 28 FEBRUARY- 5 MARCH 2001 REPORT CONTENTS SECTION A ---- EXECUTIVE SUMMARY SECTIONB-- DRAFT TECHNICAL MANUSCRIPT SECTION C ------- 1. STRATIFIED RANDOM SAMPLING CALCULATIONS 2. MODIFICATION OF SAMPLE SIZE DURING PROJECT SECTIOND-- 1. D-6, UNIT 10 DEER WINTER RANGE MAP 2. D-6, UNIT 10 DEER WINTER CONCENTRATION AREA MAP 3. D-6, UNIT 10 DEER WINTER SEVERE Rf'NGE MAP SECTION E --------- 1. SAMPLING FRAME AND SAMPLING STRATA MAP 2. SAMPLING FRAME AND PRIMARY VEGETATION TYPES 3. YAMPA MONUMENT STRATA 1 AND 2 MAP 4. UTAH WHITE RIVER STRATA 3 AND 4 MAP 5. UPPER WHITE RIVER STRATA 5, 6, AND 7 MAP 6. MASSADONA- DINOSAUR STRATA 10, 11, AND 13 MAP 7. TWELVEMILE STRATA 12 MAP (Clarification: No strata numbered 8 & 9; 11 total strata) SECTION F - - - - 1. SURVEY FLIGHT PROTOCOLS 2. SURVEY DATA FORM 3. SURVEY OBSERVER HELP SHEET SECTION G -------- SURVEY FLIGHT QUADRAT SAMPLE UNIT MAP INDEX SECTION H ---- 1. STRATIFIED RANDOM SAMPLE POPULATION ESTIMATE DATA AND CALCULATIONS 2. LETTER FROM IDAHO DEPARTMENT OF FISH & GAME SHOWING POPULATION ESTIMATE USING IDAHO SIGHTABILITY CORRECTIONS 3. COMPLETE DATA LISTING FOR AERIAL SURVEY 4. MAP SHOWING WHERE DEER WERE COUNTED 5. MAP SHOWING WHERE ELK WERE COUNTED 6. SUMMARY OF SPORTSMEN AND CDOW DEER POPULATION ESTIMATES FOR WESTERN COLORADO, DECEMBER 2000 �120 SECTION A - EXECUTIVE SUMMARY DEER AERIAL SURVEY POPULATION ESTIMATION RANGELY DEER DATA ANALYSIS UNIT D-6, GAME MANAGEMENT UNIT 10 A. Sportsmen in Colorado alleged that estimates for numbers of mule deer in western Colorado were substantially over-estimated by the Colorado Division of Wildlife (CDOW). Sportsmen believed there were only 128,000 deer in Colorado in areas west of the Continental Divide where CDOW estimated 409,000 deer. This level of discrepancy also existed for specific deer populations such as in the Rangely Deer Analysis Unit D-6 where sportsmen estimated 1,750 deer compared to CDOW estimates of 7,000 deer. 8. A series of meetings between CDOW and sportsmen from September 2000 through February 2001 did not resolve fundamental issues of sportsmen's mistrust of estimated deer population status. C. On February 16, 2001 CDOW Director Russell George authorized the Terrestrial Section to implement aerial surveys to estimate numbers of deer in Rangely Unit D-6 in accordance with survey methodologies agreed to by all interested parties, including participation in surveys by individuals independently representing sportsmen's concerns. Financial costs for the survey were paid primarily with Wildlife Commission Discretionary Funds with additional contributions from the Colorado Mule Deer Association and Colorado Bowhunters Association. D. CDOW conducted an aerial survey to estimate numbers of deer in D-6 using Colorado quad rat survey techniques that incorporated adjustments in estimated population size based on Idaho mule deer sightability models as requested by sportsmen. The survey was conducted 28 February to 5 March, 2001. E. Estimated numbers of deer in D-6 were 6,782 ± 2,497 based on Colorado quadrat system and 11,052 ± 3,503 when adjusted for the Idaho mule deer sightability model. Population estimates based on CDOW computer models were 7,000 to 7,312 deer. All estimates were substantially higher than the 1,750 deer estimated by sportsmen. F. Financial and personnel costs to design, implement, and analyze survey results likely exceed $50,000. Final costs estimates are not yet available. G. This validation exercise challenged the credibility of CDOW personnel and methodologies and the credibility of sportsmen groups. All parties participated within a certain level of risk. Not to be overlooked was a near fatal helicopter incident that threatened lives of personnel involved in an aerial survey conducted to alleviate mistrust among interested parties. H. We interpret this validation exercise as a potential forerunner of the public's interest in either challenging or understanding methods used to estimate status of wildlife populations. We can only caution that wildlife agencies should gather information using methodology that can withstand public scrutiny. We would hope that this exercise would restore a certain level of public confidence in the CDOW's efforts to manage wildlife in Colorado. �121 SECTION B - DRAFT TECHNICAL MANUSCRIPT April 18, 2001 Draft David J. Freddy Colorado Division of Wildlife 317 West Prospect Road Fort Collins, CO 80526 970-472-4346, FAX 970-472-4457 RH: Deer Population Estimates ESTIMATING MULE DEER POPULATION SIZE USING COLORADO QUAD RAT SYSTEM CORRECTED FOR IDAHO MULE DEER SIGHTABILITY: A SPORTSMEN'S ISSUE. DAVID J. FREDDY1. Colorado Division of Wildlife, Wildlife Research Center, 317 West Prospect Road, Fort Collins, CO 80526, USA GARY C. WHITE, Department of Fishery and Wildlife Biology, Colorado State University, Fort Collins, CO 80523, USA MARY C. KNEELAND, Colorado Division of Wildlife, Wildlife Research Center, 317 West Prospect Road, Fort Collins, CO 80526, USA VAN K. GRAHAM, Colorado Division of Wildlife, 711 Independent Avenue, Grand Junction, CO 81505, USA WILLIAM J. deVERGIE, Colorado Division of Wildlife, P.O. Box 1181, Meeker, CO 81641, USA JOHN H. ELLENBERGER, Colorado Division of Wildlife, 711 Independent Avenue, Grand Junction, CO 81505, USA JAMES W. UNSWORTH, Idaho Department of Fish and Game, 3101 South Powerline Road , Nampa, ID 83686, USA CHARLES H. WAGNER, Colorado Division of Wildlife, 346 Count Road 362, Hot Sulphur Springs, CO 80451 PAMELA M. SCHNURR, Colorado Division of Wildlife, 711 Independent Avenue, Grand Junction, CO 81505, USA V. W. HOWARD, JR., 1025 Hickory Drive, Las Cruces, New Mexico 88005, USA TOMMY S. BICKLE, P.O. Box 750, Hatch, New Mexico 87937, USA 1 Corresponding author. Abstract: Sportsmen expressed concerns about the credibility of Colorado's survey sampling methodology to estimate numbers of mule deer (Odocoileus hemionus) in specific populations. We therefore conducted an aerial survey in Colorado Deer Analysis Unit D-6 which was an area of concern to sportsmen. We used helicopters from 28 February to 5 March 2001 to count mule deer on randomly selected quadrats 0.25-mi 2 or 1.00-mi2 in size distributed within 11 strata encompassing 364 mi2 of deer winter range composed of sagebrush (Artemisia tridentata) and pinyon-juniper (Pinus edulis-Juniperous osteosperma) habitats. From these counts, we estimated population size using standard stratified random sample estimators and the Idaho mule deer sightability model. Stratified population estimate was 6,782 .±. 2,497 (90% Cl) deer. Counts corrected for sightability increased the estimate to 11,052 .±. 3,503 (90% Cl) deer. Both aerial survey estimates buttressed population estimates of 7,000 to 7,300 deer derived from computer models and were substantially greater than sportsmen's estimate of 1,750 deer. Cost of this validation exercise exceeded 50,000 $US. We interpreted this exercise as a forerunner of the public's interest in challenging agency integrity or methods used to estimate status of ungulate populations. We caution agencies to use tested methodology that can withstand dispassionate public scrutiny. Key Words: bias, Colorado, helicopter surveys, Idaho, mule deer, Odocoi/eus hemionus, population estimates, sightability �122 Sportsmen in Colorado alleged that estimates for numbers of mule deer (Odocoileus hemionus) in western Colorado were substantially over-estimated by the Colorado Division of Wildlife (CDOW). For example during post-hunting season 2000, sportsmen believed there were only 128,000 deer in Colorado in areas west of the Continental Divide where CDOW estimated 409,000 deer. This level of discrepancy also existed for specific deer populations such as in the Rangely Deer Analysis Unit D-6, where sportsmen estimated 1,750 deer compared to CDOW estimates of 7,000 deer after hunting season 2000 (Pers. comm. Colorado Mule Deer Association). These 4-fold differences in estimated numbers of deer explained why perceptions about the status of mule deer in Colorado varied between some sportsmen and CDOW. Sportsmen focused their concerns on the credibility of Colorado's quadrat survey sampling methodology to estimate numbers of deer in specific populations. This methodology, based on stratified random sampling theory (Thompson et al. 1998), was initially developed for helicopter counts of mule deer on 1-mi2 sample quadrat units used to estimate total numbers of deer in a population inhabiting extensive sagebrush habitats during winter (Gill 1969). This system was later expanded to estimate size of selected deer populations inhabiting pinyon-juniper habitats in western Colorado where quadrat sample unit size was reduced to 0.25-mi2 to compensate for the detrimental effects that dense pinyonjuniper canopy cover had on detecting and counting mule deer (Kufeld et al. 1980, Bartmann 1983, Bartmann et al. 1986). Aerial counting of deer using random quadrats provided estimates of deer numbers sufficiently suitable for herd management decisions but implementation costs prevented such systems from being employed in most deer management units in western Colorado (Gill et al. 1983). Alternative approaches to estimating trends in numbers of deer in every population included: intensively estimating numbers of deer, age and sex ratios, and survival rates in a few populations whose trends in population parameters could represent many deer populations inhabiting ecologically similar areas (White and Bartmann 1998, Bartmann 2000, Bowden et al. 2000); and, using computer modeling that incorporated measured parameters from appropriately similar ecological core areas in conjunction with less intense measurements of age and sex ratios and hunter harvests that could be obtained yearly for nearly every deer population (CDOW 1991, White and Bartmann 1998, Bartholow 2000,). The discrepancy in perceived numbers of deer in western Colorado more accurately reflected a concern about modeled as opposed to aerial survey estimates of deer population size because only about 10% of the deer populations were monitored using aerial quadrat sampling protocols. Nevertheless, sportsmen focused their concerns on aerial survey sampling fearing that such techniques inflated estimates of deer numbers and therefore, misrepresented the declining plight of mule deer in western Colorado. Furthermore, sportsmen desired to assess the Idaho Mule Deer Sightability survey system (Ackerman 1988, Unsworth et al. 1994) as an alternative to Colorado's approach to estimating numbers of deer on the premise that the Idaho system would provide more acceptable estimates of deer numbers. This project was prompted by sportsmen's concerns about the legitimacy of deer population estimates based upon aerial surveys employing random sampling and counts of deer on sample quad rats. We conducted an aerial quadrat survey in a deer population unit of concern to sportsmen using Colorado quadrat survey techniques incorporating adjustments in estimated population size based on Idaho mule deer sightability models (Unsworth et al. 1994). Our survey and results were monitored by participating individuals independently representing sportsmen's concerns. We then compared estimates based on aerial surveys to ongoing population models used to guide management of deer. STUDY AREA We estimated numbers of mule deer inhabiting winter range in the Rangely Deer Analysis Unit D-6 consisting of Game Management Unit 10 in northwestern Colorado near the town of Rangely. D-6 includes 837 mi2 with large expanses of public lands administered by the U. S. Department of Interior Bureau of Land Management and National Park Service. Deer typically move onto winter range in November and begin returning to summer ranges in April. The project area is semi-desert with yearly precipitation ranging from 8 to 20 inches and winters having moderate temperatures and snow depths. Deer winter range occurs between 5,000 and 7,200 feet elevation and is a mixture of pinyon-juniper, sagebrush, and greasewood (Sarcobatus vermicu/atus) - desert shrub habitats and is shared with domestic sheep, cattle, and elk (Cervus elaphus). During �123 winters with low snow depths, deer distribution could encompass 612 mi2 but under severe snow depth conditions, deer distribution may collapse to 106 mi2 (CDOW 2001 ). Deer have been managed in D-6 under a limited permit hunting system since 1991 resulting in average yearly harvests of 118 bucks (range 70-252) and 60 antlerless does and fawns (range 2-132). Helicopter surveys of population ratios post-hunting season in December have shown 9-26 bucks:100 does and 29-64 fawns: 100 does. Previous efforts to estimate population size of deer in D-6 involved assessing the practicality of using helicopter line transects (White et al. 1989) on a trial basis in 1990 and 1991 with resulting estimates of 21,630 .±. 12,321 and 13,596 .±. 5,427 (90% Cl} deer, respectively (CDOW 1991 ). METHODS Sampling Protocols We estimated deer population size using stratified random sampling and counts of deer on randomly selected quad rat units (quad rats) (Thompson et al. 1998). Counts of animals on random units assume that units are completely searched by observers and all'animals present are detected and counted. These assumptions, therefore, assume 100% sightability of target anim·aIs and resulting estimates would not incorporate correction factors for animals not counted. We delineated our sampling area (frame) based upon the distribution of deer observed during systematic strip-surveys of potential winter range in the project area conducted with a Hiller 12-E Salay helicopter on 6 February 2001. Using ESRI ArcView©, we delineated a frame of 364 mi2 that encompassed the distribution of deer observed during the survey flight. Each cadastral square-mile within the frame was subjectively rated by flight observers as to high, medium, or low expected deer densities. Guidelines for relative deer densities were: >20 for high, 5-20 for medium, and < 4 deer/mi2 for low. We then defined 11 strata based upon expected deer densities for the purpose of distributing quadrats through out the frame. Low, medium, and high density strata encompassed 113 mi2 (31 %}, 157 mi2 (43%}, and 94 mi2 (26%}, respectively, within the frame (Table 1). Proportions of vegetation types within each strata were estimated using ArcView© and Colorado GAP® vegetation coverage (Schrupp et al. 2000). We used 1- mi2 quadrats in strata where open sagebrush-type habitats comprised > 50% of a stratum (Gill 1969) and 0.25-mi2 quadrats where pinyonjuniper habitats comprised > 50% of a stratum recognizing that tree canopy would hinder detection of deer (Bartmann 1983) (Table 1). We allocated quadrats among strata using optimum allocation (Thompson et al. 1998, pages 341342) with estimated variances based upon variance to mean ratios derived from quadrat sample units previously flown in Colorado since 1968 (Expected Standard Deviation = 3.6379 + 1.0891 " [mean deer density], n = 1, 192 qua drats). We calculated number of quad rats needed to achieve precision of.±. 20% of the mean population estimate with a = 0.10 for potential population sizes ranging from 2,000 to 8,000 deer. We selected a sample size of 161 quadrats distributed among 11 strata for an expected population of 6,000 deer. We assumed the cost of flying 1-mi2 and 0.25-mi2 quadrats was the same based on a proportional per area basis (Table 1). We established a grid of point coordinates (UTM [x, y]; NAO 27, all standardized to Zone 13) every 0.25 mi within the frame using ArcView©. We then used the random number option in MS Excel97© to assign a random number to each grid point. Grid point random numbers were then ordered from low to high to initiate the process of randomly selecting locations for quadrats within strata, with quadrat location selection beginning with the lowest ordered grid point and continuing until all quadrats were assigned within each strata. We restricted locations of quadrats by defining a minimum distance between randomly selected grid points of 0.50 mi in strata using 0.25-mi2 quadrats and 1 mi in strata using 1-mi2 quadrats to reduce quadrats having common boundaries and to reduce clustering of quadrats within strata. Quadrats were irregular in shape with boundaries following terrain ridges and gullies or cultural features such as roads or trails that could be discerned on USGS 1:24,000 topographic maps and recognized by observers while flying in a helicopter (Freddy 1994, Unsworth et al. 1994). Quadrat polygons were digitized on digital topographic maps using ArcView© with quadrat area and perimeters calculated by ArcView©. Range of actual areas for 0.25-mi2 units was generally 0.22-0.28 mi2 and for 1mi2 units, 0.9-1.1 mi 2 , shaped with the intent to minimize perimeter to area ratios (Thompson et al. 1998). Randomly selected grid points had to be included within the defined quad rat and preferably centered �124 within the quadrat. Flight path starting latitude-longitude coordinates, back-corrected for UTM Zone 12 or Zone 13 as necessary, were defined for each quadrat and labeled on flight navigation digital topographic maps printed in color using ArcView© layouts and MS PowerPoint97©. Flight Protocols We used Hiller 12-E Salay (Hiller) and Bell Jet Ranger Ill (Ranger) helicopters to count deer on quadrats. While searching for deer, helicopters were flown at 35-50mph at 50-100 feet AGL. Observer, navigator, and pilot comprised flight crews, with observer and navigator having primary responsibilities to detect and count deer with the observer tape-recording all pertinent data. In the Hiller, the observer was seated in the starboard outside seat with the navigator seated in the middle. In the Ranger, the observer was seated in the port outside seat with the navigator in the port rear seat. Crews first flew boundaries of quadrats and then systematically searched the interior of quadrats using strips or strip-contours depending on steepness of terrain following standard procedures (Gill 1969, Kufeld et al. 1980, Freddy 1998, Unsworth et al. 1994). To optimize the visual scanning position of the observer when flying quadrat boundaries, the Hiller crew flew boundaries clock-wise and the Ranger crew flew boundaries counter-clockwise. Navigators and pilots determined proper starting locations of quadrats using previously calculated latitude-longitude coordinates entered into on-board Garmin Pilot Ill© global positioning units (GPS). Navigators then directed pilots along quadrat boundaries and suitable search paths within the quadrat using topographic maps and real-time flight traces recorded on GPS units. Navigators, observers, and pilots constantly adjusted flight speed, altitude and angle of attack to optimize viewing for the observer. Objectives were to fly quadrats to obtain 100% search coverage. Observers and navigators collectively detected and counted groups of deer on quadrats with the highest count by either person recorded by the observer. Observers and navigators collectively made decisions on whether to count deer detected near quadrat boundaries: groups moving onto quadrats when detected were considered outside quadrats; groups moving off quadats when detected were considered on quadrats; one-half of the deer in groups detected on boundaries were considered on quadrats. Observers and navigators also collectively kept mental track of group locations, movements and presence of unique antlered deer in groups to reduce chances of counting groups more than once. Flights were conducted when weather conditions were favorable. Flights were conducted only when wind speeds were low enough in the judgement of pilots to fly safely at desired slow airspeeds and low AGL. Lighting conditions varied from overcast to hazy or bright sunshine while snow cover background varied from Oto 100 percent. Flights continued through short episodes of snow flurries provided safety was not compromised. For each quadrat, observers recorded flight conditions and total flight time. Idaho Sightability Protocols Sightability models correct for undercounting, or negative bias, that is generally associated with counts of ungulates (Caughley 1974, Bartmann et al. 1986, Samuel et al. 1987, Steinhorst and Samuel 1989, Unsworth et al. 1990, Otten et al. 1993, Pojar et al. 1995, White et al. 1989, Anderson and Lindzey 1996, Anderson et al. 1998, Cogan and Diefenbach 1998, Freddy 1998). To correct for potential negative bias in deer detected and counted on quadrats, we obtained values for sighting variables on each group of deer counted following guidelines for the Idaho mule deer sightability model (Unsworth et al. 1994). Sighting variables were total group size, behavior, vegetation type, and percent snow cover. Behavior of the most active deer when a group was first detected was recorded as bedded, standing, or moving. Although deer could have been detected in several vegetation types, we reduced types to broad categories to simplify the process of classifying vegetation: agricultural fields/open meadows; sagebrush, representing all low brush types; and pinyon-juniper, representing all pinyon or juniper dominated areas. Deer were not detected in tall conifer, aspen, or tall mountain brush habitats. Percent snow cover on the ground where each group was detected was classified as a categorical value of low (21-79%) or high~ 80%). The Idaho mule deer model most appropriate to correct undercounting deer was the spring sightability model which contained the following sighting variables (Unsworth et al. 1994: µ = -0.254 +activity+ vegetation class+ snow cover+ 0.047 * (group size) �125 Coefficients for each variable were developed in Idaho in similar but different vegetation and terrain types than might occur in Colorado. Knowing that Idaho coefficients may only approximate coefficients suitable for use in Colorado, we used the following Idaho coefficients for sighting variables (Unsworth et al. 1994): Activity: Bedded = 0.000, Standing = 1.56, Moving = 4.43 Vegetation: Agriculture/meadow= 0.00, Sagebrush = -0.88, Pinyan-Juniper (Idaho Juniper/Mountain mahogany) = -2.383 Snow Cover: Low 21-79% = -1.37, High?_ 80% = -0.60 Estimates of population size based on counts of deer on quadrats were corrected for sightability of each group using Idaho Aerial Survey© program for Windows© beta-version (Unsworth et al. 1994). Program Aerial Survey was limited to accepting only 10 defined strata from which to calculate population size. Our survey design incorporated 11 strata so we therefore, combined strata 1 and strata 2 (Table 1) into 1 strata to accommodate the program. We compared quadrat and quadrat sightability corrected population estimates using a standard z -test (Thompson et al. 1998). Both Idaho and Colorado systems were predicated on using stratified random sampling and thus, these systems complimented each other in conceptual design and application (Gill 1969, Unsworth et al. 1994). Modeling Protocols Estimating trends in deer population size over several years in D-6 was an ongoing CDOW management evaluation process based upon computer modeling using POP-II software (Bartholow 2000). Computer models were constructed independently of data obtained during our aerial survey and by personnel who did not participate in the aerial survey. This model used yearly hunter harvest estimates (Steinert et al. 1994), deer survival rates (White et al. 1987), estimated post-season December doe:fawn and buck:doe ratios collected during yearly helicopter surveys (CDOW 1991 ), and winter severity values to estimate trends in population size. Such models provided an assessment of deer status independent of aerial quadrat surveys, and conversely, aerial quadrat surveys provided point estimates of population size to evaluate models. RESULTS We estimated deer density from 28 February to 5 March 2001 using about 35 hours of helicopter flight time to complete the survey (Table 2). Mechanical malfunctions with the Hiller resulted in using the Ranger more extensively than anticipated, altered availability of survey navigators and observers, and in conjunction with impending unfavorable weather, caused us to reduce sampling intensity in 3 strata in order to complete the survey (Table 1). Adjustments in flight crew members and survey sampling were completed with the approval of independent evaluators. Survey Population Estimates Estimated deer population size was 6,782 with 90% CL of 4,285 to 9,279 for Colorado quadrats assuming 100% sightability of deer (Table 3). Reduced sampling in strata 10, 11, and 13 (Table 1) likely contributed to increasing variability resulting in wide Cl of.:!:. 36% of the mean estimate. Additionally, deer also became more concentrated in their distribution after the ·sampling frame flight of 6 February due to increasing snow depths at upper elevat_ion limits of winter range. This shift in deer distribution contributed markedly to not detecting deer on 66% of the quadrats. Using a 100% Idaho sightability model for deer, a point estimate of population size was 6,481 (Table 5). Colorado and ldhao 100% sightability population estimates would have been equivalent except the Idaho estimate was based only on 1O strata (strata 1 and 2 were combined, Tables 1, 4) instead of 11 Colorado strata due to limitations of program Aerial Survey©. Estimated deer population size was 11,052 with 90% CL of 7,549 to14,555 for Colorado quadrats corrected for the Idaho sightability mule deer model (Table 4). Confidence intervals represented.:!:. 32% of the mean estimate. Idaho sightability increased the standard Colorado quadrat estimate by 1.63x resulting in population estimates tending to be statistically different (z = 1.63, P = 0.103). Within individual strata, sightability increased estimates by 1.37 to 6.26x (Table 4 ). The highest correction 6.26x occurred in strata 11-MDL and should be viewed cautiously and may reflect the sensitivity of sightability correction factors to low counts of deer on few sample units (Table 3). �126 The considerable increase in estimated population size due to sightability corrections reflected that 62% of the deer groups contained ~ 5 deer, 34% of the groups were in pinyon-juniper vegetation and 66% were in sagebrush-type vegetation, and 82% were detected in areas having low and broken snow cover on the ground (Table 2). In essence, many groups of deer were associated with a factor that· decreased the sightability, or probability of detecting a group. Model Population Estimates Computer modeled point estimates of population size for post-season deer populations in 2000 ranged from 7,000 to 7,312 deer. Modeled estimates were similar in magnitude to aerial survey estimates and were within or nearly within the confidence intervals of all aerial survey estimates of population size. Modeled and aerial survey estimates were substantially larger than the population estimate promoted by sportsmen (Table 5). Flight Survey Variables Search times on quadrats were acceptable and comparable to previous surveys in Colorado. Flight crews spent 20.2 :t. 1.1 (SE, n = 38) and 6.4 :t. 0.2 (SE, n = 105) minutes on 1-mi2 and 0.25-mi2 quadrats, respectively. Search times were relatively proportional to total area searched. Wind and lighting conditions were conducive to effectively searching quadrats. Low percent snow cover or broken snow ground cover on many quadrats reduced the probability of detecting deer and made observers more dependent on deer movement to detect groups (Table 2). Compared to the Hiller, flight crews in the Ranger collectively had reduced visibility primarily because the navigator seated in the rear seat had a limited scanning view and could not as effectively help the primary observer detect or count deer. We would expect counts from the Ranger to be more negatively biased than from the Hiller (pers. comm. J. W. Unsworth). Conversely, the 200-shaft horsepower advantage of the Ranger allowed effective slow and low flying in steep and variable terrain. DISCUSSION We conducted an aerial survey in response to demands by sportsmen who strongly believed that methods used to estimate numbers of mule deer over-estimated deer numbers in Colorado. Resulting survey estimates of deer numbers, whether based on the Colorado quadrat system (Gill 1969, Kufeld 1980, Bartmann 1983, Bartmann et al. 1986) or quadrats adjusted for the Idaho mule deer sightablilty model (Unsworth et al. 1994), strongly indicated that sportsmen's estimates of deer numbers were substantially below likely true population size. Furthermore, aerial survey estimates supported population estimates derived in computer population models, and as such, supported the concept that models can provide reasonable estimates of population size to adequately guide decisions for managing mule deer. Colorado quadrats, as expected, provided lower estimates of deer numbers (6,782) than quadrats corrected for sightabiltiy factors (11,052). The Idaho mule deer model (Unsworth et al. 1994) increased estimates by 1.63x compared to the correction factor of 1.51 x developed for deer in pinyon-juniper habitats in Colorado (Bartmann et al. 1986). We are confident that Colorado quadrats, without sightability corrections, will provide conservative estimates of deer numbers· when- proper and adequate sampling procedures and flight protocols are followed. Although applying a calibrated correction factor (Bartmann et al. 1986) would improve accuracy of population estimates, we question whether higher estimates would be more palatable to some sportsmen's groups. We fully recognize the limitations of our population estimates generated from this validation exercise. Our estimates were of low precision for which there a 2 primary reasons. Our efforts to estimate a sampling frame were based on only 1 aerial survey conducted quickly in response to time constraints invoked by pressure to obtain a population estimate. Normally, quadrat sampling frames are determined with several years of deer distribution data obtained when winter sriow conditions would optimize counting deer. In Colorado, quadrat surveys would normally be flown in January when deer distribution and snow cover tend to be stable. In our specific case, major shifts in distribution of deer occurred after the distribution flight and prior to conducting the survey reducing the effectiveness of our sampling allocations. We then reduced sampling intensity in 3 strata to complete the survey with impending unfavorable weather conditions. �127 Observations also suggested that some deer moved off or outside of the frame which would inherently lower estimates of population size. Appropriateness of applying sightability correction factors developed in Idaho for Colorado deer in different habitats can be argued and therefore, the legitimacy of the resulting higher estimates may elicit even less confidence from concerned sportsmen. We foresee worthwhile potential research efforts emanating from this survey effort. Colorado and Idaho both use stratified random sampling procedures in their respective survey systems. However, Idaho often uses sampling units or quadrats having search areas > 3-mi2 while Colorado uses quadrats .'.': 1-mi2 . Cooperative experiments designed to compare effects of sample unit size on population estimates and precision, especially if simultaneously compared against robust mark-resight estimators (Bartmann et al. 1987, Neal et al. 1993, Bowden and Kufeld 1995), may provide valuable insight into designing more efficient aerial survey systems. We believe lessons from this exercise apply more appropriately to human dimension rather than biological issues. Sportsmen demanded a validation process of aerial survey protocols based on their perceptions of deer numbers and not on technical demerits of the survey system in use or reasonably obtained estimates of deer population size. Such demands were not tempered by discussions between CDOW and sportsmen over several months that attempted to resolve mistrust by explaining population estimation procedures, limitations, and likely biases. The result was CDOW spending approximately $50,000 in operating and personnel expenses to estimate numbers of deer in a management unit having low priority for spending limited deer inventory resources. We suspect that our survey exercise minimally mediated concerns of some sportsmen. MANAGEMENT IMPLICATIONS Sportsmen challenged estimates of mule deer populations provided by CDOW and demanded a validation exercise to compare sportsmen's estimates of deer numbers in a specific population with estimates on record with CDOW. Subsequent aerial surveys conducted with sportsmen approval and independent oversight provided deer population estimates that substantiated previous CDOW estimates and that were at least 4x greater than the estimate provided by sportsmen. We interpret this validation exercise as a forerunner of the public's interest in either challenging or understanding methods used to estimate status of ungulate populations. We can only caution that if estimates of population status are part of a routine management process, that estimates should be based on tested methodology that can withstand public scrutiny. ACKNOWLEDGMENTS This project was funded by Colorado Division of Wildlife Federal Aid in Wildlife Restoration Project W153-R, Colorado Wildlife Commission game cash funds, and Colorado Mule Deer Association. We thank R. Kahn, R. Velarde, G. Miller, T. Wygant, and F. Pusateri and their staffs of Colorado Division of Wildlife for project support. We also thank Idaho Department of Fish and Game for allowing the cooperative efforts of J. W. Unsworth. We thank New Air Aviation and Dynamic Aviation for providing survey helicopters. This paper dedicated in memory of M. W. Gratson who contributed significantly to developing helicopter sightability correction protocols for aerial surveys in Idaho. �----- --- 128 Table 1. Characteristics of sampling strata for estimating mule deer population size in Rangely Deer Analysis Unit 0-6, Colorado, Februa~-March 2001. Strata 1 Strata Name No. Yampa Monument 1 Yampa Monument 2 Utah White River 3 Utah White River 4 Upper White River 5 Upper White River 6 Upper White River 7 Massadona Dinosaur 10 Massadona Dinosaur 11 Massadona Dinosaur 13 Twelvemile 12 Totals 11 Density Rank Low Medium Low Medium High Medium Low Medium Low High Medium Strata Area 2 PJ Area 3 Open Mi 2 (%} (%} 28.74 30.40 31.64 22.75 54.16 30.96 31.96 44.09 20.07 40.46 28.93 364.15 66 59 10 35 78 40 20 69 62 54 47 34 41 90 65 22 60 80 31 38 46 53 Sample Unit Size Mi 2 0.25 0.25 1.00 1.00 0.25 1.00 1.00 0.25 0.25 0.25 1.00 Total Quadrat Units 115 122 32 23 217 31 32 176 80 162 29 1018 Sample 4 Quadrats 9 15 5 8 41 10 5 22 6 31 10 161 Sampled 5 Quadrats 9 15 5 8 41 10 5 15 1 21 10 143 Strata numbered 8 and 9 did not exist; there were 11 total strata. Percent of strata area having pinyon-juniper canopy vegetation types. 'Percent of strata area in sagebrush or low brush vegetation types. 'Number of sample quadrat units assigned to each strata based on optimum allocation formulas. 5 Represents number of sample quad rat units actually flown. Quad rats flown in strata 10, 11, and 13 were reduced by random selection to allow completion of aerial survey considering impending weather conditions. 1 2 Table 2. Summary of aerial survey characteristics for Rangely Deer Analysis Unit 0-6, Colorado, February-March 2001. Survey Characteristic Data Summary Aerial Survey Flight Dates Total Sample Quadrats Flown Search Minutes Per Quadrat Observers for Counting Deer Navigators for Counting Deer Flight Wind Speed on Quadrats Flight Lighting on Quadrats Snow Cover on Quad rats Time Period Quadrats Flown 28 February - 5 March 2001; about 35 hours of helicopter flight time. 143; 38 sized 1-mi 2 ; 105 sized 0.25-mi2; 129 flown by Ranger (90%), 14 flown by Hiller (10%) 1-mi2 quadrats = 20.2 .:!: 1.1 (SE); 0.25-mr quadrats = 6.4 .:!: 0.2 (SE) deVergie = 112 quadrats (78%), Graham= 14 (10%), Ellenberger= 17 (12%) Freddy= 82 quadrats (57%), Bickle= 30 (21%), Graham= 17 (12%), Howard= 14 (10%) Low= 136 (95%), Moderate= 7 (5%), High= O (0%) Bright sunshine= 92 (64%), Dull sunshine= 7 (5%), Hazy sunshine= 44 (31%) Fresh snow= 4 (3%}, Old snow= 139 (97%) 7 AM - 12 PM= 65 (45%), 12PM - 5PM = 78 (55%) Total Deer Counted on Quadrats Total Deer Groups Detected Deer Group Size by Quadrat Size Frequency of Group Sizes 1,180 seen on 48 of 143 sample quad rats 179; Average group size= 6.6 .:!: 0.6 (SE), range= 1 - 58 On 1-mi2 quadrats = 6.2 .:!: 0.6 (SE) (n = 94); On 0.25-mi 2 quadrats = 7.1 .:!: 1.0 (SE) (n = 85) (1, n=27, 15%), (2, n=25, 14%}, (3-5, n=59, 33%}, (6-9, n=36, 20%}, (10-19, n=22, 12%), (20-58, n=10, 6%) In sagebrush= 7.2 .:!: 0.7 (SE) (n = 118), In pinyon-juniper = 5.4 =. 0.9 (SE) (n = 61) In Ranger= 7.0 .:!: 0.8 (SE)(n = 127), In Hiller= 5.6 .:!: 0.6 (SE) (n = 52) Group Size by Vegetation Class Group Size by Helicopter Type Deer Group Vegetation Type Deer Group Snow Cover 123 (69%) Moving; 55 (31%) Standing; 1 (<1%) Bedded 118 (66%) sagebrush-type; 61 (34%) pinyon-juniper; (0%) agriculture/meadows 146 (82%) low snow cover; 33 (18%) high snow cover Total Elk Counted 1,297 approximately; seen on 32 of 143 sample quadrats Deer Group Behavior at Detection ---- �129 Table 3. Summary for stratified random sample of mule deer counted on sample unit quadrats in Rangely Deer Analysis Unit D-6, Colorado, February-March 2001. Strata Number With Abbreviated Name and Density Ranking 2 Summary Statistics 3 4 YML YMM UWLUWM Quadrat Sampled Units (uh) 9 Deer Counted Per Stratum(Nh) 114 Mean Deer Per Quadrat (Nhtoa,) 12.67 Quad rat Unit Variance (S 2Nh) 1116 Estimated Deer Per Stratum (N\) 1456 Stratum Variance (Var")(N\) 1510126 Stratum Quadrat Size (Mi 2) 0.25 Total Stratum Quadrat Units (Uh) 115 Stratum Area ( Mi2) 29 Quadrats With 0 Deer Counted 7 Percent Quadrats With 0 Deer 78 5 6 7 10 11 13 12 WRH WRM WRL MDM MDL MDH TMM 15 5 8 41 10 5 15 4 21 10 11 31 66 322 178 112 10 9 133 194 0.73 6.20 17.80 22.40 0.67 2.25 6.33 19.40 54 8.25 181 7.85 3 214 957 760 7 20 274 1020 89 196 188 1702 551 716 118 181 1025 561 2156 9137 7607 198905 62121 131021 12645 31002 297438 55874 0.25 1.00 1.00 0.25 1.00 1.00 0.25 0.25 0.25 1.00 122 32 23 217 31 32 176 80 162 29 30 32 23 54 31 32 44 20 40 29 11 2 5 27 4 1 14 3 15 6 73 40 63 66 40 20 93 75 71 60 Total Deer Counted All Strata (sum Nh) Total Estimated Deer All Strata (sum N\) Total Variance All Strata (sum Var"[N\]) Coefficient of Variation CV"(N") % 90% Confidence Interval for N" (Lower][Upper] 95% Confidence Interval for N" [Lower][Upper] 1,180 6,782 2,318,031 22.45 4,285 9,279 ±. 36% of Population Estimate 3,798 9,766 ±. 44% of Population Estimate Table 4. Estimates of mule deer numbers in individual strata compared between Colorado quadrat and quadrats corrected for Idaho mule deer sightability model and Idaho sightability estimate summary statistics for Rangely deer Analysis Unit D-6, Colorado, February-March 2001. Numbers of Mule Deer Estimated in Each Strata Numbered with Names Abbreviated 1 2 3 4 5 6 7 10 11 13 12 Estimator YML YMM UWL UWM WRH WRM WRL MDM MDL MDH TMM Colorado Quadrats Sightability Corrected Sightability Increase 1456 1422 1 0.92 89 196 293 1.49 188 375 1.99 1702 3183 1.87 551 851 1.54 716 1397 1.95 118 154 1.31 181 1133 6.26 1025 1403 1.37 561 6,782 841 11,052 1.50 1.63 Idaho Sightability Estimate Summary Statistics Total Deer Counted All Strata (sum Nh) Total Estimated Deer All Strata (sum N\) Total Variance All Strata (sum Var"[N\]) 1,180 11,052 4,534,872 Coefficient of Variation CV"(NA) % 90% Confidence Interval for N" [Lower][Upper] 19.27 7,549 Due to Sampling (4,365,014), Sightability (149,173). Model (20,685) 14,555 .:!: 32% of Population Estimate 1Sightability corrected estimate based on pooling strata 1 and 2 resulting in no estimate for strata 2-YMM. All Total �130 Table 5. Summary of computer modeled, aerial helicopter survey, and sportsmen estimates of December post-hunting season mule deer population size in Rangely Deer Analysis Unit D-6, Colorado, 1990 - 2001. Year 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2000 2000 2001 Computer1 Model 8,017 8,016 7,563 7,917 9,141 9,171 8,409 7,801 7,856 8,176 7,312 Computer3 Model 7,0004 Aerial Survey 90 Percent Confidence Interval 21,6305 13,5965 9.309 - 33,951 8,169- 19,023 6,7826 6,481 7 11,0528 4,285 - 9,279 Sportsman Estimate9 1,750 7,549 - 14,555 6,9892 1Computer model constructed in February 2001 with POP-II software (Bartholow 2000) . 2Estimate represents a value projected for December 2001 given assumptions about likely deer recruitment and harvest June to December 2001. 3Computer model constructed in February 2000 with POP-II software (Bartholow 2000). Estimate represented a value projected for December 2000 given assumptions about likely deer recruitment and harvest June to December 2000. Estimate represents value contested by Colorado Mule Deer Association. 5Estimate from helicopter line transects (White et al. 1989) conducted on a trial basis. 6 Estimate from Colorado helicopter quad rat survey technique assuming 100% deer sightability. • 7Estimate from Colorado helicopter quadrat survey technique assuming 100% Idaho mule deer sightability model (Unsworth et al. 1994). 8Estimate from Colorado helicopter quadrat survey technique adjusted for Idaho mule deer sightability model incorporating sightability correction factors (Unsworth et al. 1994). 9Estimate provided by Colorado Mule Deer Association on behalf of sportsmen. 4 �131 LITERATURE CITED Ackerman, B. B. 1988. Visibility bias of mule deer aerial census procedures in southeast Idaho. Dissertation, University of Idaho, Moscow, Idaho USA. Anderson, C.R., Jr., and F. G. Lindzey. 1996. A sightability model for moose developed from helicopter surveys. Wildlife Society Bulletin 24:247-259. Anderson, C.R., Jr., D.S. Moody, 8. L. Smith, F. G. Lindzey, and R. P. Lanka. 1998. Development and evaluation of sightability models for summer elk surveys. Journal of Wildlife Management 62: 1055-1066. Bartholow, J. 2000. Pop-II for Windows©, version 1.0. Fossil Creek software. Fort Collins, CO USA. Bartmann, R.M. 1983. Appraisal of a quad rat census for mule deer in pinyon-juniper vegetation. Colorado Division of Wildlife Game Information Leaflet 109. Colorado Division of Wildlife, Fort Collins, CO USA Bartmann, R.M. 2000. Colorado mule deer population monitoring system procedures, user's manual. Colorado Division of Wildlife, Fort Collins, CO USA. Bartmann, R.M., L.H. Carpenter, R.A. Garrott, and D.C. Bowden. 1986. Accuracy of helicopter counts of mule deer in pinyon-juniper woodland. Wildlife Society Bulletin 14:356-363. Bartmann, R.M., G.C. White, L.H. Carpenter, and R.A. Garrott. 1987. Aerial mark-recapture estimates of confined mule deer in pinyon-juniper woodland. Journal of Wildlife Management 51 :41-46. Bowden, D.C., and R.C. Kufeld. 1995. Generalized mark-sight population size estimation applied to Colorado moose. Journal of Wildlife Management 59:840-851. Bowden, D.C., G.C. White, and R.M. Bartmann. 2000. Optimal allocation of sampling effort for monitoring a harvested mule deer population. Journal of Wildlife Management 64: 1013-1024. Caughley, G. 1974. Bias in aerial survey. Journal of Wildlife Management 38:921-933. Cogan, R.D., and D.R. Diefenbach. 1998. Effect of undercounting and model selection on a sightabilityadjustment estimator for elk. Journal of Wildlife Management 62:269-279. Colorado Division of Wildlife. 1991. POPII simulation model, DEAMAN database manager, and POPMOD parameter estimation workshop manual. Colorado Division of Wildlife. Fort Collins, CO USA Colorado Division of Wildlife. 2001. WRIS database. Colorado Division of Wildlife, Grand Junction, CO,USA Freddy, D.J. 1994. Estimating survival rates of elk and developing techniques to estimate population size. Colorado Division of Wildlife Game Research Report. July: 27-42. Colorado Division of Wildlife, Fort Collins, CO USA. Freddy, D.J. 1998. Estimating survival rates of elk and developing techniques to estimate population size. Colorado Division of Wildlife Game Research Report. July: 177-206. Colorado Division of Wildlife, Fort Collins, CO USA. Gill, R. 8. 1969. A quadrat count system for estimating game populations. Colorado Game, Fish, and Parks Game Information Leaflet 76. Colorado Division of Wildlife, Fort Collins, CO USA Gill, R.8., L.H. Carpenter, and D.C. Bowden. 1983. Monitoring large animal populations: the Colorado experience. Transactions North American Wildlife Conference 48:330-341. Kufeld, R.C., J.H. Olterman, and D.C. Bowden. 1980. A helicopter quadrat census for mule deer on Uncompahgre Plateau, Colorado. Journal of Wildlife Management 44:632-639. Neal, AK., G.C. White, R.B. Gill, D.F. Reed, and J.H. Olterman. 1993. Evaluation of mark-resight model assumptions for estimating mountain sheep numbers. Journal of Wildlife Management 57:436-450. Otten, M. R. M., J. 8. Haufler, S. R. Winterstein, and L. C. Bender. 1993. An aerial censusing procedure for elk in Michigan. Wildlife Society Bulletin 21 :73-80. Pojar, T.M., D.C. Bowden, and R.8. Gill. 1995. Aerial counting experiments to estimate pronghorn density and herd structure. Journal of Wildlife Management 59: 117-128. Samuel, M.D., E.O. Garton, M.W. Schlegel, and R.G. Carson. 1987. Visibility bias during serial sruveys of elk in northcentral Idaho. Journal of Wildlife Management 51 :622-630. Schrupp, D.L., W.A. Reiners, T.G. Thompson, LE. O'Brien, J.A. Kindler, M.B. Wunder, J.F. Lowsky, J.C. Buoy, L. Satcowitz, A.L. Cade, J.D. Stark, K.L. Driese, T.W. Owens, S.J. Russo, and F. D'Erchia. 2000. Colorado Gap Analysis Program: A geographic approach to planning for biological diversity- Final Report. USGS Biological Resources Division, Gap analysis Program and Colorado Division of Wildlife, Denver, CO USA. �132 Steinert, S.F., H.D. Riffel, and G.C. white. 1994. Comparison of big game harvest estimates from check station and telephone surveys. Journal of Wildlife Management 57:336-341. Steinhorst, R.K., and M.D. Samuel. 1989. Sightability adjustment methods for aerial surveys of wildlife populations. Biometrics 45:415-425. Thompson, W.L., G.C. White, and C. Gowan. 1998. Monitoring vertebrate populations. Academic Press, Inc., San Diego, California, USA. Unsworth, J.W., L. Kuck, and E.O. Garton. 1990. Elk sightability model validation at the National Bison Range, Montana. Wildlife Society Bulletin. 18:113-115. Unsworth, J.W., F.A. Leban, D.J. Leptich, E.O. Garton, and P. Zager. 1994. Aerial survey: user's manual, with practical tips for designing and conducting aerial big game surveys. Idaho Department of Fish and Game, Boise, ID USA. White, G.C., and R.M. Bartmann. 1998. Mule deer management-what should be monitored? Pages 104-118 in J.C. deVos, Jr., editor. Proceedings of the 1997 deer/elk workshop, Rio Rico, Arizona. Arizona Game and Fish Department, Phoenix, Arizona, USA. White, G.C., R.A. Garrott, R.M. Bartmann, L.H. Carpenter, and AW. Alldredge. 1987. Survival of mule deer in northwest Colorado. Journal of Wildlife Managemetn 51 :852-859. White, G.C., R.M. Bartmann, L.H. Carpenter, and R.A. Garrott. 1989. Evaluation of aerial line transects for estimating mule deer densities. Journal of Wildlife Management 53:625-635. ��..... .i::. w i i CALCULATIONS FOR TOTAL NUMBER SAMPLE OUAD~ATS AND SAMPLE SIZES FOR EACH STATUM SAMPLE SIZES BASED ON FORMULAS ON PAGES 341 AND 342 FOR STRATIFIED RANDOM SAMPLING IN: THOMP=s'""o;-,.N,....,- - - - - , 1 - - - - + - - - - · · · · · · · •• • • ···-···-t-----l .,,w,.,._,....L.-,o=.-=c,...,.W~H=IT=Ec-A-,-,N'""D=-c-=-.--,G=o=w"'"'A.,..,N..,.._-1~9=98.,.._--,M..,..O=N'""1=To-=-R=1=N=G..,..v=E=R=TE=s=RA=-c=T=E-=p=o-=p.,..,u,...LA'""T"'"1o""N,..,.S"".'--,A'-,,,C',-A-,,,.DE=,M,-,l=c'--P=R=E=s-=-s,-,,,..,.Nc-=-.-.-=s..,..A,-cN-'------+------+-----+-·----··-+------l oiEGo CALIFORNIA. usA. ·- •••• •••• r · · · - - - - · T - · · ' - ' - ' - ' : . C . : . : : . ; _ ; . _ , ; = - . c : : . = ~ . c . . . . c . . : = - ~ - ' - " - ' ~ ~ - - - - - + - - - - - + - - - - - 1 - - - - - - i - - - - - l ! ------1------+----f----··•······l------+---I 1--------1----+----+----+------+-----+- -----+ ESTIMATED SAMPLE VARIANCES FOR STRATA°f3ASECfON REGRESSION OF VARIANCE STANDARD DEVIATION (S) i ON MEAN DEER DENSITY FOR.i°192 QLJADRAT DEER SAMPLE UNITS PREVIOUSLY FLOWN IN COLORADO. ····•·····1-------i'i-----l ANALYSIS CONDUCTED BY G.C. WHITE. STRATA VARIANCE TO MEAN REGRESSION WAS S = 3.638 + 1.089 (MEAN DENSITY). , r········1 r -··· ••••••••• ,cc..:_.::.=..;.;c.;.'--'J..--1:----+------+i-----l----1 1------- ···---t----;----+----t------t------+-----,....---------+---------;------+----l j i BECAUSE SAMPLE SIZE CALCULATIONS ARE BASED ON ESTIMATES OF VARIANCES PER STRATA, ANO VARIANCE IS DEPENDENTLY RELATED TO DEER DENSITY, THEN WE SHOULD CALCULATE SAMPLES SIZES REci"OiRED FOR_A_ _ _ _ _--+1------····-··· ·-·, RANGE OF POPULATION SIZES. SPORTSMEN ESTIMATE ABOUT 2000 DEER IN UNIT 10 WHILE CDOW COMPUTER 1 MODEL ESTIMATES ARE 7-8000DEER.• THEREFORE THIS IS THE RANGE OF POPULATION SIZE T6ESTIMATE · • • - - 1 ; - - - - - + - - - - 1 - - - - - - + - - - - - - - - - < -·- ·~···· ~f.~F'~E S)ZE REQUIREMENTS. j i t-:Sc-:-A-:-:M=P'"'Ll""'NG=-:A=REA:-.,-F=RA=-,-,-M=E,...,B,..,A-=s=Eo·oiiJLF=Es'-"'R""'U"C"CAc=Rc'ccYc--=S,-,.,U=R.,.,VE=Y7 F=L-c,IG""'H-=T07W"'"'A.,..,S=-3c-o6cc-4-c-M=1L=E=sA~2--,0~Fc-!-W~H=1=cH~LO"WD~E~N=s=1TY~S=TRA=--='cT=A----1---------+-----1-----1 ENCOMPASS!:[) 113 MILESA2 (31%), MEDIUM DENSITY 157 MILESA2 (43%) AND HIGH DENSITY 94 MILES"2 (26%). i ESTIMATED DEER DENSITIES BELOW ENCOMPASS RANGES OF DENSITIES SEEN ON PINYON-JUNIPER WINTER RANGES SAMPLED WITH QUADRAT SAMPLE UNITS IN COLORADO. VALUES ALLOW ADJUSTMENTS IN SAMPLE SIZE CALCULATIONS BASED ON BEST ESTIMATES OF VARIANCES (S) ASSOCIATED WITH DIFFERENT EXPECTED DEER DENSITIES AND 2 DIFFERENT SIZED SAMPLE UNITS. DENISITES ON 1/4 Ml"2 UNITS EQUAL DENSITY P;_~c...M ...cl_"2c.../_4-'-.- · · - - - - + - - - - - + - ;- - - - - · 1 - ·----+----•·➔ ··~~~~~~~~~:~~~~~--i :VALUES IN TABLE ARE LINKED AND USED IN CALCULATIONS ONNEXTPAGES • ·\··--..... __,___ _ _ _ _ _ _ _ : ·- ·······-· I '"",_ _ _ __,__ __, 1----------=G""'u=e=ss=E!,,,o=-o=,E=E=,R=-D='cE=N=s'""1TY==,Mc+.-c-A""2..,.,,s=T=RA~TA-,--j ESTIMATED VARIANCES ISi PER SAMPLE UNIT SIZE PER RELATIVE DENSITY LOW MEDIUM HIGH PROJECTED LOW DENSITY ! MEDIUM DENSITY HIGH DENSITY J 1-----------11-:--:3:--::M-:-:11"'"A2:-t-1:--::s=1-=-M::-::1,.c-=2-t-:::9..,.4-=-M::-::1,.c-=2-+=-Po=-p=u-:-:LA_,..,,,T;:-::10,.,.>N=-+.-,-1,"'"'4..,.~.,.,.,A'""2-,-u=N=1r""' .... ,.,.1-=-M"'11Ac-=2-=-u"""N:-:::1T=-'1"c-c1,cc--4"'"'M~1,.-=-2 ON!i' 1 MfA2 UNIT ... 114 .~tz uN11 Ml"2 UN1T_-+_ _ _ _ _ 1- - - - l IF TRUE POP. SIZE = t ··-· 2000 TOTAL DEER 1.5 5 1 10 1895 5.00 6.36 14.53 4.05 5.27 9.08 I I 1-----:-40=0-=-o=rotAL ot:=E=R+------:4--+--1,--,1-+i--=-21-:----+----,4,-,-15=3,----+---,--=---l--,c-:,--=---+--=-=---+--c-=-::-:--t---cc-=c---j-----:,-,-,,..,.-4.73 15.62 9.36 26.51 7.99 6.63 i I 6000 TOTAL DEER 5 16 31 5991 =.=--l--="-'--------+----1 21.06 12.08 5.00 7.99 37.40 9.08 26.51 14.80 48.29 8000 TOTAL DEER 7 21 41 7942 5.54 11.26 9.36 I A ..... ! : �CALCULATION OF TOTAL NUMBER OF SAMPLE U!"'ITl~.-~"=)A::-::S-=-cS:-:::U=M=IN..,.G=-=-=D-:-clFc-=-F:-:::E:-:::R=EN_,.T,,_.,D...,,E:-:::E:-:::R,..,P,..,O.,.,P:-::U,e-LA,,,_,.,T,.,.,IO::-::N,...,S,,..,l=-ZE=S·----+------+-----t------l-----+-----; ~~.J?.J.'.SSOCIATED AVERAGE STRATA DENSITIES AND ESTIMATED VARIANCES (S) PER SAMPLE IJ..flll,T-'-S-'-IZ--'E-;;:--:;;-;--'-----:==----t----.;;---;;-;-'-=~--+----:--..,.-----':,-;;c:-::---t PonSize =2000 PonSlze =4000 PonSlze = 6000 PonSlze= mno STRATA DENSITY TOTAL TOTAL(U) (Uh,SNh) (Uh,S 2Nh) (Uh,SNt,) (Uh,S2Nh) (Uh,SNh) (Uh.s 2;,-· ···cuh.sNhl (Uh,S 2Nh) I i 1---YA-M-PA~A_dl,MR,i;OE,i.NALu-M=E-NT-+iG~l~S_.i..N~O,.,;,!.:!~.h.. .... ~RALA·o·NWK,-f.._.,M...,l..,LE..,A_2__ sA_M_P_L.,.E_U..,N...IT..,s_--"'v=A=LU=E'=---l---"'V..,A=LU=E.._ VALUE t VALUE VALUE 1 29 115 465 1882 --~543 2569 575 YAMPA-MONUMENT 2 MEDIUM 30 122 608 30391 806 5349 972 UTAH-WHITE RIVER 3 LOW 32 32 167 879 253 2022 287 UTAH-WHITE RIVER I 4 MEDIUM 23 23 207 1876! 151 1001 479 UPPER WHITE RIVER ; 5 HIGH 54 217 1378 8765 2027 18962 2617 UPPER WHITE RIVER 6 MEDIUM 31 31 281 f5541 483 7551 652 UPPER WHITE RIVER 7 LOW 32 32 168 888 255 2042 290 MASADONA-DINO 10 MEDIUM 44 176 882- - 4407 1170 7758 1410 MASADONA-DINO 11 LOW 20 80 325 1314 379 1794 401 h;--=rw~E,..,LV""E=M.,..,l'""'LEc--'--+-----..,.12cc---t-.,-,M-c=E~D~IU~M--2~9-+---~2=9--+---~26-,3-j---=23'-,0=7+-----4-5=2+----7-0~56 •••••• --- - 609 • MASADONA-DINO 13 HIGH 40 162 1029 6548 1514 14165 1955 TOTALS SUMS TOTALS SQUARED 11 364 1018 6773 33324611 34540 70268 8036 64653941 f(j:z47 106006872 VALUE ! 2873 7770 2610 10090 31604 13734 2636 11269 2006 12835 23609 VALUE 637 1137 356 603 3207 821 360 1650 445 767 2395 VALUE 3533 10641 4012 15981 47458 21752 4052 15434 2467 20329 35452 121036,· ••••• 12378 - 181112 163217028 i SAMPLE SIZE FORMULA FROM COCHRAN 1977 AND PAGE 341 THOMPSON ET AL. 1998 - - - - - - - - - t - - - - i ; - - - - - - r----1 - - - - - - - - + - - - - - + - - - - ·······--··•····--·-····- + - - - - - - f - - - - - - - + - - - - + - - - - - - + - - - - - - + - - - - - - + - - - - - + - - - - - - l - - - - 1 Total Sample 1-Q-'--u_a_dr_a_tU_n_lts___,_(u__,)_=_ _.,__·_··_··_--_·_··_·_-_-_ ••_••_••_••_••_••_••____~_-··_·_ _ _ _--l-_ _ _ _-+-_ _ _ _ 1--_ _ _ _+-_ _ _ _+ - - - - - -·--········ ··-··-··-··••1------1-----1 2 (eN/t)A2 + [sum1 toH (Uh*S Nh)] - - - - - - - - + - - - - - t - - - - t ' - - · · · · - · · · · - - - + - - - - - - + ' - - - - - t - - - - - t - - - - - - t - - - - - f - - - - •······························--+-----1----- I I .... w u, �w O'I CALCULATION OF J n, "'' •--~- •~ S"""PLE UNITS CONTINUED I I SAMPLES SIZES CALCULATED FOR ESTIMATED ERROR OF +/-20% OF MEAN ESTIMAT~._,E,__ _ _ _- + - - - - - - - - - - - - - - - 1 - - - - - - + - - - - - - - + - - - - - - + AT A.!P.ha = 0.10 T;;·rs45di>:f:fo ••••••• T • I ••••••••••••• "T AT Alpha= 0.05 t• 1.96 df>120 ··'······---t------+l------+-------+-----1----------11------+------+--------+------1 ······!·· ·-· POTENTIAL DEER POPI II T"" S121= ;mt O • 4000 6000 8000 Error B()und(e) +i: 20% 400 800 ·---'!1~2~0~0+-----'!'1~6~00~t===========t=========jf-_-_~----~----~---.,~~---_-_-_-_-_-_-_-_~""+f--~~~----~------"1-------~~~------~-_4,_-::_-::_-::_-::_-::_-::_-::_-::_-::_-::_:-r-----l WHEN t = 1.645 1.645 1.645 1.645 1.645 f--·•Error_Term~...,..(e"'7N./"'t):":2=•~·====-=5_=,9_1~2:'.7~__ 2_3_65_0_9_,__5_3_2_14_6-+--::_~~~~~~9""4"'6""0"'3"'7-t------+-------+------+-------+------1--------1-------il-----l WHENt=1.96 1.96 1.86 1.96 1.96 1---e=rr-o-r=Term ·1eN/-t1-,..-2---+---4-1-e4-9.,..._1_6_6_5_9=7+--3=7-4=s-4-4+----.6.6-'6.:38"'9act--------+-------11-------;--------11-------+------;-------+-----1 I Total Sample Units Cul ·samoleUnlts Aloha=0.10 SampleUnlta AIP!:tlll:"!J,06 1 ··----=+-------1! •-----t------+------t·-----t------>------1-----1 366 2101 ·1·e1 136 I 437 273 212 181 I ! - - - - - - - - - - - - - ; - '- - - - - t - - - - + - - - - + - - - - - - - ; - - - - - - + - - - - - t - - - - - - - 1 - - - - - - t - - - - - - ; - - - - - - - ; - - - - - - - + - - - - - - 1 Where For Each Pop. •s::,.,;;. Bound +/- 2n•L Iii> Aloha ,.n n<: Size Total Sample Quadrat Units (u) • 33324511 i 2 (eN/t)"2 + [sum1 to H (Uh*S Nh)J 41849 + 34640 ,~----------'T-~----,._·-----+------il------b--,~--- ~ - L ___ . . ------; Total Sample Units= 437 i I l ,------+------+-------1-------+-------+------1 I ! I �L __ . ! - i ] j ! CALCULATING NUMBER OF SAMPLE UNITS I l . - FOR INDIVIDUAL STRATA ·! I ' I ASSUME POPULATION OF 6000 DEER, Alpha= 0.10, AND TOTAL SAMPLE UNITS a 161, SEE PREVIOUS PAGE FOR CALCULATIONS I ASSUME COSTS OF FLYING QUADRATS IN EACH STRATUM ARE THE SAME i • I I COST TO FLY 1 Ml"2 QUADRAT IN OPEN VEGETATION ABOUT 25 MINUTES OF HELICOPTER TIME ! f COST TO FLY 1/4 Ml"2 QUADRAT IN PINYON JUNIPER VEGETATION ABOUT 20 MINUTES OF HELICOPTER TIME i I j i i ... _______ ! ·---·· TOTAL (U) SAMPLE I % UNITS i %AREA Ml"2 AREA Ml"2 Strata Test (Uh,SNhl .. __ J STRATA_ I DENSITY' TOTAL ! ' SAMPLED FLOWN MILE"2 !SAMPLE i.TNITS VALUE UNITS Sum Column GI AREA IGIS NO.(hl RANK I SAMPLED 2 115 7.9 9 YAMPA-MONUMENT l 1 LOW 29 575! t 9i 8! 12.6 YAMPA-MONUMENT 2 122 972! 15 4 I MEDIUM 30 13 151 i 51 i 14,3 5 ··--··· ••••••••••••••••••••••.. -·-·········· UTAH-WHITE RIVER 3 LOW i 32 32 2871 ...5 141 I 8 33.11 8 UTAH-WHITE RIVER 4 23 23 479! 33 8 I MEDIUM 10 UPPER WHITE RIVER 217 41' 41 5 HIGH i 54 2617' 19 19.0I I ! 101 I MEDIUM 31 33.1 i 10 10 UPPER WHITE RIVER 6 31 6521 331 ····· I 14.3\ UPPER WHITE RIVER 7 32 32 51 14 5 5 LOW 2901 I 12.6! 22 MASADONA-DINO MEDIUM 44 176 1410 22 13 6 10 i I ! -s'i ! 6 7.9: 2 6 MASADONA-DINO 11 LOW 20 80 40_1 I I I i 609; 10 TWELVEMILE 10i 33 33.1 I 10; 12 MEDIUM 29 29 I ! 19.0 MASADONA-DINO 162 31 I 19 8 31 13 HIGH 40 19551 _, i I j 16 161 ! 162!=actual sum I I ...... ··-----! ianomalv See Note Below ! FORMULA FOR ALLOCATING SAMPLE UNITS T9 .. ::§J.f.3ATAJ'_(N~Y.'!1~~-Allocation after Cochran 1977; In ThCl1_!1pson et al. 199~1 .P.~~e 342) : • I STRATA ALLOCATION FORMULA TOTAL SUMS ! 364.149 11 102471 1018 t Uh = U[ UhSNhl / (sum 1 to H UhSNhJ)] I EXAMPLE CALCULATION! ···-·· jUh -- i 68 162 I I I i I ! I ! i I i' i i I t I I I = 161*[575/ 10247] lfor Yampa-Monument Low Density Strata from above. !sample unlts._in this stratum = 9 i ···-· -----! i I i ' ' i i ' i l ! iI i I i NOTE: Due to weather conpromising ability to comolete necessarv fl'.'.f.ng of units on March 3 and 4, 2001, number of sample units flown In the Massadon-Dlnosaur strata were reduced to 21_, 15, and 4 in the high, m~~l. 1:1.r.n., and low dens._l_l}.t_S.tf!~, res.eectlvelv. This reduced total sample units flown to 143 (162.~~-9}'. __ ! . ____ ._________ L__ The 161 sum in column G Is an arithmetic anomalv that must have somethlnci to due with rounding corrections In automatic calculations .. The strata sample sizes when added as whole numbers in column L sum to 162 which was the number of quadrats org_lnallv drawn to be samoled. This problem discovered when we reduced sampling In Massadona strata by 19 quadrats and the total flown for the entire Unit 10 was 143 lie--161-19 did not eaual 143 flownl. : I J 1···· -·· -·~•-··-···•····· 'I I I I ; i' l ;J -.J �138 2. MODIFICATION OF SAMPLE SIZE DURING PROJECT lRandom Subsam~le of lll!t\.~SADONA-DINO Unit 10 Strata ! ---,-+- ·-----·-- -·· ... ·- --· -·., 1· .. -' iMarch 3, 2001 ··---... ····-----•-'"·--Reduce total sample from 59 quadrat units to 40 quadrat units proportionately across strata. Minimum of 4 sample units in Low Density Strata, reduced from 6 (1/3 reduction) Reduction in Low Densitv Established the Prooortion reduced in Hiah and Medium Densitvl Decisions Made and Techniaue of Random Subsamole comoleted with VW. Howard annroval and sunnort Samole reduction was made to allow completion of a reasonable number of samole units to ----·obtain an estimate within limited time frame and flvable weather patterns .. Used Quattro Sample Tool Random function to select subsamples within each of 3 strata .. • .·, •. •• •Howard, Bickle and Freddv discussed (3/2/01 )options for comoletina the Massadona Unit with reduced • samplina, Jnclusive of droooina entire strata ~md/or orooortional reduction in samoles within • each of 3 strata. Reducinq samples in each strata was considered preferable to droooino a strata. I i Howard, Bickle and Freddv also discussed merits of doina or not doina the Yamoa-Monument Strata tinder aiven time constraints. Howard thoui:Jht this was the least orioritv of strata because • ,. • Sportsmen do not believe deer assoicated with the Monument a part of the- available huntable deer • •. 1oooulation. However. CDOW does include Monument deer in the Unit 10 1deer oooulation model. At this staae of discussion Howard and Bickle comfortable with l not comoletina the Yampa-Monument strata. ·_i i Massadona Hiah Densitv Strata; Units Sorted/Ordered bv Samole Tool Selection Order Samole Tool Set to assian random order to a samole of 3t·and took first 21 because duolicate ·•· numbers will aet assianed so kmored ties due to assianed numbers Quads Selection Order Oriainal First 21 • Hiah Numeric Number Assianed Selected Usinq Quattro ListinQ Bv Random :Densitv !Quadrats Eauivalent SamoleTool Order 20-MDH 20 1 1 26-MDH 26 3 2 15-MDH 15 3 3 ·8,-MDH 8 5 4 10 5 5 • ' 10-MDH II 21-MDH. 21 6 6 7· 30--MDH 30 7" 24--MDH 24 9 8 ,9 31-MDH 31 9 19-MDH 19 11 1-0 7-MDH. 14 7 11 16-MDH 16 15 12 3-MDH 16 3 13 ' 4 .16 14 i 4-MDH • •11-MDH 16 15 11 25 18 16 i 25--MDH 19 22 17 '' 22--MDH 1-MDH 1 21 18 13 21 13-MDH 19 9-MDH 9 23 20 29[ 29-MDH 23 21 17-MDH 17 23 12-MDH 12 24 25 23-MDH 23 14-MDH 14 26 18-MDH 26 18 27 28 ' 28-MDH 2 27 I 2-MDH 5-MDH 5 29 311 6-MDH 6 31, 27-MDH 27 I I ' ., t ; I ' l ; i I I .• l I �139 ! •··-···-----···· ···-······· ·; i ·- ·- ··-·····--------···-·····J-........ ..-------·-·· ···-·--·-·········-······---..,-.......... , ......... __ .................... ···-·· --~-~ ! -~-~ ........ ·······-···· -~---. '----~-----------···L.·--------··· : Massadona Medium Density Strata; Units Sorted/Ordered by Sample Tool Selection Order !Sample Tool Set to a~~igl}_randomorder to a samole of 22 and took first 15 because dur::>Ji<2~'E____ . _ ----····· _ !numbers will oet assioned so ionored ties due to assioned numbers ' ! ----------+-----~- - · - - - - - - - - - - + - - - - - - - - - - - - - - - - + - ~ - - - - + - - - - - - - - - - - - - --~!-------1-----+----1-----·----------·-·-t-------l ; ! ------+-------'---------+-------'--+------!-~- ··+---------+-------< LPrigiJJ.§.I . ·····1-Q~u_a_d_s_ ___,i_S_e_le_c_ti_on_O_rd_e_r_1-F_ir_s_t~15_·- - + · - - - - 1 - - - - - + - - - - - - - - - - - ' ' - - - - ......_ t Hioh Numeric i Number Assianed Selected j Densitv Listino l Usina Quattro Bv Random 1 ;,-.: Q=-=u=a:..;:;d;.;..;ra::.::t-=-s-+.:::;E..c,.ou=i;..;_va=l=ec,c_ntc...+-S-=-a=m.;..c.1£..pl'""e_T'""'o'-'o"-I---1-..::.O..:..;rd;:.::e:_;_r_ _- 1 - - - - - + - - - - + - - - - - 1 - - - - - - + - - - - - • -.....- ....10-MDM 10 1 1. 11-MDM 11 2 2 7-MDM 7i 3 3 -+-----i-·-·--·1 20-MDM 20 4 4 ,-..::::.;;__;=..:.:.;:_--+-_ __ _ _ _ . : : ; c . c . + - - - - - - - - - - - - - ' - + - - - - - ' - - 4 - - - - - l - - - - - l - - - - - + - - - - - 21-MDM 21 _5_ _ _ _ _5-'-+-------l'-----+----1-----+--'--'-----i 14-MDM 14 6 6 3-MDM 3 8 7 I 13-MDM 13 8 8 ' 1-MDM 1 8 9 9-MOM 9 11 10 12-MDM 12 15 11 . •. 18-MDM 18 15 12 17-MDM 17 17 13 22-MDM 22 17 14 4-MDM 4 17 15 16-MDM 16 18 ----=2;:_-=M=D---'M'-'---+------"2'+-------'1..::.8+-------'------1------+-··· ..··.... --+-----~----<\ 1 5-MDM 5 19 , .__----+-----+-------+-----_,__~-------1-----l-----+-----•·--, ~8_-=M=D~M~+-----'-8t-----~2=0-+----.......----+------+----+-----------+---..~ 15-MDM 15 20 ~MOM 6 ~ 19-MDM 19 21 Massadona Low Densitv Strata· Units Sorted/Ordered by Sample Tool Selection Order Sample Tool Set to assion random order to a samole of 6 and took first 4 because duolicate numbers will aet assianed so ianored ties due to assianed numbers I Oriainal Hioh Densitv Quadrats 6-MDL 2-MDL 5-MDL 4-MDL 1-MDL 3-MDL Quads Selection Order First 15 Numeric Number Assioned Selected Listina Usina Quattro BvRandom Eauivalent Samele Tool Order 6 1 1 2 3 2 5 3 3 4 4 4 4i 1 3 5 �140 SECTION D 1. D-6, UNIT 10 DEER WINTER RANGE MAP Vegetation Types Within Mule Deer Winter Range in Game Management Unit 10, DAU D-6, Colorado Vegetation based on Gap Analysis Winter Range Boundaries based on COOW WRJS Data. Gap Vegetatioo Types - - Dry Land Oops Irrigated Q-ops i1M1 Foothills and Mruntains ~ Mesic. Upland Shrub ~ Billerbrush Shrub - Big Sagebrush Saltbrush Fats and Flats - Juniper Woodland Pinycn -Jtmiper - Shrub Dcminaled Wetland N Game Management Unit 10 CJ Greascwood Fans and Flats 2. D-6, UNIT 10 DEER WINTER CONCENTRATION AREA MAP Vegetation Types Within Mule Deer Winter Concentration Areas in Game Management Unit 10, DAU D-6, Colorado Vegetation based on Gap Analysis Winter Range Boundaries based on CDOW WRJS Data GAP Vegetalion 'I)pes g- Bitterbmsh Sbrub - Irrigated Crop Big Sagebrush Sallbrusb Fms and Flals - JuniperWoodlland CJ <hmewood F'"1S and Flals - Pinyon - Juniper - Shrub dominated Wetland N Gane Management Unit 10 �141 3. D-6, UNIT 10 DEER WINTER SEVERE RANGE MAP Vegetation Types Within Mule Deer Severe Winter Range in Game Management Unit 10, DAU D-6, Colorado Vegetation based on Gap Analysis Winter Range Boundaries based on CDOW WRIS Data Gap Vcgctatioo.Typcs -IrrigatedQ-op li!llil Bittabrush Shrub Big Sagebrush Saltbiushfais and Flats D Greasewood Fans and Flats Juniper Woodland Pinyon -Jw,ipcr Elli Shrub Dominated Wetland N Game Management Unit 10 �142 SECTION E 1. SAMPLING FRAME AND SAMPLING STRATA MAP Sampling Frame Area, Sampling Strata, Quadrat Sample Units, and Random Sample Unit Points used or Helicopter Counts of Mule Deer to Esitmate Population Size in Unit 10, DAU D-6, Colorado. Grid Shows Cadastral Square Miles and Subjective Rating of Deer Density Based Upon a Helicopter Survey Flight on 6 February 2001. Sampling Slnla and Sample Quadnint ua;,,, -Yanp••Mooumeol (Low) ■ Yanpa-Moolilileat (Medium) Ulah White River (Low) Ulali White River (Medium) Upper While Rive, (High) Upper While RMr (Medium) -UpperWhileRMo-(Low) Massadona-Dioo (Medium) Mmsadooa-DiDo (Law) ■ Twelvemil.• (M.edium) Massadona- DiDo (High) D Sample Quadrat ua;,,, Sample Quadrat Units Nat Flo"'1 D Cadastnil Square Mile Grid E3:::::i=:==:=3:::=:::::fJ'10 l<lofflltSI A • Sample Unit Points NGaneMaoogementUnit 10 2. SAMPLING FRAME AND PRIMARY VEGETATION TYPES Sampling Frame Area and Distribution of Pinyan Juniper and Jnniper Woodlands Within Strata used for Helicopter Counts of Mule Deer to Estimate Population Size in Unit 10, DAU D-6, Colorado. Sampling Slrala and Sample Qumlnmt ua;,,, -Yanpa-Mooumeol· (Low) ■ Yaopa-Monumeot (Medium) Ulah While River (Low) Ulllh While River (Medium) •upperWhilelliva- (High) Upper White Rive, (Medium) -Upper While Rivei-(Low) Massadona • Dino (Medium) -Massadooa-Dioo (I.aw) ■ T-Jvanile (Medium) Massadona-Dioo (High) EE]! J\nperWoodhmd ~ Playon. JDDiper �143 3. YAMPA MONUMENT STRATA 1 AND 2 MAP Yampa - Monument Sampling Strata Showing Grid of 1/4 Mile Points, .Randomly Selected Points, Sample Unit Points, and Quadrat Sample Units 1/4 Mile Square in Size, Circle 1/2 Mile in Diameter used to Restrict Distances Between Quadrat Units. Sampling Sira!a aad Sample Quadrant Units ■ Yainpa- Mooumerit (Low) Yanpa.-Mooument. (Medium) D • • o • Quadnll Sample Units Sample Unit Points Randomly Selected Points 1/4 Mile Gridded Points 1/2 mile ciamctcr 4. UTAH WHITE RIVER STRATA 3 AND 4 MAP Utah White River Sam piing Strata Showing Grid ot 1/4 Mile Points, Randomly Selected Points, Sample Unit Points, and Quadrat Sample Units 1 Mile Square in Size. Circle 1 Mile in Diameter used to Re.strict Distances Between Quadrat Units. ...... E=:=========i=== Sampling Strala aad. Sample Quadrant Units -Ulsh White River (Low) Ulsh White Riv..- (Medium) D Quadrat Sample Units • Sample Unit Points • Randomly Selected Points o 1/4 Milo Gridded Points N Gano Management Unit 10 �144 5. UPPER WHITE RIVER STRATA 5, 6, AND 7 MAP Upper White River Sampling Strata Showing Grid of 1/4 Mile Points, Randomly Selected Points, Sample Unit Points, and Quadrat Sample Units 1/4 and 1 Mile Square in Size. Circle ·112 and 1 Mile in Diameter .used to Restrict Distances Between Quadrat Units. - - 112 mile diameter I mile di""eter A Sampling Slrala ,ad Sample Quadrant Units Upper White River (High) Upper White River (Medium) Upper While River (Low) c;;J Quadnt-Sample Units • Sample Point Location o Randomly Selected Sample Points • 1/4 &pare Mile Gridded Points Game Management Unit 10 N 6. MASSADONA- DINOSAUR STRATA 10, 11, AND 13 MAP Massadona - Dinosaur Sam piing Strata Showing Grid of 1/4. Mile 'Points, Randomly Selected Points, Sample Unit Points, and QuadratSample Units 1/4 Mile Square in Size. Circle 1/2 Mile in Diameter used to Restrict Distances Between Quadrat Units. e;;;=1:=::iaaaaaaaaaaaaaii===4• Mllel >.. Sampling Slrala mcl Sample Quadrant Uniis Massadooa.-Dino (Medium) ~.a-Dmo (Low) -Massadoaa-Dino (Bish) c:::::J ~ Sample Un~ , ~ Sample Units Not Flown • _Satq,ie Unit Points • Randanly Selected Poinls • 1/4 Mile Gridded Points Game Management Unit 10 N �145 7. TWELVEMILE STRATA 12 MAP (Clarification: No strata numbered 8 & 9;11 total strata) Twelveniile Sampling Strata Showing Grid of 1/4 Mile Points, Randomly Selected Points, Sample Unit Points, and Quadrat Sample Units 1 Mile Square in Size. Circle 1 Mile in Diameter used to Restrict Distances Between Quadrat Units. Sampling Strata and Sample Quadrant Units - D i=======E=====~3 • • o Miles A N Twelvcmile (Medimn) Quadrat Sample Units Sample UnitPoints Randomly Selected Points 1/4 Mile Gridded Points Game Management Unit 10 �146 SECTION F 1. SURVEY FLIGHT PROTOCOLS PROCEDURES FOR FLYING HELICOPTER SAMPLE QUADRATS GMU 10, DEER POPULATION ESTIMATE FEBRUARY 2001 COLORADO QUADRAT WITH IDAHO SIGHTABILITY CORRECTIONS D.J. FREDDY, MAMMALS RESEARCH, COLO. DIV. WILDLIFE J.W. UNSWORTH, IDAHO DEPT. FISH & GAME (PROVIDED SIGHTABILITY TECHNIQUE SUGGESTIONS) FEBRUARY 8, 2001 I. BACKGROUND INFORMATION: AREAS OCCUPIED BY DEER IN UNIT 10 WERE DELINEATED (364 SQUARE MILES) AND STRATIFIED ACCORDING TO RELATIVE DEER DENSITIES INTO HIGH, MEDIUM, AN LOW STRATA (11 STRATA) BASED ON A HELICOPTER SURVEY FLIGHT CONDUCTED BY V. GRAHAM (CDOW) IN FEBRUARY 6, 2001. THE OCCUPIED DEER RANGE DELINEATED REPRESENTS A CONDENSED PORTION OF THE EXTENSIVE WINTER RANGE AREA DELINEATED IN THE CDOW WRIS INVENTORY SYSTEM BECAUSE SNOW-DEPTHS HAVE CONCENTRATED DEER TO SOME EXTENT. SAMPLE UNITS, OR QUADRATS, WITH AN AREA SIZE OF 1 SQUARE MILE (USED IN OPEN VEGETATION HABITATS) OR 1/4 SQUARE MILE (USED IN PINYONJUNIPER FORESTED HABITATS) WERE SELECTED AT RANDOM WITHIN EACH STRATUM ACCORDING TO STANDARD STATISTICAL SAMPLING FORMULAS. BOUNDARIES OF ALL SAMPLE UNITS WERE BASED ON TOPOGRAPHIC OR CULTURAL FEATURES. FROM THE STRATIFIED RANDOM SAMPLE OF QUADRATS WE WILL OBTAIN 2 ESTIMATES OF POPULATION SIZE FOR THE 364 SQUARE MILE AREA SAMPLED: 1) AN ESTIMATE BASED ON UNADJUSTED COUNTS OF DEER ON QUADRATS, AND, 2) AN ESTIMATE BASED ON COUNTS. OF DEER ON EACH QUADRAT ADJUSTED FOR SIGHTING OR DETECTION PROBABILITY OF EACH GROUP OF DEER USING A SIGHTING BIAS CORRECTION FACTOR (IDAHO MULE DEER SIGHTABILITY MODEL). II. AIRCRAFT: THE PRIMARY HELICOPTER WILL BE A HILLER 12E SOLOY AND THE SECONDARY HELICOPTER WILL LIKELY BE FRENCH A-STAR OR TWIN-STAR, ALL TURBINE POWERED AIRCRAFT. THE HILLER PROVIDES THE BEST VISIBILITY PLATFORM FOR COUNTING DEER AND WILL BE USED PRIMARILY IN AREAS OF HIGHER DEER DENSITY. FLIGHT CREWS WILL CONSIST OF A PRIMARY OBSERVER, NAVIGATOR, AND PILOT. ALL 3 PERSONS SHOULD SIT ABREAST IN THE HELICOPTER WITH THE NAVIGATOR POSITIONED IN THE MIDDLE. THE HELICOPTER MUST HAVE FUNCTIONING INTERCOM HEADSETS SO THAT ALL 3 MEMBERS CAN READILY COMMUNICATE VOICE INSTRUCTIONS OR INFORMATION. SUNNY AND COOL DAYS WITH GOOD SNOW BACKGROUND AND LOW WIND SPEEDS ARE THE MOST DESIRABLE FLYING AND COUNTING CONDITIONS. HAZY OR FLAT LIGHTING ON OVERCAST DAYS IS ACCEPTABLE, BUT NOT PREFERRED. CREWS SHOULD AVOID PUSHING TO GET WORK ACCOMPLISHED IF THERE ARE CONSTANT SNOW FLURRIES OR WIND SPEEDS THAT NECESSITATE FLYING AT HIGHER SPEEDS AND ELEVATIONS ABOVE THE GROUND. Ill. SAFETY: HELICOPTERS WILL HAVE A SURVIVAL GEAR BAG FOR SUPPORTING 3 PERSONS, A FUNCTIONING ELT, FIRE EXTINGUISHER, AND PORTABLE PACKSET FOR EMERGENCY COMMUNICATIONS. IT IS RECOMMENDED FLIGHT CREWS PERIODICALLY REPORT THEIR GENERAL LOCATION TO COUNTY SHERIFF DISPATCH VIA HELICOPTER RADIO IF POSSIBLE. FLIGHT CREW MEMBERS ARE ENCOURAGED TO WEAR NOVEX FLIGHT SUITS AND CLOTHING MADE ONLY OF COTTON OR WOOL FIBERS, NO SYNTHETICS. PILOT, NAVIGATOR, AND OBSERVER MUST ALL WORK TOGETHER TO DETECT AND COMMUNICATE THE PRESENCE OF POWER LINES WITHIN THE WORKING AREA. IF BUILDINGS ARE IN THE AREA, ALWAYS ASSUME THAT A POWER LINE IS NEARBY. IF WEATHER CONDITIONS DETERIORATE, CREWS MUST RECOGNIZE THE CHANCE FOR ICING CONDITIONS OR VISIBILITY CONDITIONS THAT CAN GREATLY COMPROMISE SAFETY. PLANNING SHOULD INSURE THAT THE FUEL TRUCK AND DRIVER ARE AT A LOCATION KNOWN TO THE PILOT AND WITHIN 15-20 MINUTES FLIGHT TIME OF THE HELICOPTER'S ANTICIPATED DESTINATION AFTER FLYING FOR 1 HOUR AND 45 MINUTES. �147 IV. PERSONNEL: A. THE PRIMARY OBSERVER SHOULD BE A PERSON EXPERIENCED IN DETECTING AND COUNTING DEER FROM A HELICOPTER, CAPABLE OF CONCENTRATING AND FLYING SEVERAL HOURS/DAY FOR SEVERAL DAYS, CAPABLE OF MAKING RAPID DECISIONS REGARDING GROUP SIZE AND SIGHTABILITY VARIABLES, AND CAPABLE OF ACCURATELY RECORDING DATA ONTO A TAPE RECORDER AND TRANSCRIBING THAT DATA TO DATA FORMS. THE PRIMARY OBSERVER IS PRIMARILY RESPONSIBLE FOR DETECTING GROUPS FORWARD AND TO THE RIGHT OF THE AIRCRAFT AND TOTALLY RESPONSIBLE FOR RECORDING GROUP SIZE AND SIGHTING VARIABLES OF ANY DEER GROUP SEEN. THIS PERSON MUST WORK IN COORDINATION WITH THE NAVIGATOR AND PILOT TO POSITION THE AIRCRAFT AT A FAVORABLE ALTITUDE AND SPEED. THE PRIMARY OBSERVER IS RESPONSIBLE FOR RECORDING ALL DATA PERTINENT TO EACH QUADRAT AND SHOULD REFER TO THE OBSERVER 'CHEAT SHEET' FREQUENTLY TO INSURE THAT PROPER DATA ARE RECORDED. B. THE NAVIGATOR SHOULD BE A PERSON CAPABLE OF NAVIGATING THE HELICOPTER TO THE PROPER LOCATION OF EACH SAMPLE QUADRAT USING LAT/LONG COORDINATES AND TOPOGRAPHIC MAPS PREPARED FOR EACH QUADRAT. THE NAVIGATOR MUST BE ABLE TO DIRECT THE PILOT TO FLY THE CORRECT BOUNDARY OF THE QUADRAT BASED ON VISUAL INTERPRETATION OF THE TOPOGRAPHIC MAP AND THE TERRAIN OVER WHICH THE HELICOPTER IS FLYING. FURTHERMORE, THE NAVIGATOR IS RESPONSIBLE FOR DIRECTING THE FLIGHT PATH TRAVERSED THROUGH THE QUADRAT TO INSURE PROPER 100% COVERAGE AND COUNTING CONDITIONS. THE NAVIGATOR WORKS IN COORDINATION WITH THE PRIMARY OBSERVER AND THE PILOT. NAVIGATOR ASSISTS THE PRIMARY OBSERVER BY DETECTING GROUPS FORWARD AND TO THE LEFT OF THE AIRCRAFT, AND ASSISTS THE PRIMARY OBSERVER IN KEEPING DIFFERENT GROUPS OF DEER SEPARATE. THE NAVIGATOR SHOULD KEEP A RUNNING TALLY OF DEER COUNTED ON EACH QUADRAT USING A TALLY-WHACKER COUNTING DEVICE. THIS PROVIDES A BASIC BACKUP TO THE TAPE RECORDER OF THE PRIMARY OBSERVER. ALTHOUGH THE OBSERVER AND NAVIGATOR WILL USE THE PRESENCE OF DEER TRACKS IN SNOW AS AN INDICATOR THAT DEER ARE PRESENT, THE NAVIGATOR SHOULD MAKE SURE THAT GROUPS OF DEER ARE NOT ACTUALLY FOUND BY TRACKING DOWN INDIVIDUAL ANIMALS BY FOLLOWING SETS OF TRACKS WITH THE HELICOPTER. SIGHTABILITY MODELS ARE BASED ON VISUAL DETECTION OF ANIMALS, NOT FROM TRACKING. C. THE PILOT'S PRIMARY RESPONSIBILITY IS TO CONCENTRATE ON FLYING THE AIRCRAFT SAFELY AND IN A MANNER THAT SUPPORTS THE PRIMARY OBSERVER. THE PILOT WILL DETECT GROUPS OF DEER NOT SEEN BY THE PRIMARY OBSERVER OR NAVIGATOR. THE PILOT WILL RELAY INFORMATION TO THE NAVIGATOR ON GROUPS HE SEES AND THE OBSERVER AND NAVIGATOR WILL COLLECT THE DATA FOR SUCH GROUPS. V. DATA COLLECTED: A. QUADRAT: AT THE BEGINNING OF EACH QUADRAT SAMPLE UNIT, THE PRIMARY OBSERVER WILL RECORD THE QUADRAT IDENTIFICATION NUMBER, APPROXIMATE LAT/LONG STARTING POINT, GENERAL FLIGHT LIGHT, SNOW, AND WIND COUNTING CONDITIONS, AND CLOCK STARTING ENDING TIME FOR EACH QUADRAT. NAVIGATOR SHOULD ASSIST OBSERVER IN MAKING SURE THIS INFORMATION IS RECORDED. B. DEER: THE PRIMARY OBSERVER WILL COUNT EACH GROUP OF DEER DETECTED AND DETERMINE SIGHTING VARIABLES FOR EACH GROUP, REGARDLESS OF WHO FIRST DETECTS THE GROUP. ON THE TAPE RECORDER THE OBSERVER WILL SAY NEXT OR NEW GROUP OF DEER WHEN EACH GROUP IS DETECTED. THE OBSERVER WILL RECORD THE ACTIVITY OF THE DEER WHO IS MOST ACTIVE WHEN THE GROUP IS FIRST DETECTED, THE VEGETATION TYPE AND RELATIVE PERCENT SNOW COVER IN THE CIRCULAR AREA WITHIN 10 METERS (30 FEED OF WHERE THE INITIAL DEER WAS DETECTED, AND THE TOTAL NUMBER OF DEER IN EACH GROUP. THESE 4 VARIABLES CONSTITUTE THE IDAHO SIGHTABILITY MODEL INPUT VARIABLES. SIGHTING VARIABLES: 1. DEER ACTIVITY IS EITHER BEDDED, STANDING, OR MOVING. THE ACTIVITY IS RECORDED FOR THE DEER WITHIN THE GROUP THAT IS MOST ACTIVE WHEN THE GROUP IS FIRST DETECTED. A DEER THAT IS GETTING UP FROM A BEDDED POSITION WHEN FIRST DETECTED IS RECORDED AS A STANDING DEER BECAUSE THE MOVEMENT OF THE DEER GETTING UP IS LIKELY WHY THE DEER WAS DETECTED. A DEER MUST REMAIN BEDDED FOR A BRIEF PERIOD AFTER INITIAL DETECTION TO BE RECORDED AS BEDDED. STANDING DEER MUST REMAIN RELATIVELY �148 MOTIONLESS AFTER INITIAL DETECTED, AND MOVING DEER ARE DEER THAT ARE WALKING TO RUNNING. 2. VEGETATION TYPE WILL BE ONE OF SEVERAL CATEGORIES AND OBSERVERS SHOULD DO THEIR BEST TO CLASSIFY VEGETATION PROPERLY TO THE MOST DOMINANT TYPE AT THE LOCATION WHERE GROUP FIRST DETECTED BUT DO NOT SPEND A GREAT DEAL OF TIME TRYING TO DIFFERENTIATE LOW BRUSH VEGETATION TYPES. IN ALL LIKELIHOOD DURING DATA SUMMARIES AND ANALYSES, VEGETATION TYPES WILL BE POOLED INTO LOW BRUSH TYPES VERSUS PINYON AND JUNIPER TYPES. VEGETATION TYPES WERE BASED ON CLASSIFICATIONS USED IN THE GAP DATABASE FOR UNIT 10. VEGETATION TYPES SAGEBRUSH BITTERBRUSH GREASEWOOD FLATS SALTBUSH FLATS PINYON-JUNIPER WOODLAND JUNIPER WOODLAND AGRICULTURAL AND NATURAL CLEARINGS RIPARIAN SHRUB TALL CONIFER SIGHTABILITY CORRECTIONS FOR VEGETATION TYPES WILL MOST LIKELY BE BASED ON CORRECTION FACTORS DEVELOPED IN IDAHO FOR THEIR VEGETATION CLASSES OF 1) GRASS/OPEN/AGRICULTURE, 2) SAGEBRUSH, 3)JUNIPER/MOUNTAIN MAHOGANY, AND 4), POSSIBLY CONIFER. NO CORRECTION FACTORS HAVE BEEN DIRECTLY DEVELOPED FOR DEER IN OUR PINYON-JUNIPER OR JUNIPER WOODLAND HABITATS SO WE MUST USE IDAHO INFORMATION TO APPROXIMATE OUR VEGETATION CONDITIONS. 3. PERCENT SNOW COVER AT THE LOCATION WHERE EACH GROUP OF DEER IS FIRST DETECTED. SNOW COVER PERCENTAGES WILL BE CLASSIFIED BY THE OBSERVER AS LOW= 079%, AND HIGH= >80% OF THE GROUND COVERED BY SNOW. THESE CLASSIFICATIONS MUST BE USED TO BEST MATCH THE IDAHO SIGHTABILITY MODEL. IDAHO DOES HAVE A SNOW COVER CLASSIFICATION OF 0-19% BUT THIS SITUATION APPLIES TO COUNTS OF DEER DURING SPRING-GREENUP WHEN DEER ARE IN LARGER GROUPS IN OPEN HABITATS. 4. TOTAL NUMBER OF DEER IN EACH GROUP. A NUMERIC COUNT OF ALL DEER SEEN IN EACH GROUP. DEER WILL NOT BE CLASSIFIED TO AGE OR SEX. VI. IN-FLIGHT PROCEDURES 1. NAVIGATOR AND OBSERVER WILL OBTAIN PROPER MAPS, ARRANGE IN ORDER OF NEED, AND DECIDE ON GENERAL ROUTE TO QUADRATS TO BE FLOWN DURING EACH FUEL LOAD. PRIMARY OBSERVER MAKES SURE THAT TAPE RECORDER IS FUNCTIONING AND ADDITIONAL BATTERIES AND TAPES ARE AVAIL.ABLE ..NAVIGATOR SHOULD HAVE A SPARE TAPE ~ECORDER AVAILABLE. . . 2. OBTAIN A GPS POSITION AT STAR"J:ING POINT AND NOTE A GENERAL DESCRIPTION, ie SE OR NW CORNER, MAKE SURE CREW IS ON PROPER QUADRAT SAMPLE UNIT. 3. FLY PERIMETER OF QUAD RAT FIRST IN A CLOCKWISE MANNER SO THE INSIDE OF THE QUAD RAT IS TO THE RIGHT OF THE PRIMARY OBSERVER. FLIGHT SPEEDS SHOULD BE 40-50MPH AT ABOUT 100 FEET ABOVE TERRAIN. IF HIGHER SPEEDS ARE NEEDED TO BE SAFE DUE TO WIND SPEEDS, CONSIDER ABORTING THE FLIGHT. SOME QUADRATS INCLUDE LOWLAND PRIVATE LAND WITH HOUSES, LIVESTOCK, ETC. USE YOUR BEST JUDGEMENT AS TO WHAT AREAS YOU NEED TO FLY TO SEARCH FOR DEER AND AVOID BUILDINGS & LIVESTOCK. 4. DETERMINE STATUS OF GROUPS OF DEER ON PERIMETER. A. DEER MOVING OFF THE QUADRAT WHEN DETECTED ARE CONSIDERED ON THE QUADRAT. B. DEER MOVING ONTO THE QUADRAT WHEN DETECTED ARE CONSIDERED OFF THE QUADRAT. C. IF A GROUP IS STANDING ON THE PERIMETER BOUNDARY, COUNT THOSE DEER INSIDE THE QUADRAT. �149 D. RESIST THE TEMPTATION TO LEAVE THE PERIMETER TO COUNT A GROUP ON THE INSIDE OF THE QUADRAT BEFORE COMPLETING THE PERIMETER. USE YOUR BEST JUDGEMENT AT THE TIME OF DETECTION. 5. FLY INTERIOR OF QUADRAT. A. FLY THE INTERIOR OF THE QUADRAT SYSTEMATICALLY IN STRIPS OR STRIP-CONTOURS AT ABOVE RECOMMEND AIR SPEEDS AND AGL. USE PROMINENT TERRAIN FEATURES TO DIVIDE THE QUADRAT INTO SMALLER COUNTING BLOCKS AND SEARCH EACH SUB-BLOCK INTENSELY. IF YOU DECIDE TO FLY STRIP-CONTOURS, WORKING FROM THE LOWEST TO HIGHEST ELEVATION USUALLY WORKS BEST AS DEER ARE MORE RELUCTANT TO RUN UPHILL THAN DOWNHILL. ON OPEN FLAT TERRAIN, SYSTEMATIC STRIPS WORK WELL WHEN USING LANDMARKS OR GPS LOCATIONS. TRY TO MENTALLY KEEP TRACK OF DEER GROUPS TO AVOID DOUBLE-COUNTING. B. BE PATIENT AND STRIVE FOR 100% COVERAGE OF THE QUADRAT. 6. WHEN MULTIPLE GROUPS OF DEER ARE DETECTED SIM ULTANEOUSLY, FOLLOW THESES GUIDELINES. A. PRIMARY OBSERVER SHOULD BEGIN OBTAINING DATA ON GROUP NEAREST THE HELICOPTER AND USE HELICOPTER TO KEEP GROUPS SEPARATED. B. NAVIGATOR SHOULD FOCUS MOMENTARILY ON SECOND GROUP OBSERVED AND NOTE LOCATION, ACTIVITY AND NUMBER OF ANIMALS FIRST SEEN. C. AFTER COMPLETING DATA COLLECTION ON FIRST GROUP, PROCEED TO LOCATION OF WHERE SECOND GROUP FIRST SEEN, DETERMINE VEGETATION TYPE, AND SNOW COVER% AT SITE OF DETECTION, THEN FIND AND COUNT SECOND GROUP. 7. AFTER EACH GROUP OF DEER IS COUNTED, OBSERVER SHOULD VERBALLY NOTIFY NAVIGATOR OF GROUP TOTAL SO NAVIGATOR CAN RECORD A RUNNING SUM OF DEER COUNTED ON THE TALLY-WHACKER. AT THE END OF THE QUADRAT, OBSERVER SHOULD TAPE RECORD THE TALLY-WHACKER SUM AS A CRUDE CHECK ON THE NUMBER OF DEER DETECTED. NAVIGATOR COULD RECORD TALLY-WHACKER SUM ON THE FLIGHT MAP ON THE QUADRAT. 8. BEFORE LEAVING THE QUADRAT, MAKE SURE YOU HAVE NOT FAILED TO SEARCH ANY OBVIOUS GEOGRAPHIC PORTIONS OF THE QUADRAT. OBSERVER AND NAVIGATOR SHOULD AGREE THEY ARE DONE BEFORE PROCEEDING TO NEXT QUADRAT. VII. DATA TRANSCRIPTION 1. PRIMARY OBSERVERS WILL BE RESPONSIBLE FOR TRANSCRIBING THEIR TAPE-RECORDED DATA ONTO STANDARD DATA FORMS. ALL OBSERVERS WILL LABEL EACH OF THEIR TAPES AS TO THEIR NAME AND DATE. DATA SHOULD BE TRANSCRIBED THE SAME DAY AS COLLECTED. IT MAY BE POSSIBLE TO HAVE AN ADDITIONAL PERSON TRANSCRIBE THE TAPES TO SAVE TIME, AND SUCH PERSON WOULD NOTE ANY QUESTIONS THAT ONLY THE OBSERVER COULD ANSWER REGARDING ANY PROBLEMS ON THE TAPE RECORDINGS. 2. STANDARD DATA FORMS WILL BE AVAILABLE AT PROJECT LOCATION. END C:\DEERCENSUS\QUADPROC.MEM �150 2. SURVEY DATA FORM ~T~TARECXHlf\13 FCH\11 CEER C?MJ 10 2CXJ1 cµid4imPR:c:'deertEr&J (IVVD'Y) 1-«:msllM:: SfAAT _ _ _ _ 8'0 JI-ER: lBIP.(F} _ WN) _ _ _ UGifCXW. Sf\ONTYPE ~ WT------NtMG4.TCR Cl3SERvffi - - - - - '/:>C!E_a=_ I llM=SfAAT: QUAD# _ _ STRA.lllv1 - - - EJ'+l)Cl..W): _ _ _ TOT.TIM:: LCRAN'GPS l.CCA.llOJ arosrARTII\G PONf: TOfAL C:RCl.JP "°· DEER VEC:E foCTlVllY lYPE o/cfN:J/1/ COffi. C:RO.P Sl2E TOfALl:E.ERSEEN IHINllCNS: WNJ. Ligi=l, Mxierae--M, Slrorg=S; LJGif COO. 8igtt=B, Hazy=H, CUl::::O; Sl'ON1YPE: Fresh =F less than 48tvs; Od=Odderlhan 48tTs. DEER ACTIVITY: Bedded=B, Stardng=S, Mlvirlf'IVI; VEGE. 1YPE: Big Sagebrush=SG. 8itlerbrush=8B, Qeasewxxi Ras=GI\( Salbush Ras=SB, Pinyon-Ju,i~J. Juriper Wxxlam=J\N Agirulhre & aeairq;=('A. Riparial Shn.b=RP, Tall Caifer=TC PERCENTSt-ONCCMR L.a.v=L=~79%, Hgi=H=>80% �151 3. SURVEY OBSERVER HELP SHEET OBSERVER CHEAT SHEET QUAD STARTING POINT GPS QUAD NUMBER QUAD TIME START/END QUAD WIND, LIGHT, SNOW TYPE GROUP INITIAL DEER ACTIVITY Bedded, Standing, Moving GROUP VEGETATION TYPE Sagebrush, Bitterbrush, Greasewood, Saltbush, PinyonJuniper, Juniper Woodland, Agriculture & Clearings, Riparian Shrub, Tall Conifer GROUP PERCENT SNOW COVER 0-79%=Low >80%=High TOTAL GROUP SIZE All deer counted in each group �152 SECTIONG SURVEY FLIGHT QUAD RAT SAMPLE UNIT MAP INDEX UNIT 10 DEER POPULATION ESTIMATE FEBRUARY 2001 I INDEX TO QUAD RAT SAMPLE UNIT MAPS /QuadMaollstwb3) I FLIGHT SEQUENCE IS ORDER OF FLYING UNITS Wl11-IIN AN ENTIRE BLOCK AREA INDEPENDENT OF STRATA iOUADRAT 11" X 17" FLIGHT I ~·-STRATUM NAME UNITNO. MAP NO. SEQUENCE LOCALE DESCRIPTION STRATUM Yamoa-Monument Low Density 1YML 5 1 I Bear Draw Mantle Ranch Rd. _1 2 Yamoa-Monument Low Density 2YML 5 Bear Draw Mantle Ranch Rd. 1 3YML 5 3 Drv Woman Canvon Mantle Rd Yamoa-Monument Low Densltv 1 4YML 6 4 Yamoa-Monument Low Densltv Drv Woman Canvon Mantle Rd 1 ~ 5YML 6 6 Schoonover Pasture Mantle Rd 1 ' Yamoa-Monument Low Density t 6YML 6 5 Yampa-Monument Low Density Schoonover Pasture Mantle Rd 1 7YML 9 Yamoa-Monument Low Density 18 Sand Canvon Mantle Rd 1 9 Yamoa-Monument Low Density 8YML 20 Chew Ranch 1 Yamoa-Monument Low Density 9YML 10 22 Trail Draw Chew Ranch 1 L. 9 Units I I I t· 1.. I Medium:Densltv - - -22 - - · Yampa-Monument Yampa-Monument Medium Density Yamoa-Monument Medium Densltv Yampa-Monument Medium Densitv Yampa-Monument Medium Densltv L_2 I Yampa-Monument Medium Density 2 ' Yampa-Monument Medium Densitv 1-- ___ 2 , Yampa-Monument Medium Densitv t 2 r-···-· 2 i Yampa-Monument Medium Densitv Yamoa-Monument Medium Density 2 i 2 i Yamoa-Monument Medium Density j........ _2 ______ i Yampa-Monument Medium Densitv 1 2 I YampaoMonuinent Medium Densitv ,2 Yamoa-Monument Medium D.ensitv 2 ; Yamoa-Monument Medium Densitv 2 2 c-··t ; 1-· ____ L_ __ 10YMM 11YMM 12YMM 13YMM 14YMM 15YMM 16YMM 17YMM 18YMM 19YMM 20YMM 21YMM 22YMM 23YMM 24YMM 15 Units 6 7 6 6 7 8 9 8 8 8 9 9 9 10 9 10 10 8 10 12 11 13 14 15 16 17 21 19 23 24 Johnson Canvon Mantle Rd Johnson Draw Mantle Rd Johnson Draw South Mantle Rd West Serviceberrv Draw Mantle Rd Yampa River Overlook Mantle Rd Marthas Peak Dangerous Mantle Ranch Cave Rock Bench Verv Steep Red Rock Ranch Site Steep Pear1 Park North· Pearl Park South Sand Canvon Rim Lower Sand Canvon Pool Creek Rim Stateline-Pool Creek I - I !INDEX TO QUADRAT SAMPLE UNIT MAPS STRATUM -- ---~ - 3 3 3 3 t ·-l ·----------:- STRATUM NAME Utah White River·Low Densitv utah White River Low Densitv Utah White River Low Densitv . Utah White River Low Densitv Utah White River Low Densitv QUADRAT 11" X 17" FLIGHT UNIT NO. MAP NO. SEQUENCE 9~UL 10-UL 11-UL 12-UL 13-UL 5 Units 11 12 13 13 15 2 5 7 6 13 LOCALE DESCRIPTION Stateline W.Drinnina Rock Ck ·DriDDina Rock Ck Open Flats White River Drinnlna Rock Ridae Stateline Drionin!l Rock Ck White River Cliffs i -I ··-···-·· ------···•-:-- l l 4 4 4 .. 4 ···---· 4 4 4 4 Utah White River Medium Densitv Utah White River Medium Densitv Utah White River Medium Densitv Utah White River Medium Density Utah White River Medium Densitv Utah White River Medium Densitv Utah White Rlver Medium Density Utah. White River Medium Densltv .. 1-UM 2-UM 3-UM 4-UM 5-UM 6-UM 7-UM 8-UM 8 Units 11 12 12 14 14 15 15 15 1 3 4 8 910 11 12 SnakeJohnReef W.Dinosaur Stateline Raven Ridae Raven Ridcie Morman Gap Raven Ridge South Raven Rldae South Raven Ridae Southeast Raven Ridae Hardware Draw RavenRldae White River i i i �153 ·- -·-INDEX TO QUADRAf SAMPLE UNIT MAPS .. .. STRATUM . ------ STRATUM NAME ·--·-··--- -··-·1 IOUADRAT 11"x17" FLIGHT UNIT NO. MAP NO. SEDUENCE _j LOCALE DESCRIPTION -~ ·----,I ···-·--·-··· "'t_j';:;~er White River Hiah Densitv 56 On White River i 1-WRH 24 5 l 2-WRH 55 On White River ___ ':!Peer White River !iiah Densltv 24 5 .I ----··-··--··· .. ___ 5____ . __ UooerWhite River Hiah Densitv 3-WRH 54 North of main ridae 1 24 5 Unner White River Hiah Densitv 4-WRH 23 53 Rldoe North Slone l ! I 5-WRH 23 On Ridae 52 5 ___ J!P.ner White River Hiah Densitv 6-WRH 51 On White River 5 Uooer White River Hiah Densitv 23 7-WRH 22 48 On White River ....... 5 _-.L Ueoer White River Hiah Densltv ·8-WRH 47 On White River 22 ____5_ _ J UQQer White River Hiah Densitv 49 North Slone Road 5 ! UnnerWhite River Hiah Densitv 9-WRH 22 - -5· Unner White River Hiah Densitv 10-WRH 22 46 South Slone Road 11-WRH 45 22 On White River r:-.--5 -···. UeperWhite River Hiah Densitv ---22 44 On White River i- _____5______ ... Uooer White River Hiah Densitv 12-WRH 13-WRH North of Coal Rldae 22 50 ... __ § .. _ __ QP.ner White River Hlah Densitv ------·· I, 21 43 5 Unner White River Hiah Densitv 14-WRH North of White River 1·-5 ·-- • • Unner White River Hiah Densitv 42 On White River Cliff 15-WRH 21 16-WRH 41 On White River Dirt Road 21 ~ 5 __\:!P._eer White River Hiah Densitv Unner White River Hiah Densitv 17-WRH 40 Main Road on Ridae 21 18-WRH __!,Jeoer White River Hiah Densitv 37 On White River Cliff & Mine 19 19-WRH Steen Gullies North of Mine Uooer White River Hiah Densitv 19 38 20-WRH __!d_eper Whlte River Hlah Densitv 19 39 Steen Gullies Northwest of Mine 21-WRH 1 Uooer White River Hiah Densltv 19 32 PinvonJufnner Ridae Road Unner White River Hiah Densitv 22-WRH 19 Chase Ck Steen Cliffs Gullev Jct. ---31 23-WRH Too of'Mesa • • ....J:!.eoer White River Hiah .Densltv 19 33 ·5 Unner White River Hlah Densitv 24-WRH 19 34 Coal Oil Rim Mesa - - -5-- J_Qeoer White River Hiah Densitv 25-WRH 18 29 Coal Oil Rim East ! 19 26-WRH 36 Coal Oil Rim Mesa Northeast ........... 5 ___ --1 U_e:i:i:er White River Hiah Densitv 27-WRH 19 35 ... __ 5 ____ .. _ UQDe"r White River Hiah Densitv ChaseCk Mesa 28-WRH 5 Unner White River Hiah Densltv 28 18 Dead Doa Draw Mesa Cliff 1- • _____ 5 ____ _J!Qner White River Hiah Densitv 29-WRH 18 30 Chase Draw Mesa 30-WRH 18 27 Unner Dead Doa Draw l ........§___ - ___ y_pi:i:er_White River Hiah Densitv -----·. 5 Uooer White River Hlah Densitv 31-WRH 18 26 Dead Doa Reservoir 32-WRH 17 25 Dead Doa Reservoir ... ____ ~- _.. . _...\:!..!?.Per White River Hiah Densitv 33-WRH _ ~ _ ___ Upeer White River Hiah Densitv 17 24 Nate Snrina Reservoir ----5 Uooer White River Hiah Densitv 34-WRH 20 Nate Sorina 17 ------~ 35-WRH 17 21 Nate Sorlna East _ § .. _ ..l:!.2e~.Whlte River High Densitf -·· -·· 36-WRH 17 19 Nate Snrinn North ....§. . . . __ Uei:i:er White Riv~J Hiah Densitv 37-WRH Scullion Gulch Lincoln ReservoirUpeer White River High Densl!Y 18 22 _____ .. 5 . --- -I Upper White. River High Den§Lty . 38-WRH 18 23 Scullion Gulch North 5 ! I 18 17 Nate Snrina North Pioeline 5 i Upp~r White.River High Density 39-WRH 17 • 40-WRH Upp~r White River_ High Densl!Y. Unner Nate S~!lli!...._____ .. _. I 5 .1---1?. 41-WRH 16 Upper Whi~e River High Den~J!Y. __ 5 I --·-·----- ....-···---- .. ··_ -- .. 16 ----· -·----- Upeer Nate Sering 41 Units ... ··-·-·· --J I ----- -· .. i ~ [·! -· =R ···.::_:1 I IINDEX TO QUADRAT SAMPLE UNIT MAPS M-! STRATUM NAME -= ··---··· QUADRAT 11" X 17" FLIGHT UNJ1NO. MAP NO. SEQUENCE jmcerWhit~ River Medium Densitv ~ • 6 UQRer White River Medium Densitv ---s·--- Unner White River Medium Densitv 1-WRM 2-WRM Unner White River Medium Densitv 6 _--6--~-1--_IJ_Qoer White River Medium Densitv 6 __ Uooer White River Medium Densitv 6 : Uooer White River Medium Densitv _ 6 ___ ·1! UpoerWhite River Medium Densitv 6 . Unner White River Medium Densitv Unner White River.Medium Densitv 6 4-WRM 5-WRM 6-WRM 7-WRM 8-WRM 9-WRM 10-WRM .10 Units . ! ---·--- --, i -_ , t--· I E 7 ____ ( __ Unner.White River Low Densitv 7 _.......J __ Upoer White River Low Densitv 7 Uoner White River Low Densitv 7 __IJ_Qner White River Low Densitv 7 ___ _JJp_Q!!r White River Low Densltv ___ --. 3-WRM 1-WRL 2-WRL 3-WRL 4-WRL 5-WRL 5 Units 28A 28A 21 20 20 17 16 16 16 16 --LOCALE DESCRIPTION Hwv40 Box Elder CK Red Wash Box Edler Ck Red Wash Prairie Doa Reservoir Raven Park Dam Scullion Gulch Raven Park Dam Scullion Gulch Rock Shale Reservoir Stlnklno Water Ck Stlnklna Water Ck Stlnklria Water Ck .Stiriklna Water Ck 1 2 4 5 6 9 12 13 14 15 I I 21 29 30-A 30 16 S.Hwv40 Red Wash Hwv 40 Red Wash Red-Wash Reservoir #1 S.H\Al\/40 Skvline Reservoir W. of Stinkina Water Ck 3 7 8 10 11 ---; �154 'INDEX TO OUADRAT_S_A_M_P_L_E_U_N_IT_M_A_P_S------,----,-----.--------.---------------, I .OUADRATI 11" x 17" FLIGHT ! i-=cS-=-T'--'R'--'A-'-T-"-U"'-M'1-,-----"S~T~RA~T~U_M_N~A~M_E~---+-U_N_fr_N_O_.+-M~A~P~N~O-=-.+S~E~,Q=U~E~N~C_E-t----'L~O~C~A~L=E~D_E_S_C_R_IP_T_IO~N--1• ..-------'1--------------+--:----t----t------+---------,-----,---------; L_1_Q_ , Massadon-Dino Medium Density 1-MDM 35 57 K-Creek I 10 Massadon-Dino Medium Density 2-MDM 35 58 Miners Draw Trail Ck ~ 10 l Massadon-Dino Medium Density 3-MDM 35 59 Miners Draw Road L 1o i Massadon-Dino Medium Density 4-MDM 34 54 Middle Ck (Flv Contours 66 & 7700) 10 _ Massadon-Dino Medium Densitv 5-MDM 32 41 Twin Wash Dino Headquarters 'ti:!Q~ .. \_ Massadon-Dino Medium Density 6-MDM 28 22 Skull Ck /Flv Contour 6200) 10 Massadon-Dino Medium Density 7-MDM 28 20 Skull Ck Rim 10 Massadon-Dino Medium Density 8-MDM I 28 19 Skull Ck Rim _ _10___ Massadon-Dino Medium Density 9-MDM 27 14 Three Sprinqs --1-0 Massadon-Dlno Medium Densitv 10-MDM 27 13 Three Sprinas ____.__,.____+....:.:.:=::.==.::...=:..:..:..::c....c.cc-=-=-:=.c-'--=--"-'-'-"-='---t-'c=...cc.;-=:-..:..:.:...+---=..:....-+-......:..:::___f-_ _ _---'...:..:.:...::..::....==,'-"-'-=------7 I r• I l--_ _._10_,.____+----'M'"-a::..:s=-=s:..::a:..::d:..::o-'-'n'--'-D=-i"'ncco....cMc:-=-ed':'.i:-"u~m-'-=:-D-'-en~s=.,i~tv-t--:1:--=1-:·M:-:=D-:--M:-t_-'=":27=--i--1:--=2:--_f--_ _ _-=---"'M'-'a':"s::'-s~a=do-=-n-:-a_____ _J 10 10 Massadon-Dino Medium Densltv 12-MDM 27 11 East Massadona Massadon-Dino Medium Density 13-MDM 27 10 Horse Draw E.Massadona . 10 Massadon-Dino Medium Densitv 14-MDM 25 1 Lower 3 Sprinas Draw 10 . Massadon-Dino Medium Densitv 15-MDM 25 3 Lower Peterson Draw l----'-1o=--.---1-....:.M:;.:;a::..:s:..::s:.;;.a"'-d--'-o-'n_-D'--'l-'n_o_M_e_d_i_u_m~D_e_·n_s...,i~tv-t-...,1_6...,-M..,.D-=--M-+__2_5_--t___ 9 _ - - 1 - - ~ -_P_e_te_rs,....o_n_D_ra_w ___R_e_s_e_rv_o_i_r_ _ 1o Massadon-Dino Medium Densitv 17-MDM 25 8 Peterson Draw 10 Massadon-Dino Medium Densitv 18-MDM 26 7 The Slouqhs 1o i Massadon-Dino Medium Density I 19-MDM 26 15 Skull Ck Rim Road 10 , Massadon-Dino Medium Dens.itv 20-MDM 25 6 Petes Post The Slouahs 10 Massadon-Dino Medium Density 21-MDM 26 16 Bear Canyon Sprina 10 Massadon-Dino Medium Densitv 22-MDM 28 18 Unner Skull Ck 22 Units - ---··-·· -- - - - - - - - - - - - - - - - ~ ~ ~ - - - - - + - - - - - - + - - - - - - - - - - - - - - - ; ,__ -:,:,--· ~· 11 ·- - . ~-----· - -----~-1_ - __ , I, 11 j___ ______11... 11 Il ' ,I Massadon-Dino Low Density 1-MDL 25 2 Peterson Draw Reservoir Massadon-Dino Low Densitv 2-MDL 25 4 North Peterson Reservoir Massadon-Dino Low Densitv 3-MDL 25 5 Wolf Creek Soring Massadon-Dino Low Density I 4-MDL 26 17 Petes Post Wolf Ck __ _ Massadon-Dino Low__ D_e~n~si~ity~--t-5_-_M_D_L_,___3_4_-+--_ _5_3_-+----~B=u~ck~w~at~e~r~D~r~aw~----- _... 35- - - - 1 56 - - - - f - - - -K-Creek - Massadon-Dlno Low Densih, ! . l . -6-MDL ----,~-~---· I -~=~-=~~-~==---=~~-- __________-::__s__u_~_n_,_·t_s_~---t~---_-_-_-_-:_--+-_--_-_-_-_~--~1---_-_-_-_~---_-_-_-_-_-_-_-_-_-_-_-_-_~~--··-~·-____·_-_· __ ~INDEXTO QUADRAT SAMPLE UNIT MAPS STRATUM NAME STRATUM - - - - - - -- 12 12 12 i 12 \i-•.0••--·--··-··· 12 , 12 12 12 12 12 I - Tweivemile Medium Density Twelvemile Medium Densitv Twelvemile Medium Density Twelvemile Medium Density Tweivemile Medium Density Tweivemile Medium Density Twelvemile Medium Density Tweivemile Medium Density Twelvemiie Medium Density Twelvemile Medium Density ·- -1 QUADRAT 11" X 17" FLIGHT UNl'T NO. MAP NO. SEQUENCE 1-TMM 2-TMM 3-TMM 4-TMM 5-TMM 6-TMM 7-TMM 8-TMM 9-TMM 10-TMM 10 Units 1 1 2 2 2 3 3 4 4 4 1 2 3 4 5 6 7 8 9 10 LOCALE DESCRIPTION i ----! I -1 Yamoa River Twelvemile Gulch Rd. N.Hwv 40 Radio Tower Road N.Hwv 40 Radio Tower Road N.Hwv 40 Buffalo Gulch S.Hwv40 Sorinas Ridae N.Hwv 40 Buffalo Gulch I N.ElkSorinas Buffalo Gulch Elk Sorinas Road ··Elk Sorinas Road Bav Gulch NW.Elk Sorinas ......! �155 • --· -•--¥ ., i -- ·------- - - - - - - - - - - ~ - - - - - . - - - - - - - . - - - - - - , c - - - - - - - - - - - - - - - , INDEX TO QUADRAT SAMPLE UNIT MAPS STRATUM~! i- STRATUM NAME QUADRAT 11" x 17" FLIGHT UNIT NO. MAP NO. SEQUENCE LOCALE DESCRIPTION ____, f 13-----Massadon-Dino Hiah Densitv 1-MDH 28 21 SkullCk 13 ; Massadon-Dino Hiah Densitv 2-MDH 28 23 Massadona Hwv 40 13 Massadon-Dino Hiah Oensitv 3-MDH 28 24 South of Skull Ck Town Site I ~-....:.1.::.3---+--'-'-'M""a~ss~a_d_o_n~--D-in_o_H~ici~h-D_e_n_s_itv~-+--4--M-D,,..H--,.--i--....:.28--,1---2~5'---t---~~~S~k~u~l....:.IC"'k~H=w-'--'y4-0~~"---1 j 13 13 __ 13_____ 13 13 .---------· 13 3 Massadon-Dino Hiah Densitv 5-MDH 29 26 Miller Ck Massadon-Dino Hiah Densitv 6-MDH 29 27 Unner Jones Twist Massadon-Dino Hiah Densitv 7-MDH 29 28 Jones Twist Hwv 40 Massadon-Dino Hicih Density 8-MDH 29 29 Martin Gao Massadon-Dlno Hicih Density 9-MDH 29 30 Little Red Wash Massadon-Dino Hiah Densitv 10-MDH 30 31 Red Wash Hwv 40 Massadon-Dino Hiah Densitv 11-MDH 30 32 Martin Gao West ti Massadon-Dino Hiah Density 12-MDH 30 33 Red Wash 13 - - • Massadon-Dlno Hlcih Density 13-MDh 30 35 Skvllne Reservoir Hwy 40 13 i Massadon-Dino Hiah Densltv 14-MDH • 30 34 Uooer Red Wash •~3 Massadon-Dino Hiah bensitv 15-MDH 31 • 36 Blue Mountain T-own Site 13 i Massadon-Dino Hiah Density 16-MDH 31 37 Willow Ck Hwv 40 east ! 13 Massadon-Dino High Densltv 17-MDH 31 38 Spencer-Willow Ck Hwv 40 L_.!~------J Massadon-Dino Hiah Densitv 18-MDH 31 39 Soencer Draw Hwv 40 ! 13 , Massadon-Dino Hiah Densitv 19-MDH 32 40 Soencer Draw West Hwv 40 ___.:11__ Massadon-Dino Hiah Densitv 20-MDH 32 42 Dino Headauarters Hwv 40 13 Massadon-Dino Hiah Densltv 21-MDH 32 43 Dinosaur Quarrv 13 Massadon-Dino Hiah Densitv 22-MDH 32 44 Drionina Rock Ck-Dino Road -·- ·---~- --- ·13 Massadon-Dino Hicih Density 23-MDH 33 45 Sand Ck 1----'-'----t-~~~-----~-=--~--r--------,..,,-t-----t-----+-----,----'-"~~-------' 13 Massadon-Dino Hicih Density 24-MDH 33 46 W. Dinosaur Hwv 40 L 13 ______ Massadon-Dino Hiah Density 25-MDH 33 47 Lower Serina. NW of Dinosaur I 13 Massadon-Dino Hiah Densitv 26-MDH 33 48 Uooer Sorina r·--ii i Massadon-Dino Hlah Densitv 27-MDH 33 49 Bull Canyon Rim 13 •••• I Massadon-Dino Hiah Density 28-MDH 33 50 K-Ranch Massadon-Dlno Hicih Density 29-MDH 33 51 Buckwater Draw -~ Massadon-Dino Hiah Densitv 30-MDH 33 52 K-Creek K l I_·_ l .... j _ _-_-_-I -~~~~~~:.~---D-in_o__H_~_1g:h=D=e:n:s=itv====~=;=~=-~=n=~=~=:===-3~3===~====5=5===~~=======:S:ta:t:el:in:e::K=-C:r:e:e:k~-- �.... V, °' (/) --4 ~ --4 :;; iii C ~ z C 0 s: (/) )> s: ( 17-WRH O t---------------i---t----t---t----t---t----+---t----i---11---+---1----+-:--1---1 :::~=~ :~ I 25-MDH 2 ;::~g~ ~ 20-WRH , 40 30,MDH 0 21-WRH O l1-MDH 0 22-WRH 0 23.WRH 0 24-WRH 0 -·--+----+-2"'5~.w-R~H..-+--eo-~---, .......·-----~-....... ---·-+----;----< 26-WRH 27-WRH 28,WRH 0 0 0 O I >m 0 z --4 i )> ---+---·- m ;:::=~ .~ ◄ O•WRH. ~ - 41,WRH "'O C -:~.,![,. .-==-1-t=--==_-:t+-==--==--=~-1-==--==--=ti--=___ =_-==--=~-1,_=..·=.:~=-_=_,=i+-:-~---.._-_. __ 1 - - - - - - - - - - - - - - + - - + - - - - + - - - t - - - - t - - + - - - - + - - + - - - + - - t - - ~ + - - t - - - - t l - ·--~-Jfllt++t=- ::::j::_ "'O 0 m (/) 1 - - - - - - - - - - - - - - + - - + - - - - + - - + - - - + - - - l - - - - + - - + - - - + - - f - - - +---+---~---+----,..,,~~~:~~-H,.._+ · • - - < , - - - . - - - - ----------+--t-'----t--+---t--+-----+-===::-!·-·~-1--+--- 1---,---t---~: .. m --4 t----------------+---+----+---+----+--+----+--+---+--f---+--f---+---ll---t-~5~ci::ii=:i~-i+"'--..;~c..._+--+,----1--+----I---! ..............31-WRH "'O r- , 18 --·---i·---f---+---4 I I ---+---'""'"" I I --4 C )> ~ z )> C 0 r0 )> C s 0 z (/) (/) 0 --4 0 z :I: �~~- ··:··l _!• - :=: ~WH~~E~R::~~=,,=:3a=1,:=:•=,~=.=~•=.;=!=l•=:.=~=;=~•=.,=:m~rJ=(N=,J===~~~:=~=:=t=d:~•,=~.=~=•:=:=~e-::-i-'r~"-~~;...-o,_·-:-_,q"':"'~"':"1,~;...i;...nu,;..:!"';!"'h;n"'S"-~ac"t.Juc..hm_·;-,~----~~----·_fl---!--:--·+j - ::~ ···-:··-r-------- ~~-• . · - - + ••••••••••• • J ........ t···-t··· ..-11 1-----,-----,,-----c-'--"'----lf--'------,--~..,...---,..:..=;..._;..._,;..;...-'-'-'-''-'-~-----'-----'-----'-----"---•---+----1-----+---+-----1,---+---f..... .___M_•-•~•-Oe_,•,..•_P_,er,..s,.,•_,m,cp,..le_,d,..U_nl-'-l(;...Nc-"''r"'-)---1-•-•""-"'~g_e_nu~m_b_e_,ol_d,..eer_co_u,..n_1,_d_on_aa_c_h_sa_m-'p-ltd_u_n1t_1_n_ea_c_h_St_,a_1_um--';'--•-d_e_e,_c_ou_n_,le_d_/_n_um,,_b_e_,_,,01_1,..am_,_pl_e_un_~_,_,_1_ _ _.i....l_--l'----+---+----!---"~.....,.. f... -~----+---+ -----+---+---·Deer Per Sample Unit Varlanco (S" ,.J • ,tandatd satnple variance of deer counted per sample unk ln 111ch stratum; calculaled using Excel "VAR" function for 11mpla variance. I_ Total Sample Units Per Slfatum (UJ • lolal numbar ol polonllol ,ample unit■ or quadral1 In oach Stratum I I I ·-----~ - - - , - - - ' - - - l r - - - t - - - + - - - - t - - - - t - - - I --+---+---< Estimated Otar Per Stmtum (N",J • average number or deer counted on each 1ampled unit In each Stratum mutllplled b)'potenttal number of sample units In each Stratum 1-----=-c--,-:--,-,-,=~~---1---,:--:-c---,-,c-::-,-:---,---:---,-~-,....,"'."""--,,-----,--~-,~-...,..---r-'----.---..----1---+----+-----l---+--f-f----!---+--- ----------1----s_1ra_1u_m_v._,_1a_nc_,_v_._,.~(N_'~,J---+·-=·'=-"":cd:--aTrd_eslirna1ed Stratum variance for deer counted in each Stratum I I I t----t---+---+---+--- - · - - - - t - - - - - 1 - - - + ' - - - ! - - - f - f - - - - l 1 1 ,_______________,_ •__ : ___,:,..., ___L_ : : : +--·- --t-------++-·-·-=------t:.:.:.-_:.~t-:.:.:.:.~:.:.-_:.:.-=-i+-_:.:.:.-:.-:...::.-:..-:..:.-:..~t----==~======~=====;!::-.·.:.:::~t-=--=--=--=-~~-=--=--=--=--=-~ •_1m_"_ 1---------------!-V-•r'--'('--N---'',+-l•_ _ __,lfU_,J,.,_(1_•---'U,/U,--",J'---·(;...S_••-=-.;--'u,,;,)l'••• Th.,;;;p,on el al. 1998 poge 341 • I i l---=roi---,al""Oee-r"'eou,-n""'lod--,-A'"'"'"'S,,-tt-al,-a.,.N"(-•u-m-N"J"'."""-l-.-,u-,,;-or1art deer coun:td In Ht~ Sl,atum : ••• --- ---- --- ··: : ==r:'---------l'----,,f---+---f----+--f-f---+-------+--+----1---+---f----i,e----J 1 l--=T"ot,..al'"'E'"1'"11m-a""'1ed-:-:o=-.-,-,."g'"'s,,.,ra-1,..a""'N"''c,-(o_u_m...,H"'•-;,J---,;-•. ■u-m-cl=-1-=H-11"'11,..m_a,..le""'d-•u-m"'b,-e-,"'•1,..d=-ae-,.,.,n-•-ac-':h-_s:::,-,.,..,u~m~;-::-lh-,s""'1,-.-.~llm-a-:ltd-:-p-op.u"'11"'11=-on-,""'1z~e"'ro=-,.,.u=-oc-ll•10=--sa-m-p,-le"'d·a,-,-.--+-----+--+---+----lf---+--+------+-- ···_--_-_--_---,_,If---+----, b-c~~=-c:.=.:.;:;.:.~~-"--'~~c-:-:-l....:;:"'--":~~~-:.'-"'c"'-.;:;.;c=-c.c.=:;c-:;.=+=-=r==,:.:;::==,.;.:::.;=:..,-:-:;==;:='-----l----+--l---l---l----l--+---+--1-Tolal Elllmated VaflanceVe,'N'_ (sumVar'(N',J) •sumofolleatimalodSlralumva,lan-11 I I I I 90:;:~:;:,~j~~~~~~~tlr~~~~:1per :.• :Poo'U1ation :~:~:km 95% Conrldenee lritetval for N'· ILO'W1Jr 111 mnerl I i 1----+---+----+-----+---+---a"----!---- I ·----l'----+-----l ~~:: ~~~Atrv1~~:~r•~l~~:X~'•=•re8=~•,~••=•~•~•l!f~°i;•::•l=••~••===:~====t======~!====~~=====~=====t=====j======~======t====j======ti=====t=====--l:'---+----1 N,\l +/. <1.96 ~Dol=M••~r'~N~•-1_ _ 1 ___1~_ __.1_ _ _1 ~ - - - - - - - - - - - . . , __ _~_ ____,__ _ __.__ _.__ _..,__ ___.__ _ _.._1_~1.__ ___, _. V, --.J �158 2. LETTER FROM IDAHO DEPARTMENT OF FISH & GAME SHOWING POPULATION April 5, 200 I V.W. Howard, Jr. 1025 Hickory Drive Las Cruces, NM 88005 Tommy S. Bickle P.O. Box 750 Hatch, NM 87937 Dear Drs. Howard and Bickle: Find enclosed 3 files representing the results of my analysis of the Colorado Unit IO mule deer survey data. Colorado Division of Wildlife personnel flew the survey Feb 25 - Mar 5, 2001. The survey was flown using a stratified random sample. Eleven strata were delineated and sample unit size varied from ¼ - 1 mi. 2 . The ¼ mi. 2 sampling units were used in the high canopy closure habitat types of pinyon-juniper and/or juniper woodland. The l mi. 2 sampling units were used in more open canopy types. One hundred and forty three units were sampled across the 11 strata and 1180 mule deer were observed. Because the Idaho Aerial Survey program has a ten strata maximum, 2 strata were combined for the analysis. Based only on the stratified random sample, without correction for visibility bias, the population estimate was 6481 mule deer. Using the Idaho Department of Fish and Game's Sightability Method the population estimate was 11,052 (90% CI 7549 - 14,555). This represents a sightability factor of 1.7. Bartmann et al. (1986. Wildlife Society Bulletin 14:356-363) recommended a 1.5 .sightability factor for Colorado pinyon-juniper woodland. You can find a copy of the Idaho Aerial Survey software and a downloadable manual on the web at: http://members.nbci.com/fred_leban/survey.html. Please call with any questions. Sincerely, James W. Unsworth Principal Wildlife Research Biologist Cc. G. Miller, CDOW �159 ESTIMATE USING IDAHO SIGHTABILITY CORRECTIONS Aerial Survey for Windows, Version 1.00 Beta 6.1.4 (12-Feb-2000) Thursday, April 05, 2001 09:06 AM Model: Mule Deer, Hiller 12-E, Idaho (Spring) [Files] Title = C:\PROGRAM FILES\IDFG\AERIAL SURVEY\colo2.ttl Summary= C:\PROGRAM FILES\IDFG\AERIAL SURVEY\colo2.sum Colorado Unit 10 Mule Deer Survey, Feb 28 - Mar 5, 2001 Section 1: Summary of Raw Counts Units Stratum Sampled Total ------1 2 3 4 5 6 7 8 9 10 24 5 8 41 10 5 15 4 21 10 125 31 66 322 178 112 10 9 133 194 Total 143 1180 Section 2: Summary of Raw Counts for Perfect Visibility Model This table projects the number of animals that would have been counted if every unit had been flown and visibility bad been perfect (no animals obscured by vegetation, etc.) Strat No of Units Popn Sample --- ---- ------ Total 237 32 23 217 31 32 176 80 162 29 24 5 8 41 10 5 15 4 21 10 1234 198 190 1704 552 717 117 180 1026 563 Total 1019 == 143 6481 1 2 3 4 5 6 7 8 9 10 ---- ---- - - �------------. ---- 160 Section 3: Estimates for Total Number Total Stratum Number of Units Popn. Sample Estimate Sampling Variance Sigbtability Model - - - - -- --- --1 2 3 4 5 6 7 8 9 10 237 32 23 217 31 32 176 80 162 29 24 5 8 41 10 5 15 4 21 10 Bound 90% ---1422 293 375 3183 851 1397 154 1133 1403 841 1158993 27155 25316 654838 154826 388737 21812 1219118 572711 141508 4492 2570 8143 44009 5916 14884 376 53969 10224 4590 349 218 1529 6681 538 1333 17 8787 784 449 1774 285 308 1382 661 1047 245 1862 1257 630 11052 4365014 149173 20685 3503 ----- - - ------Total -- 1019 143 �C') 0 4 . ~ 5 8 K F,,G HIIIJ UNIT 10 OefR PG>PUlATION ESIMATE •EB 28,MARcH •·· 2 WITHCOLORADOCCO)AND.IOAHOflC SIG"1an1L , ,, N M p 0 Q u R V w X fot 10 1.TuA••AftA!lr,1 , Sla"81ntumlDlorCO:st-lor.lD. -0..rActMIYlorCO voo ..v ... TYOO ror co· orlO :Aet•ourAclM ~r or en I I I TlffiPA•M1 or P■ PM when Ouackal nown fo, CO :::! z nl 3 1 ◄' ff ~~oi!ild~-0■ ~o~ldttl!iJ•~-~~•!!!ldL=l====t====r===t===1~==::t:==+===1===r==r==:t:==::t:==+==:+===i:===t===1 rd• or 11 It H• ~Hellco tr for CO; JR•JolR■na11r3, HR•Hillat12E 20 ITel Pflk SHn on Quad rat ror co: 2721 28' 3 31 32 ~ ''35 ie' m 31 )> Colo Quod 1T2YMI.' 3YMI. Idaho SubUnH , z ◄YMI. ◄ 5 I 7 7 7 6YML I a 9 10 11 12 13 12""'-' '""" 14 15 15""'-' 1 Yl,t,1 4 1 ~ 17YMII 45 IIYM\1 ,, ◄2 ◄6 ., 18YM'.1 2DYIIM 21\'M',4 22WM 22WM 2>VM,t 52 2 ◄VPIM 53 9UL ~- .£ 41 48 -SC ii .Colo -.Sta YMI. ·y- 3 5YML l!. - 13 ,. 10 11 0 ::0 21 2◄ ,, G') 18 IS 2S 2a. r m -I m C XP•~T•=OFDATA~l!LOS I I I -auad-au1drat Unit ,o ror CO· -SubUnll•Cuadral Unit 22 03' ~ "C II 17 18. -· 20,a 21 22 22 YMI. Yt,t; YMt YMI. YMI. YMI. Thi.'. "M,1 """' """' """' klaho Sttatum 1 ,, Coio Taeer o. 0 0 0 0 0 1 1 1 .1 1 51 28 . 17 1 13. 1 ,1 0 z 0 2 0 - 0I 2 2 2 \?.t,I YIIM., 2 T- ~ -- ---~ 2 2 0 """' """' ·- 1 0 Colo Idaho . Colo Idaho COio Idaho Coro·, Colo Colo Dtll Act2 , Act 0 0 Snow SnOWCov,r Vlat IVtaCIHI ··uatt.;. Tl Timi 1 1 p a 0 0 0 0 0 58 M M za s 17 13 M 0 0 Q 3 2 3 M , 3 3 M M 3 3 0 0 0 0 0 0 0 0 0 0 3 2 3 2 2 0 24 8 2 3 10 UL 3 UL UL 3 3 0 3 0 0 2 M M 2 0 M 3 SG so 0 PJ 0 60 L 0 H-··· H -· L L .,.._,o 80 80 PJ -- 0 0 0 0 50 50 7>.l --·PJ PJ so·- .. SG 0 L 0 2 2 2 2 0 0 Q p /2001 • 03/0◄/2001 03/0◄/2001 o:w 1 1 1 1 1 1 I 1 p p p p p ' 0 3 1 1 ,1 3 I p p 03/0◄ /2001 1 1 1 p p 12001 03/04/2001 1 I 0 1 1 1 1 0 0 2 03/0◄/2001 001 ,-~~12001 OJI0◄/200 ,umm•OO Colo Dbl 1 1 1 1 1 TimeQ 9 Wind L L L L 1 I 1 1 1 3 •3 3 3 3 1 3 3 3 3 3 3 3 3 3 3 3 3 A p p p p p p A A p 3 - --~ a 8 ◄ 3 8 5 3 3 J A A A A A I I 0310,412001 Colo p 3 0 Colo 3 p 1 0 0 0 3 3 0 0 2 Colo Nav 1 1 p p p 0 3 0 0 PJ ·O 0 0 0 3 0 0 3 .3 S<> SG 60 L 0 0 0 i;o !Kl 0 0 0 M 0 0 so so 0 0 I 0 2 ,_ T- L L L L 0 2 YIIM \'1,/M UL 0 .0 r,vm YM',1 0 0 3 0 0 0 0 0 0 Yl,t,1 YW,1 2 0 0 0 0 2 2 0 0 2 2 23 101ho 3 3 3 8 8 8 8 3 10 3 3 3 3 3 3 3 3 3 3 3 5 11 8 7 7 6 6 6 7 9 3 3 1 ., ..,l,_ 3 3 3 1 1 1 15 16 18 1 23 I L L L L L L L L L L L L L L L L M L M L L L L Colo Ute Colo snw Colo PIiot COio H•llo Colo Tolk B B 0 0 0 0 0 0 0 ET ET, JR JR. JR JR 0 100 100 0 JR JR 100 JR JR JR JR JR JR 0 B B 8 B " H H H B B B B H H H H H H 0 0 0 0 0 0 0 0 0 0 ET •T ET ET ET ET er er er ET er er er ET ET __ ET - _£_ 0 ET ::0 )> r en C ~ m 4 0 -< 20 0 4 0 0 JR JR JR JR JR JR JR JR JR JR JR ·JR JR JR 0 ET 0 ET H 0 ET H 0 ET B 0 ET H 0 ET H 0 ET B 0 ET J/1 ET 8 0 JR JR L H 0 ET L 0 ET JR D ET JR L 8 0 L I ~ 0 !T J/1 L .. ,,~,- H . ' _ O _____ ET. + - ~ " 0 m 0 0 0 0 0 3◄ ·- 14 0 0 I 0 0 0 ·- - so -siJ-:-_1_ .... ,...!._,_ -+-, so .. , ·-50··----··---st :J~ 1--.J½-:.-[:JLLJ:::_~-J-·. ~- F~t~·-::f_- -+--r,::- --··1. ---1r it -· .H1:'L ,13. ., .. Jt~ I ,, -+--· lL. __ L .. . .. ,1, ,L, __ -~--L ~, ~--:-:~i:: -~: --~:-·: -·--o ,.,.. ·;:·: .. --··o·----o·, 60 I-_:-}_::.. -!r:-= ,., t -1--- ,,.,,--- --t f----~ --i---i---, --~r- i·· i--- -·-·o -'t= it , it~ : l ·; -~- ! , : ,, i· -~. I, ~ 51 57 11UL 12UL. 11 12 ..! , M 0 3 0 L_____ __ L L 1, 3. _. , ,,.1..8.. .,_ ,, .•. t ,.,,.,,. 50 """sir" "'"T" ---- 5o-···· --~--- - - · • h • .. -•" 0 • ., ., "'"'ij"' , li:::1----L 1 • 0 , , __1___ 1 1 ___ I • 1 ..... ..... °' �,_. °' t-..J .. 62 .113 A 8 3UM 3 3UM 3UM 3 3 as • s89 70. 71 SUM 7UM 7UM .73. 74 7UM 7UM ·- D 4 C 4 4 4 E 3 5 4 4 2 4 O • 4 UM • .. UM. ,,.. 4 4 7 lJM •- UM UM, UM"' UM, 4 4 4 -Nn WRH 5 5 5 78 ,wnn I "' 2WRH 1 2 11 C UM UM UM UM UM UM UM W 7 7· 7 T 7B 1 ,., Te,, 1\1M i I I ◄ o 0 4 1 2 20 1 1 a 11 0 7 2 F 3 5 4 4 2 O G M M M M N o H 3 3 3 3 3 O I H .L L L L o, 0 4 1 .2 20 1 1 I 11 0 M M M. M s M M B 7 & .J· au 50 !50 oo 50 .0 K PJ PJ PJ PJ PJ o 0 3 3 3 3 2 3 3· 1 0 .L L L l L L L L .2 L 0 50 50 ou 50 !50 oo 90 .., N 1 1 1 1 1 1 1 1 1 I A A P 1 1 1 1 I 1 11. 1 1 1 1 1 1 1 1 4 4 4 1 A 4 1, Lao -,ZOO! 1 031V312G01 03.'V3/2G01' I 1 1 I 03IW/2001 1 A A A A A A 4 4 4 4 4 4 1 1 I 1 1 1 I. L L L L L o· 0 PJ PJ PJ PJ SG PJ PJ fNi 0 50 L 3 3 3 3 3 O· SG ll 3 3 1 3 2 3 3 2 0 z- 02/Zl/200.t •-~ 0310WZ001 SOZWRH 2 2 8 2 I. "° PJ 3 D3oV3l2001 ~a~,~~~i"H-i3-~~•~➔s~-➔-~o-➔-io-~--+~o~+~--~-ot·-➔---~~o_ _ _ -nn 1-1 IH 112 •· 1'1 , 5 • .6 5 6. 9 15 13 10 4 ti 2 16 13 10 4. I 2 M M M M .M.. ..S 3 3 .• 3 3 .3 2 L L L L L L !50 50 !50 50 10 _ .. 50 SG 80 SG 80 88 SG 2 2 2 2 2 2 p P P P P P P P P P A O 1 1 1I I 1 1 1 1 1 1 1 1 1 1I. 1 1 R 23 24 11 19 27 18 I I S L L L L L L I. L L L L L L L L L L L 1' H H H H H B a H II H H H H H H D B B B U 0 O O O O O a O O O O V E ET ET ET ET !T er l!T ET ET ET O E O O O O O ET ET ET eT " l!T ET O O W JR JR JR JR JR JR JR JR JR JR JR JR JR . JR JR JR JR JR X 0 50 0 0 0 0 0 o ETJR 1 ,. • , L " o ET JR 20 I A 4 1. 5 L B. 0 l:T JR 5 +--;:-+--:.,....--,,--~<--+--~:-+---a:--+-----~~-+--..,:~+----~--+--~ETE"T~+--:,..__,_..,~_ __, .,,"'..-1.:.·-=ii-.,.H;;-,1--•.--t-'·r.WRHiif.iH---.sc-.---,1-~lr'-t-T" "-;-----,-+--;",-+---,-+--"'o.,..._+----1--or_ ru -nH 56 IRH 5 o., 0 0 .o; . 0 ~::i-t·r-ia~'-i:R;.;~-I-~ : , : : M : L. : 80 : ~~&6~=~~~~=r==7i==t= ~a7cl--=7W~R;.;H+_7i,--f,-, gg 7WRH 7 WRl:I It 7WRH I 7 -WAH. 90 7WRH I ·, ii 7WRH I ·7 WRH 1 1 1 1 1 1 1 I 0 A A " A A A I B 8 8 B B B O 0 0 O O O ET ET ET ET ET ET JR JR JR JR JR JR �A 123 •=RH 12 ◄ 20WRH Lill RH Lill 8 20 20 20 20 ~~ ffl 20 20 U 2, 23 24. C. -NM WRH.. WRH WRl:I :: • WRlf. -NM D 5 5 6 5• e : 5. 5 5, S 5 5 • F 5 8 5 5 3 5 3 : : ~ 9. 9. M 1 0 O 0 0 1 0 O 0 0 0 ' 0 0 • 0 0 0 0 28 7. 3 4 0 0 G S M M M M 13◄ , 21·. 13 3 139 14 1◄1 1 ◄ 2 ·/32=N 14' ...- N I ◄◄ 3◄WRH W :3◄WRH. 141., 1 ◄7 : H 148 :,:J◄WIUl. · 149"3SWRH.· ""' •38WRH 15 1 21: 20.· 30 31 ,32 33 34 WR.If' • WRl:I :. WRtl WRH WRH. WM·. 3◄ WRH WRH' 34 .>A 3◄ :,S. .. 3S • WRH · -NM WRH 5 5 5 5 5 5 5 5 • 5: 5 5 5 .. • ◄ 5 0 jl: !: ; ~•- ! ; l ll!l 1 1 40 ◄O ~ ◄0 ◄0 157 '-'2 ◄ IWRH .40 40 41 41 41 183 ◄ 1WRH. ◄1 ~15 1il 41WRH 81 ◄ 1WRt! · ~ _◄ tWRH 41 !Wc.JWRM+---· 1 166 2WRM 2 WRH WRH WRH WRH WRH WRH WRH. WRH,. WRH WRH WRH WRM WRM 5 5· 5 6 5 5 ~ S 5 .s 5 6 6 8 .16 1 3 5 5 7 3 1 4 1 0 I ◄ 4. 5· 0 ~ 5 16 1 3 6 5 7 3 1 4 1 D 1 J 50 50 .50 .. 50 K Sc, SG. S<> SG ; 3 3 L L L :~ 50 60 0 O 0 0. :~ PJ SG o 0 0 : Q 0 0 0. 0 28 7• 3 I L L L. L o: I:1 = - :. : : ~1 WRH. H' H R H 2 3 3 3 M M M M' s· u M M' M .S •. S M M. M M M M M M 0.. 0 O 0 0. o .0 3 .. 3 3 3 2 3 o·. ~ 3 3 ' 2 3· 3 3 3· 3 3 3 O 3 : L L L L L PJ S<J S<J SG. 60' PJ L l 50 50 50 60 50 50 50 50 50 50 60 O 60 1 so: pJ pJ: Pl· PJ. SO SO PJ SG SO SO 6G SO PJ O A A A A P ◄ ◄ : : A, ◄ ◄ I I I 1 I 1 A. P Q 1 I R 5 1 S L L 1 L ~ -~ 1 1 1 I I 1 L L T L L L H H· 8 8 8 B o 0 O F t : g O :; E O O· 0. o ET ET ET !T El ET ET.. ◄ ◄ I ◄, P ◄ P P. ◄ - P P P P. P P P A A A ,_ A P. A ◄ ◄ -f -,5 12 I : ; : : 4 1 1 0. O 0 0 0 0 3 2 2 2• 3 2· 0 1 1 1 1 1 1 1 1 1 I 1 J : ~ 3 3 3 2 2 3 2 2 • 2 2 O 3 N 1 1 I 1 ::1 : o· 0 L L. L L' L L .. L L L L L l ~ • 2 0 O 0 0 ·M • : .0 0 O 0 O., D 0 50 50. 80· oa so. L 2 2 • • 2 1 03/0212001 03I02J:1001 03I02/2001 02128/2001 0212812001 1 1 1 I I 1 1 1 1 1 1• A, A A A A A A A A A A A A ◄ 7 8 7 4 A L T L L L L T L L L L L L T H ti ti H V O O O O O B B B B 8 8 B H H H H. H H H 4 1 1 1 1 1 1 1 1 1 1• 1 1 1 ~ -j : t : ◄ 4 ◄• ◄ ◄ ◄ 4 ◄ 4· 4 ◄ 4 ◄ ◄ 4 4 ◄ 4 4 ◄ 4 4 2 2 I I I I 1 1 1 1 1 1· 1 2 2 8 g 7 8 6 9 7 20 37 L L L L L L L T L L L ,_ L.. M M I B· B B B B B B B 8 8 B H O ~ 0. 0 0 O O O O 0 O O O g 0 O O O O O O O O O O Q. O V !T .r 1IT· ET W JR, JR JR JR X 0 ---~~"-',T'-_,_-"'JJ:.,,,__.__ __, l!T ET l!T ET !T JR JR. JR JR JR JR. 0 5 D ~ 0 0 0 0 * ~~· 0 0 0 0 0 0 !l. l!T • ET • . !T ~T. ET ET ET ET JR, JR JR JR JR JR. JR JR JR JR JR JR JR JR. JR JR JR JR . JR er JR ET JR ET JR ET JR ~y JR __ll_ ., ...,Jc,;R_l----1 JR HR 0 JR HR 200 ..... O'I w �,_. 0\ .j>, m m ABC flWRM 6 WRM IIWRM • 18C 191 S 8 OE 6 5 F 5 6 ~ ~ 6 8 6 • 8 8 8 6 8 • 3 2 3 11 5 6 6 G M M HI 3 L 3 H J 50 SO K.. SQ SQ L 2 2 M M M M M M M M 3 • .3 3 3 3 3 3 50 SG sa SG SG SQ SG SQ SQ 2 2 2 2 2 2 2 2 MN 02/291200\ 1 0212812001 1 OP P 2 P 2 QR 2 2 STU L H O L H O V JR JR W HR HR P P P P P P P P 2 2 2 2 2 2 2 2 L L L L L l L L JR JR JR JR JR ~R JR JR HR HR HR HR HR HR HR HR X --·-8WRM IWRM .m IWRM 193 SWRM 19'1 ewRM 8 I WRM WRM WRM v, RM ·~ ,., 1• 8 6 W ••• •~ •~~ -·~- S • ., ~ 3 2 3 11 5 5 L .H L L L L L H •a 50 .OIi 60 50 60 eo .02/28/2001 -02/28/2001 "" 812001 ' ·02/21112001 02/28/2001 02/21112001 02/21112001 02/24/2001 1 1 1 1 1 I 1 2 2 2 2 2 2 2 2 H H H H H H H H O O O O O O O O �246 20 _,H ◄ 7MOH 8MOH 24! 2"' 251 252 2•3 254 255 2.. 257 258 255 280 281 282 1283 284 265 255 28 1288 289 21 2 1 272 21 274 215 • 4 • -·- I5MllH 15MOH 2118 28 28 289 292 293 ~ _,H MOH J,f)H MOH MOH MOH MOH MOH \Thf',4 I .,.,_. • 5 5 6 ~- 5 5 5 7 8 9 10 10 !~: t- T!,N '""' T!,N TMM TMM T!,N. '""' TMM T!,N •= ! I I ! TMM T!,N TMM WM TMM TMM T!,N T!,N TMM ·- ' T!,N TMM ' I TMM TMM TMM T~ 199 300 • 301 302 Total• 303 Subllct 304 30$ 313 F G 3 3S 2 0 0 0 0 3 s 0 1• 1 0 0 3a 2 3 0 0 2 0 0 0 0 4 3 5 & 2 M F G H u 0 0 3a M 7 M 3 0 0 2 0 0 0. 0 M • " " 16 10 3 2 15 10 3 2. 4 6 5 18 23 19 18 23 . 19 8 0 0 32 I 1 3 : .. 9- : M M M M M M ! I ; 1 7 s a • ; 2 0 0 0 0 0 0 2 3 3 0 0 3 3 3 0 0 3 0 0 0 0 3 3 3 2 2 3 3 3 3 3 2 2 2 3 3 3 2 0 0 3 3 3 3 2 0 0 0 0 2 3 a 6 : s s M• .,M ◄ ' 2 0 0 D 0 0 0 14 3 5 7 2 2 I 35 1 0 0 0 0 6 2 H 2 2 a 0 0 32 1 3 9 1 0 0 0 !I s. s s 5 M M. M s M M M M s 0 s 1 L L L l l L J 60 50 50 0 0 0 0 0 0 K SG SG SG 50 80 SG L 60 0 0 50 1. 50 L 80 0 0 50 0 0 0 0 80 50 L H l l l L L H L l l. l l 'H l L L L l L L L L so so PJ PJ PJ PJ PJ PJ PJ 50 50 50 ,SG .60. PJ PJ PJ so so . 60 so. so 50 60 50 80 50 60 so so .PJ PJ SG so SG· 50 0 0 $0 so l L I J' 0"IW2001 03/04/2001 001 0 03/04/2001 - SG SG SG SG PJ .. SG PJ -~- 3 3 3 n 0 3 0 0 0 0 3 3 3 2 2 2 3 1 .3 2 2 2 3. A A 03"0◄/2001 1 p 03/04/2001 03/002001 03/0112001 1· p I 1 A A I A 1 1 1 1 1 1 1 1 1 1 A Ii 1 f/2001 03/0112001 =112001 001 =1 001 1 1 1 1 1 1 1 1 ·,- 03/0 1 03/0 /2001 03/0112001 03/01/2001 03/01/2001 03/0112001 03/111/2001 03/0112001 0 30I12M 1 1 1 1 1 1 1 1 L M 0 0 0 2 A A A A A A A 0"M'"OOI 03/04/2001 03"0◄/2001 1 B 274 Entrlu ,. m Entrlu 274 Entries e 171 1180· 1180 171 Dter OeerGros DeorGms Deer w/Acl w/Act Counted Counted 123.' 121 1711. 179 oroun1 Grouas Grouns Ot0UDI Counltd coun1td Movlna,. 'Movlna II'• II H. II Qu1driti Qu1dr1ts GroUna orou•• w/0dNr Sl1ndlna Stind= w/0 dair 1 1. orou• ·oroun Bedded Beddtd K 179 178 179 179 x,, PHIGmt DeerOms wlSnow DetrGR>I DHrG"'S entrl•s WN""e wN.,,, m. 14'.. 117, Grouo1w/ Low Snow 33 Oroun1W/ HtnhSnow Grouo•wl GroUDI -'"*now lnSa•e 11 ·:. Groiina : In PJ. 1 Grow In 118 Oroufts nS1•tGW w/Snow 33,.. , Grouo1w/ Hl•hlnow Ort■ seWd 11 -a,ou•s 4 4 4 4 4 4 A A p p p p p p p Q I 1 1 1 1 1 1 1 1 1 1 1 4 A A •=001 I p • • A 1 1 1 1 1 1 1 1 03I0◄/2001 2 2 0 0 2 2 2 2 3 0 ·o I I I 03/04/2001 03/04/2001 n•~-1 03l04l2001 •2 N 1 1 1 1 1 1 I 1 l 1 I 2 2 2 0 0 50 50 50 50 60 0 0 0 0 60 50 - M L 2 2 2 0 D 0 0 0 0 4 4 •4• ◄ R 9 a 8 8 5 4 8 1 s 1 1 a ◄ I 5 4 1 5 ◄ I ◄ 4 4 1 1 1 7 1· 4 4 1 1 1 1 1 1 1 1 1 1 I 1 • 1 1 ·1 A A A 1 1 1 1 1 1 1 1 1 1 t , I 1 1 1 1 ·,1 1 1.· 1 1 • p A A A A • A A A A A p p p p 1 I I 1 1 1 ! 1 1 __ 1 1 p p p ' 1 1 1 1 N 0 p 274 Ent~es 151· AMcnts 123 PMcnt, 82 Blcklt p 1 1 1 1 1 1 1 - 11 p e •8 8 1 20 11 15 13 1◄ I I , p p I I ·A ' A 3~ 301. JOI. 309 310 311 312 E 0 ,. D ·'""' T!,N ; 13 13 13 13 13 13 13 13 13 13 13 13 13 13 0 MOH 1Thl',4 1 2Thl',4 2 2Tt.!-I_ ~ 2TMM 2 JT!,N 3 13 Tt.t.l MUH I 0 13 13 13 13 13 13 13 13 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 12 30 1 1 1 1 1 1 7TM\1 295 6Tt.t,4 298 9TMM 207 10T"'4 298 I0T'-'M MOH l,llH MOH I ITMM I -:~ MOH MOH 31 I 1 1 1 1 l'IM',1 1Th'M t'!M',1 1Th'M 1TMM JTMM STh\',! -- 15 18 19 20 20 21 22 24 25 28 29 15MOH lllM:lH 20M)H 20MOH 21MOH 2,-H 24MOH 2SM:>H 25MOH 2&MOH 30MOH 3IMOH 1Th'M lTl,t,I 1Th'M 11""1 1TMM 1TMM '285 MOH MOH MOH MOH 1• 15MOH 277 271 279 280 211 282 283 • 284 MOH 7 8 9 1D 11 13 15 10MOH 11MOH '7! C MOH MOH B A 4MOH 4MOH 245 15 9 14 18 13 1 ...... s L L L L L L L L L L L L L L L L L L L L L l L l L L l l L l .L L L L L L L L. L L l L L L L L L L l l L l L l T B B B 8 8 B e B 8 B B e B 8 B 5. 8. B. 8 B B 8 s· B 8 8 8 B B B u 0 0 0 V !T !T ET 0 E 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 o· 8 8 8 B 8 8 8 8 8 8 0 0 0 0 0 0 0 0 0 0 0 .. 0 0 0 0 0 8 Q R s T u 117 1'3 218 151 . au dtVtral• Qu•drot, llltVlfnd B~ltLII• OldSnow er ET E er E ET ET er ET ET ET ET ET ET ET ET ET eT ET ET er ET ET ET ET ET ET ET ET er ET ET ET ET ET ET ET :~-- w 218 0 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 30 0 ___l!L - ~ V EdTr■cv 0 a JR ET er )( JR JR JR JR JR JR JR JR JR JR JR JR JR JR JR JR . JR JR JR JR JR JR JR JR JR JR JR JR Jf\ ET ET ET ET !T ET ET ET ET 68" I 68 Flown& I II m. Tlmt ModWlnd HazyLII• P'r1hSnw JlmRlch Howud oraham 0 21 21 StraWtnd OullLlle .ar1h1m Ellenbrar 138 Freddv ' w JR JR JR JR JR JR JR JR JR JR JR JR JR JR JR JR JR JR JR JR JR JR JR -·-w0 ·40 0 0 0 15 X 1297 21B JttRnar ElkCnld H 10 Hlll6r cu1drat1 32 Quadrats with tlk 1 In PJ .. ...... °' Vl �166 4. MAP SHOWING WHERE DEER WERE COUNTED Sampled Quadrats Showing Where Deer Were Observed and Counted During Helicopter Counts of Mule Deer to Estimate .Population Size in Unit 10, DAU D;;.6, Colorado; Sampling Strata anil Sample Quadrant Units -Yampa-Monmneot (Low) ■ Yampa O Monlllllenl (Medium) Utai Wltite River (Low) Utah White River (Medium) -UpperWhiteltivet (High) Upper White River (Medium) UpperWhiteRmr(Low) Massadoiui ~ Dino (Medium) -Massadona- Dirio (Low) • ■ Twelv0111il~ (Medruui) • Mimaoonii:- Dirio (High) CJ Deer Obsavcd and Counted A - NoDeerObsenred �167 5. MAP SHOWING WHERE ELK WERE COUNTED Sampled Quadrats Showing Where Elk Were Observed and Counted During Helicopter Counts of Mule Deer to Estimate Population Size in Unit 10, DAU D-6, Colorado. Sampling Stra!a.and Sample Quadrant Units -Yampa-Monument (Low) -Yampa-Monwnent (Medium) Utah White.River (Low) Utim White River (Medium) Upper White River (High) . -UpperWhiteRiver (Medium) Upper White River.(Low) Massaoona - Dino (Mediuin) Massaoona - Dino (Low) ■ Twelvemilc (Medium) Massaoona • DiDo (High) E:=3::==::E========i:::::::::::::;==:::ie IIMes A r7 Elk Observed. - Ek Not Observed �..... en C: s:: s:: )> DEER POPULATION AND DENSITY ESTIMATES DAU 0-1 0-2 0-3 0-6 0-7 0-8 0-9 D-11 0-43 0-41 D-42 0-18 0-12 D-13 0-14 0-15 Sportsman's Population Estimate 4,100 13,300 2.700 1,750 17,500 8,000 7,800 5,100 1,750 3,700 2,300 Modeled Population Estimate• 13,500 37,800 2,700 7,900 7,300 8,300 2,200 1,200 31,500 10,000 4,900 2,650 1,750 5,900 0-53 1,100 TOTALS 128,420 408,600 0-20 2,200 1,850 0-21 0-23 1,600 0-19 6,100 0-39 0-40 0-25 0-24 0-29 0-52 1,100 0-30 0-35 0-37 900 3,800 3,600 9,200 1,970 2,400 3,300 1,750 750 24.i"98 1,800 7,000 67,000 6,000 13,500 12,000 7,700 13,800 2,400 6,700 5,000 3,900 2,400 31,000 3,000 11,800 5,300 34,000 12,500 9,300 20,800 7,600 2,000 3,200 D-38 0-51 where applicable Quadrat Population Estimate J4,857 11,016 34,619 ~ of Deer Classified 1997 1998 1999 401 3,058 306 623 3,559 2,049 3,835 524 1,963 497 3,839 464 7,076 558 407 5,361 2,099 NIA NIA 844 982 730 1,432 NIA NIA NIA 805 1,355 2,972 926 522 839 5,134 868 586 NIA NIA 2,091 1,071 740 643 1190 1122 473 615 2014 353 425 1005 170 387 187 488 539 156 128 301 18.35 5.07 10.29 9.56 12.30 5.53 15.40 14.10 9.38 8.80 3.57 9.48 14.92 5.71 2.90 4.66 4.78 t4.29 8.31 7.03 :::0 Modeled Densitylrni2 Quadrat Densitymi2 WRIS winter range WRIS winter range 11.34 21.57 33.69 3.81 11.38 17.36 33.27 17,00 33.48 31.76 11.94 45,29 35,66 42.25 14.96 58.44 64.23 64.10 38.28 19.60 4.90 28,88 N/A NIA NIA 395 716 917 352 481 490 232 124 NIA NIA NIA NIA 280 310 1,065 1,344 1,860 365 4,342 784 1,107 2,745 509 76.7 1308 NIA 7.94 50.40 4.80 2.93 3.56 2.11 11.34 18.60 18.47 15.48 5.63 32.99 7.40 23.54 998 NIA NIA 1,190 927 NIA 20,018 NIA 912 1,195 1,055 25,128 NIA 446 5,811 mi2WRIS wl nter range Sportsman's Densitylmi2 WRIS winter range 3.45 11.85 5.71 2.85 8.69 22.66 230 NIA 345 394 266 433 1,309 248 500 1,126 491 355 97 40,568 35,346 36,645 17,352 1,068 4,242 870 1,151 2,992 1,363 226 NIA NIA 1,608 2,684 824 1,202 3,719 773 167 40.32 13.93 7.74 23.70 13.04 44.36 12.24 25.97 -< :E ~ rn en -I "tJ mO ;:o :::0 z cri (") s:: Om •z 0 )> ;;o z ~c on - C ~~ me :S:m co m m ;;o ;;o "tJ NO 0 0 19.21 "tJ C: 0~ -I 0 15.29 z m en =! s:: )> -I m en 'Tl 0 :::0 • LJOW 2000 Post-hunt Population Projections ml.:d?er pop_den ail.xis -11109/00 0\ 00 � https://cpw.cvlcollections.org/files/original/ee5a4eecbb041230ff1ea11948214a12.pdf b51986c61cb4fd9596994dcabd5d57f4 PDF Text Text Colorado Division of Wildlife July 2004 – June 2005 WILDLIFE RESEARCH REPORT State of Cost Center Work Package Task No. Colorado 3430 3002 2 Federal Aid Project: N/A : Division of Wildlife : Mammals Research : Elk Conservation : Evaluation of GnRH Vaccine as a Long-term Contraceptive Agent in Female Elk: Effects on Reproduction and Behavior : Period Covered: July 1, 2004 – June 30, 2005 Author: D. L. Baker and J. G. Powers Personnel: J. Powers, M. Wild (National Park Service), L. Miller, J. Rhyan (National Wildlife Research Center), M. Conner (Utah State University), T. Nett (Colorado State University). All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without the permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT We conducted a pilot experiment to evaluate the potential of GnRH vaccine as a long-term contraceptive agent in female elk. The objectives of this preliminary investigation were to characterize the antibody response of captive female elk to GnRH vaccine, evaluate the effectiveness of dart delivery of the agent, and document the presence and severity of systemic reactions (if any) to the treatment. Intramuscular injection of GnRH vaccine was accomplished in 4 female elk. Serum antibody responses were collected each month beginning in February, 2005 and submitted for analysis. Ultrasound imaging of the injection site was conducted in conjunction with monthly blood collections. Analysis of antibody levels have not been completed, however initial results from ultrasound imaging of vaccine injection sites reveal changes in muscle fiber and muscle tissue echogenicity compared to pre-treatment conditions. All animals show some level of disruption of normal muscle fiber patterns and changes in the quality of muscle tissue. These changes began to appear approximately 2 weeks post-treatment, peaked in severity in April then diminished during July, 2005. Based on results of this trial and similar ongoing investigations with captive white-tailed deer (Odocoileus virginianus), we prepared a detailed study plan describing research to evaluate GnRH vaccine as a long-term contraceptive agent in female elk (Appendix I). The objectives of this experiment are to evaluate the effects of this fertility control agent on pregnancy rates, reproductive behavior, and neonatal health and survival. We performed a power analysis to determine the sample sizes needed to detect treatment differences for pregnancy rates and reproductive behavior for captive female elk maintained at the Colorado’s Foothills Wildlife Research Facility in Fort Collins, Colorado. Based on this analysis, a sample size of 18-26 elk (equally divided between control and treatment groups) should provide adequate statistical power to detect treatment differences in pregnancy rates and reproductive behavior. A detailed description of hypotheses, rationale, methods, and statistical analyses are provided in this report. The status of publications in process is also provided (Appendix II). 77 �WILDLIFE RESEARCH REPORT EVALUATION OF GnRH VACCINE AS A LONG-TERM CONTRACEPTIVE AGENT IN FEMALE ELK: EFFECTS ON REPRODUCTION AND BEHAVIOR DAN L. BAKER P. N. OBJECTIVE Evaluate the effects of GnRH vaccine on pregnancy rates, fetal and neonatal growth and development, and reproductive behaviors in captive female elk. SEGMENT OBJECTIVES 1. Conduct a pilot experiment to evaluate individual animal variation in antibody response to GnRH vaccine and assess any side-effects of treatment. 2. Using results from the pilot experiment prepare a study plan program narrative and submit for internal peer review and extramural funding. 3. Summarize and analyze data from previous fertility control experiments and submit manuscripts to appropriate scientific journals. INTRODUCTION Hunting and culling have traditionally been used to regulate ungulate numbers but there are a growing number of situations where these methods are not feasible. Such places include urban and suburban areas where lethal removal is often opposed because of safety concerns or on ethical grounds (Decker and Connelly 1989, McAninch 1993, Wright 1993, McCullough et al. 1997). In addition, there are many conservation areas, and state and national parks where hunting may be inconsistent with other goals of resource management or where it is proscribed by law and policy (Leopold et al.1963, Frost et al.1997, Porter and Underwood 1999). In these situations, fertility control offers a potential alternative for limiting the growth of ungulate populations (Kirkpatrick and Turner 1985, Bomford 1990, Garrott et al. 1993). Additionally, development of fertility control technology may provide resource managers benefits beyond its value as a tool for balancing ungulates and their forage resources. Fertility control may reduce the rate of disease transmission in ungulates by regulating local host densities and pathogen shedding (Rhyan and Drew 2002, Miller et al. 2004). Simulation modeling suggests that, in some situations, fertility control can be as effective as culling in reducing endemic disease or the density of susceptible hosts (Hone 1992, Barlow 1996). Extensive research has been devoted to developing anti-fertility agents that are safe, effective, reversible and economical (Fagerstone et al. 2002) and models have been developed to represent effects of fertility control on population dynamics of wild ungulates (Garrott and Siniff 1992, Seagle and Close 1996, Hobbs et al. 2000). To date, however, only modest successes have been achieved and a practical and acceptable method of controlling reproduction in free-ranging wildlife populations has not yet been attained. In previous research, we administered gonadotropin-releasing hormone (GnRH) agonist (leuprolide acetate) in a biodegradable implant to captive and free-ranging female elk and achieved 100% contraception for one breeding season, without significant behavioral or physiological side-effects (Baker et al. 2002,2004). However, despite the demonstrated efficacy and safety of this approach over existing technology, practical application is compromised by the need for annual treatments in fall, prior to the breeding season, a time when capture efficiency is low compared to winter and early spring. 78 �GnRH Vaccine An alternative approach involves immunization against GnRH. GnRH is a small, 10 amino acid, neuropeptide with an obligatory role in reproduction. It is naturally secreted in a pulsatile pattern from neurons in the hypothalamus and specifically directs gonadotropes in the anterior pituitary gland to synthesize and release luteinizing hormone (LH) and follicle stimulating hormone (FSH). These latter two hormones, in turn, control proper functioning of ovaries in females and testes in males (Hazum and Conn 1998). To successfully immunize an animal against GnRH, it is necessary to make this endogenous protein appear foreign to the host. Therefore, many copies of the peptide are coupled to the highly immunogenic carrier molecule keyhole limpet hemocyanin (KLH). When combined with a potent adjuvant the GnRH-KLH conjugate stimulates the host’s immune system to produce antibodies against GnRH as well as KLH. Anti-GnRH antibodies bind to GnRH in the hypothalamic -pituitary portal vessels and prevent the hormone from attaching to receptors on the gonadotropes. This suppresses secretion of LH and FSH, halting the hormonal cascade that is ultimately responsible for folliculogenesis and ovulation. This condition persists as long as there are sufficient antibodies to bind to all circulating GnRH. The use of GnRH vaccine as a fertility control agent is not new. It has been administered to a variety of domestic ungulates including horses (Rabb et al. 1990), cattle (Adams and Adams 1986), swine (Meloen et al. 1994), and sheep (Brown et al. 1994). It’s use as a contraceptive agent in wild ungulates has been limited, however by the need for multiple initial treatments, annual boosters, and the use of the controversial FCA and FIA to enhance the immune response of the vaccine (Miller et al. 2000b, Curtis et al. 2002). Recently, the impracticality of this approach for wildlife applications has been largely overcome by the development of a new adjuvant by scientists at the National Wildlife Research Center (NWRC) in Fort Collins, Colorado, USA. The alternative adjuvant is thought to be safer and, equally as effective in eliciting an antibody response, as FCA or FIA. The new adjuvant (AdjuVacTM) is derived from a USDAapproved Johne’s disease vaccine (MycoparTM) which has previously been approved for use in food animals by USDA/APHIS(http://www.aphis.usda.gov/ws/nwrc/research/gnrh.html) . A single application of GnRH-KLH and AdjuVacTM) may prove to be a safe, practical, and effective multi-year immunocontraceptive for wild ungulates. This approach has several potential advantages over other methods of contraception. These include: 1) a single treatment may provide long-term (2 + years) of infertility when administered to pregnant animals during winter 2) effectiveness of treatment may be > 90% during the first breeding season following immunization 3) infertility should be reversible 4) the agent should not cause significant behavioral or physiological side-effects 5) the agent should be safe for pregnant animals and the developing fetus 6) the proteinaceous nature of the GnRH-KLH immunogen should eliminate the possibility of passage through the food chain 7) the small volume required for effective contraception should facilitate administration by syringe dart 8) the agent is currently being evaluated for FDA approval as a New Animal Drug and therefore may be available for commercial use in the near future. Preliminary investigations evaluating GnRH-KLH vaccine in captive wild horses (Killian et al. 79 �2005, in preparation), bison (Miller et al. 2004) and white-tailed deer (Miller et al. 2004, unpublished data) are promising and USDA/APHIS is seeking FDA registration of the new vaccine and adjuvant (GonaCon/AdjuVacTM). However, many unanswered questions must be addressed before this potential contraceptive can be considered an effective and acceptable method of population control in free-ranging elk. Research is needed to evaluate the effectiveness and duration of this approach in elk, the effects on elk reproductive physiology, the effect on elk social structure of removing individuals from the breeding population, and the practicality/feasibility of application in wild populations. Captive Elk Experiments Rationale: The efficacy of GnRH-KLH vaccine depends on sufficient stimulation of the immune system and subsequent production of antibodies against this reproductive hormone. Thus, an initial step in assessing the potential of a single application of GnRH-KLH vaccine as a contraceptive agent in elk is to evaluate antibody response to treatment. Such studies have been conducted in wild horses (Killian et al. unpublished data), bison (Miller et al. 2004, in press), and white-tailed deer (Miller et al. unpublished data) but not in elk. Results of these studies indicate that the immunological response to GnRH-KLH vaccine is not uniform across species and highly variable within species. As a consequence, a species specific experiment is required to measure peak antibody response in female elk, time to peak response, and duration of response. Although such titers may not provide a quantitative measure of infertility, their characterization is of interest because sustained elevation of anti-GnRH antibody titers has been consistently associated with infertility in other species. Thus, the primary purpose here was to provide preliminary information on antibody response in elk, to determine optimum sample sizes for future experiments, to assess gross and clinicopathalogical side-effects of treatment (if any), and to evaluate remote delivery of the vaccine. Objectives: We conducted a controlled pilot experiment with captive elk to: 1) characterize serum antibody response of captive female elk to GnRH-KLH vaccine. 2) evaluate the effectiveness of dart delivery of GnRH-KLH vaccine. 3) evaluate presence and severity of systemic reactions or abscesses (if any) to the GnRH-KLH/AdjuVacTM vaccine treatment 4) determine if vaccination with GnRH/AdjuVacTM causes seroconversion to Johne’s disease mycobacteria. This experiment was conducted at the Colorado Division of Wildlife’s Foothills Wildlife Research Facility (FWRF) in Fort Collins, Colorado, USA with the approval of the Colorado Division of Wildlife Animal Care and Use Committee (# 1-2005) and in compliance with U.S. Federal Animal Welfare Act Regulations). METHODS We conducted an experiments with 4, non-pregnant, multiparous adult female elk beginning 7 February 2005. These elk were closely monitored into July 2005 to meet initial objectives of the pilot experiment, but the health of these elk and responses to the vaccine will be monitored until 1 August 2007. The captive elk used in this experiment were permanently maintained at FWRF and were trained to repeated handling, weighing, and blood sampling procedures. On the day before treatment (7 February), elk were moved from holding pastures (5 ha) and placed in individual isolation pens. The next day, each elk received a single injection of 1000µ :g of GnRH-KLH conjugate (0.5 ml aqueous solution) emulsified in 1.0 ml of AdjuVacTM, as a water in oil emulsion. The conjugate was be transferred into single use, 1 ml, 13-mm-diameter, barbless darts equipped with gel-collared 32-mm-long needles (Pneu-dart, Williamsport, Pennsylvania, USA). 80 �Prior to darting, individual elk were placed in a handling chute and lightly sedated with xylazine hydrochloride (15-20 mg/animal, IV). This dose allowed the animal to remain standing in the chute and minimized excitation associated with discharge of the dart gun. We examined of the reproductive tract of each elk using rectal palpation and ultrasonographic techniques, collected blood samples (20-30 ml) and measured body weight (∀ 0.5 kg). Elk were remotely injected in the biceps femoris muscle with a dart fired from a CO2-powered pistol (DanInjectTM, Wildlife Pharmaceuticals, Fort Collins, Colorado, USA) from a distance of approximately 3 meters. In order to accurately determine the precise dose of GnRHKLH delivered to each elk, darts were weighed before and after injection. If a dart failed to discharge or only partially injected the prescribed dose, additional darts were fired until the complete dose was delivered to the animal. Once the vaccine had been administered, sedation was reversed with yohimbine (30 mg, IV) (Antagonil®, Wildlife Laboratories, Fort Collins, Colorado, USA) and elk were returned to holding pastures. One of the elk (F86) used in this experiment was previously used in a Brucella abortus Strain 19 vaccination study (1998). It may still retain antibodies or immune modulation relative to this organism that could influence its immune response to the AdjuVacTM portion of the GnRH vaccine. This elk has not shown any evidence of being affected by Johne’s disease (Mycobacterium avium partuberculosis). However, the AdjuVacTM adjuvant uses small amounts of a remarkably similar killed bacterium (derived from the Johne’s vaccine MycoparTM). This could cause seroconversion indistinguishable from Johne’s disease. Pre-vaccination serum was submitted for a large animal biochemistry profile, Johne’s disease ELISA, and Strain 19 brucellosis vaccination serology. Elk were monitored for local injection site reactions (swelling, erythema, drainage) on a daily basis for 1 week, and on a biweekly basis for the following 2 months. A second biochemistry profile was submitted if elk showed symptoms of local or systemic inflammation. Ultrasound examination of the injection site may was used to evaluate abscess and granuloma formation. Serum anti-GnRH antibody production was monitored on a bimonthly basis until peak anti-body titers were determined, then on a bimonthly basis thereafter until termination of the experiment. Once a measurable (P < 0.05) decrease in anti-body levels is observed, the need to continue monitoring antiGNRH antibodies will be reevaluated. Once peak response in each elk has been achieved, a second reproductive examination will be performed to evaluate any changes in ovarian structures. Analysis: This was a descriptive experiment and no hypotheses were tested. We used descriptive statistics to examine changes in antibody titers over time. Schedule: Date Activity 12 January 2005 Submit study plan for ACUC approval 7 February 2005 Move experimental elk to individual isolation pens 8 February 2005 Perform pre-treatment exams and administer GnRH-KLH conjugate to elk 8 February to 8 March 2005 Intensive health monitoring of elk February 2005 to August 2006 Ongoing health and anti-GnRH antibody monitoring, and compile and analyze data pertinent to Expt. 2 81 �RESULTS AND DISCUSSION Intramuscular injection of GnRH vaccine was accomplished in 4 female elk. Serum antibody responses of experimental elk were collected each month beginning in February, 2005 and submitted for analysis. Initial results from ultrasound imaging of vaccine injection sites reveal changes in muscle fiber and muscle tissue echogenicity compared to pre-treatment conditions. All animals show some level of disruption of normal muscle fiber patterns and changes in the quality of muscle tissue. These changes began to appear approximately 2 weeks post-treatment and have not been resolved to date. SUMMARY Results of the pilot experiment are incomplete at this time. Initial results suggest that GnRH vaccine can be delivered via intramuscular dart injection. However, until laboratory results are completed, it is unknown if the antibody response of elk to GnRH vaccine will be sufficiently high to suppress fertility. Regardless, injection site reaction to the vaccine is a concern and warrants further evaluation. LITERATURE CITED ADAMS, T. H., AND B. M. ADAMS. 1990. Reproductive function and feedlot performance of beef heifers actively immunized against GnRH. Journal of Animal Science 68:2793-2802. BAKER, D. L., M. A. WILD, M. M. CONNER, H. B. RAVIVARAPU, R. L. DUNN, AND T. M. NETT. 2002. Effects of GnRH agonist (leuprolide) on reproduction and behavior in female wapiti (Cervus elaphus nelsoni). Reproduction Supplement 60:155-167. ____________, _________, M. M. CONNER, H. B. RAVIVARAPU, R. L. DUNN, AND T. M. NETT. 2004. Gonadotropin-releasing hormone agonist: a new approach to reversible contraception in female deer. Journal of Wildlife Diseases 40:713-724. BARLOW, N. D. 1996.The ecology of wildlife disease control: simple models revisited. Journal of Applied Ecology 33:303-314. BOMFORD, M. 1990. A role for fertility control in wildlife management. Department of Primary Industries and Energy, Bureau of Rural Resources Bulletin No. 7, Australian Government Publishing Service, Canberra, Australia. BROWN, D. B., T. T. CAI, AND A. DASGUPTA. 2001.Interval estimation for a binomial proportion. 2001. Statistical Science 16:101-133. CURTIS, P. D., R. L. POOLER, M. E. RICHMOND, L. A. MILLER, G. F. MATTFELD, AND F. W. QUIMBY. 2002. Comparative effects of GnRH and porcine zona pellucida (PZP) immunocontraceptive vaccines for controlling reproduction in white-tailed deer (Odocoileus virginianus). Reproduction Supplement 60:131-141. DECKER, D., AND A. N. CONNELLY.1989. Deer in suburbia-pleasure or pets. Conservationist 43:46-49. FAGERSTONE, K. A., M. A. COFFEY, P. D. CURTIS, R. A. DOLBEER, G. J. KILLIAN, L. A. MILLER, AND L. WILMOT. 2002. Wildlife fertility control. Wildlife Society Technical Review 02-2. The Wildlife Society, Bethesda, Maryland, USA. FROST, H. C., G. L. STORN, M. J. BATCHELLER, AND M. J. LOVALLO. 1997. White-tailed deer management at Gettysburg National Military Park and Eisenhower National Historic Site. Wildlife Society Bulletin 25:462-469. GARROTT, R. A., AND D. B. SINIFF. 1992. Limitations of male-oriented contraception for controlling feral horse populations. Journal of Wildlife Management 56:456-464. ______________, P. J. WHITE, AND C. A. VANDERBIL WHITE. 1993. Overabundance: an issue for conservation biologist? Conservation Biology 7:946-949. HAZUM, E., AND P. M. CONN. 1998. Molecular mechanism of gonadotropin releasing hormone (GnRH) action. I. The GnRH receptor. Endocrine Review 9: 379-386. 82 �HOBBS, N. T., D. C. BOWDEN, AND D. L. BAKER. 2000. Effects of fertility control on populations of ungulates: general stage-structured models. Journal of Wildlife Management 64: 473-491. HONE, J. 1992. Rate of increase and fertility control. Journal of Applied Ecology 29:695-698. KIRKPATRICK, J. F., AND J. W. TURNER, JR. 1985. Chemical fertility control and wildlife management. Bioscience 35: 485-491. LEOPOLD, A. S., S. A. CAIN, C. M. COTTAM, AND I. GABRIELSON. 1963. Wildlife management in the national parks. Transactions of the North American Wildlife and Natural Resources Conference 28:28-45. MCCULLOUGH, D. R., K. W. JENNINGS, N. B. GATES, B. G. ELLIOT, AND J. E. DIDONATO. 1997. Overabundant deer populations in California. Wildlife Society Bulletin 25: 478-483. MCANINCH, J. B., editor. 1993. Urban deer: a manageable resource? Proceedings of the 1993 Symposium of the North Central Section, The Wildlife Society, St. Louis, Missouri, USA. MELOEN, R. H., J. A. TURKSTRA, H. LANKHOF, W. C. PUIJK, W. M. M. SCHAAPER, G. DIJKSTRA, C. J. G. WENSING, AND R. B. OONK.1994. Efficient immunocastration of male piglets by immunoneutralization of GnRH using a new GnRH-like peptide. Vaccine 12:741-746. MILLER, L. A., J. C. RHYAN, AND M. DREW. 2004. Contraception of bison by GnRH vaccine: a possible means of decreasing transmission of brucellosis in bison. Journal of Wildlife Diseases 40:725730. __________ , _________, AND ___________ . 2000. Immunocontraception of white-tailed deer with GnRH vaccine. American Journal of Reproductive Immunology 44:266-274. PORTER, W. F., AND B. UNDERWOOD. 1999. Of elephants and blind men: deer management in the U.S. national parks. Ecological Applications 9:3-9. RABB, M. H., D. L. THOMPSON, JR., B. E. BARRY, D. R. COLBORN, K. E. HEHNKE, AND F. GARZA, JR. 1990. Effects of active immunization against GnRH on LH, FSH, and prolactin storage, secretion, and response to their secretagogues in pony geldings. Journal of Animal Science 68:3322-3329. RHYAN, J. C., AND M. D. DREW. 2002. Contraception: A possible means of decreasing transmission of brucellosis in bison. In: Brucellosis in elk and bison in the Greater Yellowstone Area, T.J. Kreeger (ed.). Greater Yellowstone Interagency Brucellosis Committee, Wyoming Game and Fish Department, Cheyenne, Wyoming, 99-108. SEAGLE, S. W., AND J. D. CLOSE.1996. Modeling white-tailed deer population control by contraception. Biological Conservation 76:87-91. WRIGHT, R. G. 1993. Wildlife management in parks and suburbs: alternatives to sport hunting. Renewable Resources Journal 11: 18-22. Prepared by: _____________________ Dan L. Baker, Wildlife Researcher 83 �APPENDIX I PROGRAM NARRATIVE STUDY PLAN FOR MAMMALS RESEARCH FY 2004 – FY 2007 State of: Cost Center: Work Package: Task No.: Colorado 3430 3002 2 Federal Aid Project No.: N/A : Division of Wildlife : Mammals Research : Elk Conservation : Evaluation of GnRH Vaccine as a LongTerm Contraceptive Agent in Female Elk: Effects on Reproduction and Behavior : EVALUATION OF GnRH VACCINE AS A LONG-TERM CONTRACEPTIVE AGENT IN FEMALE ELK: EFFECTS ON REPRODUCTION AND BEHAVIOR Principal Investigator Dan L. Baker, Wildlife Researcher, Mammals Research Cooperators Lowell A. Miller, USDA/APHIS, National Wildlife Research Center Jack C. Rhyan, USDA/APHIS, National Wildlife Research Center Mary M. Conner, Department of Forestry, Range, and Wildlife Science, Utah State University Terry M. Nett, Department of Biomedical Science, Colorado State University Jenny G. Powers, National Park Service Margaret A. Wild, National Park Service STUDY PLAN APPROVAL Prepared by: ____________________________ _ Date: ______________________ Submitted by: _____________________________ Date: ______________________ Reviewed by: _____________________________ Date: _______________________ _____________________________ Date: _______________________ _____________________________ Date: _______________________ Reviewed by: _____________________________ Biometrician Date:________________________ Approved by: _____________________________ Mammals Research Leader Date:_________________________ 84 �PROGRAM NARRATIVE STUDY PLAN FOR MAMMALS RESEARCH State of : Cost Center: Work Package: Study No.: Colorado 3430 3002 : : : : Division of Wildlife Mammals Research Program Elk Conservation Evaluation of GnRH Vaccine as a LongTerm Contraceptive Agent in Female Elk: Effect on Reproduction and Behavior A. STUDY TITLE: Evaluation of GnRH Vaccine as a Long-term Contraceptive Agent in Female Elk: Effects on Reproduction and Behavior B. NEED: Overabundant wild ungulate populations have become a significant problem for natural resource managers in many areas of North America. Unregulated populations can cause adverse effects that are ecological, economic, or political in scope and resolving these problems often requires managing excessive animal numbers (Jewell and Holt 1981, Garrott et al. 1993). Hunting and culling have traditionally been used to regulate ungulate numbers but there are a growing number of situations where these methods are not feasible. Such places include urban and suburban areas where lethal removal is often opposed because of safety concerns or on ethical grounds (Decker and Connelly 1989, McAninch 1993, Wright 1993, McCullough et al. 1997). In addition, there are many conservation areas, and state and national parks where hunting may be inconsistent with other goals of resource management or where it is proscribed by law and policy (Leopold et al.1963, Frost et al.1997, Porter and Underwood 1999). In these situations, fertility control offers a potential alternative for limiting the growth of ungulate populations (Kirkpatrick and Turner 1985, Bomford 1990, Garrott et al. 1993). Additionally, development of fertility control technology may provide resource managers benefits beyond its value as a tool for balancing ungulates and their forage resources. Fertility control may reduce the rate of disease transmission in ungulates by regulating local host densities and pathogen shedding (Rhyan and Drew 2002, Miller et al. 2004). Simulation modeling suggests that, in some situations, fertility control can be as effective as culling in reducing endemic disease or the density of susceptible hosts (Hone 1992, Barlow 1996). Extensive research has been devoted to developing anti-fertility agents that are safe, effective, reversible and economical (Fagerstone et al. 2002) and models have been developed to represent effects of fertility control on population dynamics of wild ungulates (Garrott and Siniff 1992, Seagle and Close 1996, Hobbs et al. 2000). To date, however, only modest successes have been achieved and a practical and acceptable method of controlling reproduction in free-ranging wildlife populations has not yet been attained. GnRH Agonist Gonadotropin-releasing hormone (GnRH) is an endogenous neuropeptide that has an obligatory role in reproduction. It is naturally secreted in a pulsatile pattern from neurons in the hypothalamus and specifically directs gonadotropes in the anterior pituitary gland to synthesize and release luteinizing hormone (LH) and follicle-stimulating hormone (FSH). These latter two hormones, in turn, control proper functioning of the ovaries in females and testes in males (Hazum and Conn 1988). 85 �The chemical structure of endogenous GnRH has been determined (Matsuo et al. 1971) and alterations in the molecule have led to the synthesis of potent GnRH agonist analogs (Karten and Rivier 1986). Long-term treatment with GnRH agonists has been shown to prevent ovulation by decreasing GnRH receptors on gonadotropes, receptor sensitivity to GnRH (Nett et al. 1981), pituitary LH content (Aspden et al. 1996), and by suppressing pulsatile secretion of LH and FSH (D’Occhio et al. 1996). Agonists of GnRH have been used in domestic animals as fertility agents for controlling ovarian activity, gonadal steroidogeneis, and reproduction (McNeilly and Fraser 1987, Montovan et al. 1990, D’Occhio et al. 2002). In previous research, the GnRH agonist, leuprolide, was administered to captive female elk (Cervus elaphus) and mule deer (Odocoileus hemionus) in a controlled release bioimplant and achieved 100% infertility for one breeding season, without significant behavioral or physiological sideeffects (Baker et al. 2002, 2003, 2004). However, despite the demonstrated efficacy and safety of this approach over existing technology, practical application is compromised by the need for annual treatments in fall, prior to the breeding season, a time when capture efficiency is low compared to winter and early spring. Immunocontraception To date, most wildlife contraceptive efforts have been directed toward development of a safe and effective immunocontraceptive vaccine. The immunocontraceptive target antigen that has received the most research and management attention is porcine zona pellucida (PZP). Porcine zona pellucida has been administered experimentally to more than 70 species of wild mammals (Kirkpatrick et al. 1997). This approach relies on host antibodies formed against PZP to block sperm receptor sites on the ovum, thereby preventing fertilization and pregnancy (Dunbar and Schwoebel 1988). The PZP vaccine has been shown to be 85-90% effective in most ungulates, can be administered by syringe dart, is reversible, does not interfere with ongoing pregnancies, and most importantly, the immunogen is proteinaceous and therefore, is not likely to pose a threat to the environment or to non-target species, including humans (Kirkpatrick et al. 1990, Turner et al. 1992, Miller et al. 2000a, Kirkpatrick and Turner 2002, Shideler et al. 2002, Naugle et al. 2002). However, despite these desirable characteristics treatment inefficiency and undesirable sideeffects have limited management application of PZP vaccine (Rudolph et al. 2000, Turner and Kirpatrick 2002, Naugle et al. 2002). Specifically, practical application is compromised by the requirement that the target animal must receive two initial injections within 1-2 months of each other (Walter et al. 2002). Second, with the exception of SpayVacTM which encapsulates PZP within a cholesterol/phospholipid formulation (Fraker et al. 2002), effective duration is typically < 1 year; consequently, annual booster inoculations are required (Kirkpatrick et al. 1996; Turner et al. 1996). Third, while no long-term health effects have been reported for animals treated with PZP (Kirkpatrick et al. 1995, Miller et al. 2000a, Turner and Kirkpatrick 2002), extended estrous cycling and associated breeding behavior have been reported for white-tailed deer (Turner et al. 1992, 1996; McShea et al. 1997), horses (Plotka et al. 1989), elk (Heilmann et al. 1997), and fallow deer (Fraker et al. 2002). By prolonging the breeding season in males and females, PZP vaccine treatments could result in late pregnancies, parturition beyond the normal early summer period, and unpredictable and abnormal behavioral consequences. Finally, for an effective immune response, the PZP antigen must be administered with an adjuvant - a substance that enhances the specific immune response to the antigen. At present, the most effective adjuvants used with PZP are Freund’s complete (FCA) and Freund’s incomplete adjuvant (FIA). In some species, however, this combination has been shown to cause severe systemic reactions, chronic pain, and abscesses at the injection site (Anderson and Alexander 1983, Stills and Bailey 1991, Leenaars et al. 1996) and, as a consequence, it is unlikely that the Food and Drug Administration (FDA) will grant approval for the use of PZP vaccine containing these adjuvants. Thus, in the near future, practical application of immunocontraception for wildlife species will depend on development and use of improved vaccines with different adjuvants. 86 �GnRH Vaccine An alternative to PZP immunocontraception involves immunization against GnRH. To accomplish this, it’s necessary to make this endogenous protein appear foreign to the host. Therefore, many copies of the peptide are typically coupled to a highly immunogenic carrier molecule such as keyhole limpet hemocyanin (KLH) (Levy et al. 2004). When combined with a potent adjuvant, the GnRH-KLH conjugate stimulates the host’s immune system to produce antibodies against GnRH as well as KLH. Anti-GnRH antibodies bind to endogenous GnRH in the hypothalamic - pituitary portal vessels and prevent the hormone from attaching to receptors on the gonadotropes. This mechanism suppresses secretion of LH and FSH and interrupts the normal cascade of hormonal events that are ultimately responsible for folliculogenesis and ovulation. The use of GnRH vaccine as a fertility control agent is not new. It has been administered to a variety of domestic ungulates including horses (Rabb et al. 1990, Turkstra et al. 2005), cattle (Adams and Adams 1990), swine (Meloen et al. 1994), and sheep (Brown et al. 1994). However, its use as a contraceptive agent in wild ungulates has been limited by the need for multiple initial treatments, annual boosters, and the use of the controversial FCA and FIA to enhance the immune response of the vaccine (Miller et al. 2000b, Curtis et al. 2002). Recently, however, the impracticality of this approach for wildlife applications has been largely overcome by the development of a new adjuvant by scientists at the National Wildlife Research Center (NWRC) in Fort Collins, Colorado, USA. This adjuvant is thought to be safer than, and equally as effective, as FCA and FIA in eliciting an antibody response. The new adjuvant (AdjuVacTM) is derived from a United States Department of Agriculture (USDA)-approved Johne’s disease vaccine (MycoparTM) which has previously been approved for use in food animals by the USDA, Animal and Plant Health Inspection Service (APHIS) (http://www.aphis.usda.gov/ws/nwrc/research/gnrh.html). A single application of GnRH-KLH and AdjuVacTM (GonaCon™) has the potential to be a safe, practical, and effective multi-year immunocontraceptive for wild ungulates. As a contraceptive for wildlife, this agent offers the following desirable characteristics: 1) A single treatment should provide long-term infertility (2 + years) when administered either to non-pregnant females prior to the breeding season or to pregnant females during gestation. 2) Treatment effectiveness should be 85-90% the first breeding season. 3) Infertility should be reversible. 4) The agent should be safe for pregnant animals and the developing fetus 5) The agent should not cause significant behavioral or physiological side-effects. 6) The proteinaceous nature of the GnRH-KLH immunogen should eliminate the possibility of passage through the food chain. 7) The small volume required for effective contraception should facilitate administration by syringe dart. 8) The agent is currently being evaluated for FDA approval as a New Animal Drug and therefore, could be available for commercial use in deer and elk. Preliminary investigations evaluating a single application of GnRH-KLH vaccine (GonaCon™) in captive female wild horses (Equus caballus) (Killian et al. 2004), bison (Bison bison) (Miller et al. 2004), white-tailed deer (Odocoileus virginianis) (Miller et al. unpublished data), California ground squirrels (Spermophilus beecheyi) (Nash et al. 2004), New Zealand white rabbits (Oryctolagus cunniculi) (Powers et al. in preparation) and domestic male cats (Felis catus) (Levy et al. 2004) are promising. All female bison (n = 5) treated with a single injection containing 1800µg GnRH-KLH and AdjuVac™ have remained infertile for 3 breeding seasons (Miller et al. 2004, Rhyan and Miller unpublished data). Similarly, mares (n = 18) treated with either 1800µg or 2800µg GnRH-KLH vaccine have been shown to be 100% infertile after one breeding season (Killian et al. 2004). While these results are encouraging, 87 �additional species specific studies are needed to confirm the safety and effectiveness of this contraceptive approach in wildlife. Our goal in this investigation is to conduct controlled experiments with captive elk to investigate the important attributes of this technology prior to management application in free-ranging wild ungulates. C. OBJECTIVES 1. To evaluate the effective duration of a single dose application of GnRH-KLH vaccine in preventing subsequent reproduction in pregnant elk. 2. To evaluate the effect of GnRH-KLH immunization on serum concentrations of LH and progesterone, corpus luteum (CL) function and viability, and neonatal health and survival. 3. To evaluate the effect of the GnRH-KLH vaccine on breeding behavior of captive elk following a contraceptive treatment applied during the second trimester of pregnancy. 4. To evaluate the physiological side-effects (if any) of GnRH-KLH vaccination on female elk, the developing fetus, and/or neonate. D. EXPECTED RESULTS OR BENEFITS The Colorado Division of Wildlife’s Strategic Plan (2002-2007), charges the agency with “finding alternatives for game management when hunting is not a viable option” (H-1.5, p 9). One of the performance measures for accomplishing this objective is to develop alternative methods of population control. Successful development and testing of the fertility control technology described in this proposal has the potential to accomplish this objective and provide resource managers with a non-lethal strategy for controlling the growth of some wild ungulate populations when sport hunting is infeasible. E. APPROACH Proposed Research: Working Hypothesis.: In this investigation, we test the general hypothesis that a single intramuscular application of a novel anti-GnRH vaccine in mid-gestation female elk will prevent pregnancy the following breeding season and may prevent pregnancy for two or more subsequent seasons. The exact duration of infertility is unknown but will be determined in this investigation. However, permanent sterility is not anticipated and we expect treated females to eventually return to normal estrous behavior and fertility as antibody titers decline. Furthermore, we don’t expect immunization against GnRH-KLH to cause substantial negative physiological or behavioral effects in peri-parturient females or neonatal calves. However, since GnRH-KLH vaccine is expected to suppress reproductive hormones, we predict diminished breeding behaviors in treated female elk compared to controls. Design: We will test the effects of GnRH-KLH vaccine treatments on pregnancy rates in elk using a Fisher’s exact test and evaluate serology of reproductive hormones, anti-GnRH antibody titers, and breeding behavior using a one-way ANOVA for a completely randomized design with repeated measures structure. Animals: Approximately, 20 adult female elk (2-12 years of age), 2 mature, and 2 sub-adult male elk will be used in this study. Elk are permanently maintained at the Colorado Division of Wildlife’s Foothills Wildlife Research Facility (FWRF) in Fort Collins, Colorado. The female elk used in these experiments have been previously trained to repeated handling, weighing, ultrasound, and blood sampling procedures. When not involved in periodic intensive sampling procedures required by this study, elk are maintained in fenced paddocks (5 ha) containing native vegetation and fed a diet consisting of ad libitum quantities of grass-alfalfa hay, grain supplement, trace mineral block, and water. 88 �Experiment 1: Effects of GnRH-KLH vaccine on pregnancy rates (objective 1) Hypothesis: One year post-vaccination, female elk vaccinated intramuscularly with 1000µg GnRH-KLH + adjuvant (AdjuVacTM) during the second trimester of gestation will have significantly (P < 0.05) higher anti-GnRH antibody concentrations and lower pregnancy rates than females treated with adjuvant alone. Rationale: Vaccination with GnRH-KLH + AdjuVacTM has successfully stimulated sufficient anti-GnRH antibody production to prevent pregnancy in a wide range of species including many ungulate species (Curtis et al. 2002, Miller et al. 2004, and Killian et al. 2004). Of particular importance to our experiment is the ongoing study with white-tailed deer (Miller, personal communication). Preliminary results suggest that multiple year infertility has been achieved with a single treatment of GnRH-KLH vaccine. Since elk and deer are taxonomically similar and share many common ecological, morphological and physiological traits, we expect to observe a similar contraceptive response for both species. Methods: Many of the measurements in this experiment (i.e. conception/parturition dates, pregnancy rates, luteal function, and hormone concentrations) will be facilitated by controlling the breeding period of female elk. To do this, we will attempt to synchronize estrous cycles of female elk by using progesterone secreting controlled internal drug release (CIDR) implants (Fennessy et al. 1990, Asher et al. 1993, Lucy et al. 2001). The CIDR implants will be placed in female elk during the last week of August 2005 (see appendix A for detailed protocol). Following CIDR removal (approximately the first week of September 2005), reproductively sound male elk will be released into the same pasture as females (see appendix B for breeding soundness exam protocol). During January 2006, we will determine pregnancy status of all females. Once pregnancy status is determined, pregnant elk will be blocked according to age and body condition, and randomly assigned to either a control or treatment group. We will determine pregnancy rates of treatment and control elk each year thereafter until differences in treatment effects can no longer be detected (P > 0.05). Treatment and control formulations will be applied in the following the manner. On the day of application (approximately mid-January, 2006), animals will be moved from paddocks, weighed (± 0.5 kg), and lightly sedated with xylazine hydrochloride (Rompun; Bayer AG, Leverkusen, Germany; 45-55 mg/animal, IM). This dose should allow animals to remain standing in the handling chute and minimize any possible stress or pain associated with blood collection, reproductive tract examination, ultrasound imaging of injection site, and dart delivery of treatments. All elk will be remotely injected in the area of the biceps femoris muscle with 1 ml, 13-mm-diameter, barbless darts equipped with gel-collared 32-mmlong needles (Pneu-dart, Williamsport, Pennsylvania, USA) fired from a CO2 – powered pistol (DanInjectTM, Fort Collins, Colorado, USA). Darts will be fired from approximately 3 m and will contain either GnRH-KLH vaccine + AdjuVacTM) (treatment) or AdjuVacTM alone (control). In order to accurately determine the precise dose delivered to each elk, darts will be weighed (0.001g) before and after injection. Once all elk have been treated, sedation will be reversed with yohimbine (30 mg, IV) (Antagonil®, Wildlife Pharmaceuticals, Fort Collins, Colorado, USA) and animals will be returned to holding pastures. Antibody titers will be measured immediately prior to treatment application and then on a monthly or bimonthly basis until maximal levels are reached. Following peak response, these measurements will be made on a less frequent basis until just prior to subsequent breeding seasons (September 2006, 2007, 2008). At that time, females will be sampled again. Except for this period, monthly sampling will be terminated following October 2006. 89 �The effective duration of GnRH-KLH vaccine in controlling fertility in elk will be determined by comparing pregnancy rates of treated and control elk during January 2007 and 2008. Once pregnancy rates are determined, pregnant elk will be aborted using a combination of prostaglandin F2 α and dexamethasone (Bates et al. 1982) (see appendix C for detailed protocol). Blood sampling procedures for antibody determination, pregnancy rates, hormone concentrations, and serum chemistry and hematology will follow methods previously described. While elk are sedated, blood samples (20-40 ml) will be collected via jugular venipuncture. Serum will be stored at – 70 ºC until analyzed for LH, progesterone, and anti-GnRH antibodies. Following the last blood collection, sedation will be reversed and elk returned to paddocks. Measurements: Anti-GnRH antibodies will be measured using an enzyme linked immunosorbent assay (ELISA) developed by scientists at the NWRC (USDA/APHIS) and/or using radioimmunoassay (RIA) techniques at Colorado State University’s Animal Reproduction and Biotechnology Laboratory (ARBL). The effect of GnRH-KLH vaccine on reproduction will determined in January 2007 and 2008 by measuring pregnancy rates using the presence or absence of pregnancy specific protein B (PSPB) (Huang et al. 2000), rectal palpation (Greer and Hawkins 1967, Hein et al. 1991) and/or ultrasound (Curran et al. 1986). Analysis: To determine the sample sizes needed to detect treatment differences for pregnancy rates, we performed a power-based sample size determination for a one-sided Fisher’s exact test using a software program (NCSS Pass 2000) (Kang and Kim 2004, Krishnamoorthy and Thompson 2002). For this analysis, we used the highest reported pregnancy rate (approximately 30%) for GnRH-KLH vaccine treated white-tailed deer, 1 year post-vaccination (Curtis et al. 2002). To represent the best and worst case scenarios for control elk, we calculated the sample size requirements for a 90% and 100% pregnancy rate. Based on this analysis, between 18−26 female elk (equally divided between control and treatment groups) should provide an adequate sample size to detect expected differences in pregnancy rates (Table 1). Because the pregnancy rates of the control and treatment groups are expected to be close to 1.0 and 0, respectively, the normal approximation invoked for testing the difference between 2 proportions is not valid (Brown et al. 2001). 90 �Table 1. Estimated sample sizes required to detect differences in pregnancy rates, in control and treatment groups, based on a Fisher’s exact test power analysis (NCSS Pass 2000). Power Control Group Pregnancy Rate Treatment Group Pregnancy Rate Total Sample Sizea 0.90 0.90 0.90 0.90 0.90 0.90 0.90 0.90 0.80 0.80 0.80 0.80 0.80 0.80 0.80 0.80 1.0 1.0 1.0 1.0 0.9 0.9 0.9 0.9 1.0 1.0 1.0 1.0 0.9 0.9 0.9 0.9 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3 8 12 14 18 12 14 20 26 8 10 12 16 10 12 16 22 a Total sample size assumes an equal number for each group, e.g. 18 means 9 treatment and 9 control female elk. Experiment 2: Effects of GnRH-KLH vaccine on luteal function and neonatesurvival (objective 2) Hypothesis: We are uncertain of the effects of GnRH-KLH vaccine treatments on LH secretion, luteal viability and fetal/neonatal survival. Little conclusive research has been conducted on these relationships in wild or domestic ungulates. Limited evidence suggests that GnRH-KLH-induced suppression of LH and progesterone levels in late gestation are not low enough to terminate pregnancy or negatively affect fetal/neonatal survival. Rationale: Corpora lutea (CL) secrete progesterone and are an essential ovarian structure for maintenance of pregenancy in all mammals (Baird 1992). Progesterone is obligatory for early embryonic development and peaks in the blood of pregnant females at different stages of gestation for different species. While progesterone is always produced by the CL in early pregnancy, its role in maintenance of pregnancy varies among species. In some species (i.e. mare, cow, ewe, and women) the CL is not needed for the entire gestational period because the feto-placental unit begins producing sufficient progestins to maintain pregnancy (Squires 1993, Stevenson 1997, Stellflug et al. 1997) In other species (i.e. sow, rabbit, whitetailed deer), surgical removal of the CL will terminate pregnancy regardless of when it occurs during gestation (Plotka et al. 1982, Tomas 1997, Tast et. al 2000). It is well-documented that progesterone secretion is regulated by several hormones, including LH, which plays a principal role in CL function during both the estrus cycle and pregnancy (Niswender et al. 1976, 1994, Rahe et al. 1980, Farin et al. 1990, Okuda et al. 1999). In contrast, however, studies in cattle (Peters et al. 1994), pigs (Buhr 1987), dogs (Onclin et al. 2000), and to some extent sheep (McNeilly and Fraser 1987) provide evidence that LH may not be essential for all aspects of luteal function, including 91 �pregnancy. For wild ungulates, LH suppression due to high doses of the GnRH agonist, leuprolide, were not sufficiently luteolytic to terminate pregnancy when administered to elk during the first 60 days of gestation (Baker et al. 2001). Likewise, bison vaccinated with GnRH-KLH during the second and third trimesters of pregnancy, maintained a viable fetus throughout gestation and delivered healthy calves at parturition (Miller et al. 2004). In this experiment, elk will be vaccinated with GnRH-KLH at approximately 120 days gestation and should develop sufficient antibody titers to suppress LH by 180-200 days of gestation (Miller et al. 2000b).Because the average gestation period in elk is 255 days (Haigh and Hudson 1993), the animals in this experiment will be in the third trimester of pregnancy before the CL is significantly affected by lack of LH. If the elk respond similarly to cattle and bison they will likely retain the pregnancy despite expected declines in progesterone. Alternatively, if elk are highly sensitive to small changes in progesterone concentrations they may abort the fetus. Methods: One day prior to GnRH-KLH vaccine treatments in January, 2006, we will collect blood for antibody titers, LH and progesterone concentrations (Niswender et al. 1969, Niswender 1973), and perform reproductive examinations on all experimental elk. Beginning approximately 4 weeks posttreatment, and in conjunction with scheduled monthly measurements of antibody titers, we will monitor changes in these parameters until 15 April, 2006. Following parturition (approximately June 1-15), we will monitor neonatal health, survival, and growth to 30 days post-parturition. Weaned calves, not needed as replacement animals in other experiments or at other captive research facilities, will be humanely euthanized. At present, we have received a proposal from scientist at USDA/APHIS National Wildlife Research Center to use surplus elk calves in a terminal experiment to develop and test orally active vaccines for managing infectious diseases such as bovine tuberculosis and brucellosis. Analysis: We will use descriptive statistical methods to analyze hormonal data. Hormone concentrations, fetal and CL measurements will be reported as arithmetic means ± ŜE . We will estimate the correlation coefficient between antibody titers, hormone concentrations, CL measurements, and test whether these relationships are significantly different from zero (Zar 1984). We will compare the differences in growth rates (g/da) of calves born to treatment and control females from birth to 30 days of age using a two sample t-test. Experiment 3: Effects of GnRH-KLH vaccine on breeding behavior (objective 3) Hypothesis: The effectiveness of GnRH-KLH vaccine as a contraceptive agent is dependent on the suppression of ovulation and steroidogeneis. Because GnRH-KLH vaccine is expected to suppress estradiol and therefore sexual receptivity during estrus, we predict that 1) rates of male precopulatory, female precopulatory, and copulatory behavior will be lower for treated females compared to untreated controls, and 2) that rates of general breeding behavior (i.e. herding, establishing and/or defending a harem) will be similar for both treated and untreated females. Rationale: In previous research (Baker et al. 2002), we reported that breeding behavior rates of female elk treated with GnRH agonist were not different from those of untreated elk. We attributed this response to basal estradiol concentrations inducing reproductive receptivity in animals that had been exposed to progesterone earlier in the breeding season or during a “silent estrus” (Harder and Moorhead 1980, Asher 1985). However, in the present experiment, GnRH-KLH vaccine should suppress progesterone secretion, 92 �estrogen, folliculogenesis and ovulation well in advance of the onset of the September 2006 breeding season. Therefore, there should be no progesterone “priming” effect and no estrous behavior in treated females. Limited observations of male elk engaged in general breeding behaviors related to establishing and/or defending a harem suggest that they don’t discriminate between cycling and non-cycling females (Baker et al. 2002). If true, general breeding behavior rates of females treated with GnRH-KLH vaccine in the present experiment should not be different from untreated females. Methods: We will test these hypotheses by examining the effects of GnRH-KLH vaccine on reproductive behaviors of female elk during the breeding season (15 September to 31 October 2006). Our experimental unit for analysis will be individual females within each treatment group. On 15 August 2006, two male elk will be placed with treated and control female elk in the same paddocks. All females will be individually identified with color numeric-coded neck collars. Observers will not know which elk are treatments or controls. Behavior observations will be made from a distance of 50-250 m from an elevated tower using binoculars and spotting scope during day, and a spotlight and night vision scope at night. Selected behaviors (Geist 1982) will be recorded using a lap-top computer with a behavioral software program. All-animal all occurrences sampling procedures will be used to sample reproductive behaviors of all experimental animals over a 24 h period (Leaner 1996). Time-of-day sampling periods will be assigned each week using a randomized block design. Each sampling period will consist of at least 2 h of continuous observations. We will group individual sexual behaviors into four general categories (Table 2).Because behavioral interactions are generally short duration (< 30 sec) relative to the sampling interval, we will record the number of occurrences of each behavior rather than the length of time, and calculate rates of sexual interactions as behaviors per animal per hour, then multiply hourly behavior rates by 24 for a daily rate. Table 2. Elk reproductive behavior and associated behavior categories (Baker et al. 2002). Behavior Category General Breeding Male Precopulatory Female Precopulatory Copulatory Reproductive Behavior Male directed behavior related to establishing, maintaining and defending a group or harem of female elk (i.e. Herding, guarding, tending) Male courtship behavior directed toward an individual female to induce or detect estrus or ovulation (i.e. urine testing, flehmen, tongue flick, smell or rub) Female courtship behavior directed toward dominant male to arouse copulatory behavior (i.e. lick and rub male, mount, lordosis, twitch hocks) Male behavior directed toward a receptive female in estrus (i.e. intromission) Analysis: Based on the sample sizes required to detect differences in pregnancy rates (Table 1), we conducted a simulation to estimate the power to detect differences in behavior rates. To complete this simulation, we bootstrapped data from a previous study which examined the effects of an alternate fertility control agent, leuprolide, on female elk reproductive behavior during the breeding season (Baker et al. 2002). Each sample was run through Proc Mixed (SAS 1996) using repeated measures mixed effects structure. The following parameters were used to estimate power based on total experimental sample sizes of 18, 20, or 26 female elk (Table 3). 1. Male pre-copulatory behavior rate was previously shown to be higher than other reproductive behavior rates (Baker et al. 2002). Consequently, this measurement will likely be the most sensitive to detection of treatment effects. Therefore, we used the previously reported male precopulatory behavior rates to estimate power for our simulation. 2. The peak of breeding season is approximately one month in length. Therefore, we estimated power using 60 total observation periods. [4wks x 3 observation periods/da x 5 da/wk = 60 obs. periods] 93 �3. We estimated power for 3 different sample sizes using 60 observation periods and bootstrapping data from the previous leuprolide experiment. Ten control elk were randomly selected (with replacement) from the 5 control female elk of the previous leuprolide experiment (thus some elk were used multiple times in a sample). The behavioral response (male precopulatory behavior rate) for each elk was recorded; thus a complete sample consisted of 10 behavior rates for each of the 60 observation periods. Behavior data from control elk was considered the benchmark for comparison. Hence, to estimate power for treatment elk in the current experiment, we followed the same procedure except that the response was multiplied by the effect size. To represent a 50% decrease in the male precopulatory rate directed toward treatment elk, we multiplied the control response by 0.5. 4. Although our behavioral hypothesis predicts that treated female elk will have reduced reproductive behaviors compared to controls, we also estimated sample size for the possibility of increased behavior rates. Thus, we estimated sample sizes for a 75% and 50% reduction in male precopulatory behavior rates toward treated female elk compared to controls, as well as a 50% increase in male precopulatory behavior rates. 5. Power results are based on the number of times an effect was detected during 100 simulations. Because the variance is larger for higher behavior rates, there is less power to detect a 50% increase compared to a 50% decrease. From this analysis we conclude that a sample size of 20; 10 treatment and 10 control animals, and 60 observation periods would provide adequate power (>90%) to detect most of the differences in reduced reproductive behavior rates as well as reasonable power (>75%) to detect a 50% increase in behavior rates between treated and untreated elk. Table 3. Power results for detecting differences in male precopulatory behavior rates directed toward treated and untreated female elk based on 100 simulations and 60 observation periods. Difference Between Treatment and Controlsa 0.25 0.25 0.25 0.50 0.50 0.50 1.50 1.50 1.50 Total Sample Sizeb 18 20 26 18 20 26 18 20 26 a Power α=0.05 1.00 1.00 1.00 0.99 1.00 1.00 0.70 0.76 0.92 Power α=0.10 1.00 1.00 1.00 0.99 1.00 1.00 0.81 0.84 0.95 Effect size Total sample size assumes an equal number for each group, e.g. 18 means 9 treatment and 9 control female elk. b 94 �Experiment 4: Effects of GnRH-KLH vaccine on maternal behavior, neonatal survival and growth, blood chemistry and hematology (objective 4) Hypothesis: GnRH-KLH vaccine treatments will not result in significant secondary negative behavioral or physiological side-effects in female elk. Rationale: To date, the GnRH-KLH vaccine formulated with AdjuVacTM has produced few reported behavioral or physiological side-effects in any species in which it has been tested (Levy et al. 2004, Miller et al. 2004, Killian et al. 2004). However, it’s not clear from these studies how extensively the side-effects of this agent have been evaluated. In this investigation, we will evaluate the effects of GnRHKLH vaccine on maternal/neonatal behavior, neonatal growth and development, serum biochemistry, and injection site reactions. Methods: Injection site reactions. On the day prior to treatment application (early January 2006) and while elk are lightly sedated (see pages 6-7 for details), we will perform ultrasound examination of the area of the expected injection site. After dart delivery of the vaccine, we will grossly monitor the injection site on a daily basis for one month for signs of inflammation or drainage. In addition, we will use ultrasound imaging each month for 6 months in conjunction with scheduled animal handling and blood collections to monitor changes in muscle echogenicity that would indicate sub-clinical abscesses or granuloma formation. Blood Chemistry and hematology. The physiological side-effects of GnRH-KLH treatment will be assessed by comparing serum chemistry, hematology, and body weight dynamics of treated and control elk. Blood samples will be collected in conjunction with previously described measurements just prior to GnRH-KLH vaccination and one week post-vaccination for evidence of short-term inflammation or infection. At three months post-vaccination, additional blood will be collected and analyzed for biochemistry profiles and evidence of abnormal organ function. These samples will be submitted for analysis to Colorado State University, Veterinary Teaching Hospital, Clinical Pathology Laboratory, Fort Collins, Colorado for analysis. Maternal Bonding, neonatal survival and growth. We will compare maternal/neonatal bonding and neonatal survival and growth of treated and control female elk for 30 days post-parturition during approximately 1 June to 1 July 2006. Parturition behavior of elk will be monitored daily beginning in late May and early June. We will document evidence of dystocia for each adult female, calf birth weight and health, acceptance or rejection by the dam, and growth to 30 days of age. For the purpose of this experiment, we assume that calf survival after 30 days is no longer a function of GnRH-KLH vaccine treatment and multiple factors other than maternal bonding will influence neonatal survival and body weight dynamics. Analysis: Means and standard errors of blood parameters, and neonatal growth rates will be estimated using least–squares ANOVA. Hypothesis tests will be based on type III generalized equations that account for correlations in repeated measures. 95 �Project Schedule: Date 1 May – June 2005 1 Sept 2005 7 Sept 2005 1 Jan 2006 1 Jan 2006 1 Feb – Sept. 2006 1 Feb – June 2006 1 June – July 2006 15 Sept. 2006 – 31 Oct. 2006 Jan 2007 Jan 2008 Jan 2009 Mar – July 2009 Activity Submit study plan for CDOW peer review and ACUC approval Semen evaluation and CIDR’s in all experimental female elk. Remove CIDR and combine males and females. Determine pregnancy status of females and assign to experimental groups. Apply contraceptive and monitor short term health effects. Monitor antibody titers of experimental elk. Monitor hormone levels Monitor birth rates, calf survival, calf weights and cow/calf behavior. Evaluate reproductive behavior Evaluate pregnancy rates 1 year post- vaccination. If funding is available, evaluate pregnancy rates 2 year post-vaccination If funding is available, evaluate pregnancy rates 3 year post-vaccination and/or reversibility of contraceptive treatments. Analyze data and prepare final report. Budget: This research proposal has been submitted to the Morris Animal Foundation for possible funding during the period June 2006 to Jan 2008. Category 2005-‘06 2006-‘07 Total Personal Services 0 0 0 1. Co-Investigator(s) 2. Biometrician 3. Wildlife Technicians (TBA) 5,000 6,675 2,500 11,175 7,500 17,850 Total Salaries and Wages 11,675 13,675 25,350 3,600 3,600 500 2,160 1,600 1,000 1,200 1,200 500 2,160 0 1,000 4,800 4,800 1,000 4,320 1,600 2,000 12,460 6,060 18,520 Operating Supplies & Services 1.Hormone serology LH analysis (240 x $20) Progesterone (240 x $20) PSPB (40 x $25) 2. GnRH Antibody Assays (360 x $12) 3. Biochemistry profile and CBC’s (40 x $40) 4. Miscellaneous veterinary supplies Total Supplies & Expenses Animal Maintenance Total Animal Care 5,760 5,760 11,520 Subtotal of All Categories 29,895 25,495 55,390 2,391 2,039 4,430 32,286 27,534 59,820 *Maximum of 8% - Indirect Costs (University Overhead) Grant Total 96 �F. LOCATION: This study will be conducted at the Colorado Division of Wildlife’s Foothills Wildlife Research Facility in Fort Collins, Colorado, USA. G. RELATED FEDERAL AID PROJECTS: N/A H. LITERATURE CITED ADAMS, T. H., AND B. M ADAMS. 1990. Reproductive function and feedlot performance of beef heifers actively immunized against GnRH. Journal of Animal Science 68:2793-2802. ASHER, G. W., AND J. L. ADAM. 1985. Reproduction of farmed red and fallow deer in northern New Zealand. Pages 217-224 In: P.F. Fennessy and K. R. Drew (eds.), Biology of deer production. Bulletin No. 22, Royal Society of New Zealand, Wellington, New Zealand. __________, M. W. FISHER, P. F. FENNESSY, C. G. MACKINTOSH, H. N. JABBOUR, AND C. J. MORROW. 1993. Oestrous synchronization, semen collection, and artificial insemination of farmed red deer (Cervus elaphus) and fallow deer (Dama dama). Animal Reproduction Science 33:241-265. ANDERSON, D. J., AND N. J. ALEXANDER.1983. A new look at antifertility vaccines. Fertility and Sterility 40: 557-571. ASPDEN, W. J., A. RAO, P. T. SCOTT, I. J. CLARK, T. E. TRIGG, J. WALSH, AND M. J. D’OCCHIO . 1996. Direct actions of the luteinizing hormone-releasing hormone agonist, deslorelin, on anterior pituitary contents of luteinizing hormone (LH) and follicle stimulating hormone (FSH), LH and FSH subunit messenger ribonucleic acid, and plasma concentrations of LH and FSH in castrated male cattle. Biology of Reproduction 55:386-392. BAIRD, D. T. 1992. Luteotrophic control of the corpus luteum. Animal Reproduction Science 28:95-102. BAKER, D. L., M. A. WILD, AND T. M. NETT. 2001. Technical support for elk and vegetation management in Rocky Mountain National Park. Wildlife Research Report. Colorado Division of Wildlife, Fort Collins, Colorado, USA. __________, ________, M. M. CONNER, H. B. RAVIVARAPU, R. L. DUNN, AND T. M. NETT. 2002. Effects of GnRH agonist (leuprolide) on reproduction and behavior in female wapiti (Cervus elaphus nelsoni). Reproduction Supplement 60:155-167. ___________, _________, M. D. HUSSAIN, R. L. DUNN, AND T. M. NETT. 2003. Evaluation of remotely delivered leuprolide formulation as a contraceptive agent in captive female elk. Technical report for elk and vegetation management for Rocky Mountain National Park, Wildlife Research Report, Colorado Division of Wildlife, Fort Collins, Colorado, 125 pp. ____________, _________, M. M. CONNER, H. B. RAVIVARAPU, R. L. DUNN, AND T. M. NETT. 2004. Gonadotropin-releasing hormone agonist: a new approach to reversible contraception in female deer. Journal of Wildlife Diseases 40:713-724. BARLOW, N. D.1996.The ecology of wildlife disease control: simple models revisited. Journal of Applied Ecology 33:303-314. BARTH, A. D. 1986. In: Current Therapy in Theriogenology. D. A. Morrow (ed.) W. B. Saunders Co. pp205-209 ____________1997. Evaluation of potential breeding soundness of the bull. In: Current Therapy in Large Animal Theriogenology. R.S Youngquist (ed.) W.B. Saunders Co. pp222-236 BATES, G. N., J. BROOKS, AND J. CALL.1982.The use of prostaglandin to induce abortion in American elk (Cervus Canadensis). Zoo Animal Medicine 13:125-126 BOMFORD, M. 1990. A role for fertility control in wildlife management. Department of Primary Industries and Energy, Bureau of Rural Resources Bulletin No. 7, Australian Government Publishing Service, Canberra, Australia. BROWJN, B. W., P. E. MATTNER, P. A. CARROLL, E. J. HOLLAND, D. R. PAULL, R. M. HOSKINSON, AND R. D. G. RIGBY.1994. Immunization of sheep against GnRH early in life: Effects on reproductive function and hormones in rams. Journal of Reproduction and Fertility 101:15-21. 97 �BROWN, D. B., T. T. CAI, AND A. DASGUPTA. 2001.Interval estimation for a binomial proportion. 2001. Statistical Science 16:101-133. BUHR, M. M. 1987. Effect of lipoproteins and luteinizing hormone on progesterone production by large and small luteal cells throughout the porcine estrous cycle. Journal of Animal Science 65:10271033. CURRAN, S., R. A. PIERSON, AND O. J. GINTHER.1986.Ultrasonographic appearance of the bovine conceptus from days 20 through 60. Journal of the American Veterinary Medical Association 189:1295-1302. CURTIS, P. D., R. L. POOLER, M. E. RICHMOND, L. A. MILLER, G. F. MATTFELD, AND F. W. QUIMBY. 2002. Comparative effects of GnRH and porcine zona pellucida (PZP) immunocontraceptive vaccines for controlling reproduction in white-tailed deer (Odocoileus virginianus). Reproduction Supplement 60:131-141. DECKER, D., AND A. N. CONNELLY. 1989. Deer in suburbia-pleasure or pets. Conservationist 43:46-49. D’OCCHIO, M. J., W. J. ASPDEN, AND T. R.WHYTE.1996. Controlled, reversible suppression of oestrous cycles in beef heifers and cows using agonist of luteinizing hormone-releasing hormone. Journal of Animal Science 74:218-225. ____________ , G. FORDYCE, T. R. WHYTE, T. F. JUBB, L. A. FITZPATRICK, N. J. COOPER, W. J. ASPDEN, M. J. BOLAM, AND T. E. TRIGG. 2002. Use of GnRH agonist implants for long-term suppression of fertility in extensively managed heifers and cows. Animal Reproduction Science 74: 151-162. DUNBAR, B. S., AND E. SCHWOEBEL.1988. Fertility studies for the benefit of animals and human beings: Development of improved sterilization and contraceptive methods. Journal of the American Veterinary Medical Association 193: 1165-1170. FAGERSTONE, K. A., M. A. COFFEY, P. D. CURTIS, R. A. DOLBEER, G. J. KILLIAN, L. A. MILLER, AND L. WILMOT. 2002. Wildlife fertility control. Wildlife Society Technical Review 02-2. The Wildlife Society, Bethesda, Maryland, USA. FARIN, C. E., T. M. NETT, AND G. D. NISWENDER. 1990. Effects of luteinizing hormone on luteal cell populations in hypophysectomized ewes. Journal of Reproduction and Fertility 88:61-70. FENNESSY, P. F., C. G. MACKINTOSH, S. G. SHACKELL. 1990. Artificial insemination of farmed red deer (Cervus elaphus). Animal Production 51:613-621. FROST, H. C., G. L. STORN, M. J. BATCHELLER, AND M. M. LOVALLO. 1997. White-tailed deer management at Gettysburg National Military Park and Eisenhower National Historic Site. Wildlife Society Bulletin 25:462-469. FRAKER, M. A., R. G. BROWN, G. E. GAUNT, J. A. KEER, AND B. POHAJDAK. 2002. Long-lasting, singledose immunocontraception of feral fallow deer in British Columbia. Journal of Wildlife Management 66:1141-1147. GARROTT, R. A., P. J. WHITE, AND C. A. VANDERBIL WHITE. 1993. Overabundance: an issue for conservation biologist? Conservation Biology 7:946-949. ___________ AND D. B. SINIFF. 1992. Limitations of male-oriented contraception for controlling feral horse populations. Journal of Wildlife Management 56:456-464. GEIST, V. 1982. Adaptive behavioral strategies. In Elk of North America: Ecology and Management pp 219-277. J. W. Thomas and D. E. Toweill (Eds). Stackpole Books, Harrisburg, Pennsylvania, USA. GREER, K. R., AND W. W. HAWKINS. 1967. Determining pregnancy in elk by rectal palpation. Journal of Wildlife Management 31:145-149. HAIGH, J. C. AND R. J. HUDSON. 1993. Farming Wapiti and Red Deer. Mosby-Year Book, Inc. pp.121125. ________, A. D. BARTH, W. F. CATES, AND G. J. GLOVER. 1985. Electro-ejaculation and semen evaluation of wapiti. Royal Society of New Zealand. Bulletin, No. 22. p. 197-203 _________, W. F. CATES, G. J. GLOVER, AND N. C. RAWLINGS. 1984. Relationships between seasonal changes in serum testosterone concentrations, scrotal circumference and sperm morphology of male wapiti (Cervus elaphus). Journal of Reproduction and Fertility 70:413-418 98 �_________, M. CRANFIELD, AND R. G. SASSER. 1988. Estrus synchronization and pregnancy diagnosis in red deer. Journal of Zoo Animal Medication 19202-207. HARDER, J. D., AND D. L. MOORHEAD. 1980. Development of corpora lutea and plasma progesterone levels associated with the onset of the breeding season in white-tailed deer (Odocoileus virginianus). Biology of Reproduction 22:185-191. HAZUM, E., AND P. M. CONN. 1998. Molecular mechanism of gonadotropin releasing hormone (GnRH) action. I. The GnRH receptor. Endocrine Review 9: 379-386. HEILMANN, T. J., R. A. GARROTT, L. L. CADWELL, AND B. L. TILLER. 1998. Behavioral response of freeranging elk treated with an immunocontraceptive vaccine. Journal of Wildlife Management 62:243-250. HEIN, R. G., J. L. MUSSER, AND E. F. BRACKEN. 1991. Serologic, parasitic and pregnancy survey of the Colockum elk herd in Washington. Northwest Science 65:217-222 HOBBS, N. T., D. C. BOWDEN, AND D. L. BAKER. 2000. Effects of fertility control on populations of ungulates: general stage-structured models. Journal of Wildlife Management 64: 473-491. HONE, J. 1992. Rate of increase and fertility control. Journal of Applied Ecology 29:695-698. HUANG, F., D. C. COCKRELL, T. R. STEPHENSON, J. H. NOYES, AND R. G. SASSER. 2000. A serum pregnancy test with a specific radioimmunoassay for moose and elk pregnancy-specific protein B. Journal of Wildlife Management 64:492-499 JEWELL, P. A., AND S. HOLT. 1981. Problems in management of locally abundant wild animals. Academic Press, New York, New York, USA. KAND, S. H., AND S. J. KIM. 2004. A comparison of the three conditional exact tests in two-way contingency tables using the unconditional exact power. Biometrical Journal 46:320−330. KARTEN, M. J., AND J. E. RIVIER. 1986.Gonadotropin-releasing hormone analog design. Structurefunction studies toward the development of agonist and antagonist: rationale and perspectives. Endocrine Review 7:44-46. KEEN, J. E., G. P. RUPP, P. A. WITTENBERG, AND R. E. WALKER. 1999. Breeding soundness examination of North American bison bulls. Journal of the American Veterinary Medical Association 214:1212-1217 KILLIAN, G., L. MILLER, N. K. DIEHL, J. RHYAN, AND D. THAIN. 2004. Evaluation of three contraceptive approaches for population control of wild horses. Proceedings of the Vertebrate Pest Conference 21:263-268 KIRKPATRICK, J. F., AND J. W. TURNER, JR. 1985. Chemical fertility control and wildlife management. Bioscience 35: 485-491. ____________, AND ___________. 1996. Applications of pig zona pellucida immunocontraception to wildlife fertility control. Journal of Reproduction and Fertility 51:183-189. ____________, AND ___________. 2002. Reversibility of action and safety during pregnancy of immunization against porcine zona pellucida in wild mares (Equus caballus). Reproduction Supplement 60:197-202. ____________, ___________, I. K. M. LIU, R. FAYER-HOSKEN, AND A. T. RUTBERG. 1997. Case studies in wildlife immunocontraception: wild and feral equids and white-tailed deer. Reproductive Fertility and Development 9: 105-110. ___________, I. K. M. LIU, AND J. W. TURNER, JR. 1990. Remotely-delivered immunocontraception in feral horses. Wildlife Society Bulletin 18:326-330. ___________, R. NAUGLE, I., K. M. LIU, M. BERNOCO, AND J. TURNER. 1995. Effects of seven consecutive years of porcine zonae pellucidae contraception on ovarian function in feral mares. Biology of Reproduction Monograph Series 1:411-418. KRISHNAMOORTHY, K. AND J. THOMSON. 2002. Hypothesis testing about proportions in two finite populations. American Statistician 56:215−222. LEENAARS, M., E. CLAASSEN, AND C. F. M. HENDRIKSEN.1996. Considering the side-effects of adjuvant products in immunization procedures. Laboratory Animals 30:40-43. LEHNER, P. N. 1996. Handbook of ethological methods. 2nd ed. Cambridge University Press, Cambridge. 99 �LEOPOLD, A. S., S. A. CAIN, C. M. COTTAM, AND I GABRIELSON. 1963. Wildlife management in the national parks. Transactions of the North American Wildlife and Natural Resources Conference 28:28-45. LEVY, J. K., L. A. MILLER, P. C. CRAWFORD, J. W. RITCHEY, M. K. ROSS, AND K. A. FAGERSTONE. 2004. GnRH immunocontraception of male cats. Theriogenology 62(6):1116-1130. LUCY, M. C., H. J. BILLINGS, W. R. BUTLER, L. R. EHNIS, M. J. FIELDS, D. J. KESLER, JR., J. E. KINDER, R. C. MATTOS, R. E. SHORT, W. W. THATCHER, R. P. WETTEMANN, J. V. YELICH, AND H. D. HAFS. 2001. Efficacy of an intravaginal progesterone insert and an injection of PGF2α for synchronizing estrus and shortening the interval to pregnancy in postpartum beef cows, peripubertal beef heifers, and dairy heifers. Journal of Animal Science 79:982-995. MATSUO, H., Y. BABA, R. M. G. NAIR, A. ARIMURA, AND A. V. SCHALLY. 1971. Structure of the porcine LH- and FSH-releasing hormone. I. The proposed amino acid sequence. Biochemical and Biophysical Research Communication 43:1334-1339. MCANINCH, J. B., editor. 1993. Urban deer: a manageable resource? Proceedings of the 1993 Symposium of the North Central Section, The Wildlife Society, St. Louis, Missouri, USA. MCCULLOUGH, D. R., K. W. JENNINGS, N. B. GATES, B. G. ELLIOT, AND J. E. DIDONATO. 1997. Overabundant deer populations in California. Wildlife Society Bulletin 25: 478-483. MCNEILLY, A. S., AND H. M. FRASER.1987. Effect of gonadotrophin-releasing hormone agonist – induced suppression of LH and FSH on follicle growth and corpus luteum function in the ewe. Journal of Endocrinology 115:273-282. MCSHEA, W. J., S. L. MONFORT, S. HAKIM, J. KIRKPATRICK, I. LIU, J. W. TURNER, JR., L. CHASSY, AND L. MUNSON. 1997. The effect of immunocontraception on the behavior and reproduction of white-tailed deer. Journal of Wildlife Management 61:560-569. MELOEN, R. H., J. A. TURKSTRA, H. LANKHOF, W. C. PUIJK, W. M. M. SCHAAPER, G. DIJKSTRA, C. J. G. WENSING, AND R. B. OONK. 1994. Efficient immunocastration of male piglets by immunoneutralization of GnRH using a new GnRH-like peptide. Vaccine 12:741-746. MILLER, L. A., B. E. JOHNS, AND G. J. KILLIN. 2000a. Long-term effects of PZP immunization on reproduction in white-tailed deer. Vaccine 18:568-574. __________ , _________, AND ___________ . 2000b. Immunocontraception of white-tailed deer with GnRH vaccine. American Journal of Reproductive Immunology 44:266-274. __________, J. C. RYAN, AND M. DREW. 2004. Contraception of bison by GnRH vaccine: a possible means of decreasing transmission of brucellosis in bison. Journal of Wildlife Diseases 40:725730. MONTOVAN, S. M., P. P. DAELS, J. RIVIER, J. P. HUGHES, G. H. STABENFELDT, AND B. L. LASLEY. 1990. The effect of potent GnRH agonist on gonadal and sexual activity in the horse. Theriogenology 33: 1305-1321. NASH, P. B., J. K. JAMES, L. T. HUI, AND L. A. MILLER. 2004. Fertility control of California ground squirrels using GnRH immunocontraception. Proceedings of the 21st Vertebrate Pest Conference 274-278. NAUGLE, R. E., A. T. RUTBERG, H. B. UNDERWOOD, J. W. TURNER, JR., AND I. K. M. LIU. 2002. Field testing of immunocontraception on white-tailed deer (Odocoileus virginianus) at Fire Island National Seashore, New York, USA. Reproduction Supplement 60:143-153. NETT, T. M., M. E. CROWDER, G. E. MOSS, AND T. M. DUELLO.1985.GnRH-receptor interaction. V. Down-regulation of the pituitary receptors for GnRH in ovariectomized ewes by infusion of homologous hormone. Biology of Reproduction 24:1145-1155. NISWENDER, G. D. 1973. Influence of the site of conjugation on the specificity of antibodies to progesterone. Steroids 22:413-424. __________, J. L. JUENGEL, W. J. MCGUIRE, C. J. BELFIORE, AND M. C. WILTBANK. 1994. Luteal function: The estrous cycle and early pregnancy. Biology of Reproduction 50:239-247. __________, L. E. REICHERT, JR., A. R. MIDGLEY, AND A. V. NALBANDOV. 1969. Radioimmunoassay for bovine and ovine luteinizing hormone. Endocrinology 84:1166-1173. 100 �___________, T. J., REIMERS, M. A. DIEKMAN, AND T. M. NETT. 1976. Blood flow: a mediator of ovarian function. Biology of Reproduction 14: 64-81. OKUDA, K., Y. UENOYAMA, C. NAITO, Y. SAKABE, AND N. KAWATE. 1999. Luteinizing hormone receptors in the bovine corpus luteum during the oestrous cycle and pregnancy. Reproduction, Fertility, and Development 11:147-151. ONCLIN, K., J. P. VERSTEGEN, AND P. W. CONCANNON. 2000. Time-related changes in canine luteal regulation: in vivo effects of LH on progesterone and prolactin during pregnancy. Journal of Reproduction and Fertility 118:417-424. PETERS, K. E., E. G. BERGFELD, A. S. CUPP.1994. Luteinizing hormone has a role in development of fully functional corpora lutea (CL) but is not required to maintain CL function in heifers. Biology of Reproduction 52 (Suppl) 1:35. PLOTKA, E. D., U. S. SEAL, L. J. VREME, AND J. J. OZOGA. 1982. Reproductive steroids in white tailed deer. IV. Origin of progesterone during pregnancy. Biology of Reproduction 26:258-262 __________, D. N. VEVEA, T. C. EAGLE, D. B. SINIFF, AND J. R. TESTER. 1989. Ovarian function in captive feral mares. Journal of Wildlife Diseases 25:574-579. PORTER, W. F., AND B. UNDERWOOD. 1999. Of elephants and blind men: deer management in the U.S. national parks. Ecological Applications 9:3-9. POWERS, J. G., P. B. NASH, J. C. RHYAN, AND L. A. MILLER. Comparison of AdjuVac™ and Freund’s Complete adjuvants with GnRH-KLH antigen in New Zealand white rabbits. (In preparation). RABB, M. H., D. L. THOMPSON, JR., B. E. BARRY, D. R. COLBORN, K. E. HEHNKE, AND F. GARZA, JR. 1990. Effects of active immunization against GnRH on LH, FSH, and prolactin storage, secretion, and response to their secretagogues in pony geldings. Journal of Animal Science 68:3322-3329. RAHE, C. H., R. E. OWENS, J. L. FLEEGER, H. J. NEWTON, AND P. G. HARMS. 1980. Pattern of plasma luteinizing hormone in the cyclic cow: dependence upon the period of the cycle. Endocrinology 107:498-503. RHYAN, J. C., AND M. D. DREW. 2002. Contraception: A possible means of decreasing transmission of brucellosis in bison. In: Brucellosis in elk and bison in the Greater Yellowstone Area, T.J. Kreeger (ed.). Greater Yellowstone Interagency Brucellosis Committee, Wyoming Game and Fish Department, Cheyenne, Wyoming, 99-108. ROPSTAD, E., AND D. LENVIK. 1991. The use of cloprostenol and prostaglandin F2α to induce luteolysis in reindeer calves (Ranger tarandus). Rangifer 11:13-16. RUDOLPH, B. A., W. F. PORTER, AND H. B. UNDERWOOD. 2000. Evaluating immunocontraception for managing suburban white-tailed deer in Irondequoit, New York. Journal of Wildlife Management 64:463-473. SAS INSTITUTE. 1997. SAS/STAT® users guide 6.03 edition. SAS Institute, Inc., Cary, North Carolina, USA SEAGLE, S. W., AND J. D. CLOSE.1996. Modeling white-tailed deer population control by contraception. Biological Conservation 76:87-91. SHIDELER, S. E., M. A. STOOPS, N. A. GEE, J. A. HOWELL, AND B. L. LASLEY. 2002. Use of porcine zona pellucida (PZP) vaccine as a contraceptive agent in free-ranging tule elk (Cervus elaphus nannodes). Reproduction Supplement 60:169-176. STELLFLUG, J. N. Y. S. WEEMS, AND W. W. WEEMS. 1997. Clinical reproductive physiology of ewes. In: Current Therapy in Large Animal Theriogenology. R.S Youngquist (ed.) W.B. Saunders Co. pp594-598 STEVENSON, J. S. 1997. Clinical reproductive physiology of the cow. In: Current Therapy in Large Animal Theriogenlogy. R.S Youngquist (ed.) W.B. Saunders Co. pp257-267 STILLS, H. F., JR., AND M. Q BAILEY. 1991. The use of Freund’s complete adjuvant. Laboratory Animals 20:25-30. SQUIRES, E. L.1993. Endocrinology of pregnancy. In: Equine Reproduction A. O. McKinnon and J. L. Voss (Eds). Williams and Wilkins, Pennsylvania, USA. pp495-508 TAST, A., R. J. LOVE, I. J. CLARKE, AND G. EVANS. 2000. Effects of active and passive gonadotrophin- 101 �releasing hormone immunization on recognition and establishment of pregnancy in pigs. Reproduction, Fertility and Development 12:277-282. TOMAS, P. G. A. 1997. Induced abortion. In: Current therapy in large animal theriogenology, R. S. Youngquist (ed.).W.B. Saunders Co. 303-306 TURKSTRA, J. A., F. J. U. M. VAN DER MEER, J. KNAAP, P. J. M. ROTTIER, K. J. TEERDS, B. COLENBRANDER, AND R. H. MELOEN. 2005. Effects to GnRH immunization in sexually mature pony stallions. Animal Reproduction Science 86:247-259. TURNER, J. W., JR., I. K. M. LIU, AND J. F. KIRKPATRICK. 1992. Remotely delivered immunocontraception in captive white-tailed deer. Journal of Wildlife Management 56:154-157. __________, J. F. KIRKPATRICK, AND I. K. M. LIU. 1996. Effectiveness, reversibility and serum antibody titers associated with immunocontraception in captive white-tailed deer. Journal of Wildlife Management 60:45-51. __________, AND _________. 2002. Effects of immunocontraception on population longevity and body condition in wild mares (Equus caballus). Reproduction Supplement 60:187-195. WALTER, W. D., P. J. PERKINS, A. T. RUTBERG, AND H. J. KILPATRICK. 2002. Evaluation of immunocontraception in a free-ranging suburban white-tailed deer herd. Wildlife Society Bulletin 30:186-192. WRIGHT, R. G. 1993. Wildlife management in parks and suburbs: alternatives to sport hunting. Renewable Resources Journal 11: 18-22. ZAR J. H. 1984. Biostatistical analysis. Prentice-Hall, Inc. Englewood Cliffs, New Jersey, USA. 102 �APPENDIX II BAKER, D. L., M. A. WILD, M. M. CONNER, H. B. RAVIVARAPU, R. L. DUNN, AND T. M. NETT. 2005. Evaluation of a remotely delivered formulation of leuprolide acetate as a contraceptive agent in female elk (Cervus elaphus nelsoni). Journal of Wildlife Diseases 41: in press. Abstract: Practical application of fertility control technology in free-ranging wild ungulates often requires remote delivery of the contraceptive agent. The objective of this investigation was to evaluate the potential of remote delivery of leuprolide acetate for suppressing fertility in female elk (Cervus elaphus nelsoni). Fifteen captive adult female elk were randomly allocated to one of three experimental groups. Six elk were injected intramuscularly with a dart containing leuprolide, and the remaining nine elk received the same formulation without leuprolide. We determined pregnancy rates, suppression of luteinizing hormone (LH) and progesterone concentrations, and reversibility of treatments during 1 August 2002 to 3 September 2003. Leuprolide formulation caused a decrease in concentrations of LH and progesterone, temporary suppression of ovulation and steroidogenesis, and effective contraception (100%) for one breeding season. These results extend the practical application of this contraceptive agent to include dart delivery, where in the absence of such technology, wild elk must first be captured and restrained prior to treatment. BAKER, D. L., M. A. WILD, M. M. CONNER, M. D. HUSSAIN, R. L. DUNN, AND T. M. NETT. 2006. Evaluation of leuprolide as a contraceptive agent in free-ranging elk in Rocky Mountain National Park, Colorado. Journal of Wildlife Management (in preparation). ___________., _________, M. D. HUSSAIN, R.L. DUNN, AND T. M. NETT. 2006. Leuprolide acetate as a contraceptive agent in female elk: determination of minimum effective dose. Reproduction (in preparation) LUKACS, P., J. GROSS, AND D. BAKER. 2006. Estimating confidence intervals for fawn:doe and buck:doe ratios from counts across days. Journal of Wildlife Management (in preparation). INSLERMAN, R. A., J. E. MILLER, D. L. BAKER, J. E. KENNAMER, R. CUMBERLAND, B. STINSON, P. DOERR, AND S. J. WILLIAMSON. 2005. The Wildlife Society Technical Review Committee on Baiting and Artificial Feeding of Game Wildlife Species (in preparation). 103 � https://cpw.cvlcollections.org/files/original/87f1bec9d928f16c03ecdd82c4ab0334.pdf 66f7fb0598aaede242d73226ee8869e8 PDF Text Text 191 Colorado Division of Wildlife Wildlife Research Report July 2001 and July 2002 JOB PROGRESS REPORT State of ______C=-o=l=o=ra=d...o'--------- Mammals Research Work Package ------=-30-=--0=2"--------- Elk Conservation Task No. 3 ------"---------- Estimating Calf and Adult Survival Rates and Pregnancv Rates of Gunnison Basin Elk Project No. _ _ _W--'-'---"'-1=53=----=-R,._-.:. . 14..:. . . . :1=5_ _ _ __ Research and Development Period Covered: July 1, 2000 - June 30, 2001, and July 1, 2001 - June 30, 2002 Author: D. J. Freddy Personnel: D. Cole, N. Gallowich, D. Masden, L. Gepfert, J. Olterman, R. Basagoitia, L. Spicer, B. Carochi, J. Oulton, P. Mason, W. Brown, V. Organek, J. Young, T. Beck, G. Dekleva, C. Mehaffey, D. Williams, J. Johnston, K. Buffington, K. Fanson, and R. Kahn of CDOW, Dr. G. C. White and Dr. M.C. Conner of Colorado State University, CDOW Gunnison Habitat Partnership Program, and contractors/ cooperators, Helicopters by OZ, Coulter Aviation, USFS, BLM, private land owners, and elk hunters. ABSTRACT We estimated survival rates and pregnancy rates of elk (Cervus elaphpus nelsonii) in the Gunnison Basin of Colorado during 2000 and 2001. During mid-December each year, we captured and radio-collared calves age 6 months and adult females age 2:2 years and during November and December each year, we had hunters collect and submit reproductive organs from female elk harvested during laterifle seasons. During winter-spring, survival rates of calves were 0.89 ± 0.08 (CL) (n = 71), 0.83 ± 0.09 (n = 75), and 0.86 ± 0.06 (n = 146) for 2000-01, 2001-02, and both years combined, respectively. Survival of calves was not different between years (P = 0.2965) or sexes (P = 0.1456) but tended to be different among 3 management DAUs (P = 0.0737) with survival being lowest at 0.78 ± 0.12 in DAU E41. For years combined, 21 calves died with proximate causes of death being 53% predation-related, 24% malnutrition-related, 9% accidents, and 14% unknown causes. In 2000-01, calves tended to die after mid-March while in 2001-02, mortalities occurred from early January through May. Patterns of calf mortalities were not strongly associated with calf body mass. Calf body mass at capture averaged 99.1 ± 2.2 kg, ranged from 52.0 to 133.0 kg, and was not different between years or sexes (P > 0.259), but calves were larger in DAU E-41 (P = 0.003) where calf survival was lowest. Survival rate of adult females age 2:_2 years was 1.00 during winter-spring as no deaths occurred during 2000-01 (n = 39) and 2001-02 (n = 48) and annual survival was 0.92 ± 0.08 (n = 39) including hunting and other causes of death and 0.97 ± 0.05 (n = 37) including only natural deaths. Survival for yearling female elk, age 12-17 months, during summer-fall was 0.89 ± 0. 10 (n = 38) including hunting and other causes of death and 1.00 (n = 34) including only natural deaths. Survival for the same cohort of yearling male elk during summer-fall was 0.86 ± 0.15 (n = 22) including hunting and other causes of death and 0.90 ± 0.13 (n = 21) including only natural deaths. Survival rates for both yearling female and male elk, age 18-23 �192 months, were 1.00 (n = 34 F, n = 19 M) during winter-spring as no deaths occurred. Harvest removal rates during summer-fall 2001 were 0.05 for adult females, 0.11 for yearling females, 0.08 for all adult females age ~12 months, and 0.06 for yearling males. Based on biological collections provided by hunters, pregnancy rate averaged 85% for all adult female elk age ~1 year (n = 89). Conceptions peaked 23 September, spanned 68 days, and followed an expected asymmetrical pattern in timing with 17% of the adult females likely conceiving after 10 October. Litter size was 1 in all uteri with detectable fetuses (n = 69) and female fetuses predominated with fetal sex ratio (37F:21M) deviating from 50:50 (P = 0.036). Estimated percent total body fat based on kidney fat measurements indicated 65% of the adult females age ~1 year were in moderate, 30% in low, and <5 % in very low or very good body condition. Probability (logit(P)) of adult females being pregnant was dependent on estimated percent total body fat (P = 0.033). Measures ofreproductive and survival rate parameters were consistent with predictions of performance outcomes for adult female elk having low to moderate body condition status in the fall. More than likely, marginally deficient levels of seasonal nutrition were depressing optimal reproductive performance of adult female elk. All information in this report is preliminary and subject to further evaluation. �193 JOB PROGRESS REPORT ESTIMATING CALF AND ADULT SURVIVAL AND PREGNANCY RA TES OF GUNNISON BASIN ELK POPULATIONS David J. Freddy P. N. OBJECTIVE Estimate survival rates of calf, adult female, and adult male elk and estimate pregnancy rates of adult female elk in Gunnison Basin elk populations for 3 years. SEGMENT OBJECTIVES 1. Prepare study plan program narrative. 2. Estimate calf, adult female, and adult male survival rates during winter, December-June. 3. Estimate adult male and female survival rates during summer-fall, June-November. 4. Estimate harvest removal rates for yearling and adult males and females. 5. Estimate pregnancy rates, fetal rates, conception dates, and body condition of female elk collected in December. 6. Summarize data in Research Progress reports and prepare peer-reviewed publications. INTRODUCTION The elk (Cervus elaphpus nelsonii) resource has many benefits but frequent social, political, and economic conflicts suggest elk can reach "social" if not "biological" carrying capacities (Freddy et al. 1993). Recent controversy surrounding elk in the Gunnison Basin of Colorado (Roath et al. 1999) exemplifies conflicting social and biological agendas regarding appropriate numbers of elk. The core of conflict in elk management often centers on establishing management objectives for numbers of elk that are agreeable to competing interests and then monitoring elk populations to demonstrate that objectives are achieved. This type of conflict is paramount in Colorado Division of Wildlife (CDOW) elk population Data Analysis Units (DA Us) E-25, E-41, and E-43 in the Gunnison Basin where a combination of resource carrying capacity objectives for elk on winter ranges and difficulties associated with knowingly achieving those objectives has fostered argumentative distrust among public groups and management agencies. Accomplishing management by population objective can depend on reliably estimating elk population size which is expensive and intensive (Samuel et al. 1987, Bear et al. 1989, Unsworth et al. 1990, Anderson et al. 1998, Cogan and Diefenbach 1998, Eberhardt et al. 1998, Freddy 1998). Alternatively, population size and trend can be estimated using computer models that incorporate harvest, age and sex ratios, and survival rates (White 1992, Bartholow 1999). • Model outputs are extremely sensitive to estimates of survival rates such that, reliable measurements of survival can greatly enhance the quality of models (Nelson and Peek 1982). We chose to estimate survival rates of calf and adult elk during winter and adults year-around to aid in developing improved population models for the Gunnison Basin elk. The Gunnison Basin in southcentral Colorado encompasses the entire headwaters of the main Gunnison River and the centrally �194 located town of Gunnison. Between 12-16,000 elk and 8-10,000 mule deer (Odocoileus hemionus) are thought to exist within the Basin. Elk are managed as 3 populations representing DA Us E-25 (Game Management Units [GMU] 66, 67), E-41 (GMU 54), cµid E-43(GMUs 55,551). The 3 DAUs encompass about 9,291 km2 of which 3,648 km2 are considered potential winter range for elk (CDOW WRIS database). DA Us are contiguous with few major geographic barriers separating DA Us that would absolutely prevent interchange of elk among DA Us (see Program Narrative [PN] Appendix I Figure 1). The Basin represents a high altitude, cold winter range for both elk and mule deer which is similar to ecosystems in North Park, Middle Park, and the San Luis Valley, Colorado. The sagebrush (Artemisia tridentata) steppe winter ranges (2,250- 2,700 m elevation) can receive both extreme snow depths and cold temperatures that cause severe mortality among ungulates (Carpenter et al. 1984) while the conifer meadow and alpine summer ranges (3,000 - 4,200 m elevation) can be lush sources of forage subjected to periodic drought. Overall, these ranges collectively are thought to be less productive and nutritious for elk than the milder climate oakbrush-pinyon-juniper winter ranges and aspen and subalpine summer ranges of the Grand Mesa, Colorado where elk survival was measured from 1993-2000 (Freddy 2000). METHODS Capture Adult female (age 2:,2 years) and calf (age 6 months) male and female elk were captured and radiocollared using helicopter net-gunning from 16-22 December, 2000 and 16-20 December 2001 (Freddy 1993, see PN Appendix III). All radio-collars were l 72-l 76 l\.1Hz and contained 4-6 hour mortality sensors (Lotek, Inc., see PN Appendix I). Calf collars were expandable allowing collars to remain on elk as they matured to adults (see PN Appendix I). Objectives were to capture and radio-collar 78 calves each year with 13 calves of each sex in each of the 3 Gunnison Basin DA Us. For adult females, objectives were to capture and radio-collar 39 during the first year with 13 in each DAU and in subsequent years, capture sufficient adult females to maintain 2:,13 adult radioed females in each DAU (see PN, Sample Sizes). Prior to capturing elk, the 3 DA Us demarcating the entire Gunnison Basin were divided into 10 geographic trap-zones (Figure 1 A-J, see PN Figure 1). Numbers of elk to be captured in each trap-zone within a DAU were based upon the proportion of elk observed in each trap-zone within each DAU during elk sex and age composition surveys conducted with a helicopter during December-January post-harvest 1995-1999. We attempted, therefore, to distribute our sample of radioed elk across the landscape in proportion to relative elk numbers during early winter within each DAU. Elk were captured within a 3-km radius of.39 processing sites with some sites common to both years (Figure 1). Capture sites were systematically distributed within trap-zones within DAUs to radio-collar elk representing multiple segments of the entire Gunnison Basin population. Although our capture sites were not based on previously selected random coordinates, we believe we achieved a representative sample of elk from the population to provide relatively unbiased estimates of survival rates. Capture sites were accessed via vehicles when possible or by ferrying capture crews in the helicopter to more remote locations inhabited by elk. Calves were ferried by helicopter from individual elk capture locales to processing sites where body mass (kg), total body length (cm), hind foot length (cm), and rectal body temperature (F) were measured and then calves were radio-collared and released (see PN Appendix ill). All body measurements were made with the same instrumentation by the same individuals both years. Adult females were captured, aged as 2-4 years, 5-9 years, and >9+ years old based on incisor replacement or relative wear, radio-collared, and released at location of capture. We avoided capturing yearling 18-month-old females. Photographs of �195 incisor replacement and wear by elk age-class were provided to handlers responsible for judging the age of adult elk prior to releasing adults. No ear-tags were applied to calf or adult elk. Calf body measurements were compared among years, sexes, and DAUs using SAS (1988, PROC FREQ, PROC GLM[ANOVA]). All capture protocols were approved by the CDOW Animal Care and Use Committee. Telemetry Monitoring We monitored life or death status ofradioed elk daily from December through June from accessible roads using a truck equipped with magnetic-mounted omni-directional and 3-element hand-held Yagi antennas and at 1-4 week intervals from December through November using a Cessna 185 or 182 equipped with strut and/or belly mounted 'H' antennas. We used Lotek® SRX400 and Telonics® TR2 receiverscanners for monitoring telemetry signals. Elk survival data were compiled using the RADIOS module of the CDOW program DEAMAN® (White 1991). Mortality Assessments All suspected mortalities based on telemetry mortality signals were confirmed using ground searches. Once carcasses were located, criteria for assigning probable cause of death followed standardized written procedures that included assessment of body position and body condition, presence of bite or claw marks and sub-dermal hemorrhaging or gunshot wounds, presence of tracks or drag marks, and collection of organ, muscle, and femur marrow samples for laboratory analyses, if available (Wade and Browns 1982, Freddy 1998). Multiple photographs were taken of the carcass along with any potential evidence for assessing cause of death and when appropriate, an outside expert (T. D. I. Beck, CDOW) was consulted to assess evidence. Field necropsies were performed to the extent possible depending on completeness of carcass. We routinely collected muscle samples from large muscle groups in the hind- and forequarters of carcasses when available to assess for evidence of capture myopathy (Lewis et al. 1977, Spraker 1982, Haigh and Hudson 1993). Histopathology assessments of organ and muscle samples were completed by the Colorado State University Veterinary Diagnostic Laboratory and analyses of percent femur marrow fat on a dry-matter basis were conducted by the CDOW research laboratory. Field technicians provided a standardized written summary for each death. The principal investigator made the final assessment for probable cause of death based upon field summaries, photographs, and laboratory analyses. Potential causes of death included malnutrition, predation by black bears (Ursus americanus), mountain lions, (Felis concolor), coyotes (Canis latrans), and domestic dogs (Canis familiaris), legal and illegal hunter harvest, accidental trauma, plant poisoning, capture-induced , and unknown (Freddy 1997). Cause~ of death were broadly summarized as malnutrition, predation, suspected malnutrition, suspected predation, accident,.unknowp., hunter harvest, and capture-induced. Mortalities classed as malnutrition were usually nearly intact carcasses with little or no evidence of predator presence whereas mortalities classed as predation usually had evidence of bite wounds and subdermal hemorrhaging indicating bites were inflicted on a live animal. In those cases classed as suspected malnutrition or suspected predation a preponderance of collective evidence was used to assign cause of death to the most likely class. Telemetry collars that prematurely slipped-off elk causing a mortality signal to be emitted were confirmed by locating and retrieving the collar. Elk were subjected to multiple hunting seasons during fall 2000 and 2001. These seasons with approximate dates were: archery, 25 August-23 September; muzzleloading, 8-16 September; elk-only, 1317 October; deer-elk first combined, 20-26 October; deer-elk second combined, 3-9 November; deer-elk third combined, 10-14 November; late antlerless elk only, 24 November - 16 December in GMUs 54 and 55 (portions ofDAUs E-41 and E-43); and, late antlerless elk only, 1-31 December in GMU 66 (portion of DAU E-25). �196 Survival Rates Survival rates of radioed elk were calculated for this report using the binomial estimator and in final analyses will be calculated using a Kaplan-Meier estimator without staggered entry (White and Garrott 1990). Binomial estimates of survival rates were calculated as mean survival (s ) = [Alive / Alive+Dead collared elk], with a variance of [VAR (s) =(s)*(l- s) / n collars], and 95% confidence intervals of (s) ± [t a.-o.os, n-J aJ *J°( VAR (s))]. Survival rates were estimated for time intervals of winter-spring (15 December - 14 June), summer,-fall (15 June - 14 December), and yearly (15 December - 14 December) which coincided with capturing and radio-collaring elk and thus represented a biological year. By definition, calf elk became 12-month-old yearlings on 15 June and calves surviving to this date were considered to be recruited into the population. For adult elk during time intervals that included hunting seasons, we calculated survival rates inclusive of natural and hunting related mortalities, exclusive of hunting mortalities, and exclusive of natural mortalities. Excluding, or censoring hunting mortalities, provided estimates of natural survival rates, while censoring natural mortalities but including hunting mortalities provided estimates of hunting removal rates calculated as f = (1 - s), with (s) being survival rate with natural mortalities censored. Chi-square contingency tests were initially used for comparing calf survival (alive or dead categories) between sexes, years, and DA Us (White and Garrott 1990, SAS 1988 PROC FREQ). Parameter estimates were expressed as means± 95% confidence limits unless otherwise noted. Elk dying of suspected captured-induced trauma were censored from survival estimates. Deaths of calves or adults occurring within 1 week of capture were likely to be classed as capture-induced deaths unless field evidence strongly suggested a natural cause of death independent of capture. Capture-induced trauma could affect animals for up to 2-4 weeks post-capture so we routinely attempted to assess whether deaths were potentially capture-induced. We also censored elk having telemetry collars that electronically failed or slipped-off the elk (White and Garrott 1990). Elk with failed or slipped collars were censored for an entire seasonal time interval for binomial survival estimates and will be censored on the date they were last known alive based on telemetry signals in Kaplan-Meier estimates of survival. Elk whose telemetry signals disappeared during hunting seasons continued to be monitored for several subsequent months over large geographic areas until such time these elk were judged to have likely been removed during hunting seasons. Radioed elk that disappeared during hunting seasons were assumed to be legally harvested. Reproductive Collections Fecundity of adult female elk (age ,2:1 year) was estimated by examining reproductive organs of antlerless elk harvested during limited-entry late-hunting seasons that occurred from mid-November through December in portions of GMUs 54, 55, and 66. About 2-3 weeks prior to the beginning of seasons, we mailed permitted hunters collection packets· ~xplaining procedures f~r ·obtaining reproductive organs and incisor teeth (for dental cementum aging) from harvested elk as done previously in Colorado (Freddy 1992). Additionally, we asked hunters to collect kidneys with associated fat from these elk to allow calculation of kidney-fat indices and estimates of percent total fat to better assess body condition of adult females in relation to reproductive status (Kohlmann 1999, Cook et al. 2001a, 2001b). Hunters were instructed to place samples in collection bags and leave specimens in protected containers that kept samples cool at several drop-off sites within the Gunnison Basin from which samples were picked-up almost daily by project personnel. Dental cementum aging was completed by the CDOW research laboratory. Pregnancy status of elk was determined as: pregnant was uterus with embryo, fetus or fetal membranes; not pregnant was no evidence of fetus, no active uterine tissue, and no active corpora lutea of pregnancy; suspected pregnant was active corpora lutea, apparently active uterine tissue but no visible embryo or fetus; and unknown was either incomplete sample or no sample available. Fresh fetuses were sexed, �197 weighed, and measured (Armstrong 1950, Morrison et al. 1959). Fetuses and questionable uteri were stored in I 0% buffered formalin for reference examination. Morrison et al. (1959) graphically presented a logarithmic relationship between fetal crown-rump length (Y, dependent variable) and known fetal-age in days (X, independent variable) but did not present a standardized equation. To develop a standardized arithmetic equation for predicting fetal age, traditionally the unknown variable of interest, from fetal crown-rump length, traditionally the variable measured, we first used Morrison et al. data in MS-Excel® curve-fitter to develop 2 predictive polynomial equations. These equations were: (a) y(fetalcrown-nnnpmm) = 0.0085X\fetalageclays) + l.7603X 57.034, r2 = 0.9969, for the complete 8 data points presented by Morrison et al. (1959), and (b) Ycreta1crownrumpmm) = 0.0194X 2creta1agec1ays) + 0.2521X - 17.51, r2 = 0.9987, for 7 Morrison et al. ( 1959) data points with their late March data point excluded. We found equation (b) reduced the error in predicted crown-rump measurements by 2::_50% over equation (a) when predicted crown-rump lengths were compared with Morrison et al. actual crown-rump lengths, especially for the critical 60-90 day fetal-age stage that was associated with fetal collections occurring in November-December. Because fetal age in days is what is estimated from measured crown-rump values, we input polynomial equations (a) and (b) into program DERNE® to solve for fetal-age days (X) in terms of crown-rump measurements (Y). These DERNE® equations were: for polynomial equation (a) Xcre1aJageciays) = {[-f(3400000*Ycreta1crown-rumpmm) + 503781209) -17603] / 170}, and for polynomial equation (b) Xcreia1ase days)= {[-!(7760000* y(fetalcrown-nnnpmm) + 142256321) -2521] / 388}. We used DERNE® equation (b) to estimate elk fetal ages from crown-rump measurements. Fetal age in days was subtracted from date of hunter collection and then converted to calender and Julian dates of estimated conception. We measured total kidney fat mass and trimmed kidney fat mass after Riney ( 1955), Kohlmann ( 1999), and Cook et al. (200la,b). We calculated modified total and trimmed kidney fat indices after Anderson et al. (1990), Kohlmann (1999), and Cook et al. (200la,b). We estimated percent total body fat from measurements of kidney fat using simple linear equations that predicted percent body fat from kidney total fat mass (TFM), full kidney fat index (TF-KFI), and trimmed kidney fat index (TRF-KFI) presented by Cook et al. (2001a). We commonly received only I kidney fat mass submitted with reproductive samples so for those elk for which we received both kidney masses, we averaged kidney fat measurements to produce I value per elk (Cook et al. 2001a). Although TFM was potentially the best predictor of percent body fat of the measurements we made (see Cook et al. 2001a) we had no control on how well hunters collected all fat associated with kidneys so we conservatively viewed percent body fat estimates derived from trimmed kidney fat might be more accurate because we standardized the amount of fat measured among samples. All reproductive measurements were compiled in MS-Excel® and analyzed with SAS (1988, PROC FREQ, PROC UNNARIATE, PROC REG, PROC GLM[ANOVA], PROC LOGISTIC). General Elk Movements During aerial flights to monitor survival status of elk, we interpreted signal strength and direction to judge general locations of telemetry signals for each elk as to primary drainages or topographic features to describe general movements of elk to and from seasonal ranges. Elk that made large or unique movements, such as across main highways or DAU boundaries, were located relatively precisely from the airplane with the radius of location error likely <1,000 m. This location data was not gathered to assess specific habitats used but rather to describe the major movement patterns of these elk. Data will be summarized in future reports using ArcView 3.2®. �198 RESULTS AND DISCUSSION Capture In 2000, we radio-collared 78 calves, of which 38 were males and 40 were females, and 39 adult females age ::::2 years. Frequency of age classes for adult females were: 2-4 years= 3, 5-9 years = 30, and >9 years= 6. We achieved our objective of capturing 13 adult females and 26 calves in each DAU and nearly balanced sex ratios among calves in each DAU (Table 1). Elk were captured at 20 different sites representing a broad geographic area in the Gunnison Basin (Figure 1, Appendix A). There were no acute deaths of calves during capture but 2 adult females died while being blindfolded and hobbled prior to radio-collaring. Upon necropsy, both adults had extensive hemorrhaging in the thoracic cavity but no hemorrhaging in the abdominal cavity and no obvious indications of cervical injuries. We surmised that the heart-lung complex received extensive shock from capture. In neither case did we believe the animals had experienced e>,..1reme physical exertion nor aspiration of rumen contents. At time of capture, snow depths were about 25 cm and chase times appeared reasonable while ambient temperatures were 15 C and -2 C. Therefore, of 41 adult females captured and handled, 2 or 5% died of capture-induced injuries. Both adult elk were donated for human consumption. Subsequent to capture and radio-collaring, 1 male and 1 female calf died likely within 7 days of capture and were classified as capture-induced deaths and censored from the radioed-collared population of calves. Histopathology confirmed capturemyopathy in the male calf which had been killed by a mountain lion (Appendix B). At capture, rectal temperatures were 106.6 F (41.4 C) for the male and 105.5 F (40.8 C) for the female calves. Therefore, of 78 calves captured and handled, 2 or 2.6% died of capture-induced injuries resulting in a net sample of 76 radio-collared calves at the beginning of winter 2000. In 2001, we captured 80 calves, 40 males and 40 females, and 12 adult females age ::::_2 years. Frequency of age classes for adult females were: 2-4 years= 0, 5-9 years= 10, and >9 years= 2. We achieved our objectives of 26 calves of nearly balanced sex ratios in each DAU and maintained ::::_13 radio-collared adult females in each DAU (Table 1). Elk were captured at 19 new and 3 previously used sites (Figure 1, Appendix A). There were no acute deaths of adult females or calves during capture. However, 2 female calves caught from the same group in trap-zone E and radio-collared died within 48-hours of capture and were classified as capture-induced deaths (Appendix C). One calf became entangled in a fence about 1 km from the capture site and the second calf was euthanized by gun-shot about 0.5 km from the capture site because of obvious weakness following capture. At time of capture, snow depths were about 20 cm and capture chase times were < 1 minute at an ambient temperature of -4 C. Rectal temperatures of these calves at capture were 106.7 F (41.5 C) and 105.4 F (40.8 C), respectively. Two additional replacement female calves were captured from the same area prior to completing all capture activities. An additional male calf died within 3 days of capture and was classified as a capture-induced death even though a mountain lion had probably killed the calf (Appendix C). At capture, rectal temperature of-the male calf was 103.8 F (39.9 C), capture chase time seemed reasonable, and snow depths were about 20 cm. All 3 calves were censored from the collared population of calves. Therefore, of 80 calves captured and handled, 3 or 3.8% died of capture-induced injuries resulting in a net sample of 77 radio-collared calves at the beginning of winter 2001. For both years, capture-induced deaths occurred in 3.2% of the calves and 3.8% of the adult females that were captured and handled. Collar Failures We experienced pre-mature expansion of 14 calf collars, 13 males and 1 female, that resulted in collars slipping off calves causing us to censor calves during winter-spring or as yearlings during summer-fall time periods. For calves collared in December 2000, 5 males slipped collars between 30 April and 7 June 2001 and 3 males successfully recruited as yearlings, slipped their collars between 20 June and 20 July 2001. For calves collared in December 2001, 2 males slipped collars between 20 May and 3 June 2002, 3 males recruited as yearlings slipped their collars between 18 June and 17 July 2002, and 1 female recruited as a yearling, slipped her collar between 17 July and 22 August 2002. �199 Evidence suggested that amber latex tubing (3/8" O.D., 3/16" I.D, 3/32" wall thickness) used as a breakaway component to allow collar expansion pre-maturely deteriorated and broke allowing collars to expand and slip over the heads of calf and yearling elk. This calf collar design using the same or similar components had been previously used on 280 calves on the Grand Mesa, Colorado (Freddy 1997) where only I 5-month-old male calf, 1 13-month-old male yearling, and 2 23- to 26-month-old females slipped collars of which none were due to pre-mature breakage of latex tubing. On Grand Mesa, evidence indicated that latex tubing deteriorated as planned 10 to 18 months post-application after males had grown spike antlers or heads of either sex had grown to retard collars from slipping over heads. Deterioration of tubing apparently occurred sooner in the Gunnison Basin, maybe because of colder temperatures or slightly higher effective UV light levels, and males were much more prone than females to slip collars. We also speculated that antlers of yearling males in Gunnison might possibly be shorter than Grand Mesa yearling males during early summer thus allowing collars to slip more readily in Gunnison. After 2000-01, we changed brands oflatex tubing and maintained the same size of tubing for 2001-02 but the problem persisted to a lesser degree. In the future, we will change to a thicker latex tubing on male collars of either 1/8" or 3/16" wall thickness to reduce the problem of pre-mature expansion and the need to censor calves or yearlings from survival estimates but still maintain expansion capability so collars will adequately fit adult elk. We used radio-collar telemetry frequencies between 172 and 17 6 MHz and found an increasing problem with white-noise interference at frequencies> 175 MHz. At times, interference prevented hearing radiocollars except at relatively close distances, especially during aerial surveys when radio-collars or interference could be heard over several kilometers of distance. Interference was most commonly associated with human developments but likely sources could not be identified. We therefore caution project leaders to assess potential interference problems when selecting collar frequencies> 175 MHz. Weather Although official NOAA weather data has not been summarized as yet, winter-spring snow depths and summer-fall precipitation for both 2000-01 and 2001-02 were well below average for the entire Gunnison Basin. Both years were considered to represent drought conditions, not only for the Gunnison Basin, but most of southwestern Colorado. On most segments of winter range, snow depths generally did not exceed 30 cm during either winter with 2001-02 having shallower average snow depths than 2000-01. During both winters, snow had melted from primary winter ranges by late-March to mid-April. Snow depths and persistence of snow cover varied greatly in the Basin. Snow depths tended to decrease from west to east and north to south such that the deepest snow occurred in E-25 (GMU 66), E-4l(GMU 54), and E-43 (GMU 55) and the shallowest snow in E-25 (GMU 67) and in E-43 (GMU 551) (Figure I). Winter temperatures were generally mild for the Gunnison Basin with daily minimums seldom below -26 C and generally >-18 C and daily maximums often >-6 C. Calf Survival Survival rates of all calves pooled among 3 DAUs during winter-spring were 0.89 ± 0.08 (±CL, n = 71 ), 0.83 ± 0.09 (n = 75), and 0.86 ± 0.06 (n = 146) for 2000-01, 2001-02, and both years combined, respectively (Table 2). Survival of all calves was not different between years (P = 0.2965, Table 4). Male calves had lower survival (0.78) than female calves (0.97) in 2000-01 (P = 0.0105) but not in 200102 or when years were combined ( P2:._ 0.1463, Tables 2, 4). The greatest discrepancy between sexes occurred in DAU E-25 where survival of males and females was 0.71 and 0.93, respectively (Table 3). In comparison, yearly calf winter-spring survival on Grand Mesa was 0.86 to 0.92 (n = 69-73) and averaged 0.89 (n = 280) during 4 consecutive winters (1993-94 - 1996-97) with no differences in survival among years or between sexes (Freddy 1997). �200 Among the 3 DA Us, survival of all calves tended to be lower in E-41 (0.78) compared to E-25 (0.84) and E-43 (0.94) (P = 0.0737, Tables 3, 4). In paired comparisons between DA Us, calf survival was lower in E-41(0.78) than E-43 (0.94) (P = 0.0213, Tables 3, 4). During winter-spring, calves died due to predation, malnutrition, suspected predation or malnutrition, accidents, and of unknown causes. In 2000-01, 8 calves died with proximate causes of deaths being 37.5% predator-related and 62.5% malnutrition-related. In 2001-02, 13 calves died with proximate causes of death being 62% predator-related, 15% accidents, and 23% unknown causes. For years combined, 21 calves died with proximate causes of death being 53%predation-related, 24%malnutritionrelated, 9% accidents, and 14% unknown causes (Figure 2). On average then, for each 100 calves entering the population on 15 December, we would expect 86 to survive to the following 15 June with 7 deaths predation-related, 4 deaths malnutrition-related, and 3 deaths from other causes. Mountain lions and black bears predated elk calves and coyotes were suspected predators in one death. Accidental deaths were associated with a haystack collapsing and trapping a calf while elk were feeding on hay and a calf apparently slipped off a deep snow-trail used by elk and became trapped upside down among deadfall trees. In comparison, estimated causes of calf mortalities (n = 31) on Grand Mesa were 65% predation-related, 26% malnutrition-related, and 9% of unknown causes (Freddy 1997). Calf mortalities tended to occur later and primarily after 16 March in the winter-spring of 2000-01 than in 2001-02 (Figure 3, A & B). Timing of deaths in 2000-01 was consistent with deaths directly associated with malnutrition or predation-related deaths of malnourished calves as winter progressed. In contrast, predation-related deaths occurred from January through May in 2001-02 with no deaths directly attributed to malnutrition in 2001-02 (Figures 2, 3). In comparison, calves died on Grand Mesa from January into late May but the majority died in March and April (Freddy 1997). Femur marrow fat of dead calves was 34% ± 25 (SD, n = 8) in 2000-01 and marginally lower (P = 0.11, t-test) than the 59% ± 35 (SD, n = 10) in 2001-02. In general, most calves dying from any cause had femur fat< 50% in 2000-01 and >55% in 2001-02 (Figure 4). For deaths attributed directly to predation, femur fat averaged 16% (n = 3) in 2000-01 and 91 % (n = 4) in 2001-02. In 2001-02, all deaths considered predation-related had femur fat >60% (n = 7), even deaths occurring in mid-May (Figure 4). In contrast, both accidental deaths in 2001-02 had femur fat <5% and likely represented calves already extremely malnourished prior to the end of February (Table 5). Importantly, Cook et al. (2000a) noted that femur fat values <85% in adult female elk were associated with total percent body fat< 5% indicating that nearly any loss of femur fat suggested an animal in poor physical condition. Although there were limited sample sizes both years, data suggested a different dynamic between years of mild winters. in 2000-01, calf survival appeared more infl_uend~d bynutrition andrela.tive.body condition, possibly representing either the previous summer or winter forage production. Predation appeared more compensatory related. In 2001-02, predation appeared more additive than compensatory because calves that died were possibly not predisposed by malnutrition. In both.years, overall calf survival remained high regardless of the proximate cause of mortality. We must also caution that predation-related deaths in early January 2002 could not be totally separated from possible captureinduced deaths as deaths likely occurred <2 weeks post-capture. Adult Survival Survival rates for adult females age ;::2 years were 1.00 during winter-spring as no deaths occurred during 2000-0l(n = 39) and 2001-02 (n = 48). During summer-fall 2001, survival was 0.92 ± 0.08 (n = 39) when natural (1) and hunting deaths (2) were included. The one natural death occurred about July 1 of unknown causes in a female age 19 years based on dental cementum resulting in a natural summer-fall survival rate of 0.97 ± 0.05 (n = 37) (Table 6, Appendix D). Annual survival rates were 0.92 ± 0.08 (n = 39) including all causes of death and 0.97 ± 0.05 (n = 37) including only natural deaths (Table 6). In �201 comparison, natural survival of adult females was ~0.97 in winter-spring and summer-fall during 7 consecutive years on Grand Mesa (1993-94 - 1999-2000) (Freddy 2000). Yearling Survival Survival rate for yearling female elk, age 12-17 months, was 0.89 ± 0.10 (n = 38) during summer-fall 2001 and because all 4 deaths were hunting-related, natural survival during summer-fall was 1.00 (n = 34). Survival of the same cohort of yearling males was 0.86 ± 0.15 (n = 22) during summer-fall when natural (2) and hunting deaths (1) were included. Two yearling males died in July 2001 of suspected predation (Appendix D) resulting in a natural summer-fall survival rate of 0.90 ± 0.13 (n = 21) (Table 7). Survival of all females age ~12 months during summer-fall was 0.91 ± 0.07 (n = 77) inclusive of natural (1) and hunting deaths (6) and natural survival for these females was 0.99 ± 0.03 (n = 71, Table 6). Survival rates for both yearling female and male elk, age 18-23 months, was 1.00 (n = 34 F, n = 19 M) during winter-spring as no deaths occurred during 2001-02 (Table 7). Survival of all females age ~18 months during winter-spring 2001-02 was 1.00 (n = 82) as no deaths occurred (Table 6). Harvest Removal Harvest removal rates (r) during summer-fall 2001 were 0.05 for adult females, 0.11 for yearling females, 0.08 for all adult females age ~12 months, and 0.06 for yearling males. Hunting mortalities for adult females consisted of 3 legally harvested (2 regular rifle, 1 late rifle) and 3 wounding losses (1 archery/muzzleloading, 2 regular rifle). Wounding loss thus equaled the legal harvest in this small sample situation. The one yearling male hunting mortality represented an illegal harvest that occurred during a late-season in December 2001 (Appendix D). With observed calf and adult female natural survival rates, computer models suggest that removal rates for adult females age ~12 months in the Gunnison Basin would need to be ~15% per year to stabilize the population. Calf Body Size Calf body mass averaged 99.1 ±2.2 kg and ranged from 52.0 to 133.0 kg for all calves and years (Table 8) with 7% of the calves having mass <80 kg (Figure 5). There were no effects of capture year, calf sex, or year-sex interaction on body mass (P.:::_ 0.259) although calves were 2.6 kg smaller in mass in 2001 and males were 1 kg larger than females. However, calf mass was different among management DAUs (P = 0.003). In simultaneous paired comparisons, calves were larger in E-41 (104.4 kg) than in E-43 (96.8 kg) and E-25 (95.9kg) with no differences between E-43 and E-25. Calf mass was reasonably consistent within trapzones within DAUs, with mass tending to be larger in trapzones I and J (E-41) than in D (E-25) and F (E-43) (P = 0.090). Similar trends among years, sex, DAUs and trapzones occurred for calf total body length and hind-foot measurements with both measurements supporting that E-41 calves were largest (P < 0.002) (Table 8). Calf mortalities occurred across the range of calf body mass classes (Figure 5). Predation-related mortalities occurred in the most :frequent mass classes between 80 and 119 kg, suggesting predators were taking calves with no particular selectivity. Except in one case, malnutrition-related mortalities occurred in calves< 99 kg in size. Calves <80 kg did not necessarily perish, although 2 of the 3 calves <60 kg died of malnutrition or accident (Figure 5). Survival of calves tended to be lower in E-41 (P = 0.0737), where calf mass was largest, compared to survival in E-25 and E-43 (Tables 3, 4, 8). In E-41, mortalities were predator-related (45%), malnutrition-related (36%, including 1 accident where calf femur marrow was< 2%), and unknown (18%). On Grand Mesa, the larger mass of male calves also did not necessarily translate to higher survival rates compared to smaller female calves (Freddy 1997). In the Gunnison Basin, male and female calves were 15% and 7% smaller, respectively than their counterparts captured and radio-collared on Grand Mesa 1993-94 - 1996-97. On Grand Mesa, body mass was 115 ±2.5 kg for males (n = 138) ·and 106 ± 2.3 kg for females (n = 136) with an overall range in size �202 of 60 to 141 kg (Freddy 1997). Furthermore, unlike Gunnison, males were significantly larger (8%) than females on the Grand Mesa. Unfortunately, we cannot distinguish whether differences in mass reflect population or year effects. Elk Reproduction Samples Participation by successful hunters in providing biological samples from adult female elk harvested during late-seasons was disappointing. Of the estimated 665 adult females harvested during 2000-2001, we received some of the requested biological samples from 19% of the elk with reproductive organ and kidney fat samples representing only 13% of the elk. hnportantly, rates of participation by hunters harvesting elk declined from 28 to 12% from 2000 to 2001 despite attempts to improve collection instructions and packets sent to hunters in 2001 (Table 9). Providing an incisor tooth from harvested elk was the most common biological sample collected by hunters. For many data summaries and analyses, reproductive data from both years was combined because 65% of the samples were obtained in 2000. Furthermore, approximately 80% of the samples came from elk harvested in GMU 66 and 20% from elk harvested in GMUs 41 and 55. Therefore, data summaries were inherently weighted to year 2000 and GMU66. Ages of adult female elk harvested ranged from 1 to 20 years with 63% estimated to be age 3 to 10 years. Yearlings (:::::age 17-18 months) and females age 215 each represented 5% of the harvest. Because hunter selectivity and animal behavior may bias vulnerability of different elk age classes to harvest, the distribution of harvested age classes may biasly represent the age structure of the elk population (Table 10). Pregnancy Rates and Conception Dates Pregnancy rate for all adult females age 21 year was 85% (n = 89, Table 11). Pregnancy rate was 92100% for female age classes 3 to 14 years. Pregnancy rate was 67% in females age 2 and 50% in females age 215. Pregnancy rates across age classes were highly similar to rates measured in the ForbesTrinchera elk population of south-central Colorado (Freddy 1993b). The 100% pregnancy rate in yearlings could be questionable because pregnancy status was unknown in 67% of the submitted yearling samples. For age classes 2 and 3-4, pregnancy status was unknown in up to 44% of the animals. We might expect that in these younger age classes, uteri may be small, in-active and non-pregnant or in the early stages of pregnancy, creating more difficult circumstances for hunters to find and collect specimens. In comparison, for age classes 2 age 5, :::_27% of the specimens were of unknown pregnancy status. Thus, there is the possibility that pregnancy rates for young elk.age 1 to 4 could be ove_restimated due to collections biased against non-pregnant elk. Estimated conception dates followed an expected asymmetrical pattern (Flook 1970, Freddy 1993b,Noyes et al. 1996). Mode, median, and mean days of conception were 23, 26, and 29 September, respectively (n = 72, Figure 6). Conceptions spanned 68 days with 75% occurring in the 26-day interval from 8 September to 3 October. This conception pattern strongly suggested that most adult females conceived during their first estrus cycle at the expected time of year. Females conceiving after 10 October (n = 12, 17%) may have had a delayed first estrus or conceived during their second estrus cycle. Patterns and dates of conception were quite similar to estimates obtained for Forbes-Trinchera elk where post-season mature bull:cow rations commonly exceeded 35:100 (Freddy 1993b). The mirror-image asymmetrical distributions for conception dates (Figure 6) and calf body mass (Figure 5) indirectly suggest that smaller calves (7% < 80 kg) may be associated :with adult females conceiving later in the fall (17% after 10 October). Females conceiving after 10 October were comprised of 9% yearlings, 18% age 3-4, 45% age 5-7, and 27% age 215 years. All pregnant females 2 age 15 conceived after 16 October (n= 3). Later breeding by youngest and oldest age classes would not be unexpected but later conception by females age 5-7 may �203 indicate some nutritional or disturbance stress affecting timing of breeding in about 8%ofthe population. Overall, conception date was not dependent on adult female age (r2 = 0.014, P = 0.173, n = 66). Pregnancy status was associated with ovarian mass of both ovaries combined. Total ovarian mass (g) was 5.37 ± 0.28 (CL, n = 59) and larger in pregnant than the 3.36 ± 1.60 (n = 5) in non-pregnant elk (P = 0.025). Larger ovarian mass reflected the presence of active corpora lutea of pregnancy. Fetal Rates, Sex. Age, and Size Litter size was 1 fetus in uteri with detectable fetuses (n = 69). Fetal sex favored females (37F:21M) for years combined which differed from a 50:50 ratio (i = 4.414, df = I, P = 0.036). Female fetuses also dominated within each year (24F: 15M, 2000; 13F:6M, 2001) but such yearly ratios were not different from 50:50 (P 2: 0.108). Fetal sex could be determined in those fetuses near 2:70 mm crown-rump length based on external genitalia as also found by Morrison et al. (1959) and Kohlmann (1999). In general, fetal sex could not be determined in fetuses estimated to have been conceived after 10 October. Fetal sex tended to be dependently associated with adult female grouped age class (Figure 7 [Right], i = 10.885, df = 5, P = 0.054). Male fetuses predominated in adult females age 8-10 while females fetuses were most common within age classes 3-4, 5-7, and 11-14. Fetal sex ratio was equal in 2-year-old females which may have been conceiving for the first time. Male fetuses were also more predominate in elk age 2:8 years in the Forbes-Trinchera elk (Freddy 1993b). Kohlmann (1999) found male fetuses were more common in adult females having high kidney fat levels, and thus good body condition, and that adult females in good body condition conceived earlier in the breeding season. We could speculate that adult females age 8-10 had male fetuses because they were in better body condition at conception than other age classes due to their age and inherent larger body size that allowed them to withstand the rigors of a previous pregnancy and calf rearing and maintain access to better matriarchal habitats (CluttonBrock et al. 1982). In 2-year-old females, male fetuses may have been more common because these females likely had not gone through a previous pregnancy and subsequent calf-rearing prior to conception and thus were in better body condition. Although male fetuses predominated in those adult females conceiving 16-20 September just prior to the peak of conception, a pattern favoring male fetuses during early conceptions was not clearly evident (Figure 7 [LEFT]). Estimated fetal age averaged 69 and 76 days in 2000 and 200land was not different between years (P = 0.139). We found that fetal age predictive equation (b) (see ME1HODS) provided estimated dates of conception that occurred about 3-days earlier than equation (a) (paired t-test, P < 0.001). Fetal size compared favorably with fetuses measured in the Forbes-Trinchera elk population and appeared to be within an acceptable range of weight and skeletal.dimensions (Freddy 1993b). There was a general pattern of fetuses in 2001 being slightly larger in crown-rump (P < 0.080), hind-leg (P < 0.033), and hind-foot ( P < 0.070) dimensions but not in body mass (P < 0.138) (Table 12). Fetal size is highly dependent on date of collection so absolute comparisons between years or among elk populations must include corrections for date of collection. Body Fat Condition and Reproduction Total fat kidney fat index values (TF-KFI) for adult females age 2:1 year averaged 106 and ranged from 29 to 306 (n = 84) which were similar to values for Oregon elk (Kohlmann 1999) (Table 13, Figure 8[LEFT]). The 25% quantile value was 63 which was slightly higher than the 50 reported by Kohlmann ( 1999). Other kidney fat values in Table 13 were presented for reference as these measurements were the basis for estimating percent total body fat (Cook et al. 200 la), or body condition, in adult female elk. TF-KFI for calves averaged 44 (n = 6). �204 Estimates of percent body fat for all adult females age 2::,1 year based on the 3 kidney fat measurements, TF-KFI, TRF-KFL and TFM, averaged between 11.l and 12.1 % with a combined range of 5.4 to 18.5% (Table 14, Figure 8 [RIGHT]). Percent body fat of all adult females tended to be 1-2% higher in 2001 than 2000 for all estimators of body fat (PS 0.073, Table 14). For yearling females, specifically, percent body fat estimates averaged 10.7 to 13.0% with minimum-maximums of 9 and 14% (n = 4) while body fat in calves was 4-7% (see Table 14 cautionary foot-note). Based on estimates of percent total body fat, about 65% of the adult females age 2::,1 year were in moderate, 30% in low, and <5% in very low or very good body condition, and 0% in excellent condition (Figure 8 [RIGHI']). Relative condition class ratings were very low= <7% body fat, low= 7-10%, moderate= 10-15%, very good= 15-20%, and excellent= 20-25% (Cook, J. G. 2001 unpublished data). Kidney fat measurements provide the best predictive accuracy of percent total body fat at moderate levels of body condition and less accuracy at very high or very low levels of body condition (Cook et al. 2001a) so that our ability to detect outliers in body condition status may have been limited by measuring only kidney fat. Furthermore, percent total body fat levels in the excellent category may be rarely found in wild elk as these values were associated optimum nutrition in captive elk and likely represent the physiological maximums attainable by elk (Cook J. G., 2001 unpublished data). Probability of an adult female elk age 2::,1 year being pregnant was dependent on estimated percent total body fat when body fat was based on TFM (g): (logit (Pregnancy))= -2.2835 + 0.3704*(X-percent body fat); P = 0.033, n = 68). Pregnancy probability was predicted to be 2::,0.90 when percent total body fat was 2::,12%, or 2: moderate body condition (Figure 9). Similar dependent relationships (logit P) could not be detected between pregnancy status and percent total body fat based on TF-KFI (P = 0.130) and TRFKFI (P = 0.099), or on direct TF-KFI (P = 0.206) values. Cook et al. (2001a) indicated that TFM was the superior predictor of body condition within the kidney fat measurement alternatives. Kohlmann ( 1999), however, did find that probability of pregnancy (lo git P) increased with increasing TF-KFI values in Oregon elk (n = I 152). Similarly, Cook et al. (2001c) documented that low quality nutrition prior to breeding prevented or delayed conception in adult female elk and, furthermore, high pregnancy rates could be associated with marginally deficient nutritional conditions. Standard ANOVA results were consistent with logistic regression results. Estimated percent body fat was higher for pregnant than non-pregnant elk for body fat estimates based on TFM (P = 0.038) but not for body fat estimates based on TF-KFI and TRF-KFI (P 2: 0.112, Table 14 ). All estimates of percent body fat were not different among pregnant, non-pregnant and pregnancy status-unknown adult females (P 2: 0 .149, Table 14 ). Furthermore, using linear regression, all estimates of percent body fat were not dependent on adult female age 'Yithin pregnant (r S 0.014, P 2: 0.315, n = 54) or non-pregnant elk (r S 0.008, P 2: 0.543, n = 8). • • Probability of conceiving before or after the median date of conception was not dependent on percent total body fat based on TFM: ( logit (Before) P =0.221, n = 53), indicating there was no detectable increased probability to conceive before the median date based on percent total body fat. Using linear regression, conception date was not dependent on percent body fat based on TF-KFI, TRF-KFI, or TFM (r S 0.025, P > 0.141, df= 49) or dependent on 2-variable combinations of adult female dental cementum age and percent body fat (R2 S 0.031, P > 0.178, df= 49). Estimates of body fat for adult females conceiving after IO October was about 11.1 % for all 3 estimates of body fat. Overall, measured population performance of elk in the Gunnison Basin generally followed the predictions of proposed performance models for adult female elk in moderate or low body fat condition (Band C, below; Cook J.G. 2001 unpublished data). �205 The measured performance of elk in the Gunnison Basin could be summarized as: Pregnancy rates were 85%, about 17% of the females conceived after 10 October, adult female survival during mild winters was 100%, average mass of 6-month old calves was 99 kg with 7% of the calves weighing <80 kg and 21 % weighing> 110 kg, and survival of calves during mild winters was >83%; and compared to: Model A: If adult females in very good body fat condition. then we should expect: Pregnancy rates >90%, significant early breeding, high adult survival in harsh winters, > 110 kg calves in November; <5% of adult female Gunnison elk were classified in very good body fat condition. Model B: If adult females in moderate body fat condition, then we should expect: Pregnancy rates ~90%, some delayed breeding, with high adult winter survival depressed somewhat in harsh winters, 90-110 kg calves in November; 65% ofadult female Gunnison elk were classified in moderate body fat condition. Model C: If adult females in low body fat condition, then we should expect: Pregnancy rates ~70-90%, more delayed breeding, with markedly lower adult winter survival in harsh winters, 70-100 kg calves in November; 30% of adult female Gunnison elk were classified in low body fat condition. Model D: If adult females in very low body fat condition, then we should expect: Pregnancy rates 90%, delayed breeding up to 6 weeks, with low adult winter survival in harsh winters, 6090 kg calves in November; <5% of adult female Gunnison elk were classified in very low body fat condition. General Movements of Elk Insights into the general distribution and movements of elk in Gunnison Basin DA Us (see PN Figure 1,) were obtained from approximate locations of radio-collared elk obtained during 44 aerial survey flights between December 2000 and June 2002. Maps of elk distribution are in process so only verbal descriptions will be presented at this time. Elk wintered in segments of winter range near where they were trapped in December as elk did not usually make large movements during winter. Movement from winter areas towards summer ranges began in April, proceeded in earnest in mid-May after snow had melted at higher elevations, and ended with elk arriving.on highest.elevation summer ranges in July after cc!lving and subsequent to snow melting on alpine ranges. Movements from summer to winter ranges began in early September and continued through November with rates of movement most likely affected by hunting season activities and increasing snow depths. Elk essentially did not cross U.S. Highway 50 (Hy50) which separated the south (trap-zones A-E) and north (trap-zones F-J) (Figure 1) portions of the Gunnison Basin. Only 2 elk were known to cross this highway: an adult female in September 2001 moved from the Tomichi Dome area (trap-zone F) southwest to Sawtooth Mountain; and, a 24-month-old male in June 2002 moved from Tomichi Dome area (trap-zone F) to the southeast onto Sargents Mesa and then proceeded south over the La Garita Mountains to Alder Creek in the Rio Grande River drainage near South Fork, Colorado. Elk did move beyond the boundaries of the Gunnison Basin during winter and summer but only 2, at this time, have likely dispersed from the Gunnison Basin; an 18-month-old female moved from trap-zone H �206 north to Paonia Reservoir in November 200 I and a 24-month-old male moved trap-zone F south into the Rio Grande River drainage during June 2002. During summer, radioed elk were commonly found along the higher elevation divides associated with the boundaries of the Gunnison Basin DAUs, often in subalpine or alpine habitats from which they vacated in September while moving towards their winter ranges within the Gunnison Basin. These boundary areas included: upper and lower Cimarron and Little Cimarron rivers (west of trap-zone A); upper Rio Grande river in Rat Creek and near Continental and Rio Grande reservoirs (south of trap-zones A, B, C); Saguache Park (east of trap-zone D); Sargents MesaCameron Park (east of trap-zone E); upper Chalk creeks (northeast trap-zone F); upper forks of North Fork of Cottonwood, Lake Fork, Clear, and Castle creeks (east-northeast of trap-zone G); Anthracite creeks, Snowshoe and Cliff creeks, Coal Creek Basin, and Willow and Minnesota creeks (north oftrapzones H, I, J); and, upper Smith Fork, Dyer, and Crystal creeks (west-northwest of trap-zone J). The greatest overlap of Gunnison Basin elk with elk from other management DAUs occurred during summer in the Big Blue Wilderness (west trap-zone A), in the upper Rio Grande River near Slumgullion-Spring Creek Pass and west of Continental Reservoir (south trap-zones A, B, C), and in the West Elk Wilderness (north trap-zones I, J). During winter, only a few elk remained outside of the Gunnison Basin DA Us, mainly in lower Cimarron creeks (west trap-zone A), just east of North Pass and Old Cochetopa passes (east of trap-zone E), near Paonia Reservoir (north trap-zone I, J), and in Smith Fork and Doug creeks near Crawford, Colorado (west trap-zone J). Some consideration should be given to re-aligning elk management DA Us in the Gunnison Basin based on observed movements of elk. There was a continuum of elk interchange among trap-zones on an west to east basis, especially during summer and to a lessor degree in winter. South ofHy50, elk in trap-zones A through E interacted with elk in adjacent trap-zones such that there was no clear demarcation of separate elk sub-populations across this area (Figure 1). Similarly, north ofHy50, elk in trap-zones J through F interacted with elk in adjacent trap-zones such that there was no clear demarcation of separate elk sub-populations across this area, although elk in trap-zones G and F only interacted with elk from trap-zone Hin areas near Gothic, Colorado in the upper East and Slate rivers (Figure 1). Therefore, all areas south ofHy50 from Monarch pass on the east to the Cimarron Divide on the west (GMUs-part 551, 67, 66, and adding 65) could be treated as one DAU. Likewise, all areas north ofHy50 from Monarch Pass on the east to at least the Curecanti divide on the west (GMUs- part 551, 55, 54, and potentially adding 53 and 63) could be treated as one DAU. Elk that winter in GMUs 53 and 63 to the northwest of the Gunnison Basin likely have high interchange with elk from GMU 54 during summer in the West Elk Wilderness. SUMMARY In the Gunnison Basin, Colorado during winter-spring 2000-01 and 2001-02, survival rates of calves averaged 83-89% and tended to vary among elk management DAUs while survival rates of all age classes of adults were 100%. During summer-fall, survival rates were ;:::97% for adult females, 100% for yearling females, and 90% for yearling males when hunting deaths were excluded. Survival rates were comparable to survival rates previously estimated for elk inhabiting the Grand Mesa, Colorado. Harvest removal rates during summer-fall 2001 were 5% for adult females, 11 % for yearling females, 8% for all adult females age ;:::12 months, and 6% for yearling males. Measures ofreproductive and survival parameters were consistent with predictions of performance outcomes for adult female elk having low to moderate body condition status during fall. More than likely, marginally deficient levels of seasonal nutrition were depressing optimal reproductive performance of adult female elk. Consideration should be given to re-aligning management DA Us with observed distribution and movements of radioed elk. �207 LITERATIJRE CITED Anderson, A.E., D.C. Bowden, and D.E. Medin. 1990. Indexing the annual fat cycle in a mule deer population. Journal of Wildlife Managment 54:550-556. Anderson, C.R., Jr., D.S. Moody, B.L. Smith, F.G. Lindzey, and R. P Lanka. 1998. Development and evaluation of sightability models for summer elk surveys. Journal of Wildlife Management 62:9933991055-1066. Armstrong, R.A. 1950. Fetal development of northern white-tailed deer. American Midland Naturalist 43:650-666. Bartholow, J. 1999. POP-II system documentation Windows™ Version 1.0. Fossil Creek Software, Fort Collins, Colorado USA. Bear, G.D., G.C. White, L.H. Carpenter, RB. Gill, and DJ. Essex. 1989. Evaluation of aerial markresighting estimates of elk populations. Journal of Wildlife Management 53:908-915. Carpenter, L.H., R. B. Gill, D. L. Baker, and N.T. Hobbs. 1984. Colorado's big game supplemental winter feeding program. Colorado Division of Wildlife, Fort Collins, Colorado USA. Cogan, RD., and D.R. Diefenbach. 1998. Effect ofundercounting and model selection on a sightabilityadjustment estimator for elk. Journal of Wildlife Management 62:269-279. Cook, R..C., J.G. Cook, D.L. Murray, P. Zager, B.K. Johnson, and M.W. Gratson. 2001a. Development of predictive models of nutritional condition for Rocky Mountain elk. Journal of Wildlife Management 65: 973-987. Cook, R..C., J.G. Cook, D.L. Murray, P. Zager, B.K. Johnson, and M.W. Gratson. 2001b. Nutritional condition models for elk: which are the most sensitive, accurate, and precise? Journal of Wildlife Management 65: 988-997. Cook, RC., D.L. Murray, J.G. Cook, P. Zager, and S.L. Monfort. 2001c. Nutritional influences on breeding dynamics in elk. Canadian Journal of Zoology 79:845-853. Clutton-Brock, T.H., F.E. Guinness, and S.D. Alban. 1982. Red deer behavior and ecology of two sexes. University of Chicago Press, Chicago, Illinois, USA. Eberhardt, L.L., RA. Garrott, PJ. White, and PJ. Gogan. 1998. Alternative approaches to aerial censusing of elk. Journal of Wildlife Management 62:1046-1055. Flook, D.R. 1970. Causes and implications of observed sex differential in the survival of wapiti. Canadian Wildlife Service Report Series 11. Ottawa, Ontario, Canada. Freddy, DJ. 1992. Effect of elk harvest systems on elk breeding biology. Colorado Division of Wildlife Research Report July:45-70. Fort Collins, Colorado, USA. July:45-70. Freddy, DJ. I 993. Program Narrative for Estimating survival rates of elk and developing techniques to estimate population size. Colorado Division of Wildlife Research Report July:83-117. Fort Collins, Colorado, USA. Freddy, DJ. 1993b. Effects of elk harvest systems on elk breeding biology. Colorado Division of Wildlife Research Report July:50-65. Fort Collins, Colorado, USA. Freddy, D.J. 1997. Estimating survival rates of elk and developing techniques to estimate population size. Colorado Division of Wildlife Research Report July:47-73. Fort Collins, Colorado, USA. Freddy, DJ. 1998. Estimating survival rates of elk and developing techniques to estimate population size. Colorado Division of Wildlife Research Report July: 177-206. Fort Collins, Colorado, USA. Freddy, DJ. 2000. Estimating survival rates of elk and developing techniques to estimate population size. Colorado Division of Wildlife Research Report July (2):239-258. Fort Collins, Colorado, USA. Freddy, DJ., D.L. Baker, RM. Bartmann, and RC. Kufeld. 1993. Deer and elk management analysis guide, 1992-1994. Colorado Division of Wildlife Division Report 17, Fort Collins, Colorado USA. �208 Haigh, J.C., and R.J. Hudson. 1993. Fanning wapiti and red deer. Mosby, Saint Louis, Missouri, USA, pages 187-192. Kohlmann, S.G. 1999. Adaptive fetal sex allocation in elk: evidence and implications. Journal of Wildlife Managment 63: 1109-1117. Lewis, R.J., G.A. Chalmers, M.W. Barrett, and R Bhatnagar. 1977. Capture myopathy in elk in Alberta, Canada: a report of three cases. Journal of The American Veterinary Medical Association 171:927-932. Morrison, J.A., C.E. Trainer, and P.L. Wright. 1959. Breeding seasons in elk as determined from known-age embryos. Journal of Wildlife Management 23:27-34. Nelson, L.J ., and J.M. Peek. 1982. Effect of survival and fecundity on rate of increase of elk. Journal of Wildlife Management 46:535-540. Noyes, J.H., B.K. Johnson, L.D. Bryant, S.L. Findholt, and J.W. Thomas. 1996. Effects of bull age on conception dates and pregnancy rates of cow elk. Journal of Wildlife Management 60:508-517. Roath, R., L. Carpenter, B. Riebsame, and D. Swift. 1999. Gunnison Basin habitat assessment projectFinal Report March 1999. Colorado State University, Department of Range Science, Fort Collins, Colorado USA. Riney, T. 1955. Evaluating condition of free ranging red deer (Cervus elaphus), with special reference to New Zealand. New Zealand Journal of Sciene and Technology 36:429-463. Samuel, M.D., E.O. Garton, M.W. Schlegel, and R G. Carson. 1987. Visibility bias during aerial surveys of elk in northcentral Idaho. Journal of Wildlife Management 51: 622-630. SAS. 1988. SAS/STAT user's guide, release 6.03. SAS Institute, Inc. Cary, North Carolina USA. Spraker, T.R. 1982. An overview of the pathophysiology of capture myopathy and related conditions that occur at the time of capture of wild animals. Pages 82-118 in L. Nielsen, J.C. Haigh, and M.E. Fowler, editors. Chemical immobilization of North American Wildlife. Wisconsin Humane Society, Inc., Milwaukee, Wisconsin, USA. Unsworth, J.W., L. Kuck, and E.O. Garton. 1990. Elk sightability model validation at the National Bison Range, Montana. Wildlife Society Bulletin 18:113-115. Wade, D.A., and J.E. Browns. 1982. Procedures for evaluating predation on livestock and wildlife. Texas-Agricultural Experiment Station Publication B-1429. White, G.C. 1991. DEAMAN database manager and population modeling procedures; Colorado Division of Wildlife User's manual and reference. Colorado State University, Fort Collins, Colorado USA. White, G.C. and RA. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic Press, Inc., San Diego, California USA �209 Table 1. Number of male (M) and female (F) calf and adult female elk radio-collared in each DAU and tra:e-zone in the Gunnison Basin, December 2000 and 2001. Calf Elk Collared 2000 Adult Female Elk 2001 2000-01 DAU-(GMUs) Trapzone M F Total M F Total M F Total E-25 (66, 67) A- Lake Fork 4 7 11 2 4 6 6 11 17 5 B-Cebolla 3 4 7 2 0 2 5 4 9 3 2 2 2000 2001 6 16 7 5 10 4 5 8 14 3 5 7 19 39 32 3 5 7 4 5 6 6 12 10 7 17 3 2 3 2 4 6 4 5 9 2 ll 26 ll ll 26 ll 27 52 ll 2 3 5• 6 2 7 9 2 6 7 4 11 3 0 3 8 6 8 2 10 26 18 1 16 39 ll 5 10 7 14 7 15 ll E- Razor 4 G-Almont 5 3 3 11 ~ 7 11 18 5 16 18 16 34 12 14 26 ll ll 28 27 27 54 ll H-Flat Top 2 4 6 3 3 6 5 7 12 4 l- Beaver 4 6 10 3 4 7 7 10 17 4 0 4 J - West Elk 7 3 10 7 6 13 14 9 23 5 2 7 15 ll ll 26 ll ll 26 26 26 52 ll 1 li 39 29 38 40 78 40 80 78 80 .ill 39 12 51 ill 92 subtotals E-43 subtotals E-41 Totals Collared 2000-01 D- Sawtooth F-Tomichi E-41 (54) 2000 2001 C-Huntsman subtotals E-25 E-43 (55, 551) Total Elk Collared All Subtotals 40 • Includes 2 female calves that died of capture-induced causes with~- 24 hours of capture for which 2 additional female calves were captured from the same area and radio-collared pnor to completing capture o all elk. The net begmning sample size was therefore I I female calves for estimating survival rates in DAU E-43 in 2001. Table 2. Survival rates of elk calves age 6-11 months for males, females, and sexes combined from 15 December to 14 June in the Gunnison Basin, Colorado, 2000, 2001, and years pooled. Binomial estimator used to calculate survival rates and confidence limits for calves combined among DAUs E-25, E-41, and E-43. Elk Calves Elk Calves 15 Dec 2000- 14 June 2001 Survival Rate All Elk Calves 15 Dec 2001 - 14 June 2002 15 Dec - 14 June 2000 - 2002 Males Females All Males Females All Males Females All 0.78 0.97 0.89 0.84 0.82 0.83 0.81 0.90 0.86 Lower 95%CL 0.63 0.92 0.81 0.71 0.69 0.74 0.72 0.83 0.80 Upper 95%CL 0.93 1.00 0.96 0.96 0.94 0.91 0.91 0.97 0.91 n Collars 32 39 71 37 38 75 69 77 146 Collars Deployed 38 40 78 40 40 80 78 80 158 Collars Censored 6· lb 7 3' 2d 5 9 3 12 Died 7 8 6 7 13 13 8 21 Non-hunting Deaths 7 8 6 7 13 13 8 21 Hunting Deaths 0 0 0 0 0 0 0 0 0 • Male calves censored: for post-capture induced mortality 173.082/00 on 12/29/00 • for slipped-collars, 173.269/00 on 4/30/01, 173.170/00 on 5/7/01, 173.250/00 on 5/25/01, 173.151/00 on 6/7/01, and 173.220/00 on 6/1 /01. Female calves censored: for post-capture induced mortality 172.379/00 on 12/26/00. ' Male calves censored: for post-capture induced. mortality 174.720/01 on 12/19/0 I; for slipped-collars, 174.099/01 on 5/20/02, and 175.221/0 I on 6/3/02. b d Female calves censored: for post-capture induced mortality 173.429/01 on 12/16/01 and 173.740/01 on 12/16/01. �210 Table 3. Survival rates of elk calves age 6-11 months for males, females, and sexes combined from 15 December to 14 JuneforDAUs E-25, E-41, and E-43 in the Gunnison Basin, Colorado, 2000-01 and 2001-02 combined. Binomial estimator used to calculate survival rates and confidence limts. Elk Calves - DAU E-25 Elk Calves - DAU E-43 Elk Calves - DAU E-41 15 Dec - 14 June 00-01, 01-02 15 Dec - 14 June 00-01, 01-02 15 Dec - 14 June 00-01, 01-02 Males Females All Males Females All Males Females All 0.71 0.93 0.84 0.92 0.96 0.94 0.77 0.80 0.78 Lower 95%CL 0.47 0.82 0.73 0.82 0.88 0.87 0.60 0.63 0.67 Upper 95%CL • 0.94 1.00 0.95 1.00 1.00 1.00 0.94 0.97 0.90 n Collars Collars Deployed 17 25 27 27 44 52 26 27 25 27 51 54 26 26 25 26 51 52 Collars Censored 8 0 8 I 2 3 0 1 1 Died 5 2 7 2 1 3 6 5 11 Non-hunting Deaths 5 2 7 2 1 3 6 5 11 Huntin~ Deaths 0 0 0 0 0 0 0 0 0 Survival Rate Table 4. Comparisons of calf survival between years, sexes, and among DAUs in the Gunnison Basin, Colorado, 2000-01 through 2001-02 based upon chi-square (i ) contingency tests. t Calf Survival Rate Comparisons 1.2 Likelihood Ratio Ratio i Probability df 0.2965 1.10 0.2942 1 0.1463 2.12 0.1456 1 0.0105 0.8009 7.07 0.0078 1 0.06 0.8008 1 0.0737 5.71 0.0577 2 Value Probability Calf Sexes & DAUs Pooled 2000-01 vs 2001-02 1.09 Calf Male vs. Female All Years & DAUs pooled 2.11 Calf Male vs. Female in 2000-01 with DAUs pooled 6.56 Calf Male vs. Female in 2001-02 with DAUs pooled 0.06 5.21 Calf Sexes Pooled DAU E-25 vs E-43 vs E-41 & Years Pooled Calf Sexes Pooled DAU E-25 vs E-43 & Years Pooled 2.52 0.1123 2.56 0.1098 1 Calf Sexes Pooled DAU E-25 vs E-41 & Years Pooled 0.49 0.4827 0.50 0.4808 1 Calf Sexes Pooled DAU E-41 vs E-43 & Years Pooled 5.30 0.0213 5.59 0.0181 1 Table 5. Percent (enmr marrow fat in e1k calv.es_dying from estimated causes of mortality during winter-spring 15 .. . December to 14 June, 2000--Gl and 2001-02 in the Gunnison Basin,·Colora~o. - Samples Average Fat n- samples SD SE Min Max Predation Malnutrition 12/15/00- 12/15/0106/14/01 06/14/02 90.7 15.8 45.3 94.4 13.2 83.6 78.5 24.8 86.8 3.0 4.0 17.8 5.5 2.7 10.3 13.2 83.6 45.3 94.4 12/15/00- 12/15/0106/14/01 06/14/02 42.8 42.8 1.0 42.8 42.8 0.0 Estimated Mortali!I Causes Suspected Suspected Predation Malnutrition 12/15/00- 12/15/0112/15/00- 12/15/0106/14/01 06/14/02 06/14/01 06/14/02 68.6 48.7 60.0 27.5 75.6 0.0 ·11.1 68.1 38.5 0.0 3.0 4.0 0.0 7.8 24.4 4.5 12.2 60.0 0.0 15.6 48.7 Accidents Unknown 12/15/00- 12/15/0106/14/01 06/14/02 t.6 5.3 12/15/00- 12/15/0106/14/01 06/14/02 27.1 0.0 3.5 2.0 2.6 1.8 1.6 5.3 0.0 27.1 1.0 27.1 27.1 �2ll Table 6. Survival rates for winter-spring (WS), summer-fall (SF), and annual (Ann) seasonal intervals from 15 December 2000 to 14 June 2002 for adult female elk age ~2 years ~l year radio-collared in December 2000 and 200 I in the Gunnison Basin, Colorado. Binomial estimator used to calculate survival rates and confidence limts for elk combined among DAUs E-25, E-41, and E-43. Adult Female Elk Seasonal Interval and Dates :li:YS SE 6cc :li:YS 12/15/00 06/14/01 06/15/0112/14/01 12/15/0012/14/01 12/15/0106/14/02 1.00 0.92 0.84 1.00 39 39 0 3• 0.92 0.84 1.00 39 39 0 3 1.00 2 2 FEMALES ~2 yrs old) Survival Rate Lower 95%CL Upper 95%CL n Collars Collars Deployed Collars Censored Died Non-hunting Deaths Hunting Deaths 39 39 0 0 0 0 FEMALES ~l yr old) Survival Rate Lower 95%CL Upper 95%CL n Collars Collars Deployed Collars Censored Died Non-hunting Deaths Hunting Deaths 0.91 0.84 0.97 77' 77 0 7 1 6 1.00 39 39 0 0 0 0 48b 48 0 0 0 0 1.00 82 82 0 0 0 0 • Adult female deaths: 172. 758/00 about 7/1/01, 174.478/00 legal rifle harvest, and 172.030/00 archery/muzzleloading wounding loss. bIncludes 12 additional adult females radio-collared 16-20 December 200 l. c Includes 38 yearling females that survived as radio-collared calves. Table 7. Survival rates for winter-spring (WS) and summer-fall (SF) seasonal intervals from 15 December 2000 to 14 June 2002 for the cohort of 6-month old elk calves radio-collared in December 2000 in the Gunnison Basin, Colorado. Binomial estimator used to calculate survival rates and confidence limits for elk combined among DAUs E-25, E-41, and E-43. MALES Survival Rate Lower 95%CL Upper 95%CL n Collars Collars Deployed Collars Censored Died Non-hunting Deaths Hunting Deaths 6-11 mos WS 12/15/0006/14/01 12-17 mos SF 06/15/0112/14/01 0.78 0.63 0.93 32 38 6 7 7 0 0.86 0.71 1.00 22 25 3• 3 2 lb Elk Age (months) and Seasonal Interval and Dates 18-23 mos 6-11 mos ws ws 12/15/0106/14/02 12/15/00 06/14/01 1.00 19 19 0 0 0 0 FEMALES Survival Rate Lower 95%CL Upper 95%CL n Collars Collars Deployed Collars Censored Died Non-hunting Deaths Hunting Deaths 0.97 0.92 1.00 39 40 1 1 1 0 12-17mos SF 06/15/0112/14/01 18-23 mos 0.89 0.79 0.99 38 38 0 4 0 4' 1.00 ws 12/15/0106/14/02 34 34 0 0 0 0 •Yearling males censored: slipped collars, 173.091/00, 173.391/00, and 173. 510/00 between 6/22/0land 7/20/01. b Yearling male illegally wounded and died about 12/10/01 during late-season for antlerless elk. < Yearlin_g females 172.619/00 and 174.360/00 wounded during regular rifle seasons and 174.560/00 173.589/00 disappeared during regular rifle and late rifle seasons respectively, and assumed to be legalfy harvested. �212 Table 8. Body mass (kg), total body length (cm), and hindfoot length (cm) of elk calves captured and radio-collared in mid-December 2000 and 2001 in the Gunnison Basin, Colorado. Summaries include only those calves contributing to estimates of survival during winter from 15 December to 14 June and exclude calves dying from capture-induced causes. Values represent average weight (Mean), sample size (n ), standard error of the mean (SE), confidence interval of the mean (CI), minimum (Min), and maximum (Max). Mass (kg) Calf Groupings Total Body Length (cm) Hindfoot Length (cm) Gunnison Mean n SE 95%CI Min Max Mean n SE 95% CI Min Max Mean n SE 95%CI Min Max All Females 98.6 74 1.54 95.5 -101.7 57.5 1215 179 0 77 115 176.7 - 181.3 1435 196.0 56.0 77 0.22 55.6 - 56.4 49.0 61.5 All Males 99.6 74 1 64 %.3 - 102.8 52.0 133.0 178.2 76 1 12 176.0 -180.4 146.0 196.0 56.4 76 0.30 55.8 - 57.0 47.5 65.0 All Calves 99.1 14 112 %.9-101.3 52.0 133.0 178 6 153 0.80 177.0 -180.2 143.5 196.0 56.2 15 0.19 55.8 - 56.6 47.5 65.0 Females2000 99.9 37 2.14 95.6 - 104.3 57.5 119.0 178.6 39 1.62 175.3 - 181.9 143.5 194.5 55.9 39 0.33 55.2 - 56.5 49.0 59.5 Males2000 100.8 37 2.56 95.6 - 106.0 52.0 124.5 1802 37 1.81 176.4 -183.8 146.0 196.0 56.1 37 0.47 55.2 - 57.1 47.5 60.5 All2000 100.4 74 1.66 971-103.7 52.0 124.5 179.3 76 1.20 176.9 - 181.7 143.5 196.0 56.0 76 0.28 55.4 - 56.6 475 60.5 2.22 92.8 - 101.8 57.5 1215 179.5 38 1.66 176.1 - 182.9 151.5 196.0 56.2 38 0.30 55.5 - 56.8 51.5 61.5 Females 2001 97.3 37 Males 2001 98.3 37 2.05 94.2 - 102.5 78.5 133.0 176 3 39 1.30 173.7 -179 0 163.0 193.5 56.6 39 0.39 55.8 - 57.4 52.5 65.0 All2001 97.8 74 1.50 94.8 - 100.8 57.5 133.0 177.9 77 1.06 175.8 -180.0 151.5 196.0 56.4 77 0.25 55.9 - 56.9 51.5 65.0 DAUE25 All Calves 95.9 48 1.67 92.5 - 99.3 64.5 118.0 176.6 50 125 174.0 - 179.l 151.5 194.5 55.4 50 0.30 54.8 - 56.0 49.0 59.0 DAUE43 All Calves 96.8 50 2.10 92.6 - l 01.0 52.0 119.0 175.4 52 1.43 172.5 - 178.3 143.5 193.5 56.3 52 0.37 55.5 - 57.0 47.5 65.0 DAUE41 All Calves 104 4 50 1.81 ]008-108] 57 5 133 0 184.0 51 ]15 181.6 - 186 3 ]57 0 196.0 569 51 025 56.4- 57.4 52.0 60.5 E25 All Calves Trapzone A 97.6 17 2 47 92.4 - ]02.9 81.0 118.0 179 6 17 1.98 1754-183.8 166.0 194.5 55.8 17 0.50 54.7 - 56.8 51.5 59.0 E25 All Calves Trapzone B 96.2 7 3.27 88.2-1042 86.0 109 0 1771 7 l 66 173.0-18]] ] 70.5 181.5 56.2 7 0.73 54.4- 58.0 530 590 E25 All Calves Trapzone C 95.l 15 3.94 86.7 - 103.6 64.5 117.0 175.8 17 2.38 170.8 - 180.9 153.0 189.0 54.9 17 0.58 53.7 - 56.2 49.0 58.5 E25 All Calves TrapzoneD 93.7 9 3.41 85.8-1015 73.0 105.0 171.8 9 3.23 164.4-179.3 151.5 182.0 54.9 9 0.66 53.4 56.5 51.5 57.5 E43 All Calves Trapzone E 100.4 6 4.68 88.4 - 112.5 87.5 117.0 178.6 7 2.62 172.2 - 185.0 167.5 188.0 55.9 7 0.38 55.0 - 56.9 55.0 58.0 E43 All Calves TrapzoneF 93.l 10 6.17 79.l - 107.l 52.0 117.0 170.8 ']] 3.90 162 l - l 79 5 1460 185.5 56.8 11 1.36 53.8 - 59.8 47.5 65.0 E43 All Calves Trapzone G 97.3 34 2.41 92.4-102.1 57.5 119.0 176.2 34 1.68 172.8 - 179.6 143.5 193.5 56.1 34 0.37 55.4- 56.9 49.0 59.5 E4 l All Calves TrapzoneH 100.] ]] 6.37 86.0 - l 14.4 57.5 133.0 181.6 11 3.56 173.7 - 189.5 157.0 193.5 56.3 11 0.71 54.7 - 57.9 52.0 60.0 E4 l All Calves Trapzone I 104.9 17 2.60 99.4 - 110.5 85.5 121.5 185.0 17 1.86 181.0 - 188.9 1670 196.0 57.3 17 040 56.5 - 58.2 54.5 60.0 E4 l All Calves Trapzone J 106.l 22 179 102.4 - 109.8 89.0 124.5 184.3 23 1.39 181.5 -187.2 169.0 196.0 56.9 23 0.33 56.3 57.6 54.0 60.5 . �213 Table 9. Numbers of adult female elk harvested during late-seasons in the Gunnison Basin, Colorado, NovemberDecember 2000 and 2001 with numbers and percent (%) of harvested adult females from which hunters provided any of the requested biological samples, reproductive organ samples, and kidney fat samples. Samples received for kidney fat expressed as fat with 1 kidney, fat with 2 kidneys; and elk with at least 1 kidney fat sample. Estimates of adult females harvested obtained from CDOW statewide harvest surveys. Late Season Year Adult Females Harvested Adult Females With Any Requested Samples Submitted Adult Females With Reproductive Organs Submitted Adult Females With Kidney Fat Samples Submitted 1 kidney; 2 kidneys; :::_ l kidney 2000 291 (100) 81 (28) 58 (19) 17 (6); 40 (14); 57 (20) 2001 374 (100) 46 (12) 31 (8) 22 (6); 5 (l); 27 (7) All 665 (100) 127 ( 19) 89 (13) 39 (6); 45 (7); 84 (13) Table 10. Frequency (%) of dental cementum ages of adult female elk harvested in the Gunnison Basin, Colorado, November-December 2000 and 2001 based on useable incisor tooth samples submitted by hunters. Age Class of Adult Female Elk Based on Dental Cementum (years) Year 3-4 5-7 8-10 11-14 15-20 All 2000 5 7 12 20 10 14 6 74 2001 1 7 15 8 9 4 0 44 All 6 (5) 14 (12) 27 (23) 28 (24) 19 (16) 18 (15) 6 (5) 118 (100) 2 Table 11. Pregnancy rates(%) by age class of adult female elk in the Gunnison Basin, Colorado, NovemberDecember 2000-2001 based on samples submitted by hunters for years combined. Age of elk based on dental cementum. Pregnancy rates based only on numbers of known pregnant and non-pregnant elk per age class. Age Class of Adult Female Elk Based on Dental Cementum (years) Pregnancy Status 11-14 3-4 5-7 1 14 (93) 0 1 1 3 4 13 21 (100) 13 (93) I2 (92) 3 (50) 5(56) 76 (85) 8-10 15-20 Unknown Adult 2 All Non- Pregnant 0 3 Pregnant 2 (100) 6 (67) Unknown 4 5 12 7 5 5 0 0 38 Total 6 14 27 28 19 18 6 9 127 Table 12. Measurements of elk fetal size in the Gunnison, Basin, Colorado during November-December 2000 and 2001. Values represent average size ( X, mean), sample size (n), standard deviation of the mean (SD), confidence interval of the mean (CI), and minimum (min) and maximum (max) values. November-December 2000 Measurements X (n) SD November-December 2001 95%CI min-max X (n) SD 119.3 (6) 111.0 (13) 4.9 (3) 95%CI min-max 95.9 18.7-219.9 46.4-289.1 85.5 3.2 59.4-162.7 0.0-12.7 22.1-310.5 1.8-8.1 106.1-171.2 116.7-162.2 13.5-75.9 111.6-186.5 Mass (g) Male 64. 7 (15) 51.0 36.5-92.9 12.4-167.0 Female Unknown Sex Crown-Rum!! (mm) Male 86.8 (24) 4.3 (7) 97.8 3.7 45.5-128.0 0.8-7.7 12.3-425.5 2.0-12.6 114.5 (15) 30.4 97.7-131.4 73.6-165.0 138.6 (6) 31.0 125.6 (24) 33.9 (8) 39.5 18.6 108.9-142.3 18.4-49.4 68.5-223.0 0.5-69.0 139.5 (13) 44.7 (3) 37.7 12.6 44.9 (15) 50.7 (24) 24.0 (1) 16.6 35.7-54.0 23.8-73.9 42.2-59.3 23.6-103.9 60.6 (6) 59.4 (13) 15.0 20.3 . 29.2 (15) 10.6 34.7 (24) 16.5 Female Unknown Sex Hind-Leg (mm) Male Female Unknown Sex Hind-Foot (mm) Male Female Unknown Sex 13.9 {12 19.0 44.9-76.4 47.9-70.9 13.4 14.8 30.8-48.7 87.9-202.0 32.3-57.4 47.0-83.4 36.1-91.2 17.4(1) 23.3-35.0 27.7-41.6 14.9-48.1 14.8-79.9 40.4 (6) 39.8 (13) 11.5 {12 26.3-54.5 29.4-61.4 21.4-64.6 �214 Table 13. Summacy values for total fat kidney fat index (TF-KFI), trimmed fat kidney fat index (TRF-KFI), kidney total fat mass (TFM g), and kidney trimmed fat mass (TRFM g) for adult female elk in the Gunnison Basin during November-December 2000-200 I. Kidney mass one and two could represent either the left or right kidney masses with kidney mass two associated with elk for which both kidney masses were collected by hunters during late antlerless-only hunting seasons. Values represent average size (X, mean), sample size (n), standard deviation of the mean (SD), confidence interval of the mean (Cl), and minimum (min) and maximum (max) values. Kidney Mass One Fat Value X" (n) TF-KFI Kidney Mass Two SD 95%CI Min-Max X" (n) SD 95%CI Min-Max 105.8 (84) 54.5 93.9-117.6 29.1-306.1 91.0 (45) 37.5 79.8-102.3 33.4-165.3 TRF-KFI 78.4 (84) 35.5 70.7-86.1 29.1-212.8 70.3 (45) 26.5 62.3-78.3 30.4-136.1 TFM (g) 218.7 (84) 103.9 196.1-241.2 57.0-551.0 202.4 (45) 90.6 175.2-229.6 72.0-486.0 TRFM (g) 163.6 (84) 70.4 148.4-178.9 48.0-383.0 156.8 (45) 66.3 136.9-176.7 69.0-388.0 Table 14. Estimates of percent total body fat in adult female elk by pregnancy status and calf elk (sexes combined) in tl1e Gunnison Basin, Colorado during November-December 2000-200 I. Percent body fat based on total fat kidney fat index (TF-KFI), trimmed fat kidney fat index 9TRF-KFI), and total kidney fat mass (TFM g) after Cook et al 2001a Comparisons among mean values shown by P-values (ANOVA). Values represent average size (X, mean), sample size (n), and confidence interval of the mean (CI). % Body Fat TF-KFI X" (n) 95%CI % Body Fat TRF-KFI % Body Fat TFM X" (n) 95%CI X" (n) 95%CI Percent Body Fat Adult Females Pregnant 11.2 (57) 10.6-11.8 12.1 (57) 11.5-12.8 11.3 (57) 10.8-11.9 7.0 (TF) 18.5 (TRF) Non-Pregnant 10.2 (11) 8.8-11.5 10.9 (11) 9.5-12.2 9.9 (11) 8.9-10.9 8.3 (TFM) 14.8 (TRF) Pregnancy Status Unknown 11.8 (16) 10.4-13.1 12.6 (16) 11.1-14.l 11.1 (16) 9.6-12.5 5.4 (TFM) 15.3 (TRF) Pregnant vs. NonPregnant P=0.143 P = 0.112 P = 0.038 Pregnant vs. NonPregnant vs. Unk. P=0.193 P = 0.198 P = 0.149 All Adult Females 11.2 (84) 10.7-11.7 12.1 (84) 11.5-12.6 11.1 (84) 10.6-11.6 5.4 (TFM) 18.5 (TRF) Aduh Females 2000 10.7 (57) 10.2-11.3 11.6(57) 11.0-12.2 10.8 (57) 10.2-11.4 5.4 (TFM) 18.5 (TRF) AdultFemales2001 12.1 (27) .11.1-13.1 13.1 (27) 12.1-14.1 11.7(27) 10,_8-12.6 7.3 (TF) 17.0 (TRF) 2000vs. 2001 P=0.009 All Calves• P =0.010 P = 0.073 7.1 (6) 4.1-10.1 7.3 (6) 4.1-10.4 4.1 (6) 1.1-7.0 0.2 (TFM) 10.4 (TF) Calves 2000 7.4 (2) 0.3-14.5 7.7 (2) 0.0-23.4 3.8 (2) 0.0-8.2 3.4 (TFM) 9.0 (TRF) Calves 2001 7.0 (4) 1.2-12.8 7.0 (4) 1.1-10.3 4.2 (4) 0.0-9.9 0.2 (TFM) ' 10.4.(TF) • Estimates of percent body fat in calves should be viewed with caution as calibration equations developed by Cook et al. 2001 were based only on adult female elk. b Minimum and maximum estimates of percent body fat obtained from either TF-KFI, TRF-KFI, or TFM estimators. �215 Fig. i. Game Management Units (54-551). trapzones (A-J), and elk capture sites in.the GunnisonBasin .in 2000 arid 2001. C3ptu~ Sites Capture: Site Vie3r AF .. "'tall Fl:,lrct U ... De=mlll:r2CIJD .Al• A.mtOl'\ITllylOr ■ De~trt:r:l.1J1 AT• .---on.1nt11•111: + . IV• lea.in CIC.Ek SWA oC a!fflllC r 2Clla Zll'II 2DD1 cc .. cm1ncreu D "ltl:IPJ:a\C5 co .. oo., Ouldl WIiow □ ouus ~ .. o,:,c~ck e,o .. klt::hO.Jch ,,,.,.,. DV• Dcad'5Cra:k DY• D,r·Creet U~r a: .. a.; 1c;:11n crri:t rT• Flal Too 8ouh IIO • Mom OUICh • H.- N<lll'ICOl.lq"alllmar . • t::K• cat1, ~·111:'_o·tJ<:h l(Z .. IC!:mr lmh ~W N A LC• LM-IOznyCJn 0Uch IP• IG'h PsNn· P~ • P'Olt CteEIC 1 PIil .. l'ctson I\Jdat kl • 111.oi.ndull I :=:in IIIC• kdC.cB. I\D • 111.aldtn Flzal Toa: llV- Ill.GIid 11:an:r Creek 8C • 8oclll Cl'Hk C~I eo--sii:cn:O\lch eu .. ~o- Creek: Tf ■ n~r"""ZOIU~ ,U ■ Ten1111ll1: El:irlrG .'10 • TofficN DCIIIII: TT• 'nlllb: "11)JI WA• IIWH1_Anl:lopc creek UIJC) • Wood:, OUd\ WM• WIii~ M~ ltw ·vuw- -Wll<nu(lrcet:: 1· WY• \l\llmuy Hane "1'0• Yl:;J!JIIU•Oulctl YP • Ye lloWPJnE Ill.II DC 20 Kilometers �216 Calf Elk Winter Mortalities Gunnison Basin, Colorado ■ Lion Predation D Black Bear Predation D Suspected Predation ~ Malnutrition ffl Suspected Malnutrition ~ [I] 2. ated s of calf mortaliti during from 15 mber h 14 in the son Colorado, 2000-01, 2001-02, and years combined. Accident Unknown Figure Estim cause elk es ,..........._~~___.____, winter ,,.....,.~~~~ Dece throug June Gunni Basin, �217 3--r------------,------....--.--------------- Calf Elk Winter Mortalities Gunnison Basin, Colorado 2-1----------.-------r--------1 D Unknown § Accident ■ Predation D Malnutrition 1---------- 0----,---....----Jan 1, 01 Feb 1 Dec16,0D Jan 16 Mar1 Feb 16 JI.DI 1, 01 Apr1 Mar 16 May 16 Apr 16 Beginning Date of Two Week Intervals December 2000 to June 2001 a----- ~ ~ 2-+---- iii (.) 1-+---- Apr1 Dec 18,01 Jan 16 Feb 18 Mar 18 Jun 1, D2 May1 Apr 18 May 18 Beginning Date of Two Week Intervals December 2001 to June 2002 Figure 3. Timing and estimated causes of calf elk mortalities during winter from 15 December through 14 June in the Gunnison Basin, Colorado, 2000-01 and 2001-02. �218 100 ~ ~ I • I x 2001-02 2000-01 : catf Elk Marrow Fat, Gunnison Basin, Colorado: ~ - ~ 'fl' [i] ~ l'u1 D ... Predation Related E Malnutrition Relaled :::, tf. C: 40 a, ea, Accident 0 Unknown 0 0.. 20 ~ \/ '1 'i7 [i] ~ ® (i] [i] ® (it) 0 I 12116 - 1/15 \";:,,.., I I 2114 3116 I t:""' I 4115 6/14 Momh and Day Figure 4. Percent femur marrow fat in calf elk mortalities by estimated cause and timing of deaths during winter from 15 December to 14 June in the Gunnison Basin, Colorado, 2000-01 and 2001-02. 30 1;:::========-=--=-.--------------;::;-------;::==============~7 ■ 2000-01 n=8 X 2001-02 n = 12 Predation Related Accident O Males n=74 0 Malnutrition Related ro ■ Females n=74 Z1 0 Q ~0 1 5 -Unknown -t------------.-,,,---., ~ iz ~ Gil 10--+------------ 50-59 60-69 70-79 80-89 90-99 100-109 calf Body Mass aass (kg) 110-119 120-129 130-139 Figure 5. Distribution of male and female calf body masses and occurrence of calf mortalities by mass class, calf sex, and estimated cause of death during winter from 15 December to 14 June, Gunnison Basin, Colorado, 2000-01 and 2001-02. �219 20----,--------------------------------------. Elk Conception Dates Gunnison Basin, Colorado 2000-2001, n = 72 - - 16-+---------- Mode= Sep23 Median 50% Quantile = Sep 26 ~ 12 - + - - - - - - - - - - Mean= Sep29 C: Q) :::J CT Q) ._ LL 8 -+-------- - 75% Quantile = Oct 3 4-+------ 0 Sep 6-10 Sep 16-20 Sep 26-30 Oct 6-10 Sep 1-5 Sep 11-15 Sep 21-25 Oct 1-5 Oct 11-15 Oct 21-25 Oct 31-Nov4 Nov 10--14 Figure 6. Frequency of estimated conception dates in 5-day intervals for elk fetuses in the Gunnison Basin, Colorado, 2000-2001. Male Elk Fetuses Per Adult Female Age Class, Gunnisoo Basin 2000-2001 Male Elk Fetuses Per Conception Date Interval, Gunnison Basin 2000---2001 Male & Female Felu:.u Perlnl~rv?I! Sho'M1As (11) Male and Femalt": Fetuses Per Age Interval ShCJIMl As (15} 0.6 I 0.6 0.5 I 0.1 01 I Sep&,O s~11.1s S.p21-25 Sepl&-20 I 0<:1-5 I 00;6-10 I I I I ' Adul Female Dental Cemenlum 3---4 S--7 8-10 Ag/: Class 11-14 I 15-20 . Figure 7. Proportions of male elk fetuses per conception date interval (LEF1) and per adult female age class (RIGHI), Gunnison Basin, Colorado, 2000-2001. �220 ., AdlA Fem11~ Elk Total K~y Fat lndn, Guin.I~ Ba~rl. n = 64 17 Esttnt,s Pemrt Body FJIMull r.rm.ie El<, Gwrison Bun.2<m-2001 ~-~ n•Mpcr '-FatTF-KFl TRF-KR. TAI E~ □ ., 16 -· " □ ■ ,0 7 6 3 3 I I ,._,, .,_,, ~0-~ I I 80--g,§I I Hl0---119 140---759 !Xl-139 "' "' 5 0 ,. -"' 8 ' I 100-19'9 HD----179 ::,00.~!1 TOl,Jj KICrleyFS!ln-6e:I: 'watues(Pen::ent) - ~ f-- 10 Jl. 1 0 ,,._,,, ...!....o, I 0--U XD-Jd9 7-U 10---1'.t a o o 1£-11.0 -BolltFJI Figure 8. Frequency of total kidney fat index values (TF-KFI) (LEFT) and percent total body fat estimates (RIGHT) for adult female elk age :::_l year during November-December, Gunnison Basin, Colorado, 2000-2001. Body fat estimates based on TF-KFI, trimmed kidney fat index (TRF-KFI), and total kidney fat mass (TFM g), respectively, after Cook et al. 2001a. Percent body fat classes 0-6.9, 79.9, 10-14.9, 15-19.9, and 20-24.9 represent body condition classes very low, low, moderate, very good, and excellent, respectively, after Cook J.G. (2001 unpublished data). 0.9 Body Condition Upper Confldence Limit VoryGood --- -.... -- 0.8 ----------- 0.7 Body Condfflon ,._ Moderate ~0.6 a, e 1'; 0.5 ,, ,, ~ :c i0.4 , c.. , I ,, , , Body CondltlOII Low 0.3 0.2 0.1 - ..... Lower Conn;i,,"' 0 0 2 4 - L:tt 6 8 W U M Estimated Pen:ent Body Fat in November-December 16 18 20 Figure 9. Probability of adult female elk age 2::.1 year being pregnant as predicted from percent total body fat based on total kidney fat mass (TFM g) measured in November-December, Gunnison Basin, Colorado, 2000-2001. Probablility curve bracketed by approximate 95% confidence limits. Relative body condition rating classes from Cook J.G_ (2001 unpublished data)_ Logistic regression was: logit (P) = -2.2835 + 0.3704 *(X-percent body fat); regression slope significantP = 0.033). �221 Appendix A. Locations of elk capture sites in the Gunnison Basin during December 2000 and 2001. All U'IM coordinates are referenced to NAD 27 datum Erojection. Site Code WY WL WM DV TM WW pp AF TT HG WG cc NP LC AL RD WA BV DC RC CG YG KZ RV SU PR DG HR KK TO yp AT EC RB FT SG DY TF SC Trap Site WINNERY HOME WlLSONGULCH WILLOW MESA BLUE DEVILS CREEK TENMILE SPRING WILLOWCKl POLECKl ALKALIFLYING M TABLE TOP HORN GULCH WOODS GULCH CABIN CK NORTH PARLIN LOST CANYON GULCH ALMONT TAYLOR REDDEN FLATTOP WEST ANTELOPE CK BEAVER CK SWA DRY CREEK RED CREEK COW GULCH WILLOW YEAGER GULCH KEZAR BASIN NW ROAD BEAVER CK SUGARCREEK POISON RIDGE DUTCH GULCH HOME GULCH RAZOR CAMP KETTLE GULCH TOMICID DOME YELLOWPINE RIDGE ALMONT TRIANGLE EAST CABIN CREEK ROUNDUP BASIN FLATTOP SOUTH STEERS GULCH DRY CREEK UPPER TENDERFOOT MESA SOAP CREEK COAL Trapzone A A A B B C C D E F F G G G G H I I J A B B C C D D E E F F G G G H J J Capture UTMx Capture UTMv Capture UTMZone 299842 299987 302286 299S07 307088 319409 329616 343809 3S265S 367S5S 3S7804 342S64 3S1341 343419 343020 331774 32S2S9 321816 313617 30S342 302170 304088 311114 321992 32S267 344624 342179 3S1874 349729 365097 3S9615 342507 344065 348075 3330S1 323101 313563 306929 299839 4229298 4232163 4248791 422124S 42S21S1 42S044S. 4246431 423826S 4249609 42S541S 4263648 4267804 426S7S7 427S048 4287228 4282420 4274696 4266364 4262208 4263S42 4248762 423371S 42S7470 423S039 42S2487 42238S0 42S61S0 4237208 42S0925 4256608 4264943 4287754 4270540 4270061 4281590 4270141 4262643 4264206 4265834 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 13 Capture Capture 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 �222 • the Gunmson Basm, 15 December to 14 June, 2000-01 AppendixB Summarvofcalf e]k mortali.t:iesm TrapNo Elk ID Sex Mass zone 1 173.082/00 M 92 B Tassue Ferrur Marrow Fat Recovered Samples Parasites Yes CM n/a 2 172.379/00 F 127 H Death Date 12/29/0Q.. 1/4/01 12/1 S-26/00 3-Jan-01 WhiteCreamy 38.7% WhiteFirm 94.7% 3 173.640/00 F 82 A 2/1-7/01 7-Feb-01 Red.Jelly 15.8% NoCM 4 172.959/00 M 125 J 2/15/01 15-Feb-01 No CM 5 173.351/00 M 52 F 3/21/01 24-Mar-01 WateryPink 48.7% Red.Jelly 27.5% NoCM 6 173.160/00 M 94 A 3/25/01 26-Mar-01 Red Jelly 42.8% No CM 7 173.300/00 M 112 I 3131/01 1-Apr-01 Red Jelly 0.0% No CM Carcass Status Death Cause Death Location UTMy Drainage UTMx 303163 4253981 Lake Gulch 335406 4282773 304608 4243376 FlatTop Lake Fork Unk.-Suspect Starvation Unk.-Suspect Starvation Stravation 304753 4261110 Red Ck. 360867 4262422 Yellow Pine 298132 4230866 Dwyer Gulch Unk.-Suspect Starvation Unk.-Suspect Starvation Lion Predation 308379 4263917 Dry Gulch 320104 4270341 Beaver Ck. 311827 4245467 Cebolla Ck. Bear Predation 300715 113 J 4128-5110/01 23-May-01 n/a Scavenged M Mass= Weight of caH (kg) at capturej CM = capture myopathy; NI = no evidence of inllarmtion around sarcocystsj n/a = sanples not available. 4268428 E. Coal Ck. !Nan-01 8 173.041/00 M 97 I 4/13-14/01 15-Apr-01 9 173.949/00 M 107 A 4/20/01 26-Apr-01 10 173.011/00 Red Jelly 77.66% RedCreamy 45.27% Red Jelly 13.15% n/a NoCM No CM n/a Moderate sarcocysts Moderate sarcocysts Nonnal sarcocysts Moderate sarax:ysts/NI Severe sarax:ysts/NI Nearly Complete Scavenged Capture Related / Capture Related Partially Scavenged Carcass Complete Scavenged Lion Predation Carcass Complete Carcass Severe sarcocysts/NI Normal sarcocysts n/a Complete Nearly Complete Scavenged Lion I ! I ! 'I i AnnendixC. sll.lllllUllyofcalf elk mortalities in the Gunnison Basm, 15 December to 14 June, 2001-02. Trap- Death Zone E Date F Mass 100 173.740/01 F 101 E 174.720/01 M 101 B No Elk ID 1 173.429/01 Sex 2 3 Tissue Samples Parasites Firm core, pink;87.45% NoCM unremark- 21-Dec-01 Firm core, pink;66.5% Mild CM n/a 26-Dec-01 Soll core, pink;88.87% NoCM Low Partially Sarcosysts Scavenged n/a Totally Scavenged unremark- Partially able Scavenged Ferrur Recovered 20-Dec-01 Marrow Fat 12/21/ 2001 12/2123/2001 011.212/2002 01/37/2002 14-Jan-02 n/a 8-Jan-02 Firm core, pink;78-50% Firm core, pink;90.71 % Firm, red; 68.56% Jelly, red: 1.57% Firm core, pink:94.36% NoCM 7-Mar-02 Soft core, red;5.27% n/a S-Mar-02 n/a 12/18/ 2001 4 173.269/01 M 98 J 5 172.350/01 F 86 H 6 173.852/01 F 101 I 1/4-9/ 2002 10-Jan-02 7 175.181/01 M 91 A 6-Feb-02 8 172.170/01 F 58 H 1/15-2/5 /2002 1/20-21/ 2002 1/25-28/ 2002 2/21-28/ 2002 2125-316 /2002 2/27.:J/6 /2002 4-Mar-02 9 173.861/01 F 102 J 10 174.770/01 M 92 A 11 172.379/01 F 81 G 12 173.300/01 M none J 13 173.632/01 14 15 16 23-Jan-02 2S-Jan-02 Carcass Status Carcass Complete UTMx 351433 4245033 Prosser Ck Carcass Complete Capture Related eulhanized 353050 4250600 E. TableTop Mt Capture Related / 302744 4230958 Skunk Ck. 300102 4267074 Pearson Pt 334033 4282350 Flat Top Min Partially Lion predation Scavenged Scavenged Link-suspect n/a heavily coyote predation ·Accident-Haystack bronchoCarcass pnuem:,nia Complete collapse 321888 4275810 W.Antelope Ck 297064 4231180 Dwyer Gulch 330243 4280937 Redden's NoCM Moderate sarcocysts Partially Scavenged Lion predation 299250 4267725 Pearson Pt NoCM 294622 4227200 Elk Ck. Unk.nOIMl 342715 4288645 AlrnontTriangle n/a n/a Carcass Complete Totally Scavenged Not FoundSnow Partially Scavenged Totally Scavenged Totally Scavenged Accident-fell, trapped n/a Moderate sarcocysts n/a Unk.nOIM'l 299323 4268454 N.Pearson Pt Lion predation 322925 4232867 RoadBeaver Ck. Unk.nOIMl 306109 4267739 Red Ck. Unk-suspect lion predation 325443 4235881 N. RoadBeaver Ck. Totally Unk-suspect lion Scavenged predation 351363 4240438 Home Gulch able n/a No CM n/a NoCM 96 C I Link-suspect lion predation Bear Predation Fern1r Marrow Fat Tissue Samples Parasites Carcass Recovered >06/22<7 /21 /01 >06/22<7/24/01 21-Jul-01 24-Jul-01 n/a 77.46% n/a n/a n/a n/a G >6/22<7/20/01 16-Aug-01 n/a n/a n/a F C >9/25<10/18/01 19-0ct-01 n/a n/a 16 mos F J > 10/13<10/18/01 20-0ct-01 n/a n/a Co~lete 174.478/00 5-9 yrs F G 10/13/2001 17-0ct-01 v,Me,solid 97.31% pink,crumbles 85.18% n/a n/a n/a Hunter Kill 7 174.360/00 17 mos F G >11/8<11/16/01 17-Nov-01 n/a n/a Co"1)1ete 8 174.140/00 18 mos M C >12/5 <12/28/01 29-Dec-01 n/a n/a Scavenged 9 173.589/00 18 mos F 8 >12/28/01 <1/3/02 n/a white/gray, solid 95.18% white/pink, firm86.27% n/a n/a n/a n/a 10 >10/30/01 <12/31/ n/a n/a 01 Age estimated using dental cementum or know, age as elk collared as calf. n/a n/a n/a Trap- Age@ Death Sex 1 2 172.758/00 173.330/00 19 yrs 12 mos F M H G 3 173.340/00 12 mos M 4 172.030/00 6 yrs 5 172.619/00 6 174.560/00 17 rros F zone Death Date J Status ! ' ' I • the Gunnison Basm 15December2000t0 141une 2002 A,01 :>emdixD S;ummarvof adult elie mortalitiesm No ElklD ' I Lion Low Sarcosysts 26-Mar-02 Firm core. NoCM Moderate 3/20-25 pink: 83.60% sarcocysts /2002 n/a 173.780/01 108 J 4/25-4/30 1-May-02 Firm core. n/a F pink; 27.08% /2002 5/15-20 21-May-02 Red, firm n/a n/a 175.240/01 M 100 C core;60.02% /2002 95 5115-20/ 22-May-02 Firm core, n/a n/a 174.180/01 M E 2002 pink;75.63% Mass = Weight of calf (kg) at capture· CM = capture myopathy; n/a = sanples not available F Death Location UTMy Drainaae Death Cause Capture Related Fence Kill Death Location Death Cause UTMx UTMy Drainage Decomposed Unk.rov,, Mortality Scavenged Un k--Suspect Predation Heavily Unk-Suspect Scavenged PredaUon Heavily Archery/Muzzle Scavenged v.ound loss 321262 354289 4294461 $. Carbon Min. Surrmerville Ck. 384766 4303460 322980 4227638 N. Cottom11oood Ck. Swnehart Gulch Rifie V><>und loss 1st season Rifie 1st season Legal Rifte 4th season w:,und loss Late rifle v.oundrillegal loss Disappear Late Rifle season Legal Disappear 3rd rifle season Legal 299161 4275164 C<mCk. 341000 4242650 Rock Ck. 346979 4279939 E. Beaver Ck. 308573 4242145 Lake City Cut-Off n/a n/a Low Cebolla Ck. n/a n/a West Elk Ck. 4290527 �223 APPENDIX I PROGRAM NARRATIVE STUDY PLAN FOR RESEARCH FY 2000-01 - FY 2003-04 State of: Colorado ProjectNo.: W-153-R-14 Work Package: __3-'0~0~2_ _ Study No.: _ _ __,3'---- Cost Center 3430 Mammals Research Program Elk Conservation Estimating Calf and Adult Survival and Pregnancy Rates of Gunnison Basin Elk Populations ESTIMATING CALF AND ADULT SURVIVAL AND PREGNANCY RATES OF GUNNISON BASIN ELK POPULATIONS Principal Investigators David J. Freddy, Wildlife Researcher, Mammals Research R Bruce Gill, Wildlife Research Leader, Mammals Research Cooperators Rick Kahn, Terrestrial Field State Coordinator John Ellenberger, State Big Game Coordinator Jim Olterman, Senior Biologist, West Region Don Masden, Gunnison Area Terrestrial Biologist Jim Young, Gunnison Area Wildlife Manager Gary C. White, Professor Wildlife Biology, Colo. St. Univ. David C. Bowden, Professor Statistics, Colo. St. Univ. STUDY PLAN APPROVAL Prepared by: _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ Date: _ _ _ _ _ _ __ Submitted by: _ _ _ _ _ _ _ _ _ _ _ _ _ __ Date: _ _ _ _ _ _ _ __ Reviewed by: _ _ _ _ _ _ _ _ _ _ _ _ _ __ Date:. _ _ _ _ _ _ _ __ Date: _ _ _ _ _ _ __ Date: _ _ _ _ _ _ __ Approved by: _ _ _ _ _ _ _ _ _ _ _ _ _ __ Biometrician Date: _ _ _ _ _ _ __ Date: _ _ _ _ _ _ _ __ Research Leader November 2000 Final �224 PROGRAM NARRATIVE STUDY PLAN State of: Colorado Project No.: W-153-R-14 Work Package: 3002 Study No.: 3 Cost Center 3430 Mammals Research Program Elk Conservation Estimating Calf and Adult Survival and Pregnancy Rates of Gunnison Basin Elk Populations A. NEED Elk (Cervus elaphus nelsoni) are a high-profile and highly valued resource throughout much of Colorado because elk provide recreation for persons who hunt, watch, and photograph wildlife (Freddy et al. 1993). The elk resource has many benefits but frequent social, political, and economic conflicts suggest elk can reach "social" if not ''biological" carrying capacities. Recent controversy surrounding management of elk in the Gunnison Basin of Colorado (Roath et al.1999) exemplifies conflicting social and biological agendas regarding appropriate numbers of elk. The core of conflict in elk management often centers on establishing management objectives for numbers of elk that are agreeable to competing interests and then monitoring elk populations to demonstrate that objectives are achieved. This type of conflict is paramount in Colorado Division of Wildlife (CDOW) elk population Data Analysis Units (DA Us) E-25, E-41, and E-43 in the Gunnison Basin (Fig. 1) where a combination of resource carrying capacity objectives for elk on winter ranges and difficulties associated with knowingly achieving those objectives has fostered argumentative distrust among public groups and management agencies. Accomplishing management by population objective can depend on reliably estimating elk population size. Estimating population size is expensive and intensive (Samuel et al. 1987, Bear et al. 1989, Unsworth et al. 1990, Anderson et al. 1998, Cogan and Diefenbach 1998, Eberhardt et al. 1998, Freddy 1998) and these factors often preclude routinely using tested inventory methodologies. Alternatively, population size and trend can be estimated using computer models that incorporate harvest, age and sex ratios, and survival rates (White 1992, Bartholow 1999). Model outputs are extremely sensitive to estimates of survival rates such that, reliable measurements of survival can greatly enhance the quality of models (Nelson and Peek 1982). Thus, estimating survival rates is fundamental to modeling elk pop~la!ions in the absence of routine measurements of population size. Estimating calf and adult female survival during winter and annual rates of survival for adult females are higher priorities than estimating adult male survival primarily because most males are harvested when they reach legal age and contribute little to long-term problems of population growth or decline. Models having valid estimates of survival along with currently obtained precise estimates of harvests and population composition would provide more defensible estimates of population size. Although small changes in adult female survival can have major effects on population growth or decline if compounded for several years, calf survival is likely more variable among years. The ability to detect changes in calf survival should be greater than detecting smaller, but important changes in adult female survival (White et al. 1987, Bartmann et al. 1992, Freddy 1998). Estimates of calf survival in Colorado during winter are limited to the Grand Mesa in west-central Colorado where yearly average survival varied between 0.86 and 0.92 from 1993-1996 (Freddy 1998, n 2::, 67 calves/year). Applying these survival rates to other Colorado elk populations, especially those populations using winter ranges higher November 2000 Final �225 in elevation, colder, and more prone to significant snow depths such as in the Gunnison Basin, may or may not be appropriate. Rates of survival on the Grand Mesa were higher than expected and considerably greater than 0.70-0.72 survival rate estimated for elk calves during winter in Yellowstone National Park (Houston 1982, Singer et al. 1997). Estimates of annual survival for radio-collared adult female elk in Colorado averaged 0.95 and ranged from 0.94-0.99, excluding hunting mortalities, for several populations inhabiting widely differing ecosystems (Petersburg and White 1998, Freddy 1999; n > 1,250 adult female-years). Because of the availability of these adult survival estimates, the need to estimate adult female survival is therefore less than the need to obtain additional estimates of calf survival, but ideally we would measure both calf and adult survival simultaneously to document relative differences in survival. A recent evaluation of existing population models for elk in the Gunnison Basin and subsequent development of new population models using estimates of calf and adult survival measured in Colorado altered population trajectories and relative size (Freddy 2000). Consequently, management objectives for Gunnison elk were amended to continue reducing numbers of elk in all DA Us. Controversy surrounding new models and management decisions reinforced the need to obtain measurements of elk survival specific to the Gunnison Basin. B. OBJECTIVE This project will obtain estimates of population parameters for elk in the Gunnison Basin. Major objectives are: I) Estimate survival rates of elk calves during winter from 15 December-14 June within ±15% of the true survival rate at the 95% confidence interval for 3 consecutive years and identify probable sources of mortality. 2) Estimate winter (15 Dec-14 Jun) and yearly (15 Dec-14 Dec) survival rates of adult females for 3 consecutive years to assess whether the true survival rate is likely ,2:0.95 and identify probable sources of mortality. 3) Estimate pregnancy rates of adult female elk harvested during November-December late hunting seasons for 3 consecutive years if late hunting seasons are scheduled. 4) Estimate hunting removal rates for adult females, yearling males, and when possible, adult males for 3 consecutive years. 5) Evaluate Gunnison elk population models using newly acquired survival rates. C. EXPECTED RESULTS OR BENEFITS This project will provide estimates of survival rates for calf and adult female elk and estimates of hunting removal rates for adult elk in the Gunnison Basin DA Us E-25, E-41, and E-43 for 3 consecutive years. These estimates will immediately assist the CDOW in refining population models for Gunnison elk and provide estimates of survival/removal that may be applicable to modeling other elk populations inhabiting similar habitats. In the process of estimating survival rates, probable causes of mortality will be identified which may provide insight into relative health status of elk. Additionally, estimates of pregnancy rates will provide documentation on the fecundity of these elk in relation to other elk populations in Colorado and other states. November 2000 Final �226 D. APPROACH EXPERIMENTAL DESIGN SURVIVAL RATES Radio-telemetry Equipment Survival rates will be estimated by marking elk with radio-telemetry collars that emit a mortality pulse code when collars remain motionless for 4-6 hours (White et al.1987, Freddy 1993). Radios provide the ability to know the fate of individual animals (alive or dead) over discrete periods of time (White and Garrott 1990). Radio-collaring does not likely bias estimates of survival by jeopardizing or enhancing the welfare of individuals when radio-collars weigh <0.8% of an ungulate's body weight (Garrott et al. 1985, White et al. 1987). Radio-collars similar to those previously designed and successfully used for calf and adult elk on the Grand Mesa, Colorado will be used in this project (Freddy 1993, Appendix I). Collars for male and female calves will allow for expansion to adult size while adult female collars will be of fixed circumference and fitted to each individual. Calf collars weigh 840 gm and represent <l % of expected calf body weight while adult collars weigh 1.1 kg and represent <0.5% of expected body weight. Collars will be white in color, have a unique black colored number/symbol embossed on bright yellow plastic material (Ritchey Manufacturing, Brighton, CO) attached to the dorsal surface of collar to enhance visual identification from helicopters (Appendix II), have unique frequencies between 172-17 6MHz, and a battery life of ,2:4 years. Animal Capture We assume survival of those elk captured provides an unbiased estimate of population survival rates recognizing that individual behavior, social behavior, trapping methods and distribution of trapping effort all potentially bias those individuals actually marked (White et al. 1982). Recognizing these problems, elk will be captured with the intent of systematically marking elk throughout the distribution of elk in the Gunnison Basin. Each of the 3 DAUs, will be divided into trap-zones having multiple trap-sites. Capture quotas for calves and adults in trap-zones within each DAU will be proportional to expected elk density as estimated from yearly sex and age ratio classification flights conducted each January throughout the Gunnison Basin. Trap-zones will be initially defined as: 1) for DAU E-25: Big Blue Creek to Gunnison River (TzA), Gunnison River to Cebolla Creek (TzB), Cebolla Creek to Gold Basin Creek (TzC), and Gold Basin Creek to Cochetopa Creek (TzD), 2) for DAU E-43: Cochetopa Creek to Tomichi Creek (TzE), Tomichi Creek to Quartz Creek (TzF), Quartz Creek to East River (TzG), and 3) for DAU E-41: Ea,st River to Ohio Creek (TzH), Ohio Creek to Dry Creek (Tzl), and Dry Creek to Curecanti Creek (TzJ). • Elk will be captured using a Hughes 500 helicopter and net-guns (contracted services) (Freddy 1994). We will attempt to collar equal numbers of male and female calves. Helicopter trapping will occur in mid-December each year. Capture and handling procedures will follow protocols used to capture 257 calves and 46 adult females on the Grand Mesa (Freddy 1993-1996) and previously approved by CDOW Animal Care and Use Committee (Appendix III). Survival Monitoring Radioed elk will be monitored daily from the ground and bimonthly with aerial surveys (Cessna 185 or equivalent) to determine life/death status of elk. During hunting seasons, aerial surveys will be occur bimonthly in September and weekly during October and November. RADIOS database program will be used to maintain animal records. November 2000 Final �227 Suspected mortalities will be confirmed using ground searches. Criteria for assigning probable cause of death will include body position, presence of bite or claw marks and sub-dermal hemorrhaging, tracks, drag marks, and tissue samples if available (Wade and Browns 1982, Freddy 1998). Potential causes of death include starvation, accidental trauma, plant poisoning, predation by black bears, mountain lions, coyotes, and domestic dogs, and legal and illegal hunter harvest (Freddy 1997). Survival Sample Sizes and Tests Each year we will radio-collar 78 calves (39 male, 39 female) with 26 calves marked in each DAU and during the initial year, 39 adult females will be radio-collared with 13 in each DAU (Table 1). We anticipate >20 radioed female calves will be recruited to yearling adults each year resulting in >50 radioed adult females in the population to estimate adult survival in subsequent years. However, by not collaring known adult females each year, we run the risk of having biased estimates of adult female survival because the age structure of collared adult females will progressively be biased to younger aged females recruited from marked calves. An alternative would be to mark enough adult females in each subsequent year to replace those adult females marked in year 1 that had died the previous year. We anticipate needing to replace 15-20 older adult females per year to achieve this goal which will be dependent on future funding. If adult females could be replaced yearly, we would be able to separate year effects from age effects on survival rates. Approximately 30 yearling males will be available each year, 2001-2003, to estimate percent of yearling males illegally removed under a hunting system using antler-point regulations to protect yearlings. Approximately 30, 2:2-year-old males will be available each year, 2002-2004, to estimate percent of branch-antlered males removed with a hunting system using antler-point regulations. We chose to mark 78 calves per year and 39 adult females during the initial year because we will have acceptable confidence intervals about mean estimates of survival each year for all DA Us pooled into 1 elk population, have the potential to detect major differences in survival between years due to changes in winter severity when all 3 DAUs are pooled, and be able to detect major differences in survival between DA Us when data are pooled within DAUs for 3 years. The ability to detect differences between DA Us within years is desirable but economically prohibitive due to numbers of collared elk required (>47 per DAU per year). We anticipate yearly calf survival to be 0.70 to 0.90 and adult survival, exclusive of hunting-related deaths, to be 0.95 to 0.99. If calf survival is 2:0.70 (n = 78 calves), 95% confidence intervals (Zar 1984, 378) will be ,:s ±15% of the yearly mean survival rate. If adult female survival is 2:0.95 (n 2: 39 adults), 95% confidence intervals will be ,:s ±10% of the yearly mean survival rate. Additionally, if adult female survival= 0.95 we expect to estimate yearly survival within ±5% of the true survival rate at alpha= 0.10 when n 2: 50 adult females. November 2000 Final �228 Table 1. Elk calves (6 months old) and adult females (2:12 months old) captured and radio-collared for Gunnison elk DAUs E-25, E-41, and E-43, December 2000-2002 (shaded cells). Adult females captured only during initial year, 2000. Numbers of radioed adult males and females in December years 2001-05 estimated by assuming survival rates between years: adult females net rate= 0.7, male and female calves to yearling adult age net rate= 0.8, yearling males to adult males net rate = 0.9, adult males net rate= 0.3. DAUE-25 Calves DAU E-41 Adults Calves DAU E-43 Adults C alvcs ALLDAUs Adults Calves Adults Totals Year M F M F M F M F \1 F M F M F M F All 2000 13 13 0 13 13 13 0 13 13 13 0 13 39 39 0 39 117 2001 13 13 10 19 13 13 IO 19 13 13 10 19 39 39 30 57 165 2002 13 13 13 23 13 13 13 23 13 13 13 23 39 39 39 69 186 2003 14 26 14 26 14 26 42 78 120 2004 4 18 4 18 4 18 12 54 66 2005 1 13 1 13 1 13 3 39 42 42' 112' 42' 112' 42' 112' 336' 696' All, -39 39 39 39 39 39 117 126' 117 -~, - ~ ·-- • Represents elk-years and not necessarily numbers of individual radioed adult elk as adults survive between years. Number of collars deployed in combination with actual survival rates determines our ability to detect differences in survival among years, DA Us, or geographic areas. When survival rates are near 0.50, variance, or precision, about the mean survival estimate is largest, and thus the sensitivity to detecting differences in survival rates is least (Zar 1984). As survival rates approach 0.0 or 1.0, precision improves for a fixed sample size of collars, and sensitivity to detecting differences in survival increases. Given our assumptions about expected average survival rates and potential higher calf survival in DAU E-25 based on computer modeling (Freddy 2000), we estimated the statistical power (Snedecor and Cochran 1967; 113, 221, 269; pers. comm. D. Bowden) to detect differences in mean survival rates given specific hypotheses. We consider detecting differences in survival of 0.20 with statistical power of 0.80 at an alpha= 0.10 to be acceptable. Generalized hypotheses (S = survival rate) and power for detecting major differences in survival among years, DA Us, age and sex classes, and geographic areas. •· • (1) Ho: scalvesyearl = scalvesyear2 = scalvesyear3 for DAUs pooled each year. HA, scalvesyearl "F scalvesyear2 "F scalvesyear3 forDAUs pooled each year. Power= 0.80 at alpha= 0.10 to detect differences in yearly survival of 0.15 between pairs of years given 78 collars per year and expected survival rates of 0.90 and 0.75. Power= 0.80 at alpha= 0.10 to detect differences in yearly survival of0.15 between 1 year with lower survival and the average higher survival of the other 2 years given 78 collars per year and expected survival rates of 0.75 and 0.90. (2) Ho: scalves DAUi = scalvos DAU2 = scalves DAUJ for years pooled for each DAU. HA: scalwsDAUl "F scalwsDAU2 "F scalwsDAUJ for years pooled for each DAU. Power= 0.80 at alpha= 0.10 to detect a difference in 3-year average survival of 0.15 between pairs of DAUs given 26 collars per year per DAU and expected yearly survival rates of 0.90 and 0.75. November 2000 Final �229 Power= 0.90 at alpha= 0.10 to detect difference in 3-year average survival of 0.15 between 1 DAU with higher survival and the average lower survival of the other 2 DA Us given 26 collars per year per DAU and expected yearly survival rates of 0.90 and 0.75. Power= 0.90 at alpha= 0.10 to detect difference in 3-year average survival of0.15 between 1 DAU with higher survival and the average lower survival of the other 2 DA Us given 26 collars per year per DAU and expected yearly survival rates of 0.90 and 0.80 and 0.70 amongDAUs. (3) Ho: Smale cah'es = sfem:lle calves for years pooled for each seL HA: Smale calves 'F sfemalecalves for years pooled for each sex. Power= 0.90 at alpha= 0.10 to detect difference in 3-year average survival of 0.15 between sexes of calves given 35 collars per year per calf sex and expected survival rates of 0.75 for one sex and 0.90 for the other sex. (4) Ho: sadultremalesyearl = sadultremalesyear2 = Sadultremalesyear3 for DAUs pooled each year. HA, sadultremalesyearl 'F sadultremalesyear2 'F Sadultremalesyear3 for DAUs pooled each year. Power = 0. 80 at alpha = 0 .10 to detect difference in survival of 0 .15 between pairs of years given 56 collars per year and expected survival rates of 0.95 and 0.80. Power = 0. 80 at alpha = 0 .10 to detect differences in yearly survival of O.15 between 1 year with lower survival and the average higher survival of the other 2 years given 50 collars per year and expected survival rates of 0.80 and 0.95. (5) Ho: scalves = sadult females Ho: Scalves 'F sadult females Power= 0.80 at alpha= 0.10 to detect difference of 0.15 between calf and adult female survival within each year given 56 collars per year per age class and expected survival rates of 0.80 for calves and 0.95 for adult females. Power= 0.90 at alpha= 0.10 to detect difference in 3-year average survival of0.10 between calves and adult females given 51 collars per year per age class and expected survival rates of 0.85 for calves and 0.95 for adult females. (6) Ho: scalves Gunnison = scalves Grand Mesa for years pooled within each area. HA: scalves Gunnison 'F scalves Grand Mesa for years pooled within each area. Power= 0.80 at alpha= 0.05 to detect difference in a 3-year average survival of 0.10 between calf survival in the Gunnison Basin and calf survival on the Grand Mesa given 66 collars per year per area and expected survival of 0.80 in the Gunnison Basin and 0.90 on the Grand Mesa. Survival will be estimated for calves and adults during winter-spring (15 Dec -14 Jun), for adults during summer-fall (15 Jun-14 Dec), and for adults during the year (15 Dec - 14 Dec). The yearly time period, or biological year, initiates with capture and release of marked elk into the population (White et al. 1987). Capture of elk will occur in mid-December instead of early December as on the Grand Mesa (Freddy 1993-1997) to accommodate capture services on other projects. We expect this change in capture dates to have minimal effects on estimates of survival as no natural deaths of calves or adults occurred during December on the Grand Mesa (Freddy 1999). We will use a staggered entry Kaplan-Meier analysis to estimate survival rates (SAS 1988, White and Garrot 1990, Bartmann et al. 1992). We will compare survival rates using chi-square analyses and conduct pair-wise comparisons using log-rank tests to compare survival of calves and adults among years for DA Us combined, between DA Us for years combined, between male and female calves, and between calves and adults. We will assess whether calf survival can be predicted from sex, body weight, hind November 2000 Final �230 foot length, total body length, and mean monthly snow depths and temperature using logistic regression (SAS 1988). Additionally, we will test for differences in survival of calf and adult elk between the Gunnison Basin and the Grand Mesa (Freddy 1998) potentially using beta-binomial distribution approaches outlined by Unsworth et al. 1999. Tests will be significant at alphaP S 0.10. PREGNANCY RA TES Fecundity of adult female elk will be determined by examining reproductive organs of antlerless elk harvested during hunting seasons from mid-November through December. Initially, late seasons are scheduled to occur in 2000, but may continue in subsequent years depending upon population management objectives. Numbers of hunters will be controlled by limited permits issued each year. During 2000, we anticipate >650 hunters will provide ~200 useable reproductive tracts from antlerless elk harvested in portions of DAUs E-25, E-41, and E-43. Hunters will be mailed packets explaining procedures for collecting reproductive organs and incisor teeth from harvested elk as done previously in Colorado for Middle Park and Forbes-Trinchera elk collections (Freddy 1992, pers. comm. C. Wagner, CDOW). Additionally, we will ask hunters to collect kidneys and associated fat from harvested elk to allow calculation of kidney-fat indices to better assess body condition of adult females in relation to reproductive status (Kohlmann 1999). Hunters will be directed to leave collected organs at drop-off sites in Lake City, Colorado, Gunnison CDOW Service Center, Gunnison commercial meat-processors, and at CDOW Roaring Judy Hatchery. Fetuses will be sexed, weighed, and measured (Armstrong 1950) with conception dates estimated from fetal measurements (Morrison et al.1959). Pregnancy status, fetal age, fetal sex, and conception dates will be related to female dental cementum age and kidney fat indices using regression analyses (SAS 1988). Additionally, comparisons to reproductive measurements on elk from Middle Park and ForbesTrinchera will be made. POTENTIAL ADDITIONAL EXPERIMENTS AND APPLICATIONS Management of elk in the Gunnsion Basin has contentiously focused on population status of elk, impacts of elk on plant communities, long-term carrying capacities for wild and domestic ungulates and seasonal patterns of habitat use (Carpenter et al. 1980, Roath et al. 1999). Expanding our understanding of these general topics can be greatly enhanced by effectively utilizing the radio-collared elk that will be available because of this project. Investigations regarding these topics could be initiated with additional funding, personnel, and agency cooperation. . . Potential investigati9ns could address:. a). Management objectives·as of 1999 are to reduce elk populations in DAUs E-25, E-41, and E43. Reductions are projected to be most severe in DAU E-25 and approach 50% over the next 5 years based on computer models. Reductions in DAUs E-41 and E-43 are projected to be <25% and completed in 2-3 years. If elk are indeed at biological carrying capacity and if reductions proceed in E-25, there may be the opportunity to conduct management experiments to assess whether calf survival and/or fecundity increase in response to lowered density. Radio-collaring and monitoring additional calves each year would be required. Estimated additional costs could approach $50,000. b). Population reductions may also create an opportunity to apply sampling systems developed to estimate elk density, including mark-resight estimators (Freddy 1998), to verify modeled population status and achievement of populations goals. Estimated additional costs would be in helicopter hours ($40,000) and additional radio-collared elk ($40,000) for mark-resight surveys. Additionally, sampling systems to estimate sex ratios could be implemented and evaluated in E25 with reallocation of existing survey monies plus an additional $10,000. November 2000 Final �231 c). Patterns of habitat use and forage removal could be investigated utilizing intensive measurements on selected range sites and monitoring ofradioed elk and their associates. This would be a major project and possibly approach $100,000 per year including additional personnel. d). Seasonal movements and patterns of spatial use to document seasonal behavior of elk would require additional personnel and aerial fixed-wing costs of $40,000 per year. PROJECT SCHEDULE Fiscal Year 2000-01 Activity/Objective Complete study plan; purchase radio-collars; Estimate pregnancy/fetal rates; Trap and radio-collar elk and estimate survival. Period Jul-Nov Nov-Dec Dec-Jun 2001-02 Estimate survival and hunting removal rates; Estimate pregnancy/fetal rates; Trap and radio-collar elk and estimate survival. Jul-Dec Nov-Dec Dec-Jun 2002-03 Estimate survival and hunting removal rates; Estimate pregnancy/fetal rates; Trap and radio-collar elk and estimate survival; assess potential for mark-resight estimates of elk density. Jul-Dec Nov-Dec Dec-Jun 2003-04 Estimate survival and hunting removal rates; complete data analyses, initiate manuscripts. Jul-Jun Estimated Annual Costs FTE Requirements PFTE = 1.00 TFfE= 0.83 TOTAL= 1.83 Budget Category (01) Personal Services (21) Operating Supplies and Services (21) Utilities (28) Travel Expenses (31) Capital Outlay Total Costs Costs $102,000 84,000 0 1,000 0 $187.000 Costs anticipated to increase 5% each year in 2001-02, 2002-03, 2003-4 for inflation. November 2000 Final �232 Personnel Program Responsibilities David J. Freddy: Wildlife Researcher, Principal Investigator responsible for final project design, organizing field personnel, obtaining and organizing data, data analyses, financial control, and coordinating publications. R Bruce Gill: Wildlife Research Leader, provides administrative support, input for study design, and liaison with other administrative sections within the Division of Wildlife. Rick Kahn, John Ellenberger, Jim Olterman, Don Masden, Jim Young: Provide coordination and support of Terrestrial managers and biologists and Area management staff and facilities. Gary C. White: Provide input for study design and statistical protocol, conduct data analyses, and provide software support. David C. Bowden: Provide input for study design and statistical protocol. E. LOCATION The Gunnison Basin in south-central Colorado was selected for this project (Fig. 1). The Basin encompasses the entire headwaters of the main Gunnison River and the centrally located town of Gunnison. Between 12-16,000 elk and 8-10,000 mule deer (Odocoileus hemionus) are thought to exist within the Basin. Elk are managed as 3 populations representing DAUs E-25 (Game Management Units [GMU] 66, 67), E-41 (GMU 54), and E43(GMUs 55,551). The 3 DAUs encompass about 9,291 km2 of which 3,648 km2 are considered winter range for elk (CDOW WRIS database). DAUs are contiguous with no major geographic barriers separating DAUs that would prevent interchange of elk among DAUs. t N ···················••••••••••••••••••••• 50 Miles The Basin represents a high altitude, cold winter range for both elk and mule deer which BO Km is similar to ecosystems in North Park, Middl~ Fig. 1. Location of the Gl:llllison Basin and elk Data Park, and the San Luis Valley, Colorado. The Analysis Units E-25, E-41, and E-43 within Colorado. sagebrush steppe winter ranges (2,250- 2,700 m elevation) can receive extreme snow depths and cold temperatures that cause severe mortality among ungulates (Carpenter et al. 1984) while the conifer-alpine summer ranges (3,000 - 4,200 m elevation) can be subjected to drought. Overall, these ranges collectively are thought to be less productive and nutritious for elk than the milder climate oakbrush-pinyon-juniper winter ranges of the Grand Mesa where elk survival was measured from 1993-99. We anticipate dependable access to both private and public lands to conduct research activities and there is local, Area, and Regional CDOW support for conducting the project in this area. Additional financial and logistical support may be available from the Gunnison Habitat Partnership Committee. The airport and other businesses in Gunnison will provide readily accessible support services. November 2000 Final �233 F. RELATED FEDERAL AID PROJECTS Calf and adult elk survival rates were measured on the Grand Mesa, Colorado from 1993-99 under Federal Aid Research Project W-153-R (Freddy 1994-1999). G. LITERATURE CITED Anderson, C.R, Jr., D.S. Moody, B.L. Smith, F.G. Lindzey, and R. P Lanka. 1998. Development and evaluation of sightability models for summer elk surveys. Journal of Wildlife Management 62:1055-1066. Armstrong, R.A. 1950. Fetal development of northern white-tailed de~i:- American Midland Naturalist 43:650-666. Bartholow, L 1999. POP-II system documentation Windows TM Version 1.0. Fossil Creek Software, Fort Collins, Colorado USA. Bear, G.D., G.C. White, L.H. Carpenter, RB. Gill, and DJ. Essex. 1989. Evaluation of aerial markresighting estimates of elk populations. Journal of Wildlife Management 5 3: 908-915. Biutrnann RM., G.C. White, and L.H. Carpenter. 1992. Compensatory mortality in a Colorado mule deer population. Wildlife Monographs 121. Carpenter, L.H., D.L. Baker, and R.B. Gill. 1980. Tests of a nutritionally based big game habitat •· ~valuation system. Colorado Division of Wildlife Unpublished Report. ColoradcfDivision of • -- Wildlife, Fort Collins, Colorado, USA. Carpenter, L.H_, R. B. Gill, D. L. Baker, and N.T. Hobbs. 1984. Colorado's big game supplemental winter feeding program. Colorado Division of Wildlife, Fort Collins, Colorado USA. Conner, M.M. 1999. Elk movement in response to early-season hunting in the White River Area, Colorado. Dissertation, Colorado State University, Fort Collins, Colorado, USA. Cogan, RD_, and D.R. Diefenbach. 1998. Effect ofundercounting and model selection on a sightabilityadjustrnent estimator for elk. Journal of Wildlife Management 62:2~9-279. Eberhardt, L.L., RA. Garrott, PJ. White, and PJ. Gogan. 1998. Altematiye approaches to aerial censusing of elk. Journal ofWildlife Management 62:1046-1055 .. Freddy, DJ. 1992. Effect of elk harvest systems on elk breeding biology. Colorado Division of Wildlife Research Report July: 45-70. Fort Collins, Colorado, USA. July:45-70. . Freddy, D .J. 1993. Program Narrative for Estimating survival rates of elk and developing techniques to estimate population size. Colorado Division of Wildlife Research Report July:83-117. Fort Collins, Colorado, USA. • Freddy, DJ. 1994. Estimating survival rates of elk and developing techniques to estimate population size. Colom.do Division of Wildlife Research Report July: 27-42. Fort Collins, Colorado, USA. Freddy, DJ. 1995. Estimating survival rates of elk and developing_ techniques to estimate population size. Colorado Division of Wildlife Research Report July: 63-79. Fort Collins, Colorado, USA. Freddy, DJ. 1996. Estimating survival rates of elk and developing techniques fo estimate population size. • Colorado Division of Wildlife Research Report July: 87-108. Fort Collins, Colorado, USA. Freddy, DJ. 1997. Estimating survival rates ofelk and developing techniques to estimate population size. Colorado Division of Wildlife Research Report July: 47-73. Fort Collins, Colorado, USA. Freddy, DJ. 1998. Estimating survival rates of elk and developing techniques to estimate population size. Colorado Division of Wildlife Research Report July: 177-206. Fort Collins, Colorado, USA. Freddy, DJ. 1999. Estimating survival rates of elk and developing techniques to estimate population size. Colorado Division of Wildlife Research Report July: (in press). Fort Collins, Colorado, USA. Freddy, DJ. 2000. Modeling elk populations in the Gunnison Basin, Colorado using POPII and POPMOD software (draft in process). Colorado Division of Wildlife Special Report No.??, Fort Collins, Colorado USA. Freddy, DJ., D.L. Baker, R.M. Bartrnann, and RC. Kufeld. 1993. Deer and elk management analysis guide, 1992-1994. Colorado Division ofWildlife Division Report 17, Fort Collins, Colorado USA. November 2000 Final �234 Garrott, RA., R.M. Bartrnann, and G.C. White. 1985. Comparison ofradio-transmitterpackages relative to deer fawn mortality. Journal of Wildlife Management 49:758-759. Houston, D.B.: 1982. _The,nqrthemYellowstone elk, ecology and management. Macmillan Publishing Company, Inc. New York, New York, USA. Kohlinann,'S.G. 1999. Adaptive fetal sex allocation in elk: evidence and implications. Journal of • Wildlife Management 63: 1109-1117. Morrison, J.A., C.E. Trainer, and P.L. Wright. 1959. Breeding seasons in elk as determined from knownage embryos. Journal of Wildlife Management 23:27-34. Nelson, L.J., and J.M. Peek. 1982. Effect of survival and fecundity on rate of increase of elk. Journal of Wildlife Management 46:535-540. _. Petersburg, M., and G. White. 1998. Kaplan-Meier survival estimates for cow elk. Colorado Division of Wildlife Terrestrial Section Unpublished Memorandwn, Fort Collins, Colorado USA. Phillips, G.E. 1998. Effects ofhwnan-induced disturbance during calving season on reproductive success of elk in the upper Eagle River Valley, Colorado. Dissertation, Colorado State University, Fort Collins, Colorado, USA. Quimby, D.C., and J.E. Gaab. 1957. Mandibular dentition as an age indicator in Rocky Mountain elk. Jqurnal of Wildlife Management 21: 134-153. Roath, R)L. Carpenter, :B, R.4epsame, and D. Swift. 1999. Gunnison Basin habitat assessment project. ,Fi¼al Report Mafch i"999. Colorado State University, Department of Range Science, Fort Collins, • 'Colorado USA. Samuel, M.D., E.O. Garton, M.W. Schlegel, and R G. Carson. 1987. Visibility bias during aerial surveys of elk in northcentral Idaho. Journal of Wildlife Management 51: 622-630. SAS. 1988. SAS/STAT user's guide, release 6.03. SAS Institute, Inc. Cary, North Carolina USA. Snedecor, G.W., and W.G. Cochran. 1967. Statistical methods; sixth edition. Iowa State University Press, Ames, Iowa, USA. Singer, F.J., A. Harting, K.K. Symonds, and M.B. Coughenour. 1997. Density dependence, compensation, and environmental effects on elk calf mortality in Yellowstone National Park. Journal of Wildlife Management 61: 12-25. Unsworth, J .W ., L. Kuck, and E. 0. Garton. 1990. Elk sightability model validation at the National Bison Range, Montana. Wildlife Society Bulletin 18:113-115. Unsworth, J.W., D.F. Pac, G.C. White, and R.M. Bartrnann. 1999. Mule deer survival in Colorado, Idaho, and Montana. Journal of Wildlife Management 63:315-326. Wade, D.A., and J.E. Browns. 1982. Procedures for evaluating predation on livestock and wildlife. Texas Agricultural Experiment Station Publication B-1429. White, G.C. 1992. DEAMAN database manager and populatj.on modeling procedures; Colorado Division of Wildlife User',s manual and reference. Colorado State University, FQrt Collins, Colorado USA. White, G.C. and R.A. Garrott. 1990. Analysis'.ofwildlife radio-tracking data. Academic Press, Inc., San Diego, California USA. White, G.C., D.R. Anderson, K.P. Burnham, and D.L. Otis. 1982. Capture-recapture and removal methods for sampling closed populations. Los Alamos National Laboratory LA-8787-NERP, Los Alamos, New Mexico, USA. White G.C., R.A. Garrott, RM. Bartrnann, L.H. Carpenter, and A.W. Alldrege. 1987. Survival of mule deer in northwest Colorado. Journal of Wildlife Management 51: 852-859. Zar, J.H. 1984. Biostatistical analysis, second edition. Prentice-Hall, Englewood Cliffs, New Jersey, USA. November 2000 Final �235 APPENDIX I SPECIFICATIONS FOR RADIO-COLLARS Manufacturer: Lotek, Inc. Pulse Rate Nonnal: 60-65 ppm Pulse Rate Mortality: 120-130 ppm Motion Sensor Delay: 4-6 hrs Batteries: 4+ year life, 1 lithium D-cell calf collars Antenna: External whip, pvc coated Collar Material: 7.6 cm (3") wide white colored smooth surfaced conveyor belting, 2 layers sewn together, 0.64 cm (1/4") total thickness Additional Material: Bright yellow with black core Ritchie All-Flex plastic material for identification symbol/number placed as a sleeve over top portion of collar (Ritchey Manufacturing, Inc., Brighton, CO). Collar Size: 61-81 cm (24-32") adult females, individually fitted; 56-69 cm (22-27") expandable for female calves; 57-89 cm (22.5-35") expandable for male calves. Collar Weight: 820 gm female calves; 840 gm male calves; 1.1 kg adult females. November 2000 Final �236 APPENDIX II VISUAL IDENTIFICATION SYSTEM FOR RADIO-COLLARS Numbers, symbols, and letters will be used in ordered combinations to quickly allow identification of individual elk primarily during aerial surveys. No more than 2 characters will be used to identify an individual. Characters will be ordered and read from left to right on the collar from the perspective of looking down on the elk from the rear of the animal when approached by an observer in a helicopter. Numbers to be Used (8): 0, 1, 2, 3, 4, 5, 6, and 7 Symbols to be Used (5): solid circle e, solid square ■, solid triangle A, solid hourglass X, plus sign +, (solid diamond potentially ♦ ) Letters to be Used (9): A, C, F, H, J, K, N, P, X (T, V, Ypotentially) Identification Combinations: Number combinations represent 56 individuals 10,20,30,40,50,60, 70 11, 21, 31, 41, 51, 61, 71 12,22,32,42,52,62, 72 13,23,33,43,53,63, 73 14,24,34,44,54,64, 74 15,25,35,45,55,65, 75 16,26,36,46,56,66, 76 17,27,37,47,57,67, 77 Each Symbol paired with each Number represents 16 identification codes when ordered symbol-number and then number-symbol. Five symbols paired with 8 numbers represents 80 individuals. Examples: 7eand e7. Each Letter paired with each Number represents 16 identification codes when ordered letter-number and then number-letter. Nine letters paired with 8 numbers represents 144 individuals. Examples: A 7 and 7A. Each Letter paired with each Symbol represents 10 identification codes when ordered letter-symbol and then symbol-letter. Nine letters paired with 5 symbols represents 90 individuals. Examples: Ae and eA. Therefore, a minimum of370 different animals can be individually marked using this system. November 2000 Final �237 APPENDIX III HELICOPTER NET-GUNNING CAPTURE PROTOCOL FOR ELK Background: Helicopter net-gunning has been successfully and safely used to capture and radio-collar elk in Colorado during both winter and summer. This success has been in part due to following accepted protocols for handling elk (Colorado Division of Wildlife Animal Care and Use Committee Reviews and Approvals). Helicopter capture of elk on the Grand Mesa, Colorado during December 1993-1996, resulted in no acute or post-capture related deaths in 46 adult females and I acute death (broken neck, 0.4%) and 2 post-capture myopathy deaths (0.7%) in 258 calves captured and handled (Freddy 1996, 1997). During early December 1994-96 near Vail, Colorado, 2.2% of 185 adult females died from effects of helicopter capture (Phillips 1998). In the White River, Colorado, <I% of 95 adult female elk captured during July and 4% of 32 adult females captured during August near Vail, Colorado died from effects of helicopter capture (Conner 1999, pers. comm. M. Conner, 1999, Phillips I 998). Capture Protocol: Capture of elk will follow procedures successfully used on the Grand Mesa (Freddy 1995). David J. Freddy is the principal investigator and will coordinate capture of elk. All persons involved in the capture operation, including the helicopter net-gunning crew, will be instructed on proper care and handling of elk to reduce stress and injury to elk. Capture Timing and Conditions: Elk are scheduled to be captured during mid-December in the Gunnsion Basin but there remains the possibility that capture could occur in early January depending on availability of contract helicopter services. During either month, cool ambient temperatures and moderate snow depths (<60cm) contribute to successfully capturing elk by reducing threats of hyperthermia potentially induced by capture chases. We anticipate capturing elk when ambient temperatures are -18 - 3°C. Temperatures < -l 8°C (0° F) may restrict human efficiencies while temperatures >3°C (38° F) may induce hyperthermia in elk. No-fly Zones: Pursuit and capture of elk will not occur within 1,000m (0.5 miles) of human residences or other cultural developments such as well traveled roads, reservoirs, etc .. Notification of Affected Parties: Local residents and federal, state, and local agencies will be notified of the time and general area of capture activities. Notification will be via newspaper articles, public meetings, and other informal verbal communications. Emergency Services: Capture personnel will be instructed that the nearest medical and emergency services are located at the Gunnison Valley Hospital in Gunnison. Capture crews will have communications radio contact to CDOW service centers and emergency Colorado State Patrol. Radio Collars/Ear Tags: Expandable collars will be placed on male and female calves to accommodate neck growth as animals become adults (Appendix I). Collar design was previously used on 285 elk calves on the Grand Mesa with no known cases of expandable collars inducing trauma for up to 4 years of age on males and 7 years of age on females (Freddy 1999). Collars of fixed size will be placed on each adult female and individually fitted usually to 69-74 cm (27-29 in). Fixed collar design was previously used on 82 adult females with no known cases of trauma (Freddy 1999). No ear-tags will be used. Command Post: The principal investigator and handling crew will establish mobile command /handling sites that will be near actual locations of capture. The handling crew will be ferried by the capture helicopter as needed. At these command posts, elk calves will be weighed, measured, collared and released while adult female elk will be captured and released at the point of capture. This will facilitate efforts of the principal investigator to remain in contact with the helicopter net-gunning crew and make all decisions regarding care and welfare of captured elk. Chase Time: The helicopter crew will locate groups of elk and determine if calves and adult females are present. If the group is >20 animals, the helicopter will splinter the group into smaller groups within 1-2 minutes of detecting the initial group. The helicopter will then spend <5 minutes maneuvering a smaller group to a suitable capture site. Once a target animal is selected it will be actively pursued for :::I minute or until active panting is observed at which time the pursuit is terminated. Total time spent disturbing the initial group and target animal should be <10 minutes. No more than 2-3 animals will be taken from an November 2000 Final �238 initial group to avoid unnecessary chase time of non-target animals. Care will be taken by the helicopter crew to avoid chasing animals into fences, roads, rivers, or unfavorable terrain. Animal Care and Handlin2: Elk calves will be hobbled and blindfolded at the point of capture and then slung under the helicopter, one calf per ferry, to a nearby command/handling site where they will be measured, weighed, collared, and released at that site. Capture locales will be within 1-2 minutes of flying time or <3,000m (2 miles) of command/handling sites. Adult females will be blindfolded, hobbled, collared, aged, and then released at the point of capture by the net-gunning crew. Adult females will be assigned to an age class based on relative wear and height of incisors: yearling, 2-4 years, 5-9 years, and >9 years (Quimby and Gaab 1957). At the handling site, 3-4 persons will handle and release calves. Calves will be gently lowered to the ground by the helicopter near the handlers at which time handlers will check calf for injuries, remove netting, and check blindfold and hobbles for proper function. Rectal temperature will then be measured using a digital thermometer (°F) while measurements of total body length (cm), hind foot length (cm) are being obtained. If rectal temperature is 2:41.9°C (107.4°F) and heavy panting evident, the calf will be only collared and released and not weighed to reduce handling time. Previous experience on the Grand Mesa indicated calves survive when rectal temperatures briefly approach 42.2°C (l08°F) (Freddy, unpubl. data). The 2 cases of capture myopathy on the Grand Mesa were males with rectal temperatures of 42.2 and 41.5°C (100 th and 90th quantiles, respectively), ambient air temperatures -2.2 and 3 .9°C (<50 th and 90th quantiles, respectively), and below average body weights <108kg. Assuming acceptable body temperature, the calf will be weighed (kg) by gently sliding the calf into a weigh-bag which will support the entire weight of the calf while the calf is hoisted by a pulley and suspended from a scale affixed to a portable steel quad-pod. Care will be taken to always support the spine and neck of the calf during the weighing process. Once weighed, calves will be lowered to the ground, slid out of the bag, radio-collared, hobbles removed, blindfold removed, and released towards the direction from which they were ferried by the helicopter. Previous experience on the Grand Mesa indicated calves readily find and join elk groups after being released. Total time to process and release calves should be :::8 minutes. If ambient air temperature exceeds 3.3°C (38 °F), capture activities will likely be halted, especially if snow is not present to help cool captured elk. Injured Animals: We expect <3% serious injury/mortality rate. Capture techniques will be constantly monitored and changed if necessary to insure that minor injuries to animals do not chronically occur. However, any debilitating injury or mortality of a captured elk will cause at least temporary suspension of capture activities to assess the cause of injury and if further injuries can be prevented. Animals having a broken leg, neck, pelvis, or other debilitating wound will be euthanized with a gunshot to the head _(0.357 or larger caliber pistol) following euthanasia protocols of the Colorado Division of Wildlife Animal Care and Use Committee. The principal investigator will make decisions regarding euthanasia but all persons involved in capture will be trained to properly euthanize appropriate animals. The helicopter net-gunning crew and ~e handling crew will both have ready access to pistols needed for euthanasia. Euth.anized animals will be processed for human cons~ption ~d donated to social service agencies. Release of Animals: While still blindfolded, hobbled, and prior to release, elk will again be examined for injuries. Superficial injuries such as abrasions and small cuts will be treated with antibiotic ointment. The release sequence will be to place elk in sternal recumbency with head pointed towards direction of capture, remove hobbles, remove blindfold, physically hold elk until elk regains eyesight and orientation, at which time handlers release elk and help elk maintain its balance and upright position. Elk will then be observed for any signs of injury while moving away from handlers. Care will be taken to avoid releasing elk towards fences or unfavorable terrain. Post-Capture Monitorin2: All radioed elk will be monitored for their life/death status 2:2 times within 10 days of capture. If a mortality occurs, the carcass will be located, necropsy performed, and cause of death estimated if possible. If available, muscle tissue samples will be collected and sent to Colorado State University Veterinary Diagnostic Laboratory to detect evidence of capture myopathy. November 2000 Final �71 JOB PROGRESS REPORT Stateof _ _ _ _ _~C~o=lo=r=a=d~o_ _ _ _ __ Work Package -~3C-'0'-"0=2'----------Task No. -----=3_ _ _ _ _ _ _ _ __ Federal Aid Project. Division of Wildlife - Mammals Research Elk Conservation Estimating Calf and Adult Survival Rates and Pregnancy Rates of Gunnison Basin Elk W-153-R-16 Period Covered: July 1, 2002- June 30, 2003 Author: D. J. Freddy Personnel: L. Gepfert, D. Masden, R. Basagoitia, L. Spicer, B. Carochi, J. Oulton, T. Beck, C. Mehaffey, D. Williams, J. Johnston, and R. Kahn of CDOW, Dr. G. C. White Colorado State University, and cooperators/contractors Gunnison Basin Habitat Partnership Program, M. Schuette of MountainScape Imaging, L. Coulter of Coulter Aviation, USFS, BLM, private land owners, and elk hunters. ABSTRACT We used aerial and ground surveys to estimate survival rates and assess sources of mortality for radio-collared adult elk (Cervus elaphpus nelsonii) in the Gunnison Basin of Colorado. Between 15 December 2000 and 14 June 2003, hunting accounted for 94% and 79% of the adult, age ::=:12 months, female and male deaths, respectively, while natural causes were attributed to 6% and 21 % of the adult female and male deaths, respectively. During 3 winter-spring intervals, 15 December - 14 June, natural survival rates for adult females, age ::=:18 months, were::=: 0.98 (n = 39-86 elk, 148-168 elk-winters). During 2 summer-fall intervals, 15 June - 14 December, natural survival rates for adult females, age 2:12 months, were ::=:0.97 (n = 37-86 elk, 98-157 elk-summers). Including hunting mortalities reduced summer-fall female survival to 0.91 ± 0.07 in 2001 (n = 77) and 0.77 ± 0.08 in 2002 (n = 112). During 2 annual intervals, 15 December to next 14 December, natural survival rates for adult females, age::=: 18 months, were ::=:0.97 (n = 33-61). Including hunting mortalities reduced annual female survival to 0.92 ± 0.08 in 2001 (n = 39) and 0.74 ± 0.09 in 2002 (n = 82). Natural survival rates for 2 cohorts of yearlings, age 12-23 months, were 1.00 for females (n = 59) and 0.93 ± 0.08 (n = 43) for males. Including hunting mortalities reduced cohort survival to 0.87 ± 0.08 (n = 68) for females and 0.82 ± 0.11 (n = 49) for males. During summer-fall, natural survival rate for male elk, age 24-29 months, was 1.00 (n = 14) which was reduced to 0.74 ± 0.22 (n = 19) by including hunting mortalities. During winter-spring, natural survival rate for male elk, age 30-35 months, was 1.00 (n = 13). Predation by mountain lions or black bears was suspected in 4 of the 5 adult elk natural deaths. Hunting removal rates for adult females, age ::=:12 months, were 0.08 ± 0.06 (n = 76) in 2001 and lower than the 0.23 ± 0.08 (n = 112) in 2002 (P = 0.006). Removal rates for yearling females, age 12-17 months, averaged 0.13 ± 0.08 (n = 68). Removal rate for yearling • males averaged 0.13 ± 0.10 (n = 48) and for legal branch-antlered males was 0.26 ± 0.22 (n = 19). Wounding loss as a percent of legal harvest was 44 for all adult females and O for branch-antlered males. All hunting deaths of yearling males were illegal harvest/wounding loss while removal rate for branchantlered males was unexpectedly low, likely representing a year effect on elk vulnerability. Apparent differences in survival of adult females between DA Us (P .::S 0.063) likely reflected geographic differences in vulnerability of elk to hunting while differences in male survival between DA Us (P = 0.046) reflected impacts of illegal harvest/wounding loss on removal of yearling males. Adult female elk body condition suggested marginally deficient levels of seasonal nutrition in 2002. �72 Distribution and movements of radio-collared elk during 3 years of monitoring revealed that elk had a relatively high fidelity to the Gunnison Basin as defined by current DAU boundaries but elk also commonly ventured into adjoining GMUs outside the Gunnison Basin. Distribution patterns revealed minimal interchange of elk between areas north and south of U.S. Highway 50 which bisected the Gunnison Basin from east to west. Movements by adult females, young females, and young males (n = 35, 48, and 76) suggested DAU elk population management boundaries might be altered to better represent elk population units. Young male and female elk tended to move greater distances and exhibit higher rates of venturing into adjoining GMUs than adult females. Patterns of dispersion suggested movement corridors that allowed for genetic linkage between Gunnison Basin and other elk populations. All information in this report is preliminary and subject to further evaluation. �Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Project No. Work Package No. Task No. Colorado 3002 3 : : : : Federal Aid Project: N/A : Cost Center 3430 Mammals Research Elk Conservation Estimating Calf and Adult Survival Rates and Pregnancy Rates of Gunnison Basin Elk Period Covered: July 1, 2003- June 30, 2004 Author: D. J. Freddy Personnel: D. Masden, R. Basagoitia, L. Spicer, B. Diamond of CDOW, Dr. G. C. White Colorado State University, and cooperators/contractors Gunnison Basin Habitat Partnership Program, M. Schuette of MountainScape Imaging, private land owners, and elk hunters. All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT During this segment, the transition of monitoring the remaining 119 radio-collared adult elk from research to management biologists was facilitated by providing databases, telemetry equipment, and other guidance as needed. Progress reports were completed, peer-reviewed publications on elk survival rates were initiated, and publications were accepted by peer-reviewed journals. 57 �JOB PROGRESS REPORT ESTIMATING CALF AND ADULT SURVIVAL AND PREGNANCY RATES OF GUNNISON BASIN ELK POPULATIONS DAVID J. FREDDY P. N. OBJECTIVE Estimate survival rates of calf, adult female, and adult male elk and estimate pregnancy rates of adult female elk in Gunnison Basin elk populations for 3 years. NOTE: Prioritization of available research funding resulted in discontinuing efforts to estimate calf survival, pregnancy rates and body condition during 2002-03 but allowed for monitoring adult elk survival through June 2003. SEGMENT OBJECTIVES 1. 2. Facilitate the transition of monitoring the remaining 119 radio-collared adult elk from research to management biologists by providing databases, telemetry equipment, and other guidance as needed. Summarize and analyze data and publish information as Progress Reports, peer-reviewed manuscripts for appropriate scientific journals, or Colorado Division of Wildlife (CDOW) technical publications. SUMMARY Progress reports were completed for the Gunnison Basin elk project (Freddy 2002, Freddy 2003) and can be obtained through the CDOW Research Center library in Fort Collins, Colorado. Publications incorporating calf and adult elk survival rates measured in the Gunnison Basin and Grand Mesa, Colorado were initiated. Two publications were accepted by the Wildlife Society Bulletin for publication during this segment with authors and abstracts provided here for reference. How many mule deer are there? Challenges of credibility in Colorado David J. Freddy, Colorado Division of Wildlife, 317 West Prospect Road, Fort Collins, CO 80526, USA, dave.freddy@state.co.us Gary C. White, Department of Fishery and Wildlife Biology, Colorado State University, Fort Collins, CO 80523, USA Mary C. Kneeland, Colorado Division of Wildlife, 317 West Prospect Road, Fort Collins, CO 80526, USA Richard H. Kahn, Colorado Division of Wildlife, 317 West Prospect Road, Fort Collins, CO 80526, USA James W. Unsworth, Idaho Department of Fish and Game, P.O. Box 25, Boise, ID 83707, USA William J. deVergie, Colorado Division of Wildlife, 2300 South Townsend Avenue, Montrose, CO 81401, USA Van K. Graham, Colorado Division of Wildlife, 711 Independent Avenue, Grand Junction, CO 81505, USA John H. Ellenberger, Colorado Division of Wildlife, 711 Independent Avenue, Grand Junction, CO 81505, USA 58 �Charles H. Wagner, Colorado Division of Wildlife, 222 South Road 1 East, Monte Vista, CO 81144, USA Abstract: Conflict resolution between stakeholder groups and management agencies is a problem in wildlife management. We evaluated our success in resolving a conflict between sportsmen and the Colorado Division of Wildlife (CDOW). Sportsmen challenged the credibility of methods used to estimate numbers of mule deer (Odocoileus hemionus) in Colorado and demanded validating surveys to verify numbers of deer. Sportsmen, other interested wildlife stakeholders, and CDOW engaged in a conflict resolution process and designed and implemented an aerial survey to estimate numbers of deer in a specific population whose previous estimated size had been contested by sportsmen. We used helicopters to count mule deer on randomly selected sample units distributed on deer winter range in March 2001. Estimated population size was 6,782 ± 2,497 (90% CL) using stratified random sample estimators and 11,052 ± 3,503 (90% CL) when counts of deer were adjusted using the Idaho mule deer sightability model. Both aerial survey estimates supported computer modeled population estimates of 7,000 to 7,300 deer that had been contested by sportsmen and all estimates were greater than the sportsmen’s estimate of 1,750 deer which was determined from their casual observations. After the survey, sportsmen did not accept survey estimates despite their involvement in design, analysis, and interpretation of the validation survey. By failing to support results of a validation survey they had demanded, the credibility of sportsmen plummeted among other stakeholders, the Colorado Wildlife Commission, and outside public entities while credibility of CDOW managers rose. We contend that CDOW successfully met challenges of sportsmen because the aerial survey systems used to validate deer numbers were founded on credible science and applied within a resolution process that elicited trust from most stakeholders. We caution other agencies facing similar challenges to use tested methods that can withstand public scrutiny, allow ample time for planning and implementing, carefully assess technical and political risks associated with potential outcomes, and engage multiple stakeholders in planning efforts to gain trust of participants. Cost of this resolution process was about 100,000 $US. Key words: Colorado, conflict resolution, credibility, helicopter surveys, human dimensions, mule deer, Odocoileus hemionus, population estimates, sightability Wildlife Society Bulletin 32 (3):00-00. Effect of limited antlered harvest on mule deer sex and age ratios Chad J. Bishop, Colorado Division of Wildlife, 2300 South Townsend Avenue, Montrose, CO 81401, USA, chad.bishop@state.co.us Gary C. White, Department of Fishery and Wildlife Biology, Colorado State University, Fort Collins, CO 80523, USA David J. Freddy, Colorado Division of Wildlife, 317 West Prospect Road, Fort Collins, CO 80526, USA Bruce E. Watkins, Colorado Division of Wildlife, 2300 South Townsend Avenue, Montrose, CO 81401, USA Abstract: During the 1990s, in response to apparent declining mule deer (Odocoileus hemionus) numbers in Colorado, high buck harvest rates were identified as one of several factors that could be negatively affecting population productivity. Some wildlife managers and sportsmen hypothesized that increasing buck:doe ratios by limiting buck harvest would cause an increase in fawn:doe ratios, and hence, population productivity. We evaluated this hypothesis using data collected by the Colorado Division of Wildlife (CDOW) from 1983 to 1998. Beginning in 1991, CDOW reduced buck harvest in 4 deer management units to provide quality hunting opportunities while maintaining high harvests in other management units. We examined effects of limited harvest on December ratios of bucks:100 does and fawns:100 does using data obtained from helicopter surveys in limited and unlimited harvest units. 59 �Annual buck harvest was reduced by 359 bucks (SE = 133) as a result of limiting licenses in the 4 limited harvest units. Fawn:doe ratios declined by 7.51 fawns:100 does (SE = 2.50), total buck:doe ratios increased by 4.52 bucks:100 does (SE = 1.40), and adult buck:doe ratios increased by 3.37 bucks:100 does (SE = 1.04) in response to limited harvest. Evidence suggested that factors other than buck harvest were regulating population productivity with density dependence being a plausible explanation of declining fawn:doe ratios. Limiting buck harvest to enhance fawn recruitment is not justified in Colorado based on our analysis. Management for limited buck harvest should be largely framed as an issue of quality hunting opportunity rather than an issue of deer productivity. Key words: age ratio, buck:doe ratio, Colorado, fawn:doe ratio, limited harvest, mule deer, Odocoileus hemionus, productivity, quality hunting, sex ratio Wildlife Society Bulletin 33 (0):00-00. LITERATURE CITED Freddy, D.J. 2002. Estimating calf and adult survival rates and pregnancy rates of Gunnison Basin elk. Colorado Division of Wildlife Wildlife Research Report July: 191-222. Fort Collins, Colorado, USA. Freddy, D.J. 2003. Estimating calf and adult survival rates and pregnancy rates of Gunnison Basin elk. Colorado Division of Wildlife Wildlife Research Report July: In Press. Fort Collins, Colorado, USA. Prepared by _______________________________ David J. Freddy, Wildlife Researcher 60 � https://cpw.cvlcollections.org/files/original/dbd8334e4e5c3aa4faeda8663eea5d7d.pdf f07997d4a8bbba4f9cb4900c471bf04c PDF Text Text 259 / '\ Colorado Division of Wildlife Wildlife Research Report July2000 JOB PROGRESS REPORT State of Colorado _ _ _ __,,C=o=lo=rad=o=-------- Cost Center 3430 Project _ _ _ _ _ _ _W ..........-=15__3-'-R=-~1~3_ _ __ Mammals Research Work Package _ _ _ _ _-=-30;:;..;0=2'------ Elk Management Study No. ------~RMNP~~---- Technical Support for Elk and Vegetation Management Environmental Impact Statement for Rocky Mountain National Park Period Covered: July 1, 1999 - June 30, 2000 Author: Dan L. Baker \ ) Personnel: M.A. Wild, T. M. Nett, D. Finley, M. Conner, J.C. Ritchie, M, Conner, L. Wheeler ABSTRACT We conducted experiments to evaluate the effectiveness of a GnRH agonist in preventing pregnancy in captive female elk. LUPRON (luprolide acetate), administered as a subcutaneous implant was 100 % effective in preventing pregnancy in female elk treated before the breeding season. Effective duration of . LUPRON was approximately 193 to 225 days. All treated elk regained fertility following contraceptive treatments. There were no significant differences in body condition, hematology, blood chemistry, general health, or breeding behavior of treat~ and untreated elk. 1il1Di1ii1~i1ri1rn BDOWD16790 ��261 TECHNICAL SUPPORT FOR ELK AND VEGETATION MANAGEMENT ENVIRONMENTAL IMPACT STATEMENT FOR ROCKY MOUNTAIN NATIONAL PARK Dan L. Baker P. N. OBJECTIVES 1. Conduct laboratory experiments needed to reliably implement fertility control alternatives for elk management in Rocky Mountain National Park (RMNP). 2. Develop simulation models to evaluate a range of population control alternatives including fertility control. 3. Prepare a discussion of alternatives for inclusion in the Environmental Impact Statement. SEGMENT OBJECTIVES 1. Prepare scientific publications describing results ofGnRH analog and/or GnRH-toxin conjugate experiments in captive elk. 2. Develop a simulation model to evaluate the feasibility of GnRH analogs and/or GnRH-toxin conjugates to regulate elk populations. 3. Develop strategies to integrate informed local, regional, and national stakeholder input and involvement in elk management alternatives in Rocky Mountain National Park. INTRODUCTION Overabundant wild ungulates can do serious and lasting hann to many plant communities, and preventing such damage may require controlling the growth of their populations (Jewell and Holt 1981, Diamond 1992, McCullough et al. 1997). In Rocky Mountain National Park (RMNP), Colorado, the impact of herbivory by elk has emerged as a fundamentally important issue for those who manage the Park and its wildlife (Hess 1993, Singer et al. 1998). In 1968, RMNP adopted a natural-regulation policy for management of ungulates (Cole 1971, Houston 1971). The objective was to allow density dependent processes to regulate the number of ungulates within park boundaries and to use sport hunting to harvest as many elk as possible in areas surrounding the Park. Recently, however, Park managers have become concerned that possible unnatural concentrations of elk may be altering natural plant communities and ecosystem sustainability. Soil conditions and the status of willow and aspen plant communities have declined. Wet meadow, dry grasssland, and alpine and subalpine sites show evidence of deterioration from overgrazing by elk (Singer et al. 1998, White et al. 1998). As a result of the decline in these vegetation types and the diversity of the animal species that are associated with them, the Park is evaluating alternative management strategies for reducing elk densities within RMNP and the surrounding Estes Valley. Acceptable alternatives for managing elk overabundance in RMNP and adjacent lands are limited. Public hunting within National Parks is proscribed by law and policy and is not permitted without Congressional authorization and an amendment to the enabling legislation for the specific park (Wagner et al. 1995). Authorizing legislation does permit professional culling and RMNP has a long history of animal control �262 actions to hold numbers of elk to a desired level. However, opposition from animal welfare organiu1.tions and other public criticism has prevented implementation of control programs for native ungulates in national parks since 1967 (Wright 1992). When lethal control is deemed necessary, the highest priority is given to encouraging public hunting outside Park boundaries. The success of traditional hunting-based elk management depends on access to private lands. In the Estes Valley, the increase in human occupation and recreational developments during the last decade has brought populations of elk in close proximity to high densities of people. Reduced hunter access, resulting from unethical hunter behavior (Wright et al. 1988), private-land recreation liability laws, attitudes toward hunting (Decker and Gaven 1987), and health and safety concerns has limited the use and effectiveness of sport hunting as an alternative for reducing elk densities in the areas surrounding RMNP. Live capture and translocation of elk is a common management technique that is considered by many to be a more humane alternative to hunting and culling. The few attempts at large scale removals have proven to be costly, inefficient, and stressful or lethal to most of the relocated animals (O'Bryan and McCullough 1985). Two additional problems compromise the potential effectiveness of this strategy for overabundant elk in RMNP. First, there are few suitable habitats in the western United States that are not without healthy and productive elk populations. Secondly, even if release sites were available, the prevalence of chronic wasting disease in the RMNP elk herd (Spraker et al. 1997, Miller et al. 1998) and the potential for spreading this disease to uninfected populations effectively eliminates this option from consideration. Elk Fertility Control Experiments The use of fertility control to decrease birth rates is one of the most promising approaches to the long-term control of overabundant wild ungulates. During the past decade, research aimed at developing effective contraceptives for free-ranging wildlife populations has accelerated. These efforts have resulted in development and testing of a wide variety of potential contraceptive agents (Kirkpatrick and Turner 1985, Warren et al. 1995). One of the most promising new non-steroidal, non-vaccine, approaches to contraception involves synthetic analogs of gonadotropin-releasing hormone (GnRH). GnRH is a molecule produced in the hypothalamus of the brain. It directs specific cells in the pituitary gland to synthesize and secrete two important reproductive hormones; follicle stimulating hormone (FSH) and luteinizing hormone (LH). These latter two hormones, known as gonadotrophs, control the proper functioning of the ovaries in the female and testes in the male. Analogs of GnRH have the potential to either permanently or temporarily inhibit reproduction. For most free-ranging wild ungulate applications, permanent sterilization or a combination of permanent sterilization and culling have been proposed as the most efficacious approaches to population management (Hone 1992, Garrot 1995, Hobbs et al. 1999, in press). For this application, superactive analogs ofGnRH are coupled to a cytotoxin. The GnRH-toxin conjugate specifically targets the gonadotroph cells and permanently inhibits the ability of the cell to secrete FSH and LH. This approach has several potential advantages over other methods of contraception. These include: I) a single treatment should permanently sterilize an animal 2) the same treatment should be effective in both males and females and in different mammalian species 3) GnRH-toxin conjugate will be metabolized from the body within a few days of treatment 4) the proteinaceous nature ofGnRH-toxin conjugate eliminates the possibility of passage through the food chain. 5) the small volume required for effective contraception would facilitate microencapsulation and administration by syringe dart or biodegradable projectiles. �263 In other situations where wildlife managers need to maintain flexibility in the use of fertility control, reversible contraception may be desirable. Examples of these situations include I) wild ungulate populations exposed to periodic, severe, unanticipated winter mortality, 2) populations with low genetic variability, 3) populations that cannot be effectively monitored, 4) populations where public attitudes are opposed to permanent contraception, or 5) populations where non-lethal hunting recreation is a primary management objective. In these situations, superactive analogs of GnRH without the toxin subunit would be more appropriate. The inhibitory actions of long-term GnRH analog agonist on the ovulatory cycle of humans and other mammals is well-established (Casper and Yen 1979, Fraser 1983, Fraser etal. 1987, Concannon et al. 1991). Constant administration of high doses ofGnRH agonist results in down regulation of the pituitary GnRH receptors and suppression of secretion of LH and FSH. Continued treatment suppresses LH • secretion, preventing the maintenance of normal luteal function, and thus prevents viable pregnancy. Inhibition of ovulation caused by chronic administration of GnRH agonist has been successful in several species, including dogs (Vickery et al. 1989), cattle (Herschler and Vickery 1981), sheep (McNeilly and Fraser 1987), white-tailed deer (Becker and Katz 1995), and elk (Baker and Nett, unpublished data). Evidence from studies on pituitary receptors and gonadotropin content in experimental animals treated by long-term infusion of GnRH agonist shows that sustained release is the most effective approach for temporarily suppressing pituitary-gonadal function (Clayton 1982, Sandow 1982). The practicality of this approach, however, is dependent upon development of a long-acting, slow-release preparation of agonist that can be remotely delivered. Recently, a practical mode of administralion using subcutaneous implants has overcome the need for constant mechanical infusion of the analog. Slow release formulations of superactive GnRH agonist are now commercially available and have been shown to be effective in suppressing the pituitary ovarian axis for up to 6 months in a variety of mammalian species (Fraser et al. 1987, Asch et al. 1985). Proposed Research To our knowledge, only limited investigations have been conducted with either of these fertility control techniques on wild ungulates (Becker and Katz 1995), however, the minimum dose ofGnRH analog required for maximum pituitary stimulation is known for elk (Baker et al. 1995) and the minimum effective duration of controlled release GnRH agonist implants has been determined (Baker and Nett 1999, unpublished data). Additional research is needed to determine the effectiveness of these contraceptive agents in preventing pregnancy in elk and the maximum effective duration. Research is also needed to identify any nutritional, physiological or behavioral side-effects that may result from treatment. Thus, the objectives of our research are: 1) To evaluate the effectiveness and effective duration of GnRH-toxin conjugate and GnRH agonist in preventing pregnancy in elk. 2) To evaluate the effects ofGnRH-toxin conjugate and GnRH agonist on nutrition, physiology, general health and social behavior of captive elk. We will test the null hypothesis of no effect of GnRH-P AP and GnRH-Agonist on LH levels, pregnancy rates, social behavior, and general health of female elk. �264 MATERIALS AND METHODS Rocky Mountain elk exhibit highly seasonal patterns of reproduction that are controlled by photoperiod regimens. The onset of the breeding season occurs during the decreasing daily photoperiods of autumn and is preceded by a period of deep anestrous/anovulation in summer (Jopson et al. 1990). The first ovulation of the breeding season is usually preceded by one or more silent ovulations associated with the formation of short-lived corpora lutea that serve to synchronize the first overt estrus within a herd. In temperate North America, the majority of conceptions occur in late September, but recurrent estrous cycles of 21 days are possible through February if females fail to conceive. In early spring, coincidental with increasing day length, reproductive cycles cease and females remain anestrous until August. For pregnant females, parturition generally occurs in early June, after a gestation period of about 255 days (Kelly et al. 1982, Adam et al. 1992). Average conception date for elk at the Foothills Wildlife Research Facility has been estimated to be Sept 30 ± 6 days (Sept 24 - Oct 6). We conducted controlled experiments with 16 adult female and 3 adult male elk at the Colorado Division of Wildlife's Foothills Wildlife Research Facility (FWRF), Fort Collins, Colorado during August 1999 to April 2000. Treatments We compared LH responses, pregnancy rates, social behavior, and general health of the following 4 groups of female elk: Group 1: GnRH-PAP - Pre-conception' - Group 1 elk were treated with an irreversible contraceptive (GnRH-PAP) prior to conception. This was accomplished by treating all cows in this group with 3µ /50 kg BW of GnRH-P AP, IM, one week before being exposed to three adult male elk. Group 2: GnRH-PAP - Pregnant - Group 2 elk were treated with 2µ I 50 kg BW ofGnRH-PAP, IM, during the second trimester of pregnancy (approx. January 15). We confirmed pregnancy of the cows in this group using serum pregnancy-specific protein-B (PSPB) (Williard et al. 1994) prior to treatment. Group 3: GnRH-Agonist (Lupron) - Group 3 elk were treated with a reversible contraceptive (GnRHAgonist - Lupron). We administered a subcutaneous implant containing 32.5 mg ofLupron (Baker and Nett 1999, unpublished data) to female elk one week prior to being exposed to adult male elk. All treatments were applied without tranquilization by moving elk from 5 ha pastures to individual isolation pens, then into a restraining chute where treatments were applied, then relecll>ing animals back into 5 ha pastures. Group 4: Control - Group 4 elk are female elk were untreated and nonpregnant for the duration of the experiment. Four elk were assigned to each treatment group. Based on previous studies with captive elk (Baker et al. 1995, Baker and Nett 1999, unpublished data), approximately four elk per treatment is the minimum sample size needed to provide biologically significant differences among treatment means. Measurements Reprodu_ctive Status. Before assigning animals to treatment groups, we determined the reproductive status of all female elk by monitoring serum progesterone levels during late August and early September. On the day of blood sampling, elk were moved from 5 ha pastures to individual isolation pens, sedated with �265 xylazine hydrochloride (50-200 mg/animal IM), and blood samples collected (5ml). Animals were reversed with yohimbine (0.125 mg/kg, IV) and returned to the paddocks. Sedation of elk was done in order to minimize potential adrenal secretion of exogenous progesterone during handling and blood sample collections (Jopson et al. 1990). Analysis: Serum progesterone concentrations were determined using RIA procedures (Niswender 1973). Sensitivity of the progesterone assay is 0.12 ng/ml. Female elk with progesterone levels above I ng/ml were considered reproductively active (Jopson et al. 1990). Elk with progesterone levels below Ing/ml will be sampled 7 days later. If after three consecutive blood collections, progesterone levels remain lower than I ng/ml the elk were removed from the experiment. Hormonal Assessments Prior to application of contraceptive treatments, we measured the LH response of each elk in treatment Groups I, 3, and 4 to GnRH analog (Baker et al. 1995). Results from this trial provided a pretreatment baseline for comparison to future posttreatment LH responses. This and succeeding LH challenge trials were conducted as follows: On Day 1 of the trial, elk were moved from 5 ha pastures to individual isolation pens, sedated with xylazine hydrochloride (50-200 mg/animal, IM), and fitted nonsurgically with indwelling jugular catheters. Animals were reversed with yohimbine (0.125 mg/kg, IV). On Day 2, we will administer GnRH analog (lµ /50 kg BW) through the cannula and collect blood samples (5 ml) at 0, 60, 120, 180,240, 300, 360, and 480 minutes postinjection. Following the last blood collection, catheters were removed and each animal given Naxel (300 mg, I.V.). Animals were then r¢irned to 5 ha pastures. After collection, blood was held at 4 °C for 24 h until serum is obtained by centrifugation. Serum was be stored at - 20 °C until analyzed for LH. Sen.mi concentrations of LH were quantified by means of ovine LH RIA (Niswender et al. 1969). The duration of contraceptive effectiveness was assessed by conducting GnRH challenge trials each month from September 1999 to May 2000. Analysis. Responsiveness of the pituitary to GnRH challenge was assessed in three ways: I) maximum LH (ng/ml) response achieved postinjection minus baseline, 2) time required to reach maximum LH, and 3) total amount ofLH secreted (ng/ml/min). Pregnancy Rates We assessed contraceptive effectiveness by determining the pregnancy status of all experimental elk. A single blood sample (10 ml) was taken via jugular venipuncture from each animal for PSPB analysis approximately 90 days post-conception (Willard et al. 1994). Animal handling and blood collections for PSPB followed methods previously described for hormonal assessment and were collected in conjunction with these measurements. Female neonates born to any experimental cow will be incorporated into the FWRF elk herd. Male neonates born after August 15 were euthanized according to ACUC procedures. Breeding Behavior The effects of the contraceptive treatments on social behavior in elk is not known. However, if gonadotroph cells are destroyed by GnRH-PAP or desensitized by long-acting synthetic GnRH agonist, down-regulation and diminished LH and FSH can be induced. Reduced secretion of LH and FSH could decrease or disrupt ovarian function, estrus cycles and secondary sex characteristics. We provided insight into these potential effects by monitoring maintenance and sexual behavior of male and female elk during the breeding season. Each animal in each treatment and the breeding males were individually identified using color-coded neck bands or ear tags. We tested the null hypothesis that the frequency of sexual interactions between treated females and males would be similar to that of untreated females and males. �266 We tested this hypothesis using focal- animal sampling procedures and discriminant function analysis (Lehner 1987) {Appendix A). General Health Our knowledge of the effects of these contraceptive treatments on general health, nutrition. body weight dynamics, and blood chemistry of elk is limited. Previous experiments with domestic livestock and companion animals (Nett 1999, unpublished data) and captive mule deer and elk (Baker 1999, unpublished data) have shown no measurable short-term effects. We evaluated these potential side-effects by monitoring body weight, blood chemistry, and hematology of all experimental elk. We collected blood (10 ml) for blood chemistry and hematology prior to treatment with contraceptives, at I month posttreatment and at 6 month intervals thereafter. Statistical Analysis We analyzed data using least squares ANOVA for General Linear Models and the SAS Interactive Matrix Language. Response to contraceptive treatments was analyzed with a two-way factorial analysis of variance for a randomized complete block design with repeated measures structure. Levels of GnRH analog were treatments; individual animals were blocks. Factors in the analysis were treatment and time. Treatment was tested using the animal-within-treatment variance as the error term. Time was treated as a within subject effect using a multivariate approach to repeated measures. We used orthogonal contrast to test for differences among individual means (Morrison 1976, Miller 1966). RESULTS AND DISCUSSION Elk Fertility Control Experiments Pregnancy Rates and Hormonal Responses LUPRON (luprolide acetate), administered as a subcutaneous implant, was 100 % effective in preventing pregnancy in female elk treated prior to the breeding season. The antifertility effects ofLUPRON, a GnRH agonist, were associated with luteolysis and accompanied by a marked loss in ovarian luteinizing hormone and serum progesterone levels. Mean serum progesterone levels were reduced in all female elk from an average pretreatment concentration of 1.4 ng/ml (SE= 0.6) to non-detectable levels at 92 days posttreatment. Serum progesterone levels remained at these levels until approximately 193 days posttreatment (Fig. 1). We observed a similar response in serum LH values. Mean serum concentrations of LH began to decline approximately 30 days following treatment and remained at nondetectable levels for 193 days posttreatment (Fig. 2). LUPRON significantly (P ~0.002) reduced serum LH concentrations in treated elk compared to non-pregnant controls from 30 to 193 days posttreatment. Based on these measurements, anitfertility effects of LUPRON in female elk are estimated to be approximately 193 to 225 days posttreatment. We did not observe negative side-effects of LUPRON on nutrition. physiology, or general health of treated females. Breeding Behavior We collected data for 63 sampling periods: 20 morning periods totaling 45.7 hrs, 6 mid-day periods totaling 13 .5 hrs, 20 evening periods totaling 42.8 hrs, and 17 night periods totaling 32.6 hrs, for an overall total of 134.5 hrs of observation. Time of day, date, and the interactions were not significant effects and were dropped from the ANOVA model. Except for general breeding behaviors, the behavior rate of cows treated with Lupron IM was higher than for control cows or for cows treated with Lupron SQ (Table 1). �267 If Lupron had no treatment effect, then the difference between behavior rates of control and treatment groups will equal z.ero. Differences in behavior rates for cows treated with Lupron SQ and control cows were equal to z.ero for all behavior categories (fable 2). The difference in behavior rates for cows treated with Lupron IM and control cows was not equal to z.ero for male pre-copulatory behaviors and was borderline significant at a 0.10 alpha-level for female pre-copulatory behaviors (fable 2). Table 1. Mean rate of behavior and standard error for control and treatment elk groups; mean and standard error were estimated using type III least-squares. Behavior category Treatment Group Copulatory Control Lupron SQ Lupron IM Male Pre-Copulatory Control Lupron SQ Lupron IM Female Pre-Copulatory Control Lupron SQ Lupron IM General Breeding Control Lupron SQ Lupron IM ~ Mean SE (# behviors/hr') 0.022 0.027 0.045 0.014 0.017 0.020 0.263 0.298 0.453 0.058 0.026 0.080 0.047 0.040 0.271 0.033 0.017 0.139 0.324 0.062 0.330 0.279 0.050 0.045 Table 2. Difference in elk behavior rates (# behaviors/hr) between control and treatment groups, standard error of the difference, and P-value of the difference (the probability that the behavior rates were equal); statistics were estimated using type ID least-squares. Behavior category Treatment Group Copulatory Control - Lupron SQ Control - Lupron IM Lupron SQ - Lupron IM Male Pre-Copulatory Control - Lupron SQ Control- Lupron IM Lupron SQ - Lupron IM Female Pre-Copulatory Control - Lupron SQ Control - Lupron IM Lupron SQ - Lupron IM General Breeding Control - Lupron SQ Control- Lupron IM Lupron SQ - Lupron IM Mean Difference SE P-Value -0.005 -0.023 -0.018 0.023 0.024 0.027 0.829 0.346 0.503 -0.035 -0.190 -0.155 0.072 0.099 0.086 0.624 0.054 0.073 0.007 -0.224 -0.231 0.037 0.151 0.148 0.849 0.138 0.119 -0.006 0.045 0.051 0.084 0.080 0.069 0.943 0.574 0.463 �268 -e-- Lupron SC - 0- - Non-Pregnant Control 30 25 20 I \ \ \ \ ::c ....J ~ eel \ \ 15 \ Q) // \ C / \..,..,,,/ ~ eel ~ / \ a.. Q) / \ 10 5 0 Pretrmt 30 92 135 165 193 225 Days Posttreatment Figure 1. Mean serum luteinizing hormone (LH) concentrations of female elk before and after treatment with a subcutaneous implant containing 32.5 mg LUPRON. For all behavior categories, behavior rates were almost identical for the control and Lupron SQ groups. If cows cease cycling once bred, then this similarity suggests that the Lupron SQ group did not cycle or cycled once. In contrast, for all behavior categories except general breeding, behavior rates for the Lupron IM group were 1.5-6.8 times higher than for the control and Lupron SQ groups. The Lupron IM cows appeared to cycle into estrous more than one time, which resulted in higher reproductive behavior rates. There was no statistically significant difference in copulatory behavior rates between control and treatment elk. However the rate of copulatory behavior was about 2 times higher for the Lupron IM group than for the control and Lupron SQ groups. Biologically, this suggests that the Lupron IM group may have cycled and bred more than one time during the experiment. The rate of male pre-copulatory behavior was significantly higher toward the Lupron IM group than toward the control and Lupron SQ treatment groups. Male pre-copulatory behavior rates were 1.5-1. 7 times higher toward the Lupron IM group. Although the rate of female pre-copulatory behavior was 5.8-6.8 times higher for the Lupron IM group compared to the control and Lupron SQ groups, this difference was not significant. The lack of significance probably stems from temporal variations in female pre-copulatory behavior. When not in �269 --- - -0-- - Lupron SC Control 3.00 2.50 ::::::- --E 2.00 0) C -.:;f- a.. 1.50 C co (l) ~ 1.00 l G----.1- 0.50 0.00 Pretrmt 30 92 135 165 193 225 Days Posttreatment Figure 2. Mean serum progesterone (P41 concentrations of female elk before and after treatment with a subcutaneous implant containing 32.5 mg ofLUPRON. estrous, female cows showed no interest in the bulls, resulting in many sampling periods where the rate of behavior was zero. Then, when a cow came into estrous, or cycled in some way, she would spend much of her time soliciting interest from the bull, which resulted in a high rate of female pre-copulatory behavior for that sampling period. Thus, female pre-copulatory behavior rates were spiked through time, resulting in high variation and lack of statistical significance between treatment groups. Although not statistically significant, biologically, the Lupron IM group appeared to continue some sort of cycle associated with estrous. Rates of general breeding behavior were similar for all 3 treatment groups. The general breeding behaviors, approach and herding, are harem gathering and maintenance behaviors. In the wild, elk gather cows, regardless ofwhether they are in estrous at that time. Th.is may explain why the bull did not distinguish between treatment groups; he simply grouped all the cows and kept them from the other 2 bulls in the pen. Elk Fertility Control Mode/ Development of a simulation model representing dynamics of the RMNP elk herd and evaluation of the benefits and liabilities of a range of population control alternatives was not accomplished during this segment and is contingent on data not presently available from RMNP. Th.is objective will be completed during FY 2000-1. �270 Stakeholder Input on Elk Management Alternatives lb.is segment objective was not addressed during FY 1999-0 due to a delay in the initiation of the formal Environmental hnpact Statement process. LITERATORE CITED Adam, C. L., C. E. Moir, and T. Atkinson. 1985. Plasma concentrations of progesterone in female red deer (Cervus elaphus) during the breeding season, pregnancy, and anoestrus. J. Reprod. Fertil. 74:631-636. Asher, G. W. 1985. Oestrous cycle and breeding season of farmed fallow deer, Dama dama. J. Reprod. Fertil. 75:521-529. Asch, R.H., F. J. Rojas, T. R. Tice, and A. V. Schally. 1985. Studies of a controlled - release microcapsule formulation of an LH-RH agonist in the rhesus monkey menstrual cycle. Int. J. Fert. 30:19-26. Baker, D. L., M. W. Miller, and T. M. Nett. 1995. Gonadotropin-releasing hormone analog-induced patterns ofluteinizing hormone secretion in female wapiti (Cervus elaphus nelsoni) during the breeding season, anestrus, and pregnancy. Biol. of Reprod. 52: 1193-1197. Becker, S. E., Katz, L. S. 1995. Effects of gonadotropin - releasing hormone agonist on serum LH concentrations in female white-tailed deer. Small Ruminant Research 18:145-150. Casper, R. F., and S.S. C. Yen. 1979. Induction ofluteolysis in the human with a long-acting analog of luteinizing hormone-releasing factor. Science 205:408-410. Clayton, R. N. 1982. GnRH modulatidn of its own pituitary receptors: evidence ofbiphasic regulation. Endocrinology 111: 152-161. Cole, G. F. 1971. An ecological rationale for the natural or artificial regulation of native ungulates in parks. Transactions of the North American Wildlife and Natural Resources Conference 36:417426. Concannon, P. W., and V. N. Meyers-Wallen. 1991. Current and proposed methods for contraception and termination of pregnancy in dogs and cats. J. Am. Vet. Med. Assoc. 198:1214-1225. Curlewis, J. D., A. S. I. Loudon, and A. P. M. Coleman. 1988. Oestrous cycles and the breeding season of the Pere David's deer hind (Elaphurus davidianus). J. Reprod. Fert. 82: 119-126. Decker, D. J., and Gavin. 1987. Public attitudes toward a suburban deer herd. Wildlife Society Bulletin 16:53-75. Diamond, J. 1992. Must we shoot deer to save nature? Natural History August: 2-8. Fraser, H. M. 1983. Effect of treatment for one year with a luteinizing hormone-releasing hormone agonist on ovarian, thyroidal, and adrenal function and menstruation in the stumptailed monkey (Macaca arctoides). Endocrinology 112:245-253. - - - ~ M., J. Sandow, H. Seidel, and W. von Rechenberg. 1987. An implant ofa gonadotropin releasing hormone agonist (burserelin) which suppresses ovarian :function in the macaque for 3-5 months. Acta Endocrinologica 115: 521-427. Garrot, R. A. 1995. Effective management of free-ranging ungulate populations using contraception. Wildlife Society Bulletin. 23:445-452. Guinness, F. E., Lincoln, G. A., and Short, R. V. 1971. The reproductive cycle of the female red deer, Cervus elaphus L. J. Reprod. Fertil. 27: 427-438. Hess, K. 1993. Rocky times in Rocky Mountain National Park. University_ Press of Colorado, Ni wot, Colorado. Herschler, R. C., and B. H. Vickery. 1981. The effects ofLHRH ethylamide on the estrous cycle, weight gain, and feed efficiency in feedlot heifer. Arner. J. of Vet. Res. 42:1405-1408. �271 Hobbs, N. T., D. L. Baker, and RB. Gill. 1999. A general theory describing effects of fertility control on populations of ungulates. Journal of Wildlife Management (in press). Hone, J. 1992. Rate of increase and fertility control. J. Applied Ecology 29:695-698. Houston, D. B. 1971. Ecosystems of national parks. Science 172:648-65 l. _ _ _. 1982. The Northern Yellowstone Elk: Ecology and Management. MacMillan Publishing Company, New York, New York. Jewell, P. A., and S. Holt. 1981. Problems in management of locally abundant wild animals. Academic Press. 32lpp. Jopson, N. B., M. W. Fisher, and J.M. Suttie. 1990. Plasma progesterone concentrations in cycling and ovariectomiz.ed red deer hinds: the effect of progesterone supplementation and adrenal stimulation. Animal Reprod. Sci. 23:61-73. Kelly, R W., K. P. McNatty, G. H. Moore, D. Ross, and M. Gibb. 1982 . .Plasma concentrations ofLH, prolactin, oestradiol and progesterone in female red deer during pregnancy. Journal of Reproduction and Fertility 64:475-483. Kirkpatrick, J. F., and J. W. Turner, Jr. 1985. Chemical fertility control and wildlife management. Bioscience 35:485-491. Lehner, P. N. 1987. Design and execution of animal behavior research: an overview. Journal of Animal Science 65: 1213-1219. McCullough, D. R, K. W. Jennings, N. B. Gates, B. G. Elliot, and J.E. Didonato. 1997. Overabundant deer populations in California. Wildlife Society Bulletin 25:478-483. McNeilly, A. S., and H. M. Fraser. 1987. Effect of gonadotropin-releasing hormone agonist-induced suppression of LH and FSH on follicle growth and corpus luteum function in the ewe. J. Endocrinology 115:273-282. Miller, M. W., M.A. Wild, and E. S. Williams. 1998. Epidemiology of chronic wasting disease in captive rocky mountain elk. Journal of Wildlife Diseases 34:532-538. Miller, R.G. 1966. Simultaneous statistical inference. New York: McGraw-Hill Book Co. 152-168. Morrison, J. A. 1960. Characteristics of estrus in captive elk. Behavior 16:84-92. Morrison, D. F. 1976. Multivariate statistical methods. New York: McGraw-Hill Book Co. 145- 194. Niswender, G.D., L. E. Reichert, Jr., A. R Midgley, Jr., and A. V. Nalbandov. 1969. Radioimmunoassay for bovine and ovine luteinizing hormone. Endocrinology 84: 1166-1173. ____. 1973. Influence of the site of conjugation on the specificity of antibodies to progesterone. Steroids 22:413-424. _ O'Bryan, M. K., and D.R. McCullough. 1985. Survival of black-tailed deer following relocation in California. Journal of Wildlife Management 49: 115-119. Sandow, J. 1982. Inhibition of pituitary and testicular function by LHRH analogs. Pages 19-39 In: Jeffcoate S. L. and Sandlier (eds). Progress towards a male contraceptive. Wiley and Sons, Chichester. Singer, F. J., L. C. Zeigenfuss, R G. Cates, and D. T. Barnett. 1998. Elk, multiple factors, and persistence of willows in national parks. Wildlife Society Bulletin 26:419-428. Spraker, T. R., M. W. Miller, E. S. Williams, D. M. Get.zy, W. J. Adrian, G. G. Schoonveld, R A. Spowart, K. I. O'Rourke, J.M. Miller, and P.A. Merz. 1997. Spongiform encephalopathy in free-ranging mule deer (Odocoileus hemionus), white-tailed deer (Odocoileus virginianus), and Rocky Mountain elk (Cervus elaphus nelsoni) in northcentral Colorado. Journal of Wildlife Diseases 33:1- 6. Vale, W., C. River, M. Brown, J. Leppaluoto, N. Ling, M. Monaham, and J. River. 1976. Pharmacology of hypothalamic peptides. Clin. Endocrin. 5(suppl): 261-273. Wagner, F. H., R. Foresta, R. B. Gill, D.R. McMullough, M. P. Pelton, W. F. Porter, and H. Salwasser. 1995. Wildlife policies in the U. S. National Parks. Island Press. Washington, D. C. 242 pp. �272 Warren, R. J., L. M. White, W.R. Lance. 1995. Management of urban deer populations with contraceptives: practicality and agency concerns. Wildlife Society Bulletin 23: 441-444. White, C. A, C. E. Olmsted, and C. E. Kay. 1998. Aspen, elk, and fire in the Rocky Mountain national parks of North America. Wildlife Society Bulletin 26:449-462. Willard, S. T., R G. Sasser, J. C. Gillespie, J. T. Jaques, T. H. Welsh, Jr., and R. D. Randel. 1994. Methods for pregnancy determination and the effects of body condition on pregnancy status in Rocky Mountain elk. Theriogenology 42:1095-1102. Wright, R G. 1988. Wildlife issues in national parks. Pp 168-196 in W. J. Chandler (ed) Audubon Wildlife Report, National Audubon Society. _ _ _ _. 1992. Wildlife management and research in the national parks. University of Illinois Press. 224 pp. Prepared by _ _ _ _ _ __ Dan L. Baker Research Biologist �169 Colorado Division of Wildlife Wildlife Research Report July 2001 and July 2002 JOB PROGRESS REPORT State of --------"="-==-----Colorado Division of Wildlife - Mammals Research Work Package No. _____3__0___0=-2_ _ _ _ __ Elk Management Study No. ----~RMN='"'"p~----- Technical Support for Elk and Vegetation Management for Rocky Mountain National Park - Environmental Impact Statement Period Covered: July 1, 1999 -June 30, 2002 Author: Dan L. Baker, Ph.D. Personnel: M. Wild, T. Nett, D. Finley, M. Conner, J. Ritchie, L. Wheeler, E. Jones, D. Hussain ABSTRACT Fertility control offers a potential alternative to traditional methods for regulating the growth of overabundant wild ungulate populations. However, current technology is limited due to practical treatment application, undesirable side-effects, and economic considerations. A promising non-steroidal, non-immunological approach to contraception involves potent GnRH agonist. During 1999-2002, we conducted a series of experiments to evaluate the effectiveness of a GnRH agonist (leuprolide) as a contraceptive agent in captive female elk. In experiment 1, we determined the optimum dose of GnRH agonist treatment by measuring serum luteinizing hormone (LH) and progesterone (P4 ) response of female elk to 4 formulations of leuprolide administered as subcutaneous bioimplants. In experiment 2, we evaluated the effects ofleuprolide on elk pregnancy rates, duration of suppression ofLH and P4 secretion, and short-term behavioral and physiological side-effects. In experiment 3, we evaluated the effects ofleuprolide on pregnant elk, and in experiment 4, assessed the potential for delivering leuprolide remotely in a syringe dart. All concentrations ofleuprolide were equally effective in reducing serum LH and P4 to non-detectable levels for the duration of the 130 day experiment. Leuprolide administered prior to the breeding season was 100% effective in preventing pregnancy in treated females. Serum LH and P4 concentrations were reduced to baseline levels by day 92 and remained at these levels for 195-251 days posttreatment with a return to pretreatment concentrations the following breeding season. Reproductive behavior rates were similar for treated and untreated elk for all behavior categories for both the breeding and postbreeding seasons. Hematology and blood chemistry parameters of treated and untreated females were similar and seasonal intake and body weight dynamics appeared normal. Initial results indicate that leuprolide can be effectively delivered in a syringe dart but additional research is needed to confirm these observations. Thus, we conclude that leuprolide is a safe, effective contraceptive agent and has the potential for suppressing fertility in female wapiti for one breeding season. ��171 TECHNICAL SUPPORT FOR ELK AND VEGETATION MANAGEMENT FOR ROCKY MOUNTAIN NATIONAL PART ENVIRONMENTAL IMPACT STATEMENT Dan L. Baker P. N. OBJECTIVE Conduct captive elk experiments to implement fertility control as an alternative for managing elk in Rocky Mountain National Park. SEGMENT OBJECTIVES 1. Develop and test a reversible contraceptive agent for free-ranging elk. 2. Determine the duration of effectiveness of a selected contraceptive agent in captive elk. 3. Assess contraceptive effects on pregnancy, and behavioral and physiological side-effects in captive elk. 4. Develop and test a remote delivery system for administering the contraceptive agent to free-ranging elk. INTRODUCTION Overabundant wild ungulate populations have become a significant problem for natural resource managers in North America. Unregulated populations can cause adverse effects that are ecological, economic, or political in scope and resolving these issues often requires controlling animal abundance (Jewell and Holt 1981, Garrott et al. 1993, McCullough et al.1997, Smith 2001). In Rocky Mountain National Park (RMNP), Colorado, the impact ofherbivory by elk has emerged as a fundamentally important problem for those who manage the Park and its wildlife (Hess 1993, Zeignefuss et al. 1996). In 1968, RMNP adopted a natural-regulation policy for management of ungulates (Cole 1971, Houston 1971) with the objective of allowing density dependent processes to regulate elk numbers within park boundaries and use sport hunting to harvest as many animals as possible in areas surrounding the Park. Recently, however, Park managers have become concerned that possible unnatural concentrations of elk may be altering natural plant communities and ecosystem sustainability. Soil conditions and the status of willow and aspen plant communities have declined. Wet meadow, dry grasssland, and alpine and subalpine sites show evidence of deterioration from overgrazing by elk (Singer et al. 1998, White et al. 1998). As a result of the decline in these vegetation types and the diversity of the animal species that are associated with them, the Park and other natural resource agencies are evaluating alternative management strategies for reducing elk densities within RMNP and the surrounding Estes Valley. One alternative being considered is controlling the fertility of female elk. Fertility control has been widely advocated as an alternative to lethal methods of population control for wildlife and considerable research has been directed toward development of different contraceptive agents (see reviews by Kirkpatrick and Turner 1985, Fagerstone et al. 2001). Field and laboratory studies have evaluated the efficacy of delivery of contraceptives to ungulates (Jacobsen et al. 1995, DeNicola et al. 1997, �172 Kirkpatrick et al. 1997) and models have been developed to represent effects of fertility control on the population dynamics of individual species and populations (Garrott and Siniff 1992, Seagle and Close 1996, Hobbs et al. 2000). To date, most contraceptive research for wild ungulates has focused on the development of immunocontraceptive vaccines and steroidal hormonal agents. However, after more than 40 years of research, the success of these approaches have been primarily limited to captive wildlife and small localized urban populations of wild ungulates. To meet this challenge, new technologies and approaches are needed if fertility control is to become practical and acceptable management tool for controlling overabundant wildlife species. A promising new non-steroidal, non-immunological approach to contraception involves potent analogs of gonadotropin-releasing hormone (GnRH). GnRH is a molecule produced in the hypothalamus of the brain. It directs specific cells in the pituitary gland to synthesize and secrete two important reproductive hormones; follicle stimulating hormone (FSH) and luteinizing hormone (LH). These latter two hormones, known as gonadotrophs, control the proper functioning of the ovaries in the female and testes in the male. Chronic treatment with continuous, high doses of GnRH agonists results in temporary suppression of pituitary responsiveness and gonadotropin secretion. Resulting decreases in plasma LH and FSH in females leads to suppression of ovulation, estrus cyclicity, and gonadal steroidogenesis (Belchetz et al.1978, Evans and Rawlings 1994). Once GnRH agonist treatments are terminated, normal pituitary function is gradually restored (Bergfeld et al. 1996). GnRH agonists have been shown to inhibit ovulation in several domestic ungulate species including sheep (McNeilly and Fraser 1987), cattle ( D'Occhio et al. 1996, D'Occhio and Aspden 1999), and horses (Montovan et al. 1990). However, studies on wild ungulates are limited (Becker and Katz, 1995; Brown et al. 1999) and none have demonstrated their effectiveness as a contraceptive agent. GnRH agonists provide a potential biotechnology for achieving a controlled, reversible suppression of fertility in both captive and free-ranging female wild ungulates. However, their practicality as a contraceptive agent is dependent on effective inhibition of reproduction without negative behavioral or physiological side-effects. During 1999-2002, we conducted a series of experiments with sustained release formulations of GnRH agonist in captive female elk to evaluate these factors. Specifically, our objectives were: (1) to evaluate the effectiveness of GnRH agonist in preventing pregnancy, (2) to determine the duration of GnRH agonist suppression ofLH and progesterone (P 4) secretion, (3) to assess the behavioral and physiological side-effects (if any) of GnRH agonist treatments, and (4) __to develop a remote ~elivery system for administering the contraceptive agent to free-ranging animals. MATERIALS AND METHODS A. Experiment 1: Dose response 1. Objective Determine the minimum effective dose of GnRH agonist (leuprolide) that will induce halfmaximal release of luteinizing hormone in female elk during estrus and evaluate duration of effectiveness. �173 2. Methods We determined the optimum dose ofleuprolide (desGly1°-D-Leu6 -LH-RH ethylamide acetate) required for suppression of serum LH secretion in 8 female elk (6-12 years of age; 240-300 kg). Females were monitored for occurrence of oestrus cycles by measuring serum progesterone concentrations at weekly intervals beginning l November 1998 and were considered reproductively active when concentrations were greater than 1 ng m1· 1 for two consecutive sampling periods (Adam et al. 1985). Females were randomly selected to receive one of four doses (0, 45, 90, 180 mg leuprolide acetate) of 90 day sustained release leuprolide formulation using the ATRIGEL®drug delivery system (Atrix Laboratories, Inc. Ft. Collins, CO, USA) (Dunn et al.1994). These formulations at lower doses have demonstrated a sustained release and activity in rats and dogs for a period of at least 90-120 days (Ravivarapu et al. 2000). On the day before treatment application, animals were moved from paddocks, weighed(± 0.5 kg), moved to individual isolation pens (5 xl0 m), sedated with xylazine hydrochloride (Rompun; Bayer AG, Leverkusen, Germany; 25-200 mg animat1 i.m.) and fitted nonsurgically with indwelling jugular catheters. Sedation was reversed with yohimbine (30 mg) (Antagonil®, Wildlife Laboratories, Fort Collins, CO, USA). The sampling period began the next day (20 November 1998) at 0900. A patch of hair (3 cm in diameter) was shaved in the shoulder region of each female (controls did not receive a placebo formulation) and leuprolide formulations were injected under the skin using an 18 gauge needle and 3 cc syringe. Blood samples (5 ml) were collected at 0, 60, 120, 180, 240, 300,360,480 min, then at 12, 24, 36, 48, 84, and 240 h postinjection. Catheters were flushed daily with sterile saline solution. Following the last blood collection, catheters were removed and animals returned to 5 ha paddocks. We compared the effective duration ofleuprolide treatments by measuring pituitary responsiveness to an exogenous dose ofGnRH analogue (D-Ala6-GnRH-Pro9-ethylamide; Sigma Chemical Co., St. Louis, MO) administered at 35, 70, 110, and 130 days posttreatment. Animal handling and blood sampling protocols were similar to those previously described. We administered a previously determined dose (Baker et al. 1995) of GnRH analog (1 µg 50 kgBW· 1) through the jugular cannula and collected blood samples at 0, 30, 60, 90, 120, 180,240, 300, 360, 420, and 480 min postinjection. After collection, blood was held at 4 ° C for 24 h until serum was obtained by centrifugation. Serum was then stored at - 20° C until analyzed for LH. B. Experiment 2: Antifertility and behavioral effects on nonpregnant female elk 1. Objectives a. Determine the effectiveness of GnRH agonist in preventing pregnancy in female elk b. Determine the duration of GnRH agonist suppression of LH and P4 secretion c. Evaluate the behavioral and physiological side-effects (if any) of GnRH agonist treatments. 1. Methods We evaluated the effects of the optimum dose of leuprolide formulation established in Experiment 1, on elk pregnancy rates, duration of suppression ofLH and P4 secretion, blood chemistry, and reproductive behavior during 2 November 1999 to 15 May 2000. Fourteen adult female (7-13 years of age; 240-320 kg) and 3 adult male elk (4-13 years of age; 375- 400 kg) �174 were used in this experiment. Females were assigned to one of 3 experimental groups based on their tractability for handling and blood sampling. Four elk cows (Group A) were treated with 32.5 mg of leuprolide and 5 cows (Group B) served as untreated controls and were used to compare pregnancy rates, blood chemistry, and reproductive behavior to those of treated females. These two groups of females were maintained together with 3 adult male elk in adjoining paddocks (2-ha each). The remaining 4 females (Group C) served as untreated, non-pregnant controls and were placed in a separate pasture (1 ha) without direct contact with male elk. We compared LH and progesterone secretion of these females to those treated with leuprolide (Group A). a. Pregnancy rates, hormonal measurements, and blood parameters. We determined the effects ofleuprolide on pregnancy rates of treated and untreated elk by measuring pregnancy-specific protein B (PSPB)(BioTracking, Moscow, Idaho, USA) in serum at about 70, 160, and 215 days of gestation (Huang et al. 2000). We compared the effects of leuprolide on e)l..ient and duration of LH and P 4 suppression in treated and untreated, nonpregnant elk during 2 November 1999 to 11 November 2000. GnRH challenge trials were conducted prior to application ofleuprolide treatments and at 30, 90, 145, 180,225,250, and 373 days posttreatment. The final GnRH challenge trial was conducted to assess reversibility of treatment. Protocol for GnRH challenge trials followed procedures previously described in Experiment 1. We assessed physiological side-effects ofleuprolide by comparing serum chemistry, hematology, and body weight dynamics of treated (Group A) and untreated, non-pregnant elk (Group C). Blood collections and body weight measurement were made in conjunction with GnRH challenge trials. Blood samples for hematology and serum chemistry analysis were collected at 90 days posttreatment then submitted for analysis to Colorado State University, Veterinary Teaching Hospital, Clinical Pathology Laboratory, Fort Collins, Colorado, USA. Serum chemistry profiles were obtained using a Hatachi 917 autoanalyzer (Roche/Boehringer Mannheim, Indianapolis, Indiana, USA) for the following parameters: glucose, creatinine, phosphorus, calcium, magnesium, total protein, albumin, globulin, albumin/globulin ratio, bilirubin, creatinine kinase, aspartate aminotransferase, gamma-glutamyltransferase, sorbitol dehydrogenase, sodium, potassium, chloride, and biocarbonate. Values for the following hematological parameters were obtained using an ADVIA 120 autoanalyzer (Bayer Corporation, Tarrytown, New York, USA): nucleated cells, neutrophils, lymphocytes, monocytes, eosinophils, plasma protein, erythrocyte, hemoglobin, packed cell volume, mean corpuscular volume, mean corpuscular hemoglobin concentration, platelets, and fibrinogen. b. Reproductive behavior. The effectiveness of the leuprolide formulation as a contraceptive agent is dependent upon suppression of ovulation and steroidogenesis for the duration of the breeding season. Thus, we tested 2 hypotheses relative to the effects of leuprolide on reproductive behavior of elk: (I) because leuprolide was expected to suppress gonadotrophin secretion and ovulation, we predicted that sexual interactions during the breeding season would be reduced in leuprolide treated females ( Group A) compared to untreated controls (Group B), and (2) since depletion of the leuprolide implant (90 days) was expected prior to anoestrus (late March), we predicted that behavioral oestrus would resume in treated females ( Group A) and the rate of sexual interactions would be higher than that for untreated controls (Group B) �175 To test these hypotheses, we examined the effects of leuprolide on reproductive interactions of male and female elk during 2 time periods; breeding season (defined as the period 10 November- 23 December 1999) and postbreeding season (defined as the period 7 February 27 March 2000). On 2 November 1999, female elk in Group A were treated with leuprolide and released with untreated controls (Group B) into adjoining paddocks (2 ha each). Seven days later (10 November), we placed 3 adult male elk with these groups and initiated behavioral observations. All females were individually identified with color/numeric-coded neck collars. Animals selected as treatments and controls were unknown to observers. Behavioral measurements were made from a distance of 50-250 m from an elevated tower (10 m) situated between adjacent pastures using binoculars and a spotting scope during the day, and a spotlight and night vision scope at night. We recorded selected behaviors using a laptop computer with a behavioral software program. We used focal animal sampling procedures to sample reproductive behaviors of all experimental animals over a 24 -hour period (Lehner 1996). Preliminary observations indicated that elk were most active in morning (0500-0800), late day ( 1400-1700) and night (2000-2400). Thus, time-of-day sampling periods were randomly assigned each week using a randomized block design. Each sampling period consisted of at least two hours of continuous observations. Based on previously reported elk breeding behavior (Morrison et al. 1960, Geist 1982, Rapley 1985), we identified and recorded 19 sexual interactions. Because sample sizes were small, we grouped individual behaviors into 4 general categories: male copulatory, male precopulatory, female precopulatory, and general breeding (Table 1). Copulatory, male precopulatory, and general breeding were interactions of a male directed toward a specific female, while female precopulatory behaviors were actions of a specific female towards a male. Thus, our experimental unit for analyses was the individual female in each breeding group. Behavioral interactions were generally short duration (<30 sec) relative to sampling interval, therefore we recorded the number of occurrences of each event rather than length of time and calculated sexual interaction rates as acts per animal per hour, then multiplied hourly behavioral rates by 24 for a daily rate. Table 1. Description of elk reproductive behavior and associated behavior categories. Behavior category Reproductive behavior General breeding : Male directed behavior related to establishing, maintaining, and defending a group or harem of female elk (e.g. herding guarding, tending) Male precopulatory Male courtship behavior directed toward an individual female to induce or detect oestrus or ovulation (e.g. urine testing, flehmen, tongue flick, lick, smell, or rub female's body, chivy) Female precopulatory Female courtship behavior directed toward dominant male to arouse copulatory behavior (e.g. lick and rub male, mount, lordosis, twitch hocks) Copulatory Male behavior directed toward a receptive female in oestrus (e.g. precopulatory mounts, intromission, pelvic thrust) �176 c. Hormone radioimmunoassay. Serum concentrations ofLH were quantified by means of an ovine (o) LH RIA (Niswender et al. 1969). Elk serum was demonstrated to inhibit binding of 125 l-oLH to LH antiserum in a parallel manner. Likewise, when varying quantities of oLH standard (NIH-OLH-S24) were added to elk serum and samples were subjected to RIA, the values obtained were increased by the quantity of oLH added (r2 = 0.99, slope = 0.92, SEb = 0.22, P = 0.002). These data indicate that the radioimmunoassay (RIA) provided a quantitative assessment ofLH in elk serum. The limit of sensitivity of the LH assay was 0.4 ng mi- 1 . Serum concentrations of progesterone were determined by RIA (Niswender 1973). Sensitivity of the progesterone assay was 0.12 ng m1· 1 . Intra- and inter-assay coefficients of variation for each of these assays were less than 10%. d. Statistical analysis. Hormone concentrations are reported as untransformed arithmetic means ± standard error of the mean (SE). Responsiveness of the pituitary to GnRH analog challenge was assessed in two ways: 1) maximum response (highest concentration of LH (ng mJ- 1) achieved postinjection minus baseline), and 2) total amount ofLH secreted (ng mJ- 1 min•1) estimated by calculating the area under the LH response curve (Abramowitz and Stegun 1968). We analyzed differences among hormone levels using least squares analysis of variance for general linear models (SAS Institute 1993). Responses to treatments were analyzed with one-way analysis of variance for a randomized complete block design with repeated measures structure. Levels of leuprolide formulations were treatments; individual animals were blocks. Factors in the analysis were dose and time. Treatment effects were tested using the animal-within-treatment variance as the error term. Time was treated as a within-subject effect using a multivariate approach to repeated measures (Morrison 1976). A "protected" least significant difference test (Milliken and Johnson 1984) was used to separate means when the overall F-test indicated significant treatment effects (P < 0.05). We tested specific reproductive behavior hypotheses that mean behavior rate was not different between treatment and control groups for both the breeding and postbreeding seasons using an ANOVA model with a repeated measures structure. Similar to the hormonal analysis, time was treated as a within subject effect using multivariate approach to repeated measures (Morrison 1976). To test for treatment effects, we accounted for time-of-day, date effects and their interactions. PROC GENMOD (SAS Institute 1993) was used to estimate and test for differences in mean behavior rate by treatment, time-ofday, and date. Means and standard errors were estimated using least squares, and hypothesis tests were based on type III generalized ~stimating equations that accounted for correlation in repeated measuremep.ts. D. Experiment 3: Antifertility effects on pregnant elk 1. Objectives a. Evaluate the effects of GnRH agonist (leuprolide)on female elk treated during the first trimester of pregnancy. b. Assess nutritional, physiological, or behavioral side-effects that might result from treatment. �177 2. Methods a. Animals. We conducted controlled experiments with 12 adult female elk and 2 adult male elk at the Colorado Division of Wildlife's Foothills Wildlife Research Facility (FWRF), Fort Collins, Colorado during September 1, 2000 to December 15, 2001. On September 1, 2000, two intact male elk were released with 12 female elk. The purpose of mating was to confirm the reversibility ofleuprolide from previous experimental treatments (Baker et al. 2002, in press). Pregnant animals from this mating would then provide the experimental elk for the experiment described in this study plan. b. Treatments. Approximately 60 days postconception ( 1 December for cows at FWRF), all females were evaluated for pregnancy and fetal age determined using transrectal ultrasonography (Willard et al. 1994). Using ultrasound and selected measurements reported for known-age embryos (Morrison et al. 1959), we estimated fetal ages of all pregnant elk. Eight elk with embryos estimated to be 60-75 days old were randomly selected to receive a subcutaneous implant containing 32.5 mg ofleuprolide formulation. Leuprolide was injected subcutaneously on the lateral thorax using an 18 g x 4 cm needle. The remaining four pregnant elk were designated as untreated controls. Treatment and control elk were maintained in the same pastures, fed similar diets, and handled similarly throughout the study. All treatments were applied without tranquilization by moving elk from 5 ha pastures to individual isolation pens, then into a restraining chute, where treatments were applied, then returning elk back into 5 ha paddocks. Animals were observed daily by trained caretakers for general health and for signs of abortion or parturition. c. Sample size. Based on previous reproduction studies with captive elk at FWRF, 4-6 elk per treatment is the minimum sample size needed to provide biologically significant differences among treatment means ( Baker et al. 1995, Baker et al. 2002, in press). We used an unbalanced experimental design to minimize the number of untreated pregnant control elk, since most neonates will be euthanized. Pregnant, control elk were needed to insure that treatment results were not biased due to handling procedures, and/or other uncontrolled variables. d. Measurements l) Pregnancy Rates. We assessed contraceptive effectiveness by determining pregnancy status of all experimental elk. Using transrectal ultrasonography (Willard et al. 1994, 1998 ), we determined pregnancy rates of treated and control elk prior to treatment, and at 60, and 120 days posttreatment. On the day of pregnancy assessment, elk were moved from 5 ha pastures to a handling chute where they were sedated with xylazine hydrochloride (20-200 mg/animal, IM), then scanned using real-time transrectal ultrasound to determine pregnancy status and/or fetal age. Elk were then reversed with yohimbine (0.125 mg/kg, N) and returned to their original pastures. We determined the reversibility of leuprolide by releasing a epididymectomized male elk with treated female elk in October 2001 and conducting a GnRH challenge trial (Baker et al. 1995) to measure LH and P4 levels. 2) Reproductive Behavior. The effects of leuprolide on the breeding behavior of captive elk treated prior to the breeding season is known (Baker et al. 2002, in press), however, these reported effects may or may not be extended to elk treated during early pregnancy. �178 Down-regulation of gonadotroph cells by the action ofleuprolide and subsequent reduced secretion of LH could effectively inhibit progesterone secretion by the corpus luteum. If the effective action of leuprolide is luteolytic, then early embryonic loss could occur (Plotka et al. 1982, Asher et al. 1988, Flint et al. 1991). However, the efficacy of induction of luteolysis by leuprolide in Cervidae is unknown. Furthermore, it's not known if following early embryonic loss, whether female elk will regain normal estrus cycles and behavior. We evaluated these potential behavioral side-effects by monitoring maintenance and breeding behavior of male and female elk before and after leuprolide treatments. Each animal was individually identified using color-coded neck bands or ear tags. We tested the null hypothesis that the frequency of sexual interactions between males and females is similar before and after contraceptive treatment. e. Statistical Analysis. We analyzed data using least squares ANOVA for General Linear Models and the SAS Interactive Matrix Language. Response to contraceptive treatment was analyzed with a two-way factorial analysis of variance for a randomized complete block design with repeated measures structure. Factors in the analysis were treatment and time. Treatment was tested using the animal-within-treatment variance as the error term. Time was treated as a within subject effect using a multivariate approach to repeated measures. We used orthogonal contrast to test for differences among individual means (Morrison 1976). E. Experiment 4: Development of a remote delivery system 1. Objective. Begin evaluating a remote delivery system by comparing effectiveness of subcutaneous and intramuscular administration of leuprolide formulation in suppressing reproduction in female elk. 2. Methods. a. Animals and treatment. We conducted a controlled experiment with 13 adult female elk (713 years of age; 250-300 kg), lintact male elk, and 1 epididymectomized male elk at the Colorado Division of Wildlife's Foothills Wildlife Research Facility (FWRF), Fort Collins, Colorado during 15 August 2001 to 28 March, 2002. Between 15 August and 1 September, 2001, the epididymectomized male elk was released with 13 adult female elk into 2 adjoining paddocks (2 ha). Females were monitored for occurrence of estrus cycles by measuring progesterone leve'is beginning 1 September 200latid wete ·considered reproductively active when concentrations were greater than 1 ng/ml. Treatments were assigned as follows: 3 females were randomly selected to receive a subcutaneous formulation ofleuprolide (32.5 mg)(ATRIGEL, Atrix Laboratories, Inc. Fort Collins, Colorado, USA) by syringe injection; 3 elk were selected to receive an intramuscular leuprolide formulation (32.5 mg) via syringe injection; 4 elk received the leuprolide formulation via a 1 cc, PneDart dart (16 gauge, 3.39 cm. needle) fired from a CO2 -powered Dan-Inject pistol, and 3 elk were designated as untreated controls. Treatments were applied as follows. On the day before application (6 September 2001), experimental elk were moved from pastures to individual isolation pens (5 m x 10 m), weighed(± 0.5 kg), sedated with xylazine hydrochloride (Rompun; Bayer AG, Leverkusen; 25-200 mg/animal, i.m) and fitted nonsurgically with indwelling jugular catheters. The next day, treatment and placebo treatments were administered. In order to accurately determine the precise dose of leuprolide formulation remotely delivered to each elk, syringe darts were �179 weighed (0.001 g) before and after injection. Prior to darting, elk were placed in a handling chute and lightly sedated with xylazine hydrochloride (15-20 mg/animal, i.v.). This dose allowed animals to remain standing in the chute and minimized excitation associated with discharge of the dart gun. With the exception .of two animals, one dart per animal was fired from approximately 3 meters into the middle gluteus maximus muscle of the standing elk. Once all elk had been treated, sedation was reversed with yohimbine (30 mg) (Antagonil®, Wildlife Laboratories, Fort Collins, Colorado, USA) and animals were returned to individual isolation pens. b. Measurements. Approximately 1 hour following treatment applications, we measured 24hour LH response of elk treated with the leuprolide formulation and untreated control elk. Blood samples (5 ml) were collected via jugular catheters at 0, 120, 180, 240, 300,360,480 min then at 10, 16, and 24 hr after injection. Catheters were flushed after each collection with sterile saline solution. After the last blood collection, catheters were removed and animals returned to 5 ha pastures. The effect of leuprolide formulation on the duration of suppression ofLH was determined by periodically conducting pituitary stimulation trials. These trials were conducted during 29 October 2001 to 28 March 2002 to determine the capability of LH cells to respond to stimulation with an exogenous dose of GnRH analog (D-Ala6-GnRH-Pro 9ethylamide; Sigma Chemical Company, St. Louis, Missouri, USA). Pituitary stimulation trials were conducted with treated and control elk at 30, 60, 90, 120, 160, and 190 days posttreatment. Stimulation trials were conducted according to the following procedures: On the day of testing, treated and control elk were moved from 5 ha pastures to individual isolation pens, weighed, sedated (as previously described), and fitted nonsurgically with indwelling jugular catheters. GnRH analog (1 µg/50 kg body weight) was administered through the cannula and blood samples were collected (5 ml) at 0, 60, 120, 180,240, 300, 360, and 480 minutes posttreatment. After collections, blood was stored at 4° C for 24 hours until serum was obtained by centrifugation (1500 RCF for 15 minutes). Serum samples for progesterone levels were also collected from each elk on each of these trial days. Serum was stored at -20° C until analyzed for LH and progesterone. Following the last blood collection, catheters were removed and elk were returned to the holding pastures. The effect of leuprolide formulation on reproduction in treated and control elk was determined by measuring pregnancy rates using the presence or absence of pregnancy specific protein B (PSPB) (BioTracking, Moscow, Idaho, USA) in serum collected at approximately 100 and 215 days of gestation (Huang et al. 2000). c. Statistical analysis. Responsiveness of the pituitary gland to GnRH analog_ stimulation was assessed in two ways: (1) maximum response (highest concentration of LH (ng/ml) achieved after injection minus baseline), and (2) total amount ofLH secreted (ng/ml/min) estimated by calculating the area under the LH response curve (Abramowitz and Stegun, 1968). Differences among hormone concentrations were tested using least squares ANOV A for general linear models (SAS Institute, 1997). Responses to treatment were analyzed with oneway ANOV A for a randomized complete block design with repeated measures. Treatment effects were determined using the total animal-within-treatment variances as the error term. Time was treated as a within-subject effect, using a multivariate approach to repeated measures (Morrison 1976). A "protected" least significant difference test (Milliken and Johnson, 1984) was used to separate means when the overall F test indicated significant treatment effects (P < 0.05). �180 RESULTS A. Experiment 1: Dose response Administration of sustained release formulations of leuprolide to female elk resulted in an acute, transient rise in serum LH concentrations irrespective of dose. Maximum LH concentrations (15.6 ± 0.93 ng mi-1) occurred approximately 3 hours following treatment and were similar across all treatment levels (Fig. 1). Following peak response, there was a rapid decline in LH to basal levels during the next 24 hours. Total LH secretion (ng m1- 1 min-1) did not differ among treatments and all treatments resulted in higher LH secretion than controls (P s 0.002). Leuprolide reduced serum LH secretion to nondetectable levels in treated females for 130 days posttreatrnent (Fig. 2). Differences in mean maximum serum LH were significantly lower (P ~ 0.031) in treated elk compared to untreated controls at all sampling periods. For untreated females, mean maximum LH fluctuated from a high of 19.3 ± 4.2 to a low of 3.5 ± 0.06 ng m1- 1 . This variation was likely related to the phase of the oestrous cycle when control females were challenged with GnRH and the influence of fluctuating levels of estradiol and progesterone on LH secretion (Goodman and Karsch 1980). B. Experiment 2: Antifertility and behavioral effects on nonpregnant female elk a_ Pregnancy rates, hormonal measurements, and blood parameters. Because Experiment 1 did not establish a minimum effective dose ofleuprolide for LH suppression, we arbitrarily reduced the leuprolide formulation by approximately 20% below the lowest concentration tested in Experiment 1, to 32.5 mg. This dose ofleuprolide prevented pregnancy in all treated females (Group A) while the pregnancy rate of control females (Group B) was 100%. Treated females tested negative and controls positive for PSPB on all sampling dates. Estimated conception dates for pregnant elk ranged from 10 November to 19 November 1999 and parturition occurred between 12 July and 26 July 2000. Leuprolide caused a significant reduction (P ~ 0.035) in mean maximum serum LH (Fig. 3) and P4 (Fig. 4) concentrations in treated females (Group A) with a return to pretreatment levels the following breeding season (11 November 2000). Serum LH was reduced to non-detectable concentrations by 92 days posttreatment and remained at this level until day 225. In one treated female, LH remained at baseline for 250 days posttreatrnent. Maximum LH response was lower (P ~ 0.012) in treated compared to non-pregnant controls (Group C) at 30, 92, 135, 165, and 193 days following treatment. Serum LH of untreated elk declined significantly (P = 0.024) between April and May with the onset of anestrus, then returned to pretreatment levels indicative of estrus in November 2000. • Serum P4 levels of treated females followed a similar pattern to that observed for serum LH (Fig. 4). Progesterone levels were similar in treated and control elk until day 30, thereafter, serum progesterone remained at basal concentrations in treated females until day 225 of the trial, indicating that additional ovulations did not occur. Control females maintained increased serum P4 content, reflecting continued regular estrous cycles within this group until day 165 ( 18 April) when the effects of anestrus reduced P4 to basal levels. Similar to serum LH, P4 content then increased during November 2000 to pretreatment concentrations in both treated and untreated elk (Fig. 4). We evaluated 13 hematology and 19 serum chemistry parameters in treated and untreated elk females. With the exception of creatinine kinase (CK), a muscle derived enzyme, all individuals were clinically similar. Elk in the treatment group showed moderately elevated CK levels (400-702 IU L-1). Creatinine kinase levels can increase in unconditioned animals following vigorous exercise and remain elevated for 4-6 hours (Lefebvre et al. 1994). Handling procedures for blood sampling in treated females were often �181 more physically rigorous than those for controls due to the need to separate females from males. Thus, the elevated CK levels in treated elk compared to controls likely reflect a bias due to a difference in animal handling prior to blood sample collections, rather than a treatment-induced response. b. Reproductive behavior. We observed male to male dominance interactions immediately following their release into the pastures with treated and untreated females. Within 2.5 weeks, one male established dominance over the other two. Thereafter, subdominant males retreated to remote locations in the pastures and rarely interacted with females or the dominant male for the remainder of the experiment. During the breeding season, we observed reproductive interactions of males and females on 34 days during 10 November to 23 December 1999. We analyzed 63 sampling periods (134.5 h): 20 periods at dawn (45.7 h), 6 at mid-day (13.5 h), 20 at dusk (42.8 h), and 17 at night (32.6 h). The average length of the observation periods was 2.1 (SE= 0.10) h. Postbreeding observations occurred on 14 days during 7 February to 27 March. We analyzed data from 16 sampling periods (54.7 h): 6 periods at dawn (22.5 h), 2 at mid-day (7.5 h), 7 at dusk (22.2 h), and 1 at night (2.5 h). Observation periods averaged 3.4 (SE= 0.24) h. Contrary to our first hypothesis, sexual interactions during the breeding season were not diminished in leuprolide-treated females compared to controls. Instead, breeding behavior rates were similar for treated and untreated females for all behavior categories (Fig. 5). Although we did not detect a significant treatment x time interaction, copulatory (P = 0.064), male precopulatory (P = 0.083), and female precopulatory (P = 0.072) behaviors approached significance and are notable. For these 3 behavior categories, the daily behavior rate decreased over time for untreated females, but remained constant for treated elk. We also failed to reject our second hypothesis. Treated females did not resume normal oestrus cycles during the postbreeding season and reproductive behavior rates did not increase compared to untreated controls. We observed almost no sexual interactions between the dominant male and treated or untreated females during the postbreeding season. There were no copulatory or female precopulatory behaviors recorded, and too few male precopulatory (~ 0.17 day·1) and general breeding(~ 0.30 day·1) behaviors to analyze. C. Experiment 3: Antifertility effects on pregnant elk. Leuprolide administered has a 32.5 mg subcutaneous formulation to elk during the first trimester of pregnancy failed to induce fetal loss. Fetal age at the time of treatment of treated females ranged from 30 - 90 days of age and from 50 - 90 days for control elk. Treated and control females .were positive for PSPB at all sampling dates during gestation and all produced a calf at parturition. Dystocia was observed in 3 of 6 females but did not appear to be related to treatment. During the breeding season reproductive behaviors were similar (P = 0.45) for treated and control female elk. We observed almost no sexual interactions during the postbreeding season. D. Experiment 4: Development of a remote delivery system. Administration of a 90-day sustained release formulation ofleuprolide to female elk resulted in an acute, transient rise in serum LH levels irrespective of mode of delivery (Fig. 6). Maximum LH concentrations occurred approximately 3.5-4.5 h following treatment and were highest (84.9 ± 5.3 ng/ml) for the �182 intramuscular syringe treatment, followed by intramuscular dart (42.2 ± 15.8 ng/ml) and subcutaneous syringe (23.0 ± 5.1 ng/ml). Following peak response, there was a rapid decline in LH basal levels during the next 24 hours. Leuprolide reduced serum LH secretion to non-detectable levels in all treatment groups for 120 days posttreatment. Between 120 and 160 days posttreatment, LH levels in the intramuscular syringe and dart treatments increased substantially over the subcutaneous syringe treatment and control females and remained elevated for the duration of the experiment. In contrast, LH levels for the subcutaneous syringe group remained at basal levels (Fig. 6). Serum P4 levels followed a different pattern than that observed for serum LH (Fig. 7). After 60 days posttreatment, P4 concentrations in all treatment groups declined to basal levels and remained at these levels for the remainder of the experiment. Regardless of mode of delivery, leuprolide formulation prevented pregnancy in all treatment groups, whereas pregnancy rate of control females was I 00%. Leuprolide-treated females tested negative and controls positive for PSPB on both sampling dates. DISCUSSION Successful application of fertility control technology for wildlife is dependent on development of contraceptive agents that are safe, practical, and effective. Current technology is limited due to problems of treatment implementation and concerns for the health of target and non-target species. In the present study, we evaluated a promising non-steroidal, non-immunological contraceptive technology for controlling fertility in female elk. Administration of a sustained release formulation containing leuprolide to captive female elk prior to the breeding season, resulted in decreased LH and progesterone secretion, temporary suppression of ovulation and steroidogenesis, and effective contraception without detrimental behavioral or physiological side-effects. The acute increase in serum LH immediately following leuprolide treatment was consistent with previous studies in cattle (D'Occhio et al. 1996), sheep (Nett et al. 1981), horses (Montovan et al. 1990) and African elephants (Loxodonta africana) (Brown et al. 1993). There was little variation among elk in their serum LH response to different doses ofleuprolide, indicating either low variability in the amount and duration of agonist released or doses so high that any variation was masked. The minimum level of leuprolide needed to suppress estrus in female elk was not determined in this study. All doses of leuprolide were equally successful in reducing LH concentrations for the duration of the 130 trial. Additional research to establish a minimum .effective dose ofleuprolide would enhance the economic practicalify of this contracep~v~ agent. The cessation of estrous cycles in females treated with leuprolide and the return to apparently normal ovarian function after depletion of the agonist implant was consistent with findings for females in other species (D'Occhio et al. 1996; Evans and Rawlings, 1994; Fraser et al. 1989). The effectiveness of leuprolide as a contraceptive agent is dependent on suppression of ovulation from the inception of the breeding season to the onset of anestrus, a period of approximately 200 days for elk. Leuprolide inhibited ovulation for > 190 days, 2 times longer that the formulated 90 day delivery period. The prolonged suppression of gonadotrophin secretion may occur for several reasons. Among :these are that release ofleuprolide from the implant may have continued beyond the formulated 90 day period. Certainly, LH secretion remained suppressed for more that 13 0 days in Experiment 1. Likewise, leuprolide treatment may have induced prolonged suppression of gonadotroph function (i.e. extending beyond the duration of the implant). In other ruminants, if gonadotroph function is suppressed for an extended duration, a recovery period of 30-60 days following removal of the suppression is necessary before pituitary content of LH and gonadotropin secretion can return to normal levels (Nett 1987). Thus, �183 if duration of leuprolide release from the implant was 130 days and recovery of gonadotroph function requires approximately 60 days, this would be sufficient to carry the reduced secretion of LH into the normal anoestrous period when secretion of LH would be photoperiodically suppressed. If this is indeed true, then a single treatment should provide a contraceptive effect for approximately one breeding season. The effectiveness ofleuprolide in preventing pregnancy in female elk is conditional. Successful prevention of fertility was achieved by treating elk prior to the breeding season. The use of leuprolide as a contragestive in female elk during early pregnancy was unsuccessful. Since we did not measure LH responses to leuprolide treatment in pregnant elk, the mechanism for failure is unknown. We speculate that complete down-regulation of LH receptors did not occur and LH levels were high enough to stimulate an LH surge and subsequent ovulation. The overall rates of sexual interactions between treated and control elk were not different during the breeding and postbreeding seasons. During the breeding season, the dominant male established and defended a single harem of treated and untreated females. Reproductive behaviors during the breeding season between the dominant male and harem females followed a pattern similar to that described for free-ranging elk (Geist 1982). Treated and untreated females were courted, bred, and defended with equal frequency, however the pattern of reproductive interactions changed over time. Once untreated females became pregnant, reproductive behavior rates decreased, whereas, copulatory, and male and female precopulatory rates remained constant over time in treated females. These extended sexual interactions were generally intermittent and may have been related to fluctuating levels of progesterone and oestradiol. Estrus can occur with relatively low estradiol concentrations, if coupled with low progesterone content. In domestic sheep, pre-exposure to progesterone stimulates estrus behavior at much lower concentrations of estradiol once progesterone is decreased in circulation (Robinson 1954). Therefore, since these animals had ovulated prior to leuprolide treatment, they became very sensitive to low levels of estradiol, and since ovulation and corpus luteum formation were blocked they continued to show estrus behavior with basal estradiol levels. Regardless of the mechanism involved, disruption of normal behavioral patterns are not a desirable sideeffect of contraceptive treatments. However, without carefully designed large-scale investigations with larger sample sizes, and under more natural conditions, we can only speculate on the significance of these behavioral alterations on the health and social organization of treated populations. Before leuprolide can be considered a practical and efficacious approach for wildlife contraception, development of a reliable remote delivery system is needed. Our pilot efforts to develop such a system were promising, however, the small sample size (n = 3) used in this experiment support only guarded optimism. It appears that the rise in LH levels observed in females treated with syringe dart delivery of leuprolide formulation were not high enough to stimulate ovulati.on and conception. Clearly, additional research with larger sample sizes is needed to confirm or reject these findings. CONCLUSION The objective of the work reported here was to evaluate the contraceptive potential of a GnRH agonist (leuprolide) formulation in female elk, provide evidence of physiological and behavioral side effects of treatment (if any), and assess the potential for remote delivery. We conclude that leuprolide administered as a controlled release formulation prior to the breeding season, offers a new approach to reversible contraception in wild ungulates that overcomes problems associated with existing technology. �184 First, leuprolide formulation improves practical application of contraception because a single treatment can induce infertility in females without relocating and treating specific individuals each year. Second, leuprolide acetate is a neuropeptide, thus the proteinaceous nature of this agent eliminates the possibility of passage through the food chain to non-target species. Third, behavioral side-effects were minimal. Sexual interactions of treated females were extended early in the breeding season but recurrent estrous cycling and ovulation did not occur. Fourth, there were no short-term physiological side-effects of treatment. Treated animals appeared healthy and seasonal intake and body weight dynamics normal. However, before this technology can be considered a practical and efficacious approach for wildlife, additional research is needed to ascertain minimum effective dose, verify effective treatment duration, and develop a remote delivery system for administering leuprolide formulation to unrestrained animals. ACKNOWLEDGMENTS Our research was supported by the U. S. National Park Service, Rocky Mountain National Park (Grant 1520-9-9002) and the Colorado Division of Wildlife (Federal Aid to Wildlife Restoration, Project 153R4). We thank Dr. Delwar Hussain and Dr. Richard Dunn at Atrix Laboratories, Ft. Collins, Colorado for the generous donation of the leuprolide acetate formulations used in these investigations and for their technical assistance in delivery technology. We gratefully acknowledge and appreciate the technical assistance of Joan Ritchie in overall organization and execution of blood sampling protocols and animal handling. Darby Finley and Elizabeth Wheeler provided invaluable assistance with behavioral observations and general animal training and husbandry. We thank David Bowden for statistical consultation and analysis. LITERATIJRE CITED Abramowitz, M., and L. A. Stegun. 1968. Handbook ofMathematical Functions. Dover Publications, New York, USA. Adam, C. L., C. E. Moir, and T. Atkinson. 1985. Plasma concentrations of progesterone in female red deer (Cervus elaphus) during the breeding season, pregnancy, and anoestrus. Journal of Reproduction and Fertility 74:631-636. Asher, G. W. 1988. Hormonal changes during the luteal regression in farmed fallow deer, Dama dama. Journal of Reproduction and Fertility 84:379-386. Baker, D. L., M.A. Wild, M. M. Conner, H.B. Ravivarapu, R. L. Dunn, and T. M. Nett. 2002. Effects of GnRH agonist (leuprolide) on reproduction and behayior in fe~ale vyapiti (Cervus elaphu_s nelsoni). Reproduction (Suppl) 60: (in press). • _ , M. W. Miller, and T. M. Nett. 1995. Gonadotropin-releasing hormone analog-induced patterns ofluteinizing hormone secretion in female wapiti (Cervus elaphus nelsoni) during the breeding season, anestrus, and pregnancy. Biol. ofReprod. 52:1193-1197. Becker, S. E., and L. S. Katz. 1995. Effects of gonadotropin-releasing hormone agonist on serum LH concentrations in female white-tailed deer. Small Ruminant Research 18:145-150. Belchetz, P. E., T. M. Plant, Y. Nakai, E. J. Keogh, and E. Knobil. 1978. Hypophysial response to continuous and intermittent delivery of hypothalamic gonadotropin-releasing hormone. Science 202: 631-633. Bergfeld E., M. J. D'Occhio, and J.E. Kinder. 1996. Pituitary function, ovarian follicular growth, and plasma concentrations of l 7P-oestradiol and progesterone in prepubertal heifers during and after treatment with the luteinizing hormone-releasing hormone agonist deslorelin. Biology of Reproduction 54:776-782. Brown, J. L., D. L. Schmitt, A. Bellem, L. H. Graham, and J. Lehnhardt. 1999. Hormone secretion in the Asian elephant (Elaphus marimus): Characterization of ovulatory and anovulatory LH surges. Biology of Reproduction 61: 1294-1299. �185 Cole, G. F. 1971. An ecological rationale for the natural or artificial regulation of native ungulates in parks. Transactions of the North American Wildlife and Natural Resources Conference 36:417-426. D'Occhio, M. H., W. J. Aspden, and T. Whyte. 1996. Controlled reversible suppression of oestrous cycles in beef heifers and cows using agonist ofluteinizing hormone - releasing hormone. Journal of Animal Science 74:218-225. _ _, M. J., and W. J. Aspden. 1999. Endocrine and reproductive responses of male and female cattle to agonist of gonadotrophin-releasing hormone. Journal of Reproduction and Fertility Supplement 54:101-114 DeNicola, A. J., D. J. Kesler, and R. K. Swihart. 1997. Remotely delivered prostaglandin F- 2 alpha implants terminate pregnancy in white-tailed deer. Wildlife Society Bulletin 25:527-531. Dunn, R. L., J.P. English, D.R. Cowan, and D. P. Vanderbilt. 1994. Biodegradable in situ forming implants and methods of producing the same. U. S. patent no. 5,278,201. Evans, A. C. 0., and N. C. Rawlings. 1994. Effects of a long-acting gonadotrophin-releasing hormone agonist (Leuprolide) on ovarian follicular development in prepubertal heifer calves. Canadian Journal of Animal Science 74:649-656. Goodman, R. L., and F. J. Karsch. 1980. Pulsatile secretion ofluteinizing hormone: differential suppression by ovarian steroids. Endocrinology 107: 1286-1290. Fagerstone, K. A, M. Coffey, P. Curtis, R. Dolbeer, G. Killian, L.A. Miller, and L. Wilmot. 2001. Wildlife contraception. Wildlife Society Technical Review. Proceedings of the Wildlife Society 8 th Annual Conference, Reno, USA Flint, A.P.F., E. L. Sheldrick, T. J. McCann, B. R Brinklow, and A.S.I. Loudon. 1991. Prostaglandininduced secretion of oxytocin and prolactin in red deer (Cervus elaphus) and Pere David's (Elaphuro.s davidianus) deer hinds: Evidence for oxytocin ofluteal origin. General and Comparative Endocrinology 83:432-438. Garrott, R. A., P. J. White, and C. A. Vanderbilt-White. 1993. Overabundance: an issue for conservation biologist? Conservation Biology 7:946-949. ~ and D. B. Siniff. 1992. Limitations of male-oriented contraception for controlling feral horse populations. Journal of Wildlife Management 56:456-464. Geist, V. 1982. Adaptive behavioral strategies. Pages 219-277 in Elk of North America: Ecology and Management. J. W. Thomas and D. E. Toweill, editors. Stackpole Books, Harrisburg, USA. Jacobsen, N. K., D. A. Jessup, and D. J. Kesler. 1995. Contraception in captive black-tailed deer by remotely delivered norgestomet ballistic implants. Wildlife Society Bulletin 23:718-722. Hess, K. H. 1993. Rocky Times in Rocky Mountain National Park. University of Colorado Press. Niwot Colorado, USA. Huang, F., D. C. Cockrell, T. R. Stephenson, J. H. Noyes, and R. G. Sasser. 2000. A serum pregnancy test with specific radioimmunoassay for moose and elk pregnancy specific protein B. Journal of Wildlife Management 64:492-499. Hobbs, N. T., D. C. Bowden, and D. L. Baker. 2000. Effects of fertility control on populations of ungulates: general, stage-structured models. Journal of Wildlife Management 64:473-491. Houston, D. B. 1971. Ecosystems of national parks. Science 172:648-651. Jewell, P.A., and S. Holt. 1981. Problems in management oflocally abundant wild animals. Academic Press. 32lpp. Lefebvre, H.P., P. L. Toutain, J.P. Serthelon, V. Lassourd, L. Gartley, and J.P. Braun. 1994. Pharmcokinetic variables and bioavailability from muscle of creatinine kinase in cattle American Journal of Veterinary Research 55:487-497 Lehner, P. N. 1996. Handbook of Ethological Methods. Second Edition. Cambridge University Press, Cambridge, UK. Kirkpatrick, J. F., and J. W. Turner, Jr. 1985. Chemical fertility control and wildlife management. Bioscience 35:485-491. �186 ~ ~ I. K. M. Lui, R. C. Fayrer Hosken, and A. T. Rutberg. 1997. Case studies in wildlife immunocontraception: wild and feral equids and white-tailed deer. Reproduction, Fertility, and Development 9: 105-110. McCullough, D.R., K. W. Jennings, N. B. Gates, B. G. Elliot, and J.E. Didonato. 1997. Overabundant deer populations in California. Wildlife Society Bulletin 25:478-483. McNeilly, A. S., and H. M. Fraser. 1987. Effect of gonadotrophin-releasing hormone agonist-induced suppression of LH and FSH on follicle growth and corpus luteum function in the ewe. Journal of Endocrinology 115:273-282. Milliken, G. A., and D. E. Johnson. 1984. Analysis of Messy Data. Volume I. Designed Experiments. Lifetime Leaming Publications, Belmont,California, USA Montovan, S. M., P. P. Daels, J. Rivier, G. H. Hughes, G. H. Stabenfeldt, and B.L. Lasley. 1990. The effect of potent GnRH agonist on gonadal and sexual activity in the horse. Theriogenology 33: 13051321. Morrison, D. F. 1976. Multivariate Statistical Methods. McGraw-Hill Book Co., New York, USA. Morrison, J. A. 1960. Characteristics of estrus in captive elk. Behavior 16:84-92. Nett, T. M., M. E. Crowder, G. E. Moss, and T. M .. Duello. 1981. GnRH-receptor interaction. V. downregulation of pituitary receptors for GnRH in ovariectomized ewes by infusion of homologous hormone. Biology of Reproduction 24:1145-1155. Nett, T. M. 1987. Function of the hypothalamic-hypophyseal axis during the post-partum period in ewes and cows. Journal of Reproduction and Fertility Supplement 34:210-213. Niswender, G.D., L. E. Reichert, Jr., A. R. Midgley, Jr., and A. V. Nalbandov. 1969. Radioimmunoassay for bovine and ovine luteinizing hormone. Endocrinology 84: 1166-1173. _ _ . 1973. Influence of the site of conjugation on the specificity of antibodies to progesterone. Steroids 22:413-424. Plotka, E. D., U.S. Seal, L. J. Venne, J. J. Ozoga. 1982. Reproductive steroids in white-tailed deer. IV. Origin of progesterone during pregnancy. Biology of Reproduction 26:258-262. Ravivarapu, H.B., K. L. Moyer, and R. L. Dunn. 2000. Sustained activity and release ofleuprolide acetate from an in situ forming polymeric implant. AAPS PharSciTech I :23-26. Rapley, M. D. 1985. Behavior of wapiti during the rut. In Biology of Deer Production Eds. P. F. Fennessy and K. R. Drew. Bulletin No. 22, Royal Society ofNew Zealand, Wellington 357-361. Robinson, T. J. 1954. Relationship of oestrogen and progesterone in oestrous behavior of the ewe. Nature 173:870-874. SAS Institute. 1993. SAS/STAT user's guide. Release 6.03 edition. SAS Institute, Cary, North Carolina, USA. Seagle, S. W., and J. D. Close. 1996. Modeling white-tailed deer population control by contraception. Biological Conservation 76:87-91. _ Singer, F. J., L. C. Zeigenfuss, R G. Cates, and D. T. Barnett. 1998. Elk, multiple factors, and persistence of willows in national parks. Wildlife Society Bulletin 26:419-428. Smith, B. L. 200 I. Winter feeding of elk in western North America. Journal of Wildlife Management 65:173-190. White, C. A., C. E. Olmsted, and C. E. Kay. 1998. Aspen, elk, and fire in the Rocky Mountain national parks of North America. Wildlife Society Bulletin 26:449-462. Willard, S. T., R. G. Sasser, J.C. Gillespie, J. T. Jaques, T. H. Welsh, Jr., and R. D. Randel. 1994. Methods for pregnancy determination and the effects of body condition on pregnancy status in Rocky Mountain elk. Theriogenology 42:1095-1102. ~ R G. Sasser, J. T. Jaques, D. R White, D. A. Neuendorff, and R. D. Randel. 1998. Early pregnancy detection and the hormonal characteristics of embryonic-fetal mortality in fallow deer (Dama dama). Theriogenology 49:861-869. Zeignefuss, L., F. J. Singer, and D. C. Bowden. Vegetation and elk responses following release from artificial controls within Rocky Mountain National Park, 1968-1996. Report to the Superintendent, Rocky Mountain National Park. �187 Fig.1 20 T 15 -1 E .s 0) ---- 0mg -ai.- - 45 mg ----e- - 90mg -T - 180mg 10 :r: _J 5 0 2 1 0 4 3 5 7 6 8 12 24 Time (hours) Figure 1. Twenty-four hour profile of serum LH concentrations in untreated female elk ( •, n = 2) and elk treated with a sustained release formulation containing 45 (A, n = 2), 90 (o, n = 2), and 180 mg (T, n = 2) ofleuprolide. Results are shown as± SE. Fig.2 30 - - - 0 mg 45 mg -Ir- --e-- 90 mg ~ 0) C 20 I _J E E ·x ro :::J ~ C ro Q) 10 I ~ I )\ \ I 180 mg 1 /I' -1 E -T - \ J l \ 0 0 .5 1 35 75 110 130 Time (days) Figure 2. Profiles of 130--day LH concentrations in untreated ( •, n = 2) female elk and elk treated with a sustained release formulation containing 45 (A, n = 2), 90 (o, n = 2), and 180 mg {T, n = 2) of leuprolide. Results are shown as± SE. �188 Frg.l 30 - 25 E 0) C ..__, 20 I I E E ::J ·xCJ ~ T 1 "" e. "e- -- -- --1...... . . . . ...... _J 15 -0- - Control ____.,._ Leuprolide • 'T0 - - - 4\T J\ ..l 10 \ C \ CJ \ Cl) ~ 5 \ "o-.._ .L 0 Pretrmt 30 92 135 165 193 225 373 Time (days) Figure 3. Profiles of mean maximum serum LH concentrations for untreated female elk ( •, n = 4), and elk treated with a sustained release formulation containing 32.5 mg leuprolide ( o, n = 4). Results are shown as ± SE. Fig.4 2.40 -0- - Control ----- Leuprolide ,,......_ E 1.80 I 0) C ...__, Q) Q C ....Q) 0 - 120 t i) Q) ... 0 a.. C a, T T -~ C) 1 ............ 1 ,.,.,.,. . 0.60 Q) l' I /~ 'if 1 I \ I \ /J 1 I \ ~ \ \ 0.00 Pretmrt. 30 92 135 165 193 225 373 Time (days) Figure 4. Profiles of mean maximum serum progesterone concentrations for untreated female elk(•, n = 4), and elk treated with a sustained release formulation containing 32.5 mg leuprolide ( o, n = 4). Results are shown as± SE. �189 Fig 5 10 '>-(U 8 -a ~ 0 ·::;: - 6 (U ..c Q) ..Q Control Ba Leuprolide Q) ro 4 0:::: .._ 0 ·::;: Jg 2 Q) co 0 Gen. Breed. Male Precop. Female Precop. Copulatory Behavior Category Figure 5. Mean(± SE) reproductive behavior rates during the breeding season for untreated (n = 5) and leuprolide-treated (n = 5) female elk. Results are shown as± SE. Fig. 6 90 ~ - 72 1 ---- I;t\\ I \ ---.-- -0-- IM Syringe E C) C "-' I 54 ....J a.. / t\\\ //I,\ I 36 I C (ll (I) ~ SC syringe ----s::J - Control ~ (ll (I) IM Dart 18 ,,510 0 .5 '.I 30 60 90 120 160 190 Time(days) Figure 6. Profiles of mean maximum serum LH concentrations for untreated female elk (e, n = 3), and elk treated with a 32-5 mg leuprolide formulation delivered intermuscularly via syringe dart (0, n = 3), intermuscular syringe (.A., n = 3), and subcutaneous syringe (T, n = 3). Results are shown as± SE. �190 Fig. 7 - - k- • IM syringe _. •• SC syringe --e · Control ---€r IM Dart 2 E ........ C) C ...__,, Cl) 0 I 1.- .... Cl) "' -·• 1 T - C 1 1 1 - T l Ti -V Cl) C) 0 1 1.- a.. C <ti Cl) ~ 0 ---Pretrmt 30 60 90 120 160 190 Time(days) <•. Figure 7. Profiles of mean maximum serum progesterone concentrations for untreated female elk n= 3), and elk treated with a 32.5 mg leuprolide formulation delivered intermuscularly via syringe dart (0, n = 3), intermuscular syringe(.&, n = 3), and subcutaneous syringe (T, n =3). Results are shown as± SE. �57 JOB PROGRESS REPORT State of Division of Wildlife - Mammals Research Colorado Work Package No. --'3'""'0'-"0=2_ _ _ _ _ __ Elk Conservation Task No. Technical Support for Elk and Vegetation Management Environmental Impact Statement for Rocky Mountain National Park Federal Aid Project RMNP W-153-R Period Covered: July 1, 2002 - June 30, 2003 Authors: D. L. Baker, M.A. Wild, and M. M. Conner Personnel: M. Coffey, G. Dodd, B. Gill, T. Johnson, M. Monello, R. Monello, D. Plattner, J. Powers, J. Ritchie, R. Spowert, T. Nett, D. Hussain, R. Dunn, K. Zollers. ABSTRACT Overabundant wild ungulate populations have become a significant concern for natural resource managers in many parts of North America. Wild ungulates can do serious and lasting harm to many plant communities, and preventing such damage requires controlling the growth of their populations. In protected areas such as national parks, traditional methods of population control may not be feasible or publically acceptable. In these situations, alternative methods of population control are needed. One alternative is controlling the fertility of females. In this study, we evaluated the feasibility of using gonadotropin releasing hormone (GnRH) analog to control reproduction in free-ranging female elk in Rocky Mountain National Park. During fall of 2002, we captured, radio-collared and treated 34 adult elk. Seventeen elk were treated subcutaneously with a controlled release bio-implant containing 32.5 mg of leuprolide and seventeen elk were treated with the same formulation without leuprolide. We evaluated the effects ofleuprolide treatments on reproductive rates, body condition, behavior, and daily activity patterns offemale elk during September 2002 to April 2003. Leuprolide administered as a sustained release formulation was 100% effective in preventing pregnancy in female elk. Body condition of all experimental elk declined from fall 2002 to spring 2003. Changes in loin depth and body condition score were similar (P. 0.254) for both treated and control elk, whereas overwinter loss in mean percent rump fat was greater (P. 0.057) for treated elk compared to controls. There were no differences (P = 0.36) in reproductive behavior rates during the breeding season between treated and control elk. �58 TECHNICAL SUPPORT FOR ELK AND VEGETATION MANAGEMENT ENVIRONMENTAL IMPACT STATEMENT FOR ROCKY MOUNTAIN NATIONAL PARK D. L. Baker, M.A. Wild, and M. M. Conner P. N. OBJECTIVE Conduct experiments with captive and free-ranging elk to evaluate fertility control as an management alternative for controlling elk populations in Rocky Mountain National Park (RMNP), Colorado. SEGMENT OBJECTIVES 1. Capture, radio-collar, and apply fertility control treatments to a target sample of free-ranging adult female elk in RMNP during September 2002. 2. Evaluate the effects of fertility control on reproductive rates of treated and non- treated adult female elk and the reversibility of these effects if they occur. 3. Evaluate the effects of fertility control on body condition of treated and non-treated adult female elk. 4. Evaluate the effects of fertility control on reproductive behavior and daily activity patterns of treated and non-treated adult female elk. INTRODUCTION Overabundant wild ungulate populations have become a significant problem for natural resource managers in North America. Unregulated populations can cause adverse effects that are ecological, economic, or political in scope and resolving these issues often requires controlling animal abundance (Jewell and Holt 1981, Garrott et al. 1993, McCullough et al.1997, Smith 2001). In Rocky Mountain National Park (RMNP), Colorado, the impact ofherbivory by elk has emerged as a fundamentally important problem for those who manage the Park and its wildlife (Hess 1993, Zeignefuss et al. 1996). In 1968, RMNP adopted a natural-regulation policy for management of ungulates (Cole 1971, Houston 1971) with the objective of allowing density dependent processes to regulate elk numbers within park boundaries and use sport hunting to harvest as many animals as possible in areas surrounding the Park. Recently, however, Park managers have become concerned that possible unnatural concentrations of elk may be altering natural plant communities and ecosystem sustainability. Soil conditions and the status of willow and aspen plant communities have declined. Wet meadow, dry grasssland, and alpine and subalpine sites show evidence of deterioration from overgrazing by elk (Singer et al. 1998, White et al. 1998). As a result of the decline in these vegetation types and the diversity of the animal species that are associated with them, the Park and other natural resource agencies are evaluating alternative management strategies for reducing elk densities within RMNP and the surrounding Estes Valley. One alternative being considered is controlling the fertility of female elk. Fertility control has been widely advocated as an alternative to lethal methods of population control for wildlife and considerable research �59 has been directed toward development of different contraceptive agents (see reviews by Kirkpatrick and Turner 1985, Fagerstone et al. 200 I). Field and laboratory studies have evaluated the efficacy of delivery of contraceptives to ungulates (Jacobsen et al. 1995, DeNicola et al. 1997, Kirkpatrick et al. 1997) and models have been developed to represent effects of fertility control on the population dynamics of individual species and populations (Garrott and Siniff 1992, Seagle and Close 1996, Hobbs et al. 2000). To date, most contraceptive research for wild ungulates has focused on the development of immunocontraceptive vaccines and steroidal hormonal agents. However, after more than 40 years of research, the success of these approaches have been primarily limited to captive wildlife and small localized urban populations of wild ungulates. To meet this challenge, new technologies and approaches are needed if fertility control is to become practical and acceptable management tool for controlling overabundant wildlife species. A promising new non-steroidal, non-immunological approach to contraception involves potent analogs of gonadotropin-releasing hormone (GnRH). GnRH is a molecule produced in the hypothalamus of the brain. It directs specific cells in the pituitary gland to synthesize and secrete two important reproductive hormones; follicle stimulating hormone (FSH) and luteinizing hormone (LH). These latter two hormones, known as gonadotropes, control the proper functioning of the ovaries in the female and testes in the male. Chronic treatment with continuous, high doses of GnRH agonists results in temporary suppression of pituitary responsiveness and gonadotropin secretion. Resulting decreases in plasma LH and FSH in females leads to suppression of ovulation, estrous cyclicity, and gonadal steroidogenesis (Belchetz et al.1978, Evans and Rawlings 1994). Once GnRH agonist treatments are terminated, normal pituitary function is gradually restored (Bergfeld et al. 1996). GnRH agonists have been shown to inhibit ovulation in several domestic ungulate species including sheep (McNeilly and Fraser 1987), cattle ( D'Occhio et al. 1996; D'Occhio and Aspden 1999), and horses (Montovan et al. 1990). However, studies on wild ungulates are limited (Becker and Katz 1995; Brown et al. 1999) and to our knowledge, and only one study has demonstrated their effectiveness as a contraceptive agent (Baker et al. 2002 ). GnRH agonists provide a potential biotechnology for achieving a controlled, reversible suppression of fertility in both captive and free-ranging female wild ungulates. However, their practicality as a contraceptive agent is dependent on effective inhibition of reproduction without negative behavioral or physiological side-effects, and efficacious application in free-ranging elk. In previous experiments, we determined the effectiveness of GnRH agonist (leuprolide) for controlling fertility in captive female elk and assessed the physiological and behavioral side-effects of treatment (Baker et al. 2002). Leuprolide administered as a subcutaneous, controlled release formulation was 100 % effective in preventing reproduction in elk for one breeding season. Serum LH and progesterone (P 4) concentrations were reduced to baseline levels by day 30 and remained at those levels for 190-252 days posttreatment, with a return to normal fertility the following breeding season. In addition, there were no adverse physiological side-effects and behavioral effects were minimal. However, these results were obtained under controlled conditions with captive animals of known fertility and in excellent body condition. While these results provide strong inference on the potential utility of leuprolide as a contraceptive agent, studies with wild elk are needed to evaluate whether the technique is truly feasible and practical. Thus, the goal of this study was to conduct a field experiment to examine the efficacy of leuprolide as a contraceptive agent and to contribute further understanding of its effects on reproduction and behavior in free-ranging female elk. Our specific objectives were to determine in elk : 1) the effectiveness ofleuprolide in preventing pregnancy, 2) the effects ofleuprolide on reproductive behavior, 3) the effects ofleuprolide on body condition, and 4) the reversibility ofleuprolide treatments. �60 MATERIALS AND METHODS Study Area Investigations were conducted in Rocky Mountain National Park and adjacent Estes Valley on the east slope of the Continental Divide between 2000 and 2800 m elevation. Experimental elk were selected from one of two subpopulations that historically wintered in Moraine Park/Beaver Meadows or Horseshoe Park (Bear 1989). Experimental Procedures During late summer and early fall of 2002, 34 adult female elk were immobilized by darting, from the ground, with 3.0 mg of carfentanil citrate (Wildlife Pharmaceuticals, Fort Collins, Colorado, USA) and 10-20 mg xylazine hydrochloride (Rompun; Bayer AG, Leverkusen, Germany). In order to insure that reproductive failure, if it occurred, was due to contraceptive effects rather than the effects of age or diminished body condition, we attempted to select only adult females of prime reproductive age and in moderate to excellent body condition. We hoped to accomplished this in 2 ways: 1) before immobilization, we made a visual assessment of the target animal using age (calf, yearling, adult) and relative fatness and body musculature (condition). Animal condition was classified as good, medium or poor (Riney 1960) and only medium or good condition females were selected, and 2) once the animal was immobilized we estimated age using tooth wear and replacement (Quimby and Gaab 1957), lactational status, and body condition using ultrasonography (Cook et al. 2001). Captured elk were fitted with frequency-specific transmitters on neck collars containing a plastic identification sleeve marked with a unique alpha-numeric code of 76 mm-high black characters on a colored background (white for controls; yellow for treatment)(Freddy 1993). To meet U.S. Food and Drug Administration regulations, all immobilized animals were marked to prevent human consumption. Radio collars were marked with "Do Not Consume". Once sedated, female elk received a subcutaneous, sustained release leuprolide formulation (32.5 mg) using the ATRIGEL ® drug delivery system (Atrix Laboratories, Inc., Ft. Collins, CO, USA) (Dunn et al. 1994). We reversed the effects of the immobilizing drug with 300 mg of naltrexone HcL (Wildlife Pharmaceuticals, Fort Collins, Colorado, USA). To minimize any possibility of infection from immobilization, each darted elk also received a subcutaneous injection of long-lasting penicillin. We collected blood (20 ml) from each elk as baseline information for health parameters. Blood was archived by veterinarians with the National Park Service (NPS). Measurements Reproductive rates: We assessed the effects of leuprolide treatments on reproduction in elk using 4 methods: pregnancyspecific protein B (PSPB) (Noyes et al. 1997), serum progesterone (P4) (Willard et al. 1994), rectal palpation (Greer and Hawkins 1967) and fecal progesterone metabolites (FPM) (Garrott et al. 1998). We determined pregnancy status of all treated and untreated elk during late gestation (March- April) by relocating animals using radiotelemetry and recapturing them following the immobilization procedures previously described. Once immobilized, a trained wildlife veterinarian, rectally palpated each female and determined the presence or absence of a gravid uterus. A single blood sample ( 10 ml) was collected via jugular venipuncture from each animal for PSPB(BioTracking, Moscow, Idaho, USA) and P4(Niswender 1973) analysis. At the same time, a single fecal sample was collected for fecal P4determination. �61 Females having fecal P4 levels< 0.9 Og/gm were considered nonpregnant and those. 1.0 Og/gm pregnant. Discrimination for samples with concentrations between 0.90-0.99 Og/gm was regarded as inconclusive. We will evaluate the reversibility of leuprolide treatments during March-April 2004 by using the reproductive measurements described above. Reproductive behavior: We examined the effects of leuprolide on reproductive interactions of male and female elk during 2 time periods; breeding season (defined as the period 15 September to 15 November) and postbreeding season (defined as the period 15 January to 15 March). We used focal animal sampling procedures to sample reproductive behaviors of all experimental elk (Lehner 1996). Behavioral measurements were be made by locating a breeding group containing radio collared/marked elk. Depending on the environmental conditions, topography, available cover, and elk viewing restrictions in RMNP, the observer attempted to approach the group undetected to within 150-500 m. Observations were made with the aid of binoculars and 15-60X spotting scope during morning (0500-0800) and late day (1400-1700). Time-of-day sampling periods were randomly assigned each week using a randomized block design. Each sampling period consisted of at least 2 hours of continuous observations. We combined individual behaviors into 4 general categories: male copulatory, male precopulatory, female precopulatory, and general breeding (Table 1). Our experimental unit for analyses was the individually marked female in each breeding group. Because sexual interactions were generally short duration(< 30 sec) relative to sampling interval, we recorded the number of occurrences of each event rather than length of time and calculated sexual interaction rates as behaviors per animal per hour. Body condition: Recent research has correlated measures of body condition, using ultrasonography of body fat deposits, to reproductive success in elk (Cook et al. 2001 ). Using these predictive models, we estimated the body condition of all female elk using body condition scoring and ultrasonography of fat and lean body mass. We classified each female as either excellent, very good, moderate, low, or very low reproductive candidates. We selected only those females that were judged to be, at least, in the moderate (10-15 % body fat;> 90% pregnancy rate) category. Elk that met this criteria were randomly assigned to either treatment or control groups; elk that did not, were rejected from the experiment. Additionally, we measured change in rump fat and lean body of females between fall capture and spring re-capture to evaluate the effects of leuprolide treatments on body condition. Statistical analysis: Reproductive rates. In previous experiments, a sample size of 5 treated and 5 control elk was sufficient to detect significant differences (P. 0.05) in pregnancy rates of captive animals (Baker et al. 2002). However, free-ranging elk are more elusive than their captive counterparts and treatment application and measurements of response variables less certain. Uncontrolled variables such as natural mortality, hunting mortality, low pregnancy rates, relocation success, and transmitter failure increase the need for larger sample sizes. We performed a sample size analysis with Fisher's Exact Test, using a software program (NCSS PASS 2000) to estimate the number of treated and control animals needed to detect treatment differences for PSPB, fecal progesterone metabolites, and calving rates (Table 2). For PSPB and fecal progesterone metabolites, we assumed the lowest reported pregnancy rate (63 %) for elk in RMNP (Johnson and Monello, unpublished data), 90 % recapture of radio collared females, and 100% accuracy of PSPB for pregnancy determination in elk greater than 100 days of gestation (Huang et al. 2000). For estimating �62 sample sizes for calving rates, we assumed 63 % pregnancy rates and an 85 % success in confirming presence or absence of a calf. Results of these analyses indicated that a sample size as low as 10 treated and 10 control females would be sufficient to detect a significant treatment effect using PSPB, and serum and fecal P4 values. Reproductive behavior: We tested specific reproductive behavior hypotheses that mean behavior rate was not different between treatment and control groups for both breeding and postbreeding seasons using an ANOV model with repeated measures structure. Time was treated as a within subject effect using a multivariate approach to repeated measures (Morrison 1976). To test for treatment effects, we accounted for time-of- day effects, date effects, and their interactions. PROC GENMOD (SAS Institute 1993) was used to estimate and test for differences in mean behavior rate by treatment, time- of- day, and date. Means and standard errors were estimated using least squares, and hypothesis tests were be based on type III generalized estimating equations that accounted for correlation in repeated measures. Table 1. Description of elk reproductive behaviors and associated behavior categories. Behavior category Reproductive behavior Reproductive: General Breeding Male directed behavior related to establishing, maintaining, and defending a group or harem of female wapiti Male pre-copulatory Male courtship behavior directed toward an individual female to induce or detect oestrus or ovulation (e.g. urine testing, flehmen, tongue flick, lick , smell, or rub female's body, chivy) Female pre-copulatory Female courtship behavior directed toward dominant male to arouse copulatory behavior (e.g. lick and rub male, mount, lordosis, twitch hocks) Copulatory Male behavior directed toward a receptive female in oestrus (e.g. precopulatory mounts, intromission, pelvic thrust) Non-Reproductive: Feeding Head down in vegetation Idling Bedded or standing upright and not feeding Moving Ambulating �63 Table 2. Sample size estimates and power of the test for measurements of reproductive rates in female elk in RMNP. Measurement Treatment (n) Control (n) a 1-f3 PSPB/Fecal P l. 10 10 0.05 0.9386 2. 10 20 0.05 0.9890 3. 10 120 0.05 0.9996 l. 10 20 0.05 0.8613 2. 20 20 0.05 0.9685 3. 20 25 0.05 0.9829 4. 20 30 0.05 0.9865 Calving rates RESULTS AND DISCUSSION Fall: 2002 We captured, sampled, and radio collared 34 female elk in RMNP during 24 August - 7 September, 2002 Elk were captured from 5 general locations in the RMNP : Kawuneeche Valley (7), alpine tundra areas near Trail Ridge Road (4), Hidden Valley (3), Beaver Meadows (9), and Moraine Park (11) Seventeen females were given a subcutaneous formulation containing 32.5 mg of leuprolide and seventeen a placebo formulation without leuprolide. No capture-related mortalities were observed. Estimated ages of leuprolide-treated females ranged from 1-12 years of age~(= 6.9, SE= 0.82) and 1-10 years of age ( = 6.3, SE= 0 72) for untreated elk Two yearling females were included in both groups. Yearling females were included as experimental animals because they met a priori body composition criteria. and because we wanted additional information on the effects of leuprolide in this age group. Seventy percent of treated females were determined to be lactating when captured compared to 61 % of control females. Fall body condition of leuprolide-treated and control females were similar for rump fat depth ( P = 0 .56), loin depth (P = 0.91), and body condition score (BCS) (P 0.38) (Table 3). Rump fat percent of leuprolide-treated females ranged from 8.8 - 16.3 %~( = 13.1 % , SE= 0.40) and from 10.6 - 15.9 %~( = 12.7 %, SE= 0.38) for control elk. With the exception of one animal, all females in the experiment had a rump fat percentage of greater thanlO % (> 90 % pregnancy rate). �64 Table 3. Mean fat depth, percent rump fat, body condition score, and loin depth of leuprolide-treated and control female elk sampled during Aug-Sept, 2002 and Mar-Apr, 2003, in Rocky Mountain National Park, Colorado. Leuprolide Measurements Control Mean SE Mean SE 2.13 5.43 3.53 13.10 0.18 0.12 0.12 0.40 2.00 5.41 3.38 12.73 0.11 0.10 0.11 0.38 0.37 4.84 2.36 6.90 0.04 0.08 0.12 0.04 0.72 5.00 2.48 8.20 0.12 0.11 0.13 0.49 Fall (Aug-Sept 2002): Rump fat depth (cm) Loin depth (cm) Body condition score Rump fat(%) Spring (Mar-Apr 2003): Rump fat depth (cm) Loin depth (cm) Body condition score Rump fat(%) Fall - Spring ~ Rump fat depth (cm) Loin depth (cm) ~ Body condition score ~ Rump fat (%) ~ - 1.76 - 0.59 - 1.17 -6.20 - 1.28 - 0.41 -0.90 -4.50 We observed reproductive behaviors of treated and control elk in RMNP and Estes Valley during 11 September to 27 November, 2002. We recorded a total of 144, one hour observations for 16 different radio collared female elk (8 treated; 8 control). No copulatory behaviors were observed during this period, thus there was no analysis for this category. There were no differences in reproductive behavior rates (number of behaviors/hour) for general breeding (P = 0.36), female precopulatory (P = 0.13), or male precopulatory (P = 0.70) behaviors (Fig. I). In general, control females showed somewhat higher rates of general breeding (25 % higher than treated females) and male precopulatory (9 % higher than treated females) behaviors, but none of these differences were statistically significant. In addition to reproductive behaviors, we evaluated the effects of leuprolide on the daily activity patterns of treated and control female elk. These data are currently being analyzed. �65 11m Control Leuprolide a '1.00 i ............ ··•.•·· a a ,._ 0,25 0.00 I Behavior category Figure 1. Mean(± SE) reproductive behavior rates during the breeding season for control female elk (n = 8) and females treated with a sustained release implant containing 32.5 mg leuprolide formulation (n = 8), in Rocky Mountain National Park, Colorado. Columns with different lower case letters indicate significant differences between means (P 0.05). Spring 2003 During 24 March to 30 April, 2003, we evaluated the effects ofleuprolide on pregnancy rates, body condition, and reproductive behavior of treated and control female elk. Using the capture methods previously described, we recaptured 15 out of 17 treated elk and 17 out of 17 control elk. Elk were recaptured in 3 general locations: RMNP (13), Estes Park, Colorado area (16), and Loveland, Colorado area (3). Leuprolide, administered as a sustained release formulation, prior to the breeding season, effectively prevented pregnancy in all female elk for one year. Pregnancy rates of untreated females ranged from �66 64. 7 - 78.5 %, depending on the method of determination. Fecal P4 analyses for pregnancy determination have not been completed. Body condition of experimental elk declined for all measures of body composition during fall 2003 and spring 2004 (fable 3). Changes in mean loin depth (P. 0.25) and body condition score (P. 0.08) were similar for both treated and control female elk, whereas, overwinter loss in mean percent rump fat was greater (P. 0.057) for elk treated with leuprolide. Post-breeding season reproductive behaviors and daily activity patterns of control and leuprolide-treated females are currently being analyzed. SUMMARY To date, we have completed or are in the process of completing 3 out of the 4 objectives originally stated for this investigation. First, we have evaluated the effects of leuprolide on pregnancy rates of female elk using rectal palpation, PSPB, and P4 analysis and all methods support the conclusion that leuprolide is 100% effective in preventing pregnancy for at least one breeding season. The only remaining analysis for pregnancy determination is fecal P4, which will be completed during winter 2004. Second, we have evaluated the effects of leuprolide on breeding and post-breeding reproductive behavior of elk. Although neither of these data sets have been completely analyzed, leuprolide does not appear to have deleterious effects on elk reproductive behavior or daily activity patterns. Third, we assessed the effects of leuprolide on body condition dynamics of elk. We observed only minor differences in overwinter body composition changes between treated and control elk. The only objective yet to be completed is to confirm the reversibility of leuprolide treatments. This will be accomplished during March-April 2004 by comparing pregnancy rates of treated and control female elk. LITERATURE CITED Baker, D. L., M. A. Wild, M. M. Conner, H. B. Ravivarapu, R. L. Dunn, and T. M. Nett. 2002. Effects of GnRH agonist (leuprolide) on reproduction and behavior in female wapiti (Cervus elaphus nelsoni). Reproduction (Suppl) 60:155-167. Bear, G.D. 1989. Seasonal distribution and population characteristics of elk in the Estes Valley, Colorado. Colorado Division of Wildlife R-S-65-89, Special Report 65. Becker, S. E., and L. S. Katz. 1995. Effects of gonadotropin-releasing hormone agonist on serum LH concentrations in female white-tailed deer. Small Ruminant Research 18:145-150. Belchetz, P. E., T. M. Plant, Y. Nakai, E. J. Keogh, and E. Knobil. 1978. Hypophysial response to continuous and intermittent delivery of hypothalamic gonadotropin-releasing hormone. Science 202: 631-633. Bergfeld E., M. J. D'Occhio, and J.E. Kinder. 1996. Pituitary function, ovarian follicular growth, and plasma concentrations of 17~-oestradiol and progesterone in prepubertal heifers during and after treatment with the luteinizing hormone-releasing hormone agonist deslorelin. Biology of Reproduction 54: 776-782. Brown, J. L., D. L. Schmitt, A. Bellem, L. H. Graham, and J. Lehnhardt. 1999. Hormone secretion in the Asian elephant (E/aphus maximus): Characterization of ovulatory and anovulatory LH surges. Biology of Reproduction 61:1294-1299. �67 Cole, G. F. 1971. An ecological rationale for the natural or artificial regulation of native ungulates in parks. Transactions of the North American Wildlife and Natural Resources Conference 36:417426. Cook, R. C., J. G. Cook, and D. L. Murray. 2001. Development of predictive models of nutritional condition for Rocky Mountain elk. Journal of Wildlife Management 65:973-987. D'Occhio, M. H., W. J. Aspden, and T. Whyte. 1996. Controlled reversible suppression of oestrous cycles in beef heifers and cows using agonist of luteinizing hormone - releasing hormone. Journal of Animal Science 74:218-225. _ _ __, M. J., and W. J. Aspden. 1999. Endocrine and reproductive responses of male and female cattle to agonist of gonadotrophin-releasing hormone. Journal of Reproduction and Fertility Supplement 54: 101-114 DeNicola, A. J., D. J. Kesler, and R. K. Swihart. 1997. Remotely delivered prostaglandin F- 2 alpha implants terminate pregnancy in white-tailed deer. Wildlife Society Bulletin 25:527-531. Dunn, R. L., J.P. English, D.R. Cowan, and D. P. Vanderbilt. 1994. Biodegradable in situ forming implants and methods of producing the same. U. S. patent no. 5,278,201. Evans, A. C. 0., and N. C. Rawlings. 1994. Effects of a long-acting gonadotrophin-releasing hormone agonist (Leuprolide) on ovarian follicular development in prepubertal heifer calves. Canadian Journal of Animal Science 74:649-656. Fagerstone, K. A, M. Coffey, P. Curtis, R. Dolbeer, G. Killian, L.A. Miller, and L.Wilmot. 2001 Wildlife contraception. Wildlife Society Technical Review. Proceedings of the Wildlife Society 8th Annual Conference, Reno, USA Freddy, D. J. 1993. Estimating survival rates of elk and developing techniques to estimate population size. Colorado Division of Wildlife Report, July. Garrott, P. J. White, and C. A. Vanderbilt White. 1993. Overabundance: an issue for conservation biologist? Conservation Biology 7:946-949. - - - - ~ and D. B. Siniff. 1992. Limitations of male-oriented contraception for controlling feral horse populations. Journal of Wildlife Management 56:456-464. • _ _ _ _ , S. L. Monfort, P. J. White, K. L. Mashburn, and J. G. Cook. 1998. One-sample pregnancy diagnosis in elk using fecal steroid metabolites. Journal of Wildlife Management 34: 126-131. Huang, F., D. C. Cockrell, T. R. Stephenson, J. H. Noyes, and R. G. Sasser. 2000. A serum pregnancy test with specific radioimmunoassay for moose and elk pregnancy specific protein B. Journal of Wildlife Management 64:492-499. Hess, K. H. 1993. Rocky Times in Rocky Mountain National Park. University of Colorado Press. Niwot Colorado, USA. Hobbs, N. T., D. C. Bowden, and D. L. Baker. 2000. Effects of fertility control on populations of ungulates: general, stage-structured models. Journal of Wildlife Management 64: 473-491. �68 Houston, D. B. 1971. Ecosystems of national parks. Science 172: 648-651. Jacobsen, N. K., D. A. Jessup, and D. J. Kesler. 1995. Contraception in captive black-tailed deer by remotely delivered norgestomet ballistic implants. Wildlife Society Bulletin 23:718-722. Jewell, P.A., and S. Holt. 1981. Problems in management of locally abundant wild animals. Academic Press. 32lpp. Lehner, P. N. 1996. Handbook of Ethological Methods. Second Edition. Cambridge University Press, Cambridge, UK. Kirkpatrick, J. F., and J. W. Turner, Jr. 1985. Chemical fertility control and wildlife management. Bioscience 35:485-491. - - - - , I. K. M. Lui, R. C. Fayrer Hosken, and A. T. Rutberg. 1997. Case studies in wildlife immunocontraception: wild and feral equids and white-tailed deer. Reproduction, Fertility, and Development 9: 105-110. McCullough, D.R., K. W. Jennings, N. B. Gates, B. G. Elliot, and J.E. Didonato. 1997. Overabundant deer populations in California. Wildlife Society Bulletin 25:478-483. McNeilly, A. S., and H. M. Fraser HM. 1987. Effect of gonadotrophin-releasing hormone agonist-induced suppression of LH and FSH on follicle growth and corpus luteum function in the ewe. Journal of Endocrinology 115: 273-282. Montovan, S. M., P. P. Daels, J. Rivier, G. H. Hughes, G. H. Stabenfeldt, and B.L. Lasley. 1990. The effect of potent GnRH agonist on gonadal and sexual activity in the horse. Theriogenology 33:1305-1321. Morrison, D. F. 1976. Multivariate Statistical Methods. McGraw-Hill Book Co., New York, USA. Quimby, D. C., and J.E. Gaab. 1957. Mandibular dentition as an age indicator in Rocky Mountain elk. Journal of Wildlife Management 21: 134-15 3. Riney, T. 1960. A field technique for assessing physical condition of some ungulates. Journal of Wildlife Management 24:92-94. SAS Institute. 1993. SAS/STAT user's guide. Release 6.03 edition. SAS Institute, Cary, North Carolina, USA. Seagle, S. W., and J. D. Close. 1996. Modeling white-tailed deer population control by contraception. Biological Conservation 76:87-91. Singer, F. J., L. C. Zeigenfuss, R. G. Cates, and D. T. Barnett. 1998. Elk, multiple factors, and persistence of willows in national parks. Wildlife Society Bulletin 26:419-428. White, C. A., C. E. Olmsted, and C. E. Kay. 1998. Aspen, elk, and fire in the Rocky Mountain national parks of North America. Wildlife Society Bulletin 26:449-462. �69 Willard, S. T., R. G. Sasser, J.C. Gillespie, J. T. Jaques, T. H. Welsh, Jr., and R. D. Randel. 1994. Methods for pregnancy determination and the effects of body condition on pregnancy status in Rocky Mountain elk (Cervus elaphus nelsoni). Theriogenology 42: 1095-1102. White, C. A, C. E. Olmsted, and C. E. Kay. 1998. Aspen, elk, and fire in the Rocky Mountain national parks of North America. Wildlife Society Bulletin 26:449-462. Willard, S. T., R. G. Sasser, J.C. Gillespie, J. T. Jaques, T. H. Welsh, Jr., and R. D. Randel. 1994. Methods for pregnancy determination and the effects of body condition on pregnancy status in Rocky Mountain elk (Cervus elaphus nelsoni). Theriogenology 42: 1095-1102. Zeignefuss, L., F. J. Singer, and D. C. Bowden. Vegetation and elk responses following release from artificial controls within Rocky Mountain National Park, 1968-1996. Report to the Superintendent, Rocky Mountain National Park. �Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Project No. Work Package No. Task No. Colorado W-153-R 3002 RMNP Federal Aid Project: : : : : : Cost Center 3430 Mammals Research Elk Conservation Technical Support for Elk and Vegetation Management Environmental Impact Statement for Rocky Mountain National Park Period Covered : July 1, 2003 - June 30, 2004 Authors : D. L. Baker, M. A. Wild, M. D. Hussain, R. L. Dunn, T. M. Nett Personnel: E. Jones, J. Ritchie, A. Mitchell, X. Sha, M. Allen, J. Powers. All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT Practical application of fertility control technology in free-ranging wild ungulates requires remote delivery of a safe and efficacious contraceptive agent. The objective of this investigation was to evaluate the potential of a remotely delivered, sustained release, biodegradable implant formulation of leuprolide acetate, to achieve reversible suppression of ovulation and fertility in female elk (Cervus elaphus nelsoni). Fifteen, captive adult female elk were randomly allocated to one of three experimental groups. Six elk were injected intramuscularly with a dart containing the implant formulation of leuprolide, and the remaining nine elk received the same formulation without leuprolide. We measured pregnancy rates, suppression of luteinizing hormone (LH) and progesterone concentrations, and reversibility of leuprolide treatments during 1 August 2002 to 3 September 2003. The sustained release implant formulation, remotely administered by dart, resulted in decreased concentrations of LH and progesterone, temporary suppression of ovulation and steroidogenesis, and effective contraception (100%) for one breeding season. These results extend the potential for practical application of the leuprolide implant as contraceptive agent in female elk, where in the absence of such technology, wild elk must first be captured and restrained prior to treatment. 45 �JOB PROGRESS REPORT TECHNICAL SUPPORT FOR ELK AND VEGETATION MANAGEMENT ENVIRONMENTAL IMPACT STATEMENT FOR ROCKY MOUNTAIN NATIONAL PARK D. L. Baker, M. A. Wild, M. D. Hussain, R. L. Dunn, and T. M. Nett P. N. OBJECTIVE Conduct experiments with captive and free-ranging elk to evaluate fertility control as an management alternative for controlling elk populations in Rocky Mountain National Park (RMNP), Colorado. SEGMENT OBJECTIVES 1. 2. 3. Determine the effectiveness of a remotely delivered intramuscular leuprolide implant in preventing pregnancy in captive female elk. Determine the duration of effectiveness of remotely delivered leuprolide implant (if any) on luteinizing hormone (LH) and progesterone secretion in captive female elk. Determine the reversibility of remotely delivered leuprolide implant on infertility (if achieved) in captive female elk. INTRODUCTION Fundamental to practical application of contraceptives to wildlife, is a safe and effective antifertility agent that can be remotely delivered to the target species. To attain this goal, considerable research has focused on the development and testing of ballistic systems and controlled drug release formulations that can remotely administer contraceptive agents to wild ungulates (Kreeger, 1997). Contraceptive agents have been delivered via projectile dart or biodegradable implant to a variety of wild ungulate species including deer (Odocoileus spp.) (Turner et al., 1992; Jacobsen et al., 1995; DeNicola et al., 1997), elk (Cervus elaphus nannodes) (Shideler et al., 2002), wild horses (Equus caballus) (Kirkpatrick et al.,1990), burros (Equus asinus) (Turner et al., 1996), and elephants (Loxodonta africana) (Delsink et al., 2002). However, to date, no contraceptive agent that possess all of the desired attributes (Fagerstone et al., 2002) has been developed for remote delivery. The use of GnRH agonist implants to suppress short-term ovarian follicular growth and ovulation are well documented for a number of species including cattle (McLeod et al., 1991, D’Occhio et al., 1996), sheep (McNeilly and Fraser,1987), monkeys (Fraser et al., 1987), and humans (Broekmans et al., 1996). However, few studies have established the efficacy of these agents for long-term suppression of ovarian activity and contraception (Trigg et al., 2001; Baker et al., 2002, 2004; D’Occhio et al., 2002) and to our knowledge, none have previously demonstrated effective contraception by dart delivery of the implant. In previous research, we administered gonadotropin releasing hormone (GnRH) agonist leuprolide acetate by hand injection to captive female elk (Cervus elaphus nelsoni) (Baker et al., 2002), and mule deer (Odocoileus hemionus hemionus) (Baker et al. 2004), as a sustained release injectable implant, and achieved 100 % infertility for one breeding season. The implant formulation consisted of 45 % w/w 75/25 poly (DL-lactide-co-glycolide) (PLG) polymer having an intrinsic viscosity of 0.20 dL/g dissolved in N-methyl-2-pyrrolidone (NMP) and containing 6 % w/w leuprolide in the polymer solution. This formulation was designed to release the drug for a period of 3 to 4 months after subcutaneous injection (Ravivarapu et al., 2000). 46 �In these previous studies, the leuprolide formulation was demonstrated to be highly effective when delivered subcutaneously, however, it’s not known if similar effectiveness can be achieved when administered as an intramuscular (IM) injection via dart. Differences in drug pharmacokinetics and metabolism between muscle and subcutaneous tissues could affect release dynamics of the implant and possibly decrease the antifertility properties of leuprolide. Therefore, the objectives of this experiment were to determine in captive female elk (1) the effectiveness of this remotely delivered intramuscular leuprolide implant in preventing pregnancy, (2) the duration of effects (if any) on luteinizing hormone (LH) and progesterone secretion, and (3) the reversibility of infertility (if achieved). MATERIALS AND METHODS Experimental animals During 1 August 2002 to 3 September 2003, we evaluated the effects of remotely delivered leuprolide formulation on pregnancy rates, luteinizing hormone (LH), and progesterone secretion in captive female elk. Controlled experiments were conducted with 15 adult females (2-14 years of age; 220 - 275 kg BW), two intact adult male elk (3 years of age; 350-400 kg BW), and one epididymectomized adult male elk (3 years of age; 340-375 kg BW) at the Colorado Division of Wildlife’s Foothills Wildlife Research Facility in Fort Collins, Colorado, USA. Captive elk used in this experiment were permanently maintained at this facility and were trained to repeated handling, weighing, blood sampling techniques, and isolation pens. When not involved in the periodic intensive sampling procedures required by this study, elk were maintained in fenced pastures (5 ha) containing native vegetation and fed a diet consisting of ad libitum quantities of grass-alfalfa hay, grain supplement, trace mineral block, and water. In an effort to induce normal cyclic ovulatory responses and synchronize estrus, we released an epididymectomized male elk with 15 seasonally anovulatory female elk on 20 July 2002 (McComb, 1987). Four weeks later (21 August) and prior to assigning elk to experimental treatments, we assessed the reproductive status of each female by: 1) manual rectal palpation of the reproductive tract to diagnose ovarian status and identify any abnormalities, and 2) measuring the responsiveness of pituitary gonadotropes to an exogenous dose of GnRH analog. Females showing evidence of reproductive tract abnormalities or suppressed gonadotrope function were excluded from the experiment. Experimental design Fifteen female elk were randomly assigned to one of three experimental groups. Six elk (group A) were injected with a dart containing the polymeric matrix formulation of leuprolide acetate (D-Leu6GnRH-Pro9-ethylamide). Four elk (group B) were designated as pregnant controls. They received the polymer solution without leuprolide and were used to compare the effects of leuprolide formulation on pregnancy rates between treated and untreated elk. These two groups of elk were maintained together in the same pastures with two intact, adult male elk from 13 September 2002 to 10 April 2003. The remaining five elk (group C) served as non-pregnant controls and were placed in a separate pasture (2 ha) without direct contact with male elk. We compared concentrations of LH and progesterone of these females to those treated with leuprolide formulation (group A). Non-pregnant control females (group C) provided a more representative comparison to treated elk for evaluating treatment-induced hormonal responses than potentially pregnant elk, thus the need for two separate control groups. Treatments Leuprolide implant formulation. The polymer, 85/15 poly (DL-lactide-co-glycolide) (PLG) with intrinsic viscosity 0.31 dL/g (Absorbable Polymer Technologies, Pelham, Alabama, USA) and N-methyl2-pyrrolidone (NMP, International Speciality Products, Wayne, New Jersey, USA) were mixed in a ratio of 50:50 in a vial until the polymer was completely dissolved. The polymer solution was sterilized by γirradiation at a dose of approximately 25Gy (Isomedix, Morton Grove, Illinois, USA) and an appropriate amount of the sterilized polymer solution was filled into 1.2 luer-lock female syringes. For the leuprolide 47 �part of the system, calculated volume of filtered aqueous solution of leuprolide acetate (Mallinkrodt, St. Louis, Missouri, USA) was filled in 1-mL male syringe barrels (Becton-Dickenson, Franklin Lakes, New Jersey, USA) and lyophilized. This formulation was designed to deliver a 32.5 mg dose of leuprolide at a controlled rate over a 180-day therapeutic period. A similar formulation was previously shown to suppress ovulation and pregnancy for one breeding season in captive elk when delivered subcutaneously by hand-injection (Baker et al., 2002). Treatment application. On the day before treatment application (6 September 2002), experimental elk were moved from holding paddocks to individual isolation pens (5 m x 10 m), weighed (± 0.5 kg), sedated with xylazine hydrochloride (Rompun; Bayer AG, Leverkusen; 25-200 mg/animal, IM) and fitted nonsurgically with indwelling jugular catheters. The next day, and just prior to injection, separate syringes containing the polymer and the leuprolide were connected and the contents mixed with 60 back and forth mixing cycles. The resulting homogenous dispersion was drawn into the male syringe, and the formulation was transferred into single use, 1 ml, 13-mm-diameter, barb-less darts equipped with gel-collared 32-mm-long needles (Pneu-dart, Williamsport, Pennsylvania, USA). The final concentration of leuprolide was 12 % in the homogenous mixture of polymer solution and leuprolide acetate after mixing and was designed to deliver approximately 32.5 mg of leuprolide acetate to the elk. Control elk received only the polymer solution processed the same way but without leuprolide. Prior to darting, individual elk were placed in a handling chute and lightly sedated with intravenous (IV) xylazine hydrochloride (15-20 mg/animal). This dose allowed animals to remain standing in the chute and minimized excitation associated with discharge of the dart gun. All elk were remotely injected with a dart fired from a CO2-powered pistol (DanInject™ , Wildlife Pharmaceuticals, Fort Collins, Colorado, USA). In order to accurately determine the precise dose of leuprolide formulation delivered to each elk, darts were weighed before and after injection. With the exception of two animals, one dart per animal was fired from approximately 3 meters into the area of the biceps femoris muscle of the standing elk. In two animals, the dart failed to discharge or only partially injected the prescribed dose. In these cases, we re-weighed and fired additional darts until the complete dose was delivered to each animal. Once all elk had been treated, sedation was reversed with yohimbine (30 mg, IV) (Antagonil®, Wildlife Laboratories, Fort Collins, Colorado, USA) and animals were returned to individual isolation pens. Measurements 24 h LH response to leuprolide treatment. Immediately following application of treatments to groups A and group C, we determined the amount of LH released during the initial 24 h of the treatment period. Blood samples (5 ml) were collected via jugular catheters at 0, 120, 180, 240, 300, 360, 480, 600, 960, and 1440 min after drug injection. Catheters were flushed after each collection with sterile saline solution. After the last blood collection, catheters were removed and animals were returned to holding paddocks. Eight days later, two intact male elk were placed into the same pasture with these females. Duration of LH and progesterone response to leuprolide treatment. The effect of leuprolide formulation on the duration of suppression of LH and progesterone was determined by periodically conducting pituitary stimulation trials. These trials were performed prior to treatment application as an aid in the selection of animals for this experiment and periodically during 29 October 2002 to 3 September 2003 to determine pituitary responsiveness to an exogenous dose of GnRH analog (D-Ala6-GnRH-Pro9ethylamide; Sigma Chemical Company, St. Louis, Missouri, USA). Pituitary stimulation trials were conducted with elk in groups A and C elk at 50, 100, 150, 185, 215, and 361 days post-treatment. The final stimulation trial (3 September 2003) provided hormonal evidence of the reversibility of leuprolide treatment. Stimulation trials were conducted according to the 48 �following procedures: On the day of testing, elk from groups A and C were moved from 5 ha pastures to individual isolation pens, weighed, sedated (as previously described), and fitted nonsurgically with indwelling jugular catheters. A bolus dose of GnRH analog (1 g/50 kg body weight) was administered through the cannula and blood samples (5 ml) were collected at 0, 60, 120, 180, 240, 300, 360, and 480 min post-administration. After collections, blood was stored at 4 C for 24 h until serum was obtained by centrifugation (1500 RCF for 15 min). Serum for progesterone analysis was obtained from the 0 h blood sample for each animal on each of the trial days. Serum was stored at -20 C until analyzed for LH and progesterone. Following the last blood collection, catheters were removed, and elk were returned to holding pastures. Reproductive response to leuprolide treatment - The effect of leuprolide formulation on reproduction in groups A and B was determined in two ways : (1) by measuring pregnancy rates using the presence or absence of pregnancy specific protein B (PSPB) (BioTracking, Moscow, Idaho, USA) in serum collected at approximately 100 and 215 days of gestation (Huang et al., 2000), and (2) by observing the presence or absence of calves the following summer. Analyses Serum concentrations of LH were quantified by means of an ovine oLH radioimmunoassay (Niswender et al., 1969). Elk serum was demonstrated to inhibit binding of 125 I-labeled oLH to LH antiserum in a manner that paralleled the standard (NIH- oLH-S24). Similarly, when different quantities of oLH standard were added to elk serum and samples were subjected to radioimmunoassay, the values obtained were increased by the quantity of oLH added (r2 = 0.99, slope = 0.92, β1 = 0.22, P = 0.002). These data indicated that the radioimmunoassay provided a quantitative assessment of LH in elk serum. The limit of sensitivity of the LH assay was 0.02 ng /ml. Serum concentrations of progesterone were also determined by radioimmunoassay (Niswender, 1973). Sensitivity of the progesterone assay was 0.12 ng /ml. Intra-and-inter assay coefficients of variation for each of these assays were < 10 %. Hormone concentrations are reported as untransformed arithmetic means (± SE). Responsiveness of the pituitary gland to GnRH analog stimulation was determined by the total amount of LH secreted (ng /ml/ min) which was estimated by calculating the area under the LH response curve (Abramowitz and Stegun, 1968). Differences among hormone concentrations were tested using least squares ANOVA for general linear models (SAS Institute, 1997). Responses to treatment were analyzed with one-way ANOVA for a randomized complete block design with repeated measures. Treatment effects were determined using the total animal-within-treatment variances as the error term. Time was treated as a within-subject effect, using a multivariate approach to repeated measures (Morrison et al., 1976). A “protected” least significant difference test (Milliken and Johnson, 1984) was used to separate means when the overall F- test indicated significant treatment effects (P < 0.05). RESULTS Intramuscular injection of leuprolide formulation via dart, was 100 % effective in suppressing ovulation and preventing pregnancy in captive female elk for one breeding season. All leuprolide treated females (group A) tested negative and untreated controls (group B) positive for PSPB at approximately 100 and 215 days of gestation. No calves were born to treated elk, whereas the calving rate of untreated elk was 100 %. The amount of leuprolide acetate delivered to each elk ranged from 22. 6 to 38.1 mg ( = 33.1, SE = 2.4). We did not observe any unusual bleeding, swelling or trauma at the injection site nor did any of the elk show evidence of impaired mobility or post-treatment tissue necrosis or abscesses related to dart delivery of the bioimplant. Of particular interest was that the lowest individual dose delivered (22.6 mg) was equally as effective as higher doses in suppressing hormone concentrations 49 �and pregnancy, suggesting that the minimum effective dose in elk could be substantially lower than the estimated dose (32.5 mg) used in this experiment. Mean serum concentrations of LH increased (P = 0.015) in treated elk (group A) within 2 h of drug injection, peaked at 63.12 ± 10.8 ng/ml (mean ± SE) 4.3 ± 0.65 h (mean ± SE) later, then gradually declined to baseline levels by 16 h post-treatment (Fig. 1). Levels of LH in group A were greater (P = 0.032) than those of untreated controls (group C) for 2- 10 h post-treatment, after which, values decreased to baseline levels and were similar (P = 0.285) for both groups. Results of periodic GnRH challenges revealed that the leuprolide formulation reduced pituitary content of LH to basal concentrations for at least 215 days post-treatment, which was 35 days longer than the expected 180-day delivery period (Fig. 2). Concentrations of GnRH analog-induced LH secretion were lower (P = 0.022) in leuprolide - treated elk (group A) compared to non-pregnant controls (group C) at days 50, 150, 185, and 215 days after treatment. Chronic suppression of LH in treated females was followed by a return to pretreatment levels, indicative of estrus, prior to the subsequent breeding season (September 2003, Fig. 3). In contrast to leuprolide-treated elk, pituitary responsiveness of untreated elk (group C) to GnRH analog were elevated and relatively similar (P = 0.64) in magnitude during the first 185 days of the experiment, after which, these levels declined (P = 0.087), presumably with the onset of seasonal anestrus (March). Similar (P = 0.582) to treated elk, pituitary responsiveness in control elk (group C) returned to pretreatment levels in September 20003. Serum concentrations of progesterone in leuprolide - treated females (group A) followed a parallel pattern to that observed for serum LH (Fig. 3). The suppressive effects of leuprolide on corpus luteum formation and steroidogenesis was readily apparent by its effects on serum progesterone concentrations in treated elk compared to controls (group C). Progesterone levels in treated elk declined (P = 0.017) to limits of detection by 50 days post-treatment and remained at those levels for the duration of the breeding period. For untreated elk (group C), serum progesterone was more variable and consistently higher (P = 0.043) than that for treated elk at 50, 100, 150, 185, and 215 days post-treatment. As evidence of normal estrous cycles and contraceptive reversibility, progesterone concentrations in both treated and untreated elk (group C) returned to pretreatment levels (P = 0.435) at the onset of the following breeding season. DISCUSSION In the present experiment, we evaluated the effectiveness of projectile dart delivery of the GnRH agonist, leuprolide, as a potential antifertility agent in female elk. The sustained release polymeric implant formulation of leuprolide acetate, remotely delivered in a projectile dart, resulted in decreased LH and progesterone secretion, presumably suppression of ovulation and steroidogenesis, and effective contraception (100 %) without adverse effects for one breeding season. The contraceptive effects of leuprolide formulation followed a two-phase process. The first phase was characterized by an acute, transient rise in serum LH which gradually declined to basal concentrations about 16 h post-treatment. The second phase was defined by chronic inhibition of LH and progesterone secretion for the duration of the seasonal breeding period. Subsequently, normal ovarian function and fertility were re-established prior to next breeding season. We conclude from these patterns of LH and progesterone in serum that gonadotropes in female elk are down-regulated during treatment with GnRH agonist. As a consequence, long-term exposure to GnRH agonist resulted in reduction in GnRH receptors on gonadotropes (Clayton, 1989), depletion of pituitary LH and FSH content (Aspden et al., 1996), and elimination of the preovulatory LH surge (Gong et al., 1995; D’Occhio et al., 1996). These responses have been shown to result in ovulation failure and infertility which persists as long as the GnRH agonist is present in circulation at therapeutic levels (Melson et al., 1986; D’Occhio et al., 2000). 50 �Our findings here are consistent with previous observations of acute and chronic responses of sheep (Dobson, 1985), cattle (D’Occhio et al., 1989; Gong et al., 1996), horses (Montovan et al., 1990), deer (Becker and Katz, 1995), and elk (Baker et al.,2002) treated with GnRH agonist. Effective contraception in polyestrous, seasonal breeders is dependent on suppression of ovulation from the beginning of the breeding season to the onset of seasonal anestrous, a period of approximately 200 d in elk. Therefore, the timing of treatment application is an important consideration in successful contraception. Because of the acute rise in LH concentrations that occurs following GnRH agonist treatments, ovulation of growing follicles can be induced (Macmillan and Thatcher, 1991, D’Occhio and Aspden, 1999). Therefore, to ensure effective contraception in female elk, leuprolide treatments should be applied prior to the initiation of seasonal estrus. In the present study, leuprolide inhibited LH secretion and ovulation for at least 215 days which is in close agreement with previous research, in which a subcutaneous dose of leuprolide suppressed LH levels for 190-250 days (Baker et al., 2002). In other studies, implants containing GnRH agonist have been shown to suppress ovarian activity for a minimum of 150 days in mule deer and (Baker et al., 2004) and almost 400 days in cattle (D’Occhio et al., 2002). Persistent suppression of ovarian function, beyond the formulated delivery period of the implant, has been reported for a number of different species. Leuprolide suppressed LH and progesterone levels in elk in this experiment for at least 35 days longer (19 %) than the expected six month effective duration and 30 -110 days longer in deer and elk in previous studies (Baker et al., 2002, 2004). Similar observations of extended gonadotrope suppression were reported previously in cattle (Bergfeld et al., 1996; D’Occhio et al., 1996), monkeys (Fraser et al., 1987), men (Hall et al., 1999), and women (Broekmans et al., 1996). The underlying mechanism for this effect is not completely understood, but it is thought to be a associated with prolonged dysfunction of gonadotrope cells rather than direct action on the ovaries (D’Occhio et al., 2000; Aspden et al., 2003). Regardless of the mechanism involved, the extended suppression of ovarian function, as a consequence of GnRH agonist treatment, is fundamentally essential to effective contraception in deer and elk. In conclusion, intramuscular delivery of the sustained release biodegradable polymeric implant formulation of leuprolide via dart resulted in effective suppression of ovarian function and fertility in female elk for one breeding season with a return to normal reproductive function the following year. These results are particularly important for wildlife applications where, in the absence of such technology, animals must first be captured and restrained prior to treatment. LITERATURE CITED Abramowitz, M., and I. A. Stegun. 1968. Handbook of mathematical functions. Dover Publishing, Inc., New York, New York, 343 pp. Aspden, W. J., A. Rao, P. T. Scott, I. J. Clark, T. E. Trigg, J. Walsh, and M. J. D’Occhio. 1996. Direct actions of the luteinizing hormone -releasing hormone agonist, deslorelin, on anterior pituitary contents of luteinizing hormone (LH) and follicle-stimulating hormone (FSH), LH and FSH subunit messenger ribonucleic acid, and plasma concentrations of LH and FSH in castrated male cattle. Biology of Reproduction 55:386-392. __________ , A. Jackson, T. E. Trigg, and M. J. D’occhio. 2003. Pituitary expression of LHβ-and FSHβsubunit mRNA, cellular distribution of Lhβsubunit mRNA and LH and FSH synthesis during and after treatment with a gonadotrophin-releasing hormone agonist in heifers. Reproduction, Fertility and Development 15:149-156. 51 �Baker, D. L., M. A. Wild, M. M. Conner, H. B. Ravivarapu, R. L. Dunn, and T. M. Nett. 2002. Effects of GnRH agonist (leuprolide) on reproduction and behavior in female wapiti (Cervus elaphus nelsoni). Reproduction (Suppl.) 60: 155-167. ___________ , ___________ , ___________ , __________ , _____________ , _________ . 2004. Gonadotropin releasing hormone agonist: a new approach to reversible contraception in female deer. Journal of Wildlife Diseases 40: (in press). Becker, S. E., and L. S. Katz. 1995. Effects of gonadotropin-releasing hormone agonist on serum LH concentrations in female white-tailed deer. Small Ruminant Research 18:145-150. Bergfeld E. G. M., M. J. D’occhio, and J. E. Kinder. 1996. Continued desensitization of the pituitary gland in young bulls after treatment with the luteinizing hormone-releasing hormone agonist deslorelin. Biology of Reproduction 54:769-775. Broekmans, F. J., P. G. Hompes, C. B. Lambalk, E. Broeders, and J. Schoemaker. 1996. Short-term desensitization: effects of different doses of gonadotrophin-releasing hormone agonist triptorelin. Human Reproduction 11:55-60. Clayton, R. N. 1989. Gonadotropin-releasing hormone: its actions and receptors. Journal of Endocrinology 120:11-19. Delsink, A. K., J. J. van Altena, J. J. Kirkpatrick , and R. A. Fayrer-Hosken. 2002. Field application of immunocontraception in African elephants (Loxodonta africana). Reproduction (Suppl.) 60:117124. DeNicola, A. J., D. J. Kesler, and R. K. Swihart. 1997. Remotely delivered prostaglandin F2 implants terminate pregnancy in white-tailed deer. Wildlife Society Bulletin 25: 527-531. Dobson, H. 1985. Effects of chronic treatment with a GnRH agonist on oestrous behavior and on the secretion of LH and progesterone in the ewe. Theriogenology 24: 1-11. D’occhio, M. J., D. R. Gifford, C. R. Earl, T. Weatherly, and W. Von Rechengerg. 1989. Pituitary and ovarian responses of post-partum acyclic beef cows to continuous long-term GnRH and GnRH agonist treatment. Journal of Reproduction and Fertility 85:495-502. ____________, W. J. Aspden, and T. R. Whyte. 1996. Controlled, reversible suppression of estrous cycles in beef heifers and cows using agonist of gonadotropin-releasing hormone. Journal of Animal Science 74:218-225. ____________, and Aspden, W. J. 1999. Endocrine and reproductive responses of male and female cattle to agonist of gonadotrophin-releasing hormone. Journal of Reproduction and Fertility 54 (Suppl.):101-114. ____________ , G. Fordyce, T. R. Whyte, W. J. Aspden, and T. E. Trigg. 2000. Reproductive responses of cattle GnRH agonist. Animal Reproduction Science 60-61: 433-442. ______________ , _____________ , ____________ , L. A. Fitzpatrick, N. J. Cooper, W. J. Aspden, M. J. Bolam, and T. E. Trigg. 2002. Use of GnRH agonist implants for long-term suppression of fertility in extensively managed heifers and cows. Animal Reproduction Science 74: 151-162. Fagerstone, K. A., M. A. Coffey, P. D. Curtis, R. A. Dolbeer, G. J. Killian, L. A. Miller, and D. L. Wilmot. 2002. Wildlife fertility control. Wildlife Society Technical Review 02-02. The Wildlife Society, Bethesda, Maryland, USA. Fraser, H. M, J. Sandow, H. R. Seidel, W. Von Rechenberg. 1987. An implant of gonadotropin releasing hormone agonist (buserelin) which suppresses ovarian function in the macaque for 3-5 months. Acta Endocrinology 121:841-853. Gong, J. G., T. A. Bramley, C. G. Gutierrez, A. R. Peters, R. Webb. 1995. Effects of chronic treatment with gonadotrophin-releasing hormone agonist on peripheral concentrations of FSH and LH, and ovarian function in heifers. Journal of Reproduction and Fertility 105: 263-270. ___________ , B. K. Campbell, T. A. Bramley, C. G. Gutierrez, A. R. Peters, and D R. Webb. 1996. Suppression in the secretion of follicle-stimulating hormone and luteinizing hormone, and ovarian follicle development in heifers continuously infused with a gonadotropin- releasing hormone agonist. Biology of Reproduction 55: 68-74. 52 �Hall, M. C., R. J. Fritzsch, A. I. Sagalowsky, A. Ahrens, B. Petty, and C. G. Roehrborn. 1999. Prospective determination of the hormonal response after cessation of luteinizing hormonereleasing hormone agonist treatment in patients with prostate cancer. Urology 53:898-902. Huang, F., D. C. Cockrell, T. R. Stephenson, J. H. Noyes, and R. G. Sasser. 2000. A serum pregnancy test with a specific radioimmunoassay for moose and elk pregnancy specific protein B. Journal of Wildlife Management 64: 492-499. Jacobson, N. K., D. A. Jessup, and D. J. Kesler. 1995. Contraception in black-tailed deer by remotely delivered norgestomet ballistic implants. Wildlife Society Bulletin 23: 718-722. Kreeger, T. J. 1997. Overview of delivery systems for the administration of contraceptives to wildlife. In Contraception in wildlife management. T. J. Kreeger, (ed.) USDA-APHIS Technical Bulletin 1853, Washington, D. C., pp 29-48. Kirkpatrick, J. F., I. K. M. Liu, and J. W. Turner. 1990. Remotely-delivered immunocontraception in feral horses. Wildlife Society Bulletin 18: 326-330. Macmillan, K. L., and D W. W. Thatcher. 1991. Effects of gonadotrophin-releasing hormone on ovarian follicles in cattle. Biology of Reproduction 45:883-889. McComb, K. 1987. Roaring by red deer stags advances the date of oestrus in hinds. Nature 330:648-649. McCleod, B. J., S. E. Dodson, A. R. Peters, and G. E. Lamming. 1991. Effects of a GnRH agonist (Buserelin) on LH secretion in post-partum beef cows. Animal Reproduction Science 24:1-11. McNeilly, A. S., and H. M. Fraser. 1987. Effect of gonadotrophin-releasing hormone agonist-induced suppression of LH and FSH on follicle growth and corpus luteum function in the ewe. Journal of Endocrinology 115:273-282. Melson, B. E., J. L. Brown, H. M. Schoeneman, G. K. Tarnavsky, and J. J. Reeves. 1986. Elevation of serum testosterone during chronic LHRH agonist treatment in the bull. Journal of Animal Science 62:199-207. Milliken, G. A., and D. E. Johnson. 1984. Analysis of messy data. Volume 1.Designed experiments. Lifetime Learning Publications, Belmont, California, 399 pp. Montovan, S. M., P. P. Daels, J. River, J. P. Hughes, G. H. Stabenfeldt, and B. L. Lasley. 1990. The effect of potent GnRH agonist on gonadal and sexual activity in the horse. Theriogenology 33:1305-1321. Morrison, J. A., C. E. Trainer, and D P. L. Wright. 1976. Multivariate statistical methods. McGraw-Hill, New York, New York, 432 pp. Niswender, G. D, Reichert, L. E., JR., A. R. Midgley, and A.V. Nalbandov. 1969. Radioimmunoassay for bovine and ovine luteinizing hormone. Endocrinology 84:1166-1173. ___________ . 1973. Influence of the site of conjugation on the specificity of antibodies in progesterone. Steriods 22: 413-424. Ravivarapu, H. B., K. L. Moyer, and D. R. Dunn. 2000. Sustained activity and release of leuprolide acetate from an in situ forming polymeric implant. American Association of Pharmaceutical Scientist 1:1-12. SAS Institute. 1997. SAS/STAT ® user’s guide 6.03 edition. SAS Institute Incorporated, Cary, North Carolina, USA. Shideler, S. E., M. A. Stoops, N. A. Gee, J. A. Howell, and D B. L. Lasley. 2002. Use of porcine zona pellucida (PZP) vaccine as a contraceptive agent in free-ranging tule elk (Cervus elaphus nannodes). Reproduction (Suppl) 60: 169-176. Trigg, G. G, T. E., P. J. Wright, A. F. Armout, P. E. Williamson, A. Jundaidi, G. B. Martine, A. G. Doyle, and D. J. Walsh. 2001. Use of GnRH analogue implant to produce reversible long-term suppression of reproductive function in male and female domestic dogs. Journal of Reproduction and Fertility 57:255-261. Turner, J. W., JR., I. K. M. Liu, and J. F. Kirkpatrick. 1992. Remotely delivered immunocontraception in white-tailed deer. Journal of Wildlife Management 56: 154-157. ___________ , ____________ , ____________ . 1996. Remotely delivered 53 �immunocontraception in free-roaming feral burros. Journal of Reproduction and Fertility 107: 31-35. Prepared by ___________________________ Dan L. Baker, Wildlife Researcher Figure 1. Twenty-four hour serum LH concentrations (ng/ml, mean ± SE) for untreated female elk (", n = 5) and female elk (!, n = 6) treated with a 180-day sustained release implant formulation, containing approximately 32.5 mg of leuprolide acetate, remotely delivered via projectile dart. Fig. 1 - Baker et al. - 80 i I 70 50 - 60 - -E 0) C: :r: _J C: (IJ Q) T 40 30 ~ 20 10 0 - l 0 J 1 Treatmert Leuprolide dart (Group A) -e-- Control dart (Group C) l T "" .L _j_ 120 180 240 300 360 Time (min) 54 480 600 960 1444 �Figure 2. Total serum LH concentrations (ng/ml/min, mean ± SE) for GnRH analog-induced release of LH for untreated female elk (", n = 5), and female elk (!, n = 6) treated with a 180 -day sustained release implant formulation, containing approximately 32.5 mg of leuprolide acetate, remotely delivered via projectile dart. Different lower case letters indicate significant differences between means (P 0.05). Fig. 2 - Baker et al. 100 - - - e- - 80 60 I _J C 40 (0 Q) ~ 20 a b b b b b 50 100 150 185 215 0 Pretreatment 361 Time (days) Figure 3. Serum profiles of mean progesterone concentrations (ng/ml, mean ± SE) for untreated female elk (", n = 5) and female elk (!, n = 6) treated with a 180-day sustained release implant formulation, containing approximately 32.5 mg of leuprolide acetate, remotely delivered via projectile dart. Different lower case letters indicate significant differences between means. (P 0.05). Fig. 3 • Baker et al. 1.00 Leuprolide dart (Group A) --- e- - a ...-.. -E 0) C e a a -S Q) 0.60 0) 0.40 C a a l J -- f'',,L_ a ' a (0 Q) ~ i ----- ' ' '" 1 ' ---e- Q) 1n Q) e c.. Control dart (Group C) 0.80 0.20 b b b 1 b b a 50 100 150 185 215 361 000 Pretreatment Time (days) 55 � https://cpw.cvlcollections.org/files/original/f7ad38888bf9552ad312a2744148ccda.pdf 8450b161cafd0b4122e4462f6d54451f PDF Text Text Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Colorado Project No. 1 Work Package No. 3003 Task No. 1 : : : : Federal Aid Project: : N/A Cost Center 3430 Mammals Research Predatory Mammals Conservation Colorado Puma Research & Management Program Period Covered: July 1, 2003― June 30, 2004 Author: K. A. Logan Personnel: J.Apker, J. Kindler, Colorado Division of Wildlife; L. Mundy-Four Corners Houndsmen’s Association; and Safari Club International Foundation. All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT The Colorado Puma Research and Management Program started in July 2003. The goal is to improve the scientific foundation of puma management by the Colorado Division of Wildlife. The program was developed with inputs of Division researchers, managers, and biologists, and Colorado citizens interested in wildlife management, hunting, and the environment. Puma population research is scheduled to begin in November 2004. Associated projects to improve puma management in Colorado, also initiated this year include: the Colorado puma data map, prospective work for Front Range pumahuman interaction research, puma technical workshops, and Data Analysis Unit management plans and puma-human conflict guidelines. 61 �JOB PROGRESS REPORT COLORADO PUMA RESEARCH AND MANAGEMENT PROGRAM Kenneth A. Logan INTRODUCTION The Colorado Puma Research and Management Program started in July 2003. The goal is to improve the scientific foundation of puma management in Colorado. A Prospectus was developed with inputs of Division researchers, managers, and biologists, and Colorado citizens interested in wildlife management, hunting, and the environment. The major part of the program is the puma research project in the prospectus, scheduled to begin in November 2004. The initial design of the research will be clarified in a study plan in September 2004 which will pertain to the puma population research on the Uncompahgre Plateau study area. Other work associated with the development of the research program included: visiting with affected publics (private landowners, ranchers, hunters, guides and outfitters) and agency cooperators, surveying potential study areas, and two public meetings for information on the proposed puma research. Associated projects to improve puma management in Colorado included: the Colorado puma data map, puma workshops, technical advice on puma Data Analysis Unit management plans, and puma-human conflict guidelines. COLORADO PUMA DATA MAP The objective of this project is to map and quantify puma data that exists in records of the Colorado Division of Wildlife. This is the first step for Division staff to examine historical and current situations regarding puma management actions and puma mortality patterns state-wide and within management units. The map is intended to be an evolving instrument that allows comparisons with puma data gathered in the future to examine potential effects of changing puma management prescriptions, habitat, ungulate populations, and human developments. Interpretations of the map could be clarified from information on puma populations, movement patterns, habitat use, habitat characteristics, pumaungulate interactions, and hunter access to occupied puma habitat. Reliable interpretations of such maps would be useful to managers. Number and distribution of puma mortality locations and absence of mortality locations may indicate relative puma abundance or hunting pressure. High mortality areas, influenced principally by sport-hunting pressure, may indicate potential areas of high puma densities, puma population sinks (defined as areas where the average population growth rate is negative), areas of facilitated puma hunting conditions (e.g., high road density, consistent snow coverage), and liberal puma harvest objectives. Low puma mortality areas and blank areas on the map may indicate potential puma habitat with puma source populations (defined as areas where mean population growth rate is positive, and which serve as net exporters of dispersing animals), areas where few if any puma live, or areas with low hunter access or good hunting conditions. RESULTS Desired products are maps and associated tabulated data on geographical distribution and intensity of puma mortality. Puma mortality data (including sport-harvest, depredation control, public safety management, and accidental deaths) recorded by the Division on mandatory check forms since 1997 were mapped, state-wide and by Data Analysis Units and Game Management Units, and stratified by year and puma sex and age class (e.g., adult, subadult, cub). This entails about 2,423 data points statewide, from 1997―2002 (2003 data have not been entered, yet). Of that, over 90% of the mortality locations are due to sport-harvest. The remainder is due to depredation or public safety control kills, roadkills, and other recorded deaths. Maps are currently in a preliminary development stage. Mortality 62 �locations of puma will be buffered by average puma home range sizes for adult puma in western North America (195 km2 for females; 357 km2 for males) and overlaid by mule deer, elk, and bighorn sheep winter ranges. In addition, the identity of DAUs with the management objective of a stable or increasing puma population and DAUs with the management objective of a suppressed population also need to be mapped for managers to consider the number, distribution and effects of potential source and sink populations. Other map overlays that may facilitate interpretation of the puma data include: road distribution (i.e., paved, all-weather, dirt), vegetation cover types, elevation, and human developments and density. Puma mortality characteristics (i.e., location, density) might be modeled by using an analytical approach that uses habitat and biological features (e.g., ungulate distribution and relative density, elevation, roads, vegetation, terrain ruggedness) as variables. Another approach might be to distribute puma mortality maps to Division field personnel and to knowledgeable puma hunters to record their explanations about geographical puma distributions, relative densities, mortality patterns, and effects of habitat characteristics (e.g., landownership, snow conditions, access). COLORADO PUMA-HUMAN INTERACTIONS RESULTS Meetings involving Division staff, and individuals from the U.S. Geological Survey Ft. Collins Research Center and Colorado State University were held to discuss potentials for puma-human interaction research on the Colorado Front Range. Meetings were held in conjunction with field trips to explore potential study areas west Ft. Collins on October 6, 2003; west of Boulder on January 28 and February 6, 2004; and west of Colorado Springs on February 5, 2004. Division staff from the southeast and northeast expressed a great deal of interest in developing puma-human interaction of research in the next 1―2 years, and discussions indicated the need to develop a reliable funding base and connections with potential cooperators. The Division of Wildlife and Four Corners Houndsmen’s Association co-hosted four workshops in 2004 to inform hunters and other interested citizens about puma in Colorado. Workshops were held in Grand Junction (July 19), Alamosa (July 17), Denver (July 22), and Canon City (August 14). In all, the workshops were attended by about 60 people. Workshop agenda topics included: puma population characteristics, vital rates, reproductive biology, behavior, prey selection, female and cub vulnerability to hunting, gender identification in the field, aging techniques, Colorado puma data map, puma management, Data Analysis Unit plans, quota setting process, proposed puma research, and puma-human conflict management. A PowerPoint presentation was developed as the main source of information on these topics. Some attendees suggested that such workshops should be held periodically for puma hunters and other people interested in puma. Associated with this effort to bring reliable information on puma to hunters, we also produced printed guidelines for sexing puma in the wild. This information is available on the Division’s webpage: http://wildlife.state.co.us/hunt/BigGame/pdf/MtLionGender.pdf and in APPENDIX I of this progress report. Drafts of guidelines developed by Division managers for addressing incidents when puma conflict with people were discussed and reviewed and a final draft was created for review by Division staff: Draft—Human Mountain-Lion Incidents. 63 �PUMA RESEARCH AND MANAGEMENT PROSPECTUS PROBLEM STATEMENT Division of Wildlife managers need reliable information on puma biology and ecology in Colorado to develop sound management strategies that address diverse public values and the Division objective of actively managing puma while “achieving healthy, self-sustaining populations”(Colorado Div. of Wildlife 2002-2007 Strategic Plan:9). Although 4 puma research efforts have been made in Colorado since the early 1970s and puma harvest data is compiled annually, reliable information on certain aspects of puma biology and ecology, and management tools that may guide managers toward effective puma management is lacking. Members of the Division’s Mammals Research staff met with Division wildlife managers and biologists from the Northwest, Southwest, Southeast, and Northeast Regions regarding puma management issues and the resultant research needs. In addition, we consulted with other agencies, organizations, and interested publics either directly or through other Division employees. In general, Division staff in western Colorado conveyed concern about puma population dynamics, especially as they relate to their abilities to manage puma populations through regulated sport-hunting. Secondarily, (perhaps because of results from recent research findings in western Colorado), they expressed interest in puma-prey interactions. Division managers on the Rocky Mountain’s Front Range placed greater emphasis on puma-human interactions. Staff in both eastern and western Colorado cited information needs regarding effects of puma harvest, puma population monitoring methods, and identifying puma habitat and landscape linkages. Specific management needs and lines of inquiry identified by Division staff and public stakeholders are categorized as follows: Improve our ability to manage puma hunting with enhanced scientific bases, strategies, and tools ● Puma population characteristics (i.e., density, sex and age structure). ● Puma population dynamics and vital rates (i.e., birth rates, survival rates, emigration rates, immigration rates). ● Methods and models for assessing and tracking changes in puma populations. ● Relative vulnerability of puma sex and age classes to hunter harvest. Improve our understanding of puma habitat needs and interrelationships of puma management units ● Puma habitat use, movements, and use of landscape linkages. ● Puma recruitment patterns (i.e., progeny, immigration, emigration). ● Models for identifying puma habitat and landscape linkages. Improve our understanding of the puma’s role in the ecology of other species ● Relationships of puma to mule deer, elk, and other natural prey. ● Relationships of puma to species of special concern, e.g., desert bighorn sheep. Improve our understanding of puma-human interactions and abilities to manage them ● Behavior of puma in relation to people and human facilities. ● Puma predation on domestic animals. ● Effects of translocating nuisance puma. ● Effects of aversive conditioning on puma. Past Puma Research in Colorado Data from past puma research in Colorado that address the topics above are limited. Currier et al. (1977:8) studied 29 captured puma on 2 tracts― one 900 km2, and one 1,950 km2― in Fremont and Custer Counties during 1974―1977. The puma population under study was subject to sport-hunting. Hunters killed a total of 31 puma in 3 winters. Non-systematic puma track counts were used to estimate 64 �the minimum puma population at 11 on the 900 km2 tract (density = 1.2/100 km2) and 25―28 on the 1,950 km2 tract (density = 1.3―1.4/100 km2. The Petersen mark-recapture method was used to estimate 95 puma (95% CI = 35, 155; density = 4.8/100 km2) during 1977―1978; however, researchers probably did not meet 4 of the 6 assumptions needed for valid estimates of puma numbers (Anderson 1983:61, 63). Vital rate data were limited to mean litter size of 2.1 (range = 1―5, n = 14). An effort to estimate puma population density in Game Management Units (GMU) 33 (Garfield and Rio Blanco Counties) and 40 (Mesa County) was made during 1980―1983 (Brent 1981, 1982, 1983). A total of 38 puma were captured: 21 were marked and released; 8 were released unmarked; 9 were killed for livestock depredation control (8) or during handling (1). Twelve puma were captured in GMU 33 and 26 were captured in GMU 40. Crude adult puma density estimates for GMU 33 ranged from 2.7―3.1 puma/100 km2. GMU 40 crude adult puma density estimates ranged from 1.2―3.7 puma/100 km2. Anderson et al. (1992) studied 57 captured puma on 3,426 km2 of the eastern slope of Uncompahgre Plateau in Mesa, Delta, Montrose, and Ouray counties during 1981―1988. Puma density was estimated only for 1987; the minimum density (mean ± SE) of residents was 1.1 ± 0.15 pumas/100 km2. Male to female sex ratios for 26 captured puma 1―12 months old was 1:1; for 19 captured puma ≥24 months old, it was 1:1.4. Age structure in that sample was 66.7% <24 months old and 33.3% ≥24 months old (the class most likely comprised of breeding adults). Vital rates included, mean (± SD) litter size of 2.4 ± 0.80 (n = 17), birth interval of 12 months (n = 2 intervals for 1 female), estimated annual survival rate for 42 puma of both sexes of 88.0 % (90% CI = 83.1, 91.4). Humans caused 18 of 21 deaths in radio-collared puma even though the study population was supposed to be protected from human offtake. Anderson et al. (1992) examined aerial locations of 7 radio-collared puma and subjective estimates of relative deer and elk density categories and could not identify consistent relationships probably because of the small non-random sample of puma, the subjective nature of the ungulate density categories, and other non-quantified factors. Mean annual home range sizes ranged from 436―732 km2 for 3 males and 190―463 km2 for 7 females. All of 9 radio-collared subadult male puma dispersed from natal areas. Two of 6 radio-collared subadult females did not disperse. Means and extremes of dispersal distances were 86.2 km (23―151) for 8 males that were 10―13 months old and 37.0 km (17―54) for 4 females that were 11―31 months old. Data on puma―human interactions were from 17 responses to 40 questionnaires submitted to residents in the housing development on the southeastern extreme of the study area. Seven of 17 respondents reported 25 puma sightings during about 260 months of residence. Ten respondents did not observe puma in about 476 months of residence. Koloski (2002) studied 19 captured puma on the 2,758 km2 Southern Ute Indian Reservation in La Plata, Archuleta, and Montezuma Counties during 1999―2001. The puma population was not subject to sport-hunting at the time. Transect intercept probability sampling was used in 2001 to estimate the puma population at 55 (90% CI = 9.0, 114.4) and a density of independent puma at 2.7/100 km2. Male to female ratio of the captured sample of 14 independent puma was 1:2.8. Of 16 captured pumas that were aged, 31% were <24 months old and 69% ≥24 months old. Vital rates included: litter size (mean ± SD) of 2.5 ± 0.58 (n = 4), birth interval of 16 months (n = 1), annual survival rate for radio-collared males (mean ± SD) of 0.89 ± 0.19 (n = 3) and radio-collared females of 0.72 ± 0.19 (n = 8); earliest age for female reproduction at 2―3 years; annual reproductive rate for resident females (mean ± SD) of 42% ± 12%. Mean home range sizes were 252.4 km2 for 3 radio-collared males and 182.4 km2 for 8 radio-collared females. Road density within puma home ranges and core areas was lower than that on the landscape where pumas occurred (P ≤ 0.002). Distance from puma locations to nearest roads was lower than distance from random points to nearest roads (P = 0.002). 65 �Current Puma Research in Colorado Presently, researchers with the Colorado Division of Wildlife (Mike Miller, Ph.D., DVM) and Colorado State University (Caroline Krumm, Graduate Student and Dr. N. Thompson Hobbs, Advisor) are conducting puma research in Larimer County. The research goal is to test for selective puma predation on mule deer infected with chronic wasting disease (CWD) by comparing CWD prevalence in puma-killed deer to prevalence in harvested deer. The research protocol calls for 6 or more puma fitted with global positioning system (GPS) collars. RESEARCHABLE OBJECTIVES The management issues listed previously in the PROBLEM STATEMENT may be translated into a number of researchable objectives, requiring descriptive studies and field experiments (Fig 1). Our goal is to provide managers with reliable information on puma biology and ecology and to develop and test tools for their efforts to adaptively manage puma in Colorado to maintain healthy, self-sustaining populations. Researchable objectives address managers’ main needs. We propose that the Division begin to address objectives that focus on puma population dynamics, effects of harvest, and estimating puma population abundance with an intensive puma population study on the West Slope. Those objectives include: 1. Describe and quantify puma population characteristics, including: density, sex and age structure. 2. Describe and quantify puma population vital rates, including: birth rates, age or stage-specific survival rates, emigration rates, immigration rates. 3. Describe and quantify agent-specific mortality rates and vulnerability of different classes of puma to hunter harvest and quantify agent-specific mortality rates. 4. Develop and test puma population models using metrics from objectives 1―3. 5. Develop and test indices to puma abundance calibrated on an estimated puma population (i.e., puma track counts, catch per unit effort, DNA genotyping). In addition, other objectives could be partially addressed during the intensive puma population research effort (i.e., objectives 1―5). Those include: 6. Describe and quantify relationships of puma to people and human facilities on the study area. 7. Describe and quantify puma use of habitats and landscape linkages. 8. Describe and quantify relationships of puma to mule deer, elk, and other prey. 9. Describe and quantify responses of puma to aversive conditioning. 10. Describe and quantify behavior and survival of translocated puma. Data collection for primary objectives 1―5 will often have applicability to objectives 6―10. For example, GPS-collared puma will enable us to quantify puma predation rates on ungulate prey, puma use of habitats and landscape linkages, puma-human interactions, and behavior and survival of translocated puma (if puma are removed from a study area as part of an experimental manipulation). However, we cautioned that such opportunistic data gathering likely will not yield the power or confidence levels of studies specifically designed to meet those objectives. Yet, such efforts could function as pilot studies. Additional research efforts can be phased in later in the puma research program. And some, (e.g., puma relationships to people, puma use of habitats and landscape linkages) can be conducted in different areas of Colorado. 66 �GOAL: Strategies, Information, & Tools for Managing Healthy, Self-sustaining Puma Populations in Colorado Puma Population Effects of Harvest & Other Mortality Movements & Corridors Population Dynamics: Density, Sex & Age, Vital Rates, Growth Rates Vulnerability to Harvest Puma Habitat Human Development Habitat Use PumaHuman Relationships Effects of Predation Map Prey Distribution Models for Habitat Indices of Abundance for Monitoring Deer, Elk, Other Natural Prey & Species of Concern Domestic Animals Effects of Evasive Conditioning Effects of Translocation Models for Populations Puma Prey Map Habitat Model PumaPrey Relationships Fig. 1. Conceptual model of the Colorado Puma Research & Management Program. 67 �TESTING ASSUMPTIONS AND HYPOTHESES Hypotheses associated with the main objectives can be structured to test assumptions, information, and methods that may guide puma management in Colorado. 1. Lacking Colorado-specific information, managers might assume that puma population densities in Colorado are within the range of those quantified in other populations studied in Wyoming (Logan et al. 1986), Idaho (Seidensticker et al. 1973), Alberta (Ross and Jalkotzy 1992), and New Mexico (Logan and Sweanor 2001). The Division has used density ranges of 2.0―4.6 puma/100 km2 to extrapolate to Data Analysis Units to estimate a range of 3,000―7,000 puma in Colorado and to guide the quota-setting process. Likewise, managers may assume that the population sex and age structure is similar to puma populations described in the above-mentioned studies. Using capture, mark-recapture (or resight) techniques, a descriptive study will test H1a: The puma population density on the study area will vary within the range of 2.0―4.6 puma/100 km2 and will exhibit a similar sex and age structure to puma populations in Wyoming, Idaho, Alberta, and New Mexico. Yet, an experimental study that allows puma population growth to approach carrying capacity in high quality puma habitat can test if a puma population in Colorado might exceed published density estimates. H1b: Puma population density in high quality puma habitat in Colorado exceeds the high range of 4.6 puma/100 km2. 2. Background material that guides puma management in Colorado assumes a moderate rate of growth of 15% for the adult puma population. Theoretically, consideration of management changes would occur if hunter kill exceeds 15% of the low end adult puma population estimate. A field experiment, involving an increase population growth phase, is required to test H2: The estimated average annual adult puma population growth rate in high quality puma habitat in Colorado (during an increase phase) will match or exceed the hypothetical r = 0.15. 3. The same background material assumes “that when female” puma “comprise greater than 50% of the hunter harvest it is an indicator that hunting may be acting to suppress the population.” An experimental study with a decline puma population growth phase will test H3: The population of harvest-age puma (i.e., adults, subadults) will decline only when 50% or more of the harvest is comprised of harvest-age female puma (i.e., independent subadults ≈≥12―24 mo. old and adults ≥24 mo. old). 4. Colorado puma Data Analysis Units (DAUs) or other management units may behave as a demographic source-sink metapopulation structure where the puma population of a region is comprised of subpopulations each of which may have dynamics that are not necessarily correlated with other subpopulations. Source-sink metapopulation dynamics function as a result of variable puma habitat quality and management practices (e.g., prey population dynamics, harvest pressure). Sources are increasing or stable populations where recruitment via local reproduction and immigration equal or outpace mortality. These source populations produce emigrating progeny that immigrate into other subpopulations, augmenting them numerically and genetically. Comparatively, sink populations are those where mortality exceeds recruitment, and puma numbers are declining or suppressed to a relatively low density. Sink populations contribute few emigrating progeny as potential recruits for other subpopulations. Sink populations are augmented by immigrants from source populations (Sweanor et al. 2000, Logan and Sweanor 2001). This project will examine H4: The study population will exhibit characteristics of a subpopulation in a demographic source-sink metapopulation structure. The following predictions must come true to support this hypothesis. 68 �a. The majority (i.e., >50%) of male recruits in the adult segment of the population on the study area will be immigrants (Logan and Sweanor 2001). Immigrants will not be offspring of puma on the study area as determined by genetic parentage analysis. b. Up to one-third of female recruits will be immigrants (Logan and Sweanor 2001). c. Male and female immigrant recruits will produce viable young. d. The majority of male progeny from the study area will emigrate (Logan and Sweanor 2001). e. About one-third of female progeny from the study area will emigrate (Logan and Sweanor 2001). f. Movements of male and female emigrants will be large enough to carry them to other Data Analysis Units with differing management objectives (i.e., population reduction, population stability). g. Male and female emigrants will establish adult home ranges in other puma habitats in Colorado. H4A: Recruits born in the local population are the largest contributor (i.e., >80%) to the maintenance of the puma population on the study area. 5. In southern Utah, Van Sickle and Lindzey (1992) found a positive relationship (r2 = 0.73) between the number of radio-collared puma known to have home ranges overlapping dirt roads (response variable) and track-finding frequency (explanatory variable). Similarly, researchers in Montana are finding a positive relationship between the number of puma home ranges overlapping search routes and puma track density (DeSimone et al. 2002). This relationship should reflect changes in puma numbers on the survey area, and thus may be useful as an index to relative abundance. A field experiment requiring both increase and decrease puma population growth phases is required to test H5a: Puma track-finding frequency (response variable) is positively correlated to number of puma with home ranges overlapping snow-covered search routes (explanatory variable). H5b: Puma track-finding frequency (response variable) is positively correlated to the density of independent puma (explanatory variable). 6. Theoretically, the amount of effort (i.e., hunting days) that hunters spend in pursuit of finding harvest-age puma (i.e., adults and subadults) should be proportionate to the abundance of puma (Lancia et al. 1996). A relationship should exist between changes in catch (or encounter)-perunit-effort and population changes. A field experiment, involving manipulation of the study population (i.e., increase phase, decline phase), is required to test H6a: Catch-per-unit effort of the research team in the increase phase, teams in the capture-recapture occasions during the increase and decline phases, and puma hunters during the decline phase will reflect the trend in the puma population. There will be an inverse (i.e., negative) correlation between the mean number of days per capture (response variable) and the number of independent puma in the population (explanatory variable) during the increase phase and the decline phase. H6b: During the increase phase, there will be a positive correlation between the mean number of days per capture of unmarked puma (response variable) and the number of marked puma “removed” from the unmarked population per year (explanatory variable). In the decline phase, there will be an inverse correlation between the mean number of days per capture (response variable) and the number of puma killed by hunters per year (explanatory variable). 7. Relative vulnerability of puma to hunters is limited to information from 2 studies on the same area in southern Utah (Van Dyke et al. 1986, Barnhurst 1986). Van Dyke et al. (1986) quantified effort to locate 4 classes of puma by looking for their tracks on dirt roads, a method that hunters use to find puma. He found that cubs and adult females required the least effort, followed by independent subadults and adult males. In contrast, Barnhurst (1986) assessed vulnerability based on the relative road crossing frequencies of radio-collared puma in each of 7 classes that were relocated once per week. He found that the most vulnerable puma were subadult males, followed 69 �by adult resident males, subadult females, and adult females (in 4 classes― females with 0 cubs, females with cubs 0-6 mo. old, females with cubs 7-12 mo. old, females with cubs 13-18 mo. old). Mothers with 0―6 month-old cubs had the lowest road-crossing frequency of all classes. Cubs in this age class are most vulnerable to death if their mothers die (K. Logan, unpublished data). A descriptive study will test H7: Relative vulnerability of GPS-collared puma on the study area, based on road-crossing frequency per day, will reflect the results of Van Dyke et al. (1986). H7A: Relative vulnerability of GPS-collared puma on the study area, based on road-crossing frequency per day, will reflect results of Barnhurst (1986). 8. Studies on the effectiveness of puma translocation, and the behavior, survival, and agent-specific mortality of translocated puma in western North America are limited to 2 studies. Ruth et al. (1998) reported on 14 puma translocated from 338―510 km in New Mexico, and Ross and Jalkotzy (1995) reported on 3 puma that were translocated 51―94 km in Alberta. The New Mexico research found that translocation was most successful for puma that were 12―27 months old, the age at which puma naturally attempt to disperse and search for a home range or establish a home range if they are philopatric. Older adult puma attempt to move back to their original home ranges. They found that mortality rates for translocated puma were significantly higher than mortality rates of non-translocated puma in a reference population. If translocation is used to experimentally reduce the population, this research would test H8: Translocation of puma will exhibit similar characteristics to the New Mexico results. For this hypothesis to be supported, the following predictions must be true. a. Mortality rates of translocated puma will be significantly higher than mortality rates of non-translocated puma. b. Independent puma 12 to about 30 months old will establish home ranges in or near release areas and have relatively greater survival rates than older adult translocated puma. c. Adult puma about 3 years old and older will tend to move back toward their original home ranges. DESIRED OUTCOMES AND MANAGEMENT APPLICATIONS 1. Quantification of variations in puma population density, sex and age structure, growth rates, vital rates, and an understanding of factors affecting them will aid adaptive puma management by yielding population model(s) useful for estimates of puma population abundance and trends, evaluation of management alternatives, and effects of management prescriptions. 2. Indices to puma abundance or trends of known reliability will allow managers to “ground truth” modeled populations and estimate effects of management prescriptions designed to achieve specified puma population objectives. 3. Testing assumptions about puma populations, currently used by Division managers, will help those managers adapt puma management based on Colorado-specific estimated characteristics and dynamics of puma populations. 4. An understanding of relative vulnerability of the various puma sex and age classes to harvest could enable managers to better structure harvest data collection and interpretation, and to develop novel prescriptions to meet management objectives. 5. Functional relationships between population vital rates and population density could be examined. Puma life history traits in Colorado may be used to test generalized hypotheses regarding puma life history strategies in the literature, and inform managers to structure successful management strategies. 70 �6. Determining whether or not the puma population in Colorado has a source-sink demographic sourcesink structure is important to evaluating the current Game Management Unit and Data Analysis Unit structure of puma management and potential effects of the juxtaposition of puma sub-populations in Colorado managed for stability, suppression, or that may function as refugia. 7. Knowledge of the relationships of puma to deer, elk, and species of special concern would allow managers to realistically consider potential effects of puma predation on those prey in the development of management strategies and policy. In addition, such information would enable the testing of scientific hypotheses on relationships of puma to their prey. 8. In the study areas currently being contemplated, some puma home ranges will probably contain human habitations and other facilities. GPS-collared and VHF-collared puma will generate quantitative information on puma behavior in relation to human activity and assist managers to better inform people about ways of reducing potential conflicts between people and puma, and to structure puma conflict policy. 9. Habitat use data gathered during the course of this research could be used to quantify puma habitat characteristics on the study area, as well as habitats and landscape linkages used by dispersing puma. Such information could be used to structure more extensive investigations of puma habitat that contribute to habitat modeling efforts that may help identify puma habitat in Colorado. This would allow a more realistic conceptual inventory of puma habitat in the state. 10. This information could be disseminated to public stakeholders interested in pumas in Colorado, and thus contribute to informed public participation in puma management. STUDY AREAS Three potential study areas were evaluated and are under consideration: the Plateau Creek-toSouth Canyon area (in Garfield, Mesa, Gunnison, and Pitkin counties), the lower Dolores River-toDisappointment Creek area (in Dolores and Montezuma counties), the South Uncompahgre Plateau (in Mesa, Montrose, Ouray, and San Miguel Counties) (Table 1). These areas appear to have attributes conducive to an intensive puma research effort, including sufficient area (400―500 mi.2 = 1,036―1,295 km2) of puma habitat and suitable road access. Preferably, there should be a 300―400 mi.2 buffer zone around the study area to reduce the effect of puma harvest. 71 �Table 1. Potential puma study area locations and characteristics. Location Dolores River-toDisappointment Creek (GMUs 71 & 711) Area ~840 mi.2 = 2,176 km2 Junction of Interstate 70 & State Route 6 northeast to South Canyon (GMU 42) ~400 mi.2 = 1,036 km2 Area can be expanded ~155 mi.2 (~401 km2) by adding northern portion of GMU 421 (north of state routes 6 & 330, the north slope of the Plateau Creek drainage. South Uncompahgre Plateau (southern halves of GMUs 61 & 62) ~870 mi.2 = 2,253 km2 Other Attributes Large enough for core study area and buffer. Ratio of public:private land (mi.2) ~4:1. Town of Dolores is at the south end of this area. Substantial number of people live on the plateau. Domestic sheep and cattle use the area. Puma hunting pressure is moderate. Puma predation on domestic animals is low. Minimum study area size. Ratio of public:private land (mi.2) ~2:1. Substantial number of people live along the I-70 corridor and in lower Mamm, Hollow, Divide, and Battlement Creeks, and on Grass Mesa. Recent research on elk seasonal movements, survival and causespecific mortality rates (Freddy). An unknown number of cattle and horses use the area, but there are no domestic sheep. Gas exploration and development is occurring on the area. Puma hunting pressure is moderate. Puma predation on domestic animals is low. Large enough for core study area and buffer. Ratio of public:private land (mi.2) ~3:1. Ongoing mule deer research (Bishop et al. 2003), substantial “pre-treatment” data on mule deer productivity, survival, and cause-specific mortality (Pojar, Watkins, Bishop). Historical puma research (Anderson et al. 1992). Substantial number of people live in the eastern foothills and along the eastern, western, and southern edges of the plateau. Domestic sheep (~6 operators), cattle, and horses use the area. Puma hunting pressure is moderate. Puma predation on domestic animals is low. Puma can be captured year-round using 4 basic methods: trained dogs, cage traps, foot-hold snares, and hands (for small cubs). Capture efforts with dogs will be conducted mainly during the winter when snow facilitates searches for puma tracks and the ability of dogs to follow puma scent. The study area will be searched systematically multiple times per year by four-wheel-drive trucks, all-terrain vehicles, snow-mobiles, and walking. When puma tracks ≤1 day old are detected, trained dogs will be released to trail puma. Puma usually climb trees to take refuge from the dogs. Adult and subadult puma captured for the first time or requiring a change in telemetry collar will be immobilized with Telazol (tiletamine hydrochloride/zolazepam hydrochloride) dosed at 3.3 mg/kg estimated body mass (Wildlife Restraint Handbook, 1996, California Dep. of Fish and Game, Wildlife Investigation Laboratory, Sacramento). Immobilizing agent will be delivered in a Pneu-Dart® shot from a CO2-powered pistol. Immediately, a 3m-by-3m square nylon net will be deployed beneath the puma to catch it in case it falls from the tree. A researcher will climb the tree, fix a Y-rope to two legs of the puma and lower the cat to the ground with an attached climbing rope. Once the puma is on the ground, its head will be covered, its 72 �legs tethered, and vital signs monitored (Logan et al. 1986). (Normal signs: pulse ≈ 70―80 bpm, respiration ≈ 20 bpm, capillary refill time ≤2 sec., rectal temperature ≈ 101oF average, range = 95―106oF.) (Wildlife Restraint Handbook, 1996, California Dep. of Fish and Game, Wildlife Investigation Laboratory, Sacramento). A cage trap will be used to capture adults, subadults, and large cubs when puma can be lured into the trap using road-killed or puma-killed ungulates (Sweanor et al. 2004). Efficiency of the trap will be enhanced by using an automated digital call box that emits puma vocalizations (Wildlife Technologies, Windham, NH). Researchers will monitor the set cage trap from about 1 km distance by using VHF beacons on the cage and door. This allows researchers to be at the cage to handle captured puma within 30 minutes. Puma will be immobilized with Telazol injected with a pole syringe. Immobilized puma will be restrained and monitored as described above. Foot-hold snares will be used to capture adults, subadults, and large cubs as described by Logan et al. (1999). Puma will be immobilized with Telazol injected with a pole syringe and their vital signs monitored during the handling procedures. Efficiency of snares will also be enhanced with the use of an automated digital call box. Small cubs (≤10 weeks old) will be captured using our hands (covered with clean leather gloves) or with a capture pole. Cubs will be restrained inside new burlap bags during the handling process and will not be administered immobilizing drugs. Cubs at nurseries will be approached when mothers are away from nurseries (as determined by radio-telemetry). Cubs captured at nurseries will be removed from the nursery a distance of ≈100 m to minimize disturbance and human scent at nurseries. Immediately after handling processes are complete, cubs will be returned to nurseries (Logan and Sweanor 2001). All captured puma will be examined thoroughly to ascertain sex and describe physical condition and diagnostic markings. Age of adult puma will be estimated initially by the gum-line recession method (Laundre et al. 2000) and dental characteristics of known-age puma (Logan and Sweanor, unpubl. data). Ages of subadult and cub puma will be estimated initially based on dental and physical characteristics of known-age puma (Logan and Sweanor unpubl. data). Body measurements recorded for each puma will include at a minimum: mass, pinna length, hind foot length, plantar pad dimensions. Tissue collections will include: skin biopsy (from the pinna receiving the 6 mm biopsy punch for the ear-tag ), blood (30 ml from the saphenous or cephalic veins), and hair (from various body regions) for genotyping individuals, parentage analysis and disease screening; fecal for diet analyses. Universal Transverse Mercator Grid Coordinates on each captured puma will be fixed via Global Positioning System (GPS, North American Datum 27). Marking, Global Positioning System and Radio-telemetry- Objectives 1―9 Puma do not possess easily identifiable natural marking, such as tigers (see Karanth and Nichols 1998, 2002), therefore, the capture and marking of individual puma is essential to a number of program objectives. Adult and subadult puma will be marked 3 ways: radio-collar, ear-tag, and tattoo. The identification number tattooed in one pinna is permanent and cannot be lost unless the pinna is severed. A colored, numbered 25 mm diameter ear-tag will be inserted into the other pinna to facilitate individual identification during recaptures and in photos taken by field cameras (see capture-recapture methods below). Adult and subadult puma will be fitted with GPS collars (approximately 400 g each, Lotek Wireless, Canada) programmed to fix and store puma locations at least 4 times per day at 6-hours intervals to sample daytime, nighttime, and crepuscular locations. Each collar will have a color-coded identification number on each side also to facilitate identification during physical recaptures and photographic resightings. GPS locations for puma will provide precise, quantitative data for estimating 73 �puma home ranges, habitat use, quantifying road crossings (an index to vulnerability to hunting), finding ungulate kills (at location clusters), and estimating kill rates on ungulate prey (i.e., days per kill). VHF radio transmitters on GPS collars will enable researchers to find those puma on the ground in real time to acquire remote GPS data reports, facilitate recaptures for re-collaring, and to check on their reproductive and physical status. VHF transmitters will have a mortality mode set to alert researchers when puma have been immobile for at least 4 hours so that dead puma can be found to quantify survival rates and agentspecific mortality rates by gender and age. At least one cub of each sex in each litter will be fitted with small VHF transmitter mounted on an expandable collar (≈100g, MOD 210, Telonics, Inc., Mesa, Arizona). Simultaneous locations of mothers and radioed cubs enable researchers to quantify the frequency that mothers are away from cubs to assess the potential risk of orphaning by hunters and other mortality factors, and quantification of survival rates and agent-specific mortality rates. Attrition of cubs in the remainder of the litter can be estimated by periodic visual checks for other siblings by homing on radioed cubs (Logan and Sweanor 2001). Locations of GPS- and VHF-collared puma will be fixed at least once per week from light fixedwing aircraft (e.g., Cessna 182) fitted with radio signal receiving equipment (Logan and Sweanor 2001). This monitoring will enable researchers to find GPS-collared puma to acquire remote GPS location reports from the ground, monitor the status (i.e., live or dead) of individual puma, and to recover carcasses for necropsy. It will also provide simultaneous location data on mothers and cubs. GPS- and VHF-collared puma will be located from the ground opportunistically using hand-held yagi antenna. At least 3 bearings on peak aural signals will be mapped to fix locations and estimate location error around locations (Logan and Sweanor 2001). Aerial and ground locations will be plotted on 7.5 minute USGS maps and UTMs along with location attributes will be recorded on standard forms. GPS locations will be mapped using ArcGIS 8 software. Puma Abundance― Objectives 1, 4 &5 Capture-recapture estimates 1. Capture-recapture models will be used to estimate the parameters of primary interest― absolute numbers of independent puma (i.e., number of puma present in the survey area) and puma density (i.e., number of puma/100 km2) each winter― Dec. through Mar.― when snow facilitates detection and capture of puma, provided that we meet model assumptions. The Dec.―Mar. period also corresponds with Colorado’s puma hunting season. The population of interest is independent puma (i.e., adults and subadults) because those are the puma of legal harvest age. Furthermore, adults comprise the breeding segment of the population and subadults comprise potential recruits into the adult population in ≤1 year. Thus, the sampling unit is the individual independent puma (≈≥1 yr. old). General assumptions for capture-recapture models are: (1) the population is closed; (2) animals do not lose their marks during the interval; (3) all marks are correctly noted and recorded at each trapping occasion; (4) each animal has a constant and equal probability of capture on each capture occasion. Open population models allow the assumption of closure to be relaxed (Otis et al. 1978, White et al. 1982, Pollock et al. 1990). Marked puma will make it possible to acquire most of the basic statistics needed for capture-recapture models. Those include: nj (number of individually identified puma caught and released on occasion j), mj (number of previously marked puma recaptured in occasion j), uj (number of new unmarked puma captured in occasion j) (Otis et al. 1978, White et al. 1982). Attribute data of captured puma, such as age and sex, will be recorded to stratify the population in case separate analysis of different strata is necessary (if sample size allows) to meet certain assumptions of capture-recapture models (Pollock et al. 1990). Precise estimates of puma population size may also allow analyses of functional relationships between population vital rates and population size. 74 �We anticipate it may take 2 years to capture and mark the large majority of puma in the population. Our operational objective will be to have ≥90% of the independent puma marked before capture-recapture occasions commence. Capturing and marking puma is time consuming, and would lengthen the time to thoroughly search the study area for capturing and marking puma during the capture-recapture occasions, therefore, we will capture and mark puma prior to performing capture-recapture occasions. In addition, by marking puma before capture-recapture occasions begin, we will have opportunities to capture female puma at different stages of their reproductive status, and thus reduce the chance that mothers in a stage with suckling cubs and small activity areas are not detected and marked during the winter. After cubs are weaned, the mothers’ activity area expands (Logan and Sweanor 2001). The probability of females having suckling cubs in winter is naturally small; that season exhibits the lowest rate of births (Logan and Sweanor 2001). Our year-round capture efforts using trained dogs, foot-hold snares, and cage traps should help to reduce biases in capture probabilities attributed to any individual capture method (Miller et al. 1997). Thus, capture-recapture occasions may not begin until the end of the second winter. Capture-recapture occasions performed at that time will be viewed as a pilot study allowing us to examine the logistics of the field methods, the extent to which model assumptions are met, biases in field methods (relative to GPS data on collared puma), and precision of capturerecapture models used to estimate the puma population. Data gathered directly from GPS-collared puma and knowledge of the study area acquired by the research team in years 1―2 will allow us to assess if capture-recapture methods are appropriate (i.e., if basic assumptions can be met), and if they are, facilitate the exact design of the mark-recapture schemes for population and density estimates. Movements of GPS-collared puma in and out of the study area during capture occasions will also allow us to estimate corrections for such movements (White 1996). A composite range (i.e., minimum convex polygon) of all the GPS-collared puma home ranges (i.e., using locations from each of the collared puma) will be estimated and mapped to define the search area (i.e., the area inhabited by the estimated population) and allow mapping of search routes for a thorough systematic search of the area to detect puma for capture (i.e., any individual puma should not have a negligible probability of detection). Density estimates using the generated population estimates will be based on the search area (i.e., N/Area) (Miller et al. 1997). A grid will be constructed on the same search area with cells equal to the minimum home range size. There will be a minimum of 4 search routes per cell, each chosen to sample a quarter of each cell. Any spaces in habitat on the search area not occupied by collared puma will also be sampled. Capture occasions will be repeated 3―6 times each winter (i.e., t1, t2, t3,...t6) to resample the population. Unmarked puma will be marked and returned to the population to increase the precision of population estimates. Teams of trained houndsmen (4―6 teams of at least 2 persons each) will by used to systematically and thoroughly search the study area each occasion. Capture occasions will be about 1―2 weeks apart. Capture occasions will commence 1―2 days following fresh snowfall that covers the study area and last 3―5 days (i.e., this is how long it may take teams to thoroughly search the study area). But if fresh snowfall is lacking, we will attempt capture efforts anyway (although this could increase variation in capture probabilities). At puma captures, the puma identification number, sex, age, and location (U.T.M. coordinate) will be the minimum information recorded. If the same individual puma is caught more than once in the same occasion, each capture will be recorded, but only the first capture will be used for data analysis. All capture-recapture occasions will be conducted within a 2―3 month span in winter to minimize the chance of population changes (i.e., deaths, immigration, emigration). Once capture history data on puma are gathered, estimates of the number of independent puma in the population in winter can be made by using capture-recapture models that deal with variation in capture probability and closed or open populations (Otis et al. 1978, White 75 �et al. 1982, Pollock et al. 1990). In order for closed models to be valid, the population of independent puma cannot change (as a result of death, emigration, or immigration) during the 2―3 month span that contain the capture-recapture occasions. Because the precision of estimates for small populations is sensitive to the probability of capture (White et al. 1982, Pollock et al. 1990), our operational goal will be to achieve capture probabilities of about 0.6 (for 3 occasions) and 0.4 (for 6 occasions) to yield capture probabilities ≥0.9 for individual puma in the population each winter (Trolle and Kery 2003). Theoretically, capture probabilities within this range (i.e., 0.4―0.6) would tend to reduce the coefficient of variation of the estimate to about 0.20 (i.e., increase the precision of the estimate) in small populations where individuals have a survival probability of about 0.90 in 5 samples (Pollock et al. 1990:72), which is realistic for puma. In addition, behavior, movements, and survival rates of GPS-collared puma will allow direct biological examinations of assumptions of geographic and demographic closure (White et al. 1982) and variation in capture probability of individual puma and puma classes (i.e., adult females, adult males, subadult females, subadult males). If capture probabilities vary by puma class, we will examine if data stratification is necessary or possible (depending upon sample size). For example, we expect the larger home ranges of male puma to expose them to more search routes, thus, this may increase their probability of capture. If the assumption of demographic closure cannot be satisfied, then open population models may be used (Pollock et al. 1990). GPS locations (4 fixes/day) on individual puma will provide data on the probability that puma may temporarily move out of and back into the survey area between capture occasions. Unmarked puma that are subsequently GPS-collared should provide such information as well. This will allow us to determine the number of marked puma present in the search area each capturerecapture occasion and the probability that unmarked puma move in and out of the search area during each occasion. 2. Photographic captures and recaptures (i.e., camera traps) could be assessed in a 3-year pilot study (e.g., years 2―4) as an independent method for estimating puma numbers and density on the study area (Mace et al. 1994, Karanth 1995, Karanth and Nichols 1998, Karanth and Nichols 2001, Trolle and Kery 2003). The pilot study could be carried out by a graduate student and begin in the second winter, at which time we may begin estimating the puma population using capturerecapture methods (described above). 3. Genotype captures and recaptures could be assessed in coordination with the physical and photographic capture methods described above (1 and 2). This could also be initiated as a 3-year pilot study (e.g., years 2―4) done by a graduate student. Genotypes can be used to estimate minimum population size directly (Kohn et al. 1999) and in capture-recapture models (e.g., Otis et al. 1978, Boulanger et al. 2002). However, genotyping errors should be estimated and considered in population estimates (Creel et al. 2003). This project could be another independent non-invasive method for estimating puma numbers and density on the study area. Moreover, this method may be useful to monitor puma populations in Colorado (see Indices to Puma Abundance below). Vulnerability of Puma to Hunters― Objective 3 Puma hunting in Colorado normally involves hunters searching for puma tracks while driving snow-covered roads with four-wheel drive vehicles or snowmobiles. Thus, vulnerability of puma to hunters is associated with the frequency that puma cross roads (Murphy 1983, Van Dyke et al. 1986, Barnhurst 1986). Hunters active on snow may successfully catch puma >85% of the time that they release dogs on tracks, and road access influences puma hunting distribution (Murphy 1983). 76 �Road crossing frequencies of GPS-collared puma will be used to assess relative vulnerability of various sex, age, and reproductive classes of puma to detection by puma hunters. Density of roads (i.e., km/100 km2) will be estimated within each GPS-collared puma home range (i.e., 100% minimum convex polygon), each quadrat in the sampling frame, and the entire study area. Road crossings per puma per 24hour periods will be quantified (GPS collars programmed for 4 locations per day, each 6 hours apart). Comparisons will be made between road crossing frequencies of puma classes and road densities in home ranges for puma classes. Puma classes will be: adult females (≥2 yr. old) with no cubs; adult females with cubs ≤2 months old (i.e., nursling cubs), adult females with cubs 3―12 months old (i.e., weaned, carnivorous cubs), subadult females (i.e., independent females <2 yr. old), adult males (i.e., ≥2 years old), and subadult males (i.e., independent males <2 years old) (Logan and Sweanor 2001). Vulnerability will also be quantified using road crossing frequencies of different classes of puma (female, male, divided into adult and subadult age classes if individuals are known) during the track index discussed below. In addition, actual capture rates of those puma will be quantified during capturerecapture occasions (no. of captures per puma class/no. of puma in each class). Frequency of capturing known puma mothers with and without their cubs will also be tallied to quantify vulnerability of puma mothers to legal harvest (i.e., known mothers without cubs by their side). During the puma population decline (i.e., reduction) phase (years 6―10), hunters can be used to kill puma. Relative vulnerability to and selection by hunters will be quantified during hunting seasons by having hunters report number of hunter-days, number of tracks encountered and locations (fixed by GPS), number of times dogs were released on tracks, number of captures, characteristics of killed or captured puma (i.e., hunters may capture puma and release them). At the same time, researchers will have quantitative knowledge of puma available for harvest on the study area (as a result of ongoing capture and marking procedures and GPS data) to estimate capture rate per puma class. The hunter-kill will also allow a direct assessment of puma mothers in the harvest and a comparison of the fates of potential orphaned cubs with cubs in intact families. Indices to Puma Abundance― Objective 5 This project will develop and test both the efficacy and feasibility, including costs and other management considerations, of using indices to monitor changes in puma abundance. Two such indices are track counts and catch-per-unit-effort. These indices will be calibrated with the estimated puma population. Track Counts An index to puma abundance using counts and classification of puma tracks on snow-covered routes will be developed and tested. This will be done simultaneous with obtaining estimates of puma population size using capture-recapture occasions with houndsmen teams (above) and spatial analysis of home ranges of GPS-collared puma. A 3-year pilot study for this index may be started in year 2 and could be conducted by a graduate student. Information on puma gathered during years 1―2 will facilitate the exact research design. Experimental manipulation of the puma population resulting in a 5-year increase phase and a 5-year decline phase will allow testing the index through known puma population changes and assessment of the sensitivity of the index with the parameter in question― puma population size― by analyzing the statistical power of the method to detect population change (up or down) (Kendall et al. 1992, Beier and Cunningham 1996). The main operating assumption is that the frequency of finding puma tracks on snow-covered routes is related to puma numbers. We will examine the number of individual puma track sets/km (differentiated by size and direction of travel) and presence of puma tracks/km of search route to see 77 �which metric better detects actual population change (Kendall et al. 1992). In addition, we will examine how frequency of different classes of tracks (male, female, females with cubs) may relate to the known puma population changes. The most direct relationship will be a linear one between puma numbers and frequency of encountering puma tracks. Snow-tracking conditions (e.g., powder, crusted, slush, continuous, patchy) will be categorized each search day. Tracking teams, different than houndsmen teams used in capture-recapture occasions, will be used to make the puma track surveys 3 to 6 times per winter. Track surveys will be run on 4-wheel drive vehicles or snowmobiles 1―2 days following snowfall that covers the study area. Effort and costs will be quantified. Catch-Per-Unit-Effort The main operating assumption is that the amount of effort to capture puma is related to puma numbers (Lancia et al. 1996). The puma research team(s) will quantify the number of days required to capture individually identified puma with dogs during each year in the population increase phase. We will also quantify the number of days required to capture unmarked puma with dogs each year, and treat marked puma as though they have been removed from the unmarked population. Number of days per capture will also be quantified by capture teams involved in the capture-recapture occasions each winter. During the decline phase, puma hunters used to reduce the puma population will quantify the number of hunt days per puma captured each winter. Theoretically, the number of days per puma capture should increase as the puma population is reduced by 20% increments in years 6 and 7, and in possible further reductions in years 8―10. Hunters will also be asked to record (i.e., GPS location, date) the total number of tracks of independently-traveling puma (classified as male and female), and the number of tracks of female puma and cubs they encounter during their hunting periods. Male and female track categories will be distinguished by the width of the hind foot plantar pads. Hind foot plantar pad widths that are >52 mm will be classed as male; hind foot plantar pad widths ≤52 mm will be classes as female. We will explore functional relationships of these efforts to the estimated puma population on the study area. Quantifying Puma Diet and Ungulate Kill Rates― Objective 8 Data collected on puma diet and ungulate kills are not directly pertinent to a puma population study. However, they would be basic to an integrated study that involves effects of puma predation on mule deer and elk. Location clusters where puma are located for ≥2 nights will be investigated to estimate puma kill rates of ungulate prey (i.e., days/ungulate kill type/puma class) (Anderson and Lindzey 2003). Sex of animals will be determined by secondary sex characteristics and ages will be estimated from tooth eruption patterns (Quimby and Gaab 1952, Robinette et al. 1957, Dimmick and Pelton 1996) and cementum annuli of incisors (Low and Cowan 1963). Necropsies will be performed on all ungulate prey recovered in the field (Roffe et al. 1996), whether killed by puma or not, and data will be recorded on standard forms. If disease is suspected, whole carcasses or vital organ tissues will be collected and preserved by standard procedures (Roffe et al. 1996) and submitted for analysis to the Colorado Division of Wildlife’s Wildlife Health Laboratory or the Colorado State University Diagonostic Laboratory. An index to physical condition of ungulates prior to death will be estimated from percent marrow fat in femurs or metatarsi (depending on presence; femurs are preferred) (Neiland 1970, Mech and DelGiudice 1985, Fuller et al. 1986, Husseman et. al. 2003). Puma feces will be collected opportunistically year-round and stored by either freezing or oven drying (80-85oC, then stored in paper bags with a fumigant) for later macroscopic diet analysis (Big Sky Laboratory, Florence MT) to estimate frequency of occurrence of prey species (Litvaitis et al. 1996). This research component could also be carried out by a graduate student. 78 �Behavior of Puma Subject to Aversive Conditioning― Objective 9 Information on responses of puma to aversive conditioning is lacking. Individual puma with activities in residential areas on the study area might be research subjects on effectiveness of aversive conditioning. GPS collars on puma would be the primary source of behavioral response data before, during, and after aversive conditioning treatments. Behavior and Survival of Translocated Puma― Objective 10 If translocation is chosen as the method of reducing the puma population during the decline phase (years 6―10), then researchers will remove puma at rates needed to test research hypotheses. Prior to translocation, potential puma habitat areas for the release of the puma will need to be identified which are: 1) relatively remote, 2) large enough to accommodate exploratory movements up to 84 km away from release areas, and 3) not near human residential areas, domestic animal operations, or desert bighorn sheep populations (Ruth et al. 1998, Logan and Sweanor 2001). Puma will be captured alive on the study area, fit with new GPS collars, transported in wooden crates, provided food and water, and translocated by truck a minimum of 120 airline km (75 mi.) for females and 220 airline km (137 mi.) for males (Ruth et al. 1998). GPS collar locations will allow researchers to map movements of translocated puma. The VHF transmitters will allow researchers to quantify survival rates and agent-specific mortality rates. This research could also be carried out by a graduate student. ANALYTICAL Puma class survival rates and agent-specific mortality rates will be estimated by using KaplanMeier (Pollock et al. 1989a, b) and Trent and Rongstad procedures (Micromort software, Heisey and Fuller 1985). Cub survival curves for each gender will also be plotted with survival rate on age in months (Logan and Sweanor 2001:119). To analyze capture-recapture, photographic, and genetic capture-recapture data, closed population capture-recapture models are available in program CAPTURE obtainable at www.mbrpwrc.usgs.gov/software.html and program MARK obtainable at www.cnr.colostate.edu/~gwhite.). Closed population model selection can be achieved with the algorithm based on goodness-of-fit tests and between model tests in program CAPTURE (Otis et al. 1978). For open populations, programs JOLLY (for 1 age class), and JOLLYAGE (handles 2 age classes) are available at www.mbr-pwrc.usgs.gov/software.html. Programs JOLLY and JOLLYAGE contain chi-square goodness-of-fit tests of model assumptions and between model tests that enable researchers to choose the most appropriate model for the data (Pollock et al. 1990). NOREMARK (White 1996), also available at www.cnr.colostate.edu/~gwhite, has an extension that accommodates immigration and emigration; thus, it does not assume geographic closure (but demographic closure is still assumed). Finite rates of increase (Nt+1/Nt) between consecutive years and average annual rates of increase (r) for 3- to 5-year periods will be calculated (Caughley 1978, Van Ballenberghe 1983) and plotted. Graphical methods will be used to examine relationships of track counts and catch-per-unit effort (i.e., indices to puma abundance) to changes in the population of independent puma. Linear regression procedures and coefficients of determination will be used to assess functional relationships of track counts and catch-per-unit effort to changes in the population of independent puma if data for the response variable are normally distributed and the variance is the same at each level. If the relationship is not linear, data is non-normal, and variances are unequal, we will consider appropriate transformations of the data for regression procedures (Ott 1993). We will also consider non-parametric correlation methods, such as Spearman’s rank correlation coefficient, to test for a monotonic relationship between the index of abundance and the change in the puma population (Conover 1999). 79 �Statistical analyses will be performed using SYSTAT 10.2 and SAS 6.11. The risk of committing a type I error (i.e., concluding that a population change occurred when it did not) will be controlled at alpha = 0.10 because we will normally have small sample or population sizes (typical of large-carnivore studies). The higher alpha level will increase the probability of detecting a change and reduce the risk of a type II error (i.e., failing to reject a null hypothesis that is false). For managers, the risk of a type II error is probably more important. ArcView 8 geographic information system software will be used to map and analyze puma locations, movements, and home ranges. It will also be used to map and quantify attributes of the study area and sampling frame. PRELIMINARY SCHEDULE Years 1―5 (2005―2009) will be the puma population increase phase. Protecting the puma population from sport-hunting will be vital to allowing the puma population to increase within the bounds of the ecological carrying capacity of the study area. This will allow researchers to quantify baseline demographic data on the puma population and test indices to puma abundance during an increase phase. In this phase, capture-recapture occasions, track counts (for index to abundance), and photographic and genetic capture-recapture efforts will begin in about year 2 (2006). Years 6―10 (2010―2014) will be the puma population decline phase. Puma hunters (or translocation) will be used to experimentally reduce the puma population. The portion of independent puma (i.e., adults and subadults) in the population will be reduced by 20% in year 6 and 20% more in year 7 (i.e., a 40% reduction from year 5). Additional reductions may be made to test the indices to abundance or other hypotheses that may be developed and related to effects of harvest or puma predation on mule deer and elk. Those decisions can be made later in project development and as late as years 8―10. REGULATORY NEEDS Puma on the study area that may be involved in depredation of livestock or human safety incidences will not be treated any differently than other puma in Colorado, whether they are marked or not. Thus, they may be lethally controlled. Researchers that find that GPS-collared puma have killed domestic livestock will record such incidents to facilitate reimbursement to the property owner for loss of the animal(s). The increase phase in years 1―5 will require a temporary interruption of puma sport-hunting on the study area and protection of radio-collared puma that range off the study area. In years 6―10, regulated puma sport-hunting will resume. POTENTIAL COOPERATORS The Colorado Division of Wildlife will be the principal research and regulating agency in this program. The Bureau of Land Management and the Forest Service will be cooperators because the majority of the study area may be on lands under their management jurisdiction. U. S. D. A., A. P. H. I. S. Wildlife Services may provide puma capture assistance. Private landowners on the study area will be asked to cooperate with this effort. Colorado State University and other universities may cooperate by providing graduate research assistants and professors to carry out specific projects of the research program. Private individuals interested in the puma research may be asked to cooperate in puma capture and monitoring operations. 80 �Table 2. Preliminary puma research schedule. Phase Objectives: 1. 2. 3. 4. 5. 6. 7. 8. 9. Increase (Years 1―5) Decline (Years 6―10) Methods & Data: Initial capture & mark efforts of ≥90% of Capture-recapture estimates (yrs. 6―10) using independent puma (yrs. 1―2). data from physical and photographic & Capture-recapture estimates (yrs. 2―5; 3 genotype captures (if reliable). yr. minimum required for trend) using Reduce puma population using hunting or physical, photographic & genotype translocation. Reduce by 20% increments in captures. Quantify puma population sex & years 6 & 7. Puma hunting will continue, and age structure, density, & population there may be additional population reductions growth rate. in subsequent years. Quantify structure of hunter-kill. Quantify puma population vital rates as Quantify puma population vital rates as population increases. population declines. Quantify agent-specific mortality & puma Quantify agent-specific mortality & road-crossing frequency. vulnerability and selectivity of puma to hunters. Develop and test puma population models Develop and test puma population simulation validated by observed increase phase models validated by observed decrease phase puma population. puma population. Use track counts, catch-per-unit effort, & Use track counts, catch-per-unit effort, & genotype capture-recapture methods as genotype capture-recapture methods (if indices to puma abundance (yrs. 2―5). reliable) as indices to puma abundance. Use GPS data to quantify puma activity in Use GPS data to quantify puma activity in relation to people and human facilities on relation to people and human facilities on the the study area. study area. Use GPS data to quantify puma use of Use GPS data to quantify puma use of habitats habitats and landscape linkages. and landscape linkages. Estimate puma kill rates on mule deer and Estimate puma kill rates on mule deer and elk elk using GPS data. Quantify puma diet using GPS data. Quantify puma diet from from feces. feces. Describe & quantify behavior & survival of translocated puma if translocation is used to reduce the puma population. Begin final data analysis & report year 10. POTENTIAL IMPEDIMENTS Because of the relatively low densities of puma, difficulty of capture and research, obtaining needed sample sizes is expensive. Furthermore, multiple years of study are requisite to fulfill objectives. Collared puma that are killed, therefore, represent a significant effort and data loss. Minimizing such losses is a challenge that will improve the efficiency of the study. For certain projects within the program, experimental manipulations of the puma population on the primary study area, possibly ranging from extreme protection to extreme suppression at different stages of the project, are necessary to maximize reliability and scientific defensibility of findings. 81 �FINANCIAL ESTIMATES Conducting intensive puma research requires significant and steady financial support. Yearly costs during years 1-5 are estimated to range between $177,000 and $355,000 (Table 3). LITERATURE CITED Anderson, A. E. 1983. A critical review of literature on puma (Felis concolor). Colorado Division of Wildlife, Special Report No. 54. _____, D. C. Bowden, and D. M. Kattner. 1992. The puma on Uncompahgre Plateau, Colorado. Colorado Division of Wildlife, Technical Publication No. 40. Anderson, C. R., and F. G. Lindzey. 2003. Estimating cougar predation rates from GPS location clusters. Journal of Wildlife Management 67:307-316. Barnhurst, D. 1986. Vulnerability of cougars to hunting. Masters Thesis. Utah State University, Logan. Beier, P., and S. C. Cunningham. 1996. Power of track surveys to detect changes in cougar populations. Wildlife Society Bulletin 24:540-546. Bishop, C. J., G. C. White, D. J. Freddy, and B. E. Watkins. 2003. Effect of nutrition and habitat enhancements on mule deer recruitment and survival rates. Colorado Division of Wildlife, Wildlife Research Report, Federal Aid in Wildlife Restoration Project W-153-R, Progress Report. Ft. Collins, Colorado, U.S.A. Boulanger, J., G. C. White, B. N. McLellan, J. Woods, M. Proctor, and S. Himmer. 2002. A meta-analysis of grizzly bear DNA mark-recapture projects in British Columbia, Canada. Ursus 13:137-152. Brent, J. 1981. Colorado mountain lion population investigations: Game management units 33 and 40. Project Report October 27, 1980 through March 31, 1981. Colorado Division of Wildlife, Grand Junction. ______ _____. 1982. Colorado mountain lion population investigations: Game management units 33 and 40, Phase II. Project Report November 12, 1981 through March 31, 1982. Colorado Division of Wildlife, Grand Junction. _____. 1983. Colorado mountain lion population investigations: Game management units 33 and 40, Phase III. Project Report January 29, 1983 through August 30, 1983. Colorado Division of Wildlife, Grand Junction. Caughley, G. 1978. Analysis of vertebrate populations. John Wiley and Sons, New York. Colorado Division of Wildlife 2002-2007 Strategic Plan. 2002. Colorado Department of Natural Resources, Division of Wildlife. Denver. Conover, W. J. 1999. Practical nonparametric statistics. John Wiley & Sons, Inc., New York. Currier, M. J. P., and K. R. Russell. 1977. Mountain lion population and harvest near Canon City, Colorado, 1974-1977. Colorado Division of Wildlife Special Report No. 42. DeSimone, R., B. Semmens, B. Shinn, T. Chilton, J. Sikich, and B. Weisner. 2002. Garnet Mountains Mountain Lion Research. Progress Report, January 2001 to December 2002. Montana Fish, Wildlife and Parks, Helena. Dimmick, R. W., and M. R. Pelton. 1996. Criteria of sex and age. Pages 169-214 in T. A. Bookhout, editor. Research and management techniques for wildlife and habitats. Fifth edition. The Wildlife Society, Bethesda Maryland. Heisey, D. M., and T. K. Fuller. 1985. Evaluation of survival and cause specific mortality rates using telemetry data. Journal of Wildlife Management 49:668-674. Husseman, J. S., D. L. Murray, G. Power, and C. M. Mack. 2003. Correlation patterns of marrow fat in Rocky Mountain elk bones. Journal of Wildlife Management 67:742-746. Fuller, T. K., P. L. Coy, and W. J. Peterson. 1986. Marrow fat relationships among leg bones of whitetailed deer . Wildlife Society Bulletin 14:73-75. Karanth, K. U. 1995. Estimating tiger Panthera tigris populations from camera-trap data using capturerecapture models. Biological Conservation. 71:333-338. 82 �_____, and J. D. Nichols. 1998. Estimating tiger densities in India from camera trap data using photographic captures and recaptures. Ecology 79:2852-2862. _____, and J. D. Nichols. 2002. Monitoring tigers and their prey: A manual for researchers, managers and conservationists in tropical Asia. Centre for Wildlife Studies, Bangalore, India. Kendall, K. C., L. H. Metzgar, D. A. Patterson, and B. M. Steele. 1992. Power of sign surveys to monitor population trends. Ecological Applications 2:422-430. Koloski, J. H. 2002. Mountain lion ecology and management on the Southern Ute Indian Reservation. M. S. Thesis. Department of Zoology and Physiology, University of Wyoming, Laramie. Kufeld, R. C., J. H. Olterman, and D. C. Bowden 1980. A helicopter quadrat census for mule deer on Uncompaghre Plateau, Colorado. Journal of Wildlife Management 44:632-639. Laundre, J. W., L. Hernandez, D. Streubel, K. Altendorf, and C. L. Lopez Gonzalez. 2000. Aging mountain lions using gum-line recession. Wildlife Society Bulletin 28:963-966. Lancia, R. A., J. D. Nichols, and K. H. Pollock. 1996. Estimating the number of animals in wildlife populations. Pages 215-253 in T. A. Bookhout, editor, Research and management techniques for wildlife and habitats. The Wildlife Society, Bethesda, Maryland. Litvaitis, J. A., K. Titus, and E. M. Anderson. 1996. Measuring vertebrate us of terrestrial habitats and foods. Pages 254-274 in T. A. Bookhout, editor. Research and management techniques for wildlife and habitats. Fifth edition, revised. The Wildlife Society, Bethesda Maryland. Logan, K. A., E. T. Thorne, L. L. Irwin, and R. Skinner. 1986. Immobilizing wild mountain lions (Felis concolor) with ketamine hydrochloride and xylazine hydrochloride. Journal of Wildlife Diseases. 22:97-103. _____, L. L. Sweanor, J. F. Smith, and M. G. Hornocker. 1999. Capturing pumas with foot-hold snares. Wildlife Society Bulletin 27:201-208. _____, and L. L. Sweanor. 2001. Desert puma: evolutionary ecology and conservation of an enduring carnivore. Island Press, Washington, D.C. Mace, R. D., S. C. Minta, T. L. Manley, and K. E. Aune. 1994. Estimating grizzly bear opulation size using camera sightings. Wildlife Society Bulletin 22:74-83. McDaniel, G. W., K. S. McKelvey, J. R. Squires, and L. F. Ruggiero. 2000. Efficacy of lures and hair snares to detect lynx. Wildlife Society Bulletin 28:119-123. Mech, L. D., and G. D. DelGuidice. 1985. Limitations of the marrow fat technique as an indicator of body condition. Wildlife Society Bulletin 13:204-206. Miller, S. D., G. C. White, R. A. Sellers, H. V. Reynolds, J. W. Schoen, K. Titus, V. G. Barnes, Jr., R. B. Smith, R. R. Nelson, W. B. Ballard, and C. C. Schwartz. 1997. Brown and black bear density estimation in Alaska using radiotelemetry and replicated mark-resight techniques. Wildlife Monographs 133:1-55. Murphy, K. M. 1983. Relationships between a mountain lion population and hunting pressure in western Montana. Final Report, Federal Aid in Wildlife Restoration, Project W-120-R13 and 14. Montana Department of Fish, Wildlife, and Parks, Helena. Otis, D. L., K. P. Burnham, G. C. White, and D. R. Anderson. 1978. Statistical inference from capture data on closed animal populations. Wildlife Monographs 62:1-135. Ott, R. L. 1993. An introduction to statistical methods and data analysis. Fourth edition. Wadsworth Publishing Co., Belmont, California. Pojar, T. M. 2000. Investigating factors contributing to declining mule deer numbers. Colorado Division of Wildlife, Wildlife Research Report, Federal Aid in Wildlife Restoration Project W-153-R-13, Progress Report. Fort Collins, Colorado, U.S.A. Pollock, K. H., S. R. Winterstein, C. M. Bunck, and P. D. Curtis. 1989a. Survival analysis in telemetry studies: the staggered entry design. Journal of Wildlife Management 53:7-15. _____, S. R. Winterstein, and M. J. Conroy. 1989b. Estimation and analysis of survival distributions for radio tagged animals. Biometrics 45:99-109. _____, J. D. Nichols, C. Brownie, and J. E. Hines. 1990. Statistical inference for capture-recapture experiments. Wildlife Monographs 107:1-97. 83 �Quimby, D. C., and J. E. Gaab. 1952. Preliminary report on a study of elk dentition as a means of determining age classes. Proceedings Western Association of State Game and Fish Commissions. 32:225-227. Robinette, W. L., D. A. Jones, G. Rogers, and J. S. Gashwiler. 1957. Notes on tooth development and wear for Rocky Mountain mule deer. Journal of Wildlife Management 21:134-153. Ruth, Tik, K. A. Logan, L. L. Sweanor, M. G. Hornocker, and L. L. Temple (1998) Evaluating cougor translocation in New Mexico. Journal of Wildlife Management 62:1264-1275. Sweanor, L. L., K. A. Logan, and M. G. Hornocker. 2000. Cougar dispersal patterns, metapopulation dynamics, and conservation. Conservation Biology 14:798-808. _____, K. Logan, J. Bauer, and W. Boyce. 2004. Puma and humans in and around Cuyamaca Rancho State Park, San Diego County, California. School of Veterinary Medicine, University of California, Davis. Trolle, M., and M. Kery. 2003. Estimation of ocelot density in the pantanal using capture-recapture analysis of camera-trapping data. Journal of Mammalogy 84:607-614. Van Ballenberghe, V. 1983. Rate of increase of white-tailed deer on the George Reserve: a re-evaluation. Journal of Wildlife Management 47:1245-1247. Van Dyke, F. G., R. H. Brocke, and H. G. Shaw. 1986. Use of road track counts as indices of mountain lion presence. Journal of Wildlife Management 50:102-109. Van Sickle, W. D., and F. G. Lindzey. 1992. Evaluation of road track surveys for cougars (Felis concolor). Great Basin Naturalist 52:232-236. White, G. C., D. R. Anderson, K. P. Burnham, and D. L. Otis. 1982. Capture-recapture and removal methods for sampling closed populations. Los Alamos National Laboratory Publication LA-8787NERP. Los Alamos, NM, U.S.A. _____. 1996. NOREMARK: population estimation from mark-resighting surveys. Wildlife Society Bulletin 24:50-52. Prepared by ______________________ Kenneth A. Logan, Wildlife Researcher 84 �Table 3. Estimated project costs for years 1―5 only. Budget Item Personnel: -DOW Researcher -Houndsman -Project Technician -Temporary Technician Volunteers Support: Lodging, food, fuel for 8―12 Vehicles: -4x4 Trucks (2) -all terrain vehicles (3) -snowmobiles (1)a -utility trailers (1)a -dog sled & trailer Gasoline Vehicle Maintenance GPS- & Radio-telemetry: -GPS-collars -VHF-collars (cub) -cub collar material -command unit (1) -receivers, type (2) -H-antennae (3) -omni antennae (3) -coaxial cables (2) -coaxial cables (2) -antenna switch-box (2) -intercom system (1) -head sets (2) Capture Equipment: -drugs -cage trap -snares -call box -miscellaneous (darts, vials, syringes, needles, envelopes, gloves, tapes, calipers, thermometers, ear-tags, tattoos, etc...) Dog Care: -veterinary -food Aerial Support: Work tack: -backpacks, climbing gear, nets, ropes, office materials, etc... Laboratory: -genetics -carcass or tissue analysis -fecal analysis Photographic: -trail cameras (~40) -film (~200) Total a Year 1 Year 2 Year 3 Year 4 Year 5 K. Logan’s support not incl. 12,500 ($2,500/mo.*5 mo.) 35,000 33,000 12,500 35,000 33,000 13,125 35,000 33,000 13,125 35,000 33,000 13,781 35,000 33,000 0 15,000 15,000 15,000 15,000 50,000 (2*$25,000) 18,600 (3*$6,200) 5,300 1,900 1,000 4,800 (2*$2,400) 3,000 0 0 0 0 0 4,800 3,000 0 0 0 0 0 4,800 3,000 60,000 0 0 0 0 4,800 3,000 0 0 0 0 0 4,800 3,000 114,750 (25*$4,590) 3,192 (12*$266) 100 4,500 5,390 (2*$2,695) 636 (3*$212) 234 (3*$78) 56 (2*$28) 29 (2*$14.50) 112 (2*$56) 285 300 (2*150) 0 0 0 0 0 0 0 0 0 0 0 0 22,950 1,596 50 0 0 0 0 0 29 0 0 0 22,950 0 0 0 0 0 78 56 0 56 0 0 22,950 1,596 50 0 0 0 0 0 29 0 0 0 650 (25*$26/bottle Telazol) 2,000 (2*$1,000) 1,000 850 1,000 650 0 0 0 1,000 650 0 0 0 1,000 650 0 0 0 1,000 650 0 0 0 1,000 2,000 600 (120/mo.*5 mo.) 40,000 ($200/hr x 4 hr x 50) 2,000 600 40,000 2,000 600 40,000 2,000 600 40,000 2,000 600 40,000 2,000 1,000 1,000 1,000 1,000 4,650 (25*$186) 600 4,650 4,650 600 4,650 4,650 600 4,650 4,650 600 4,650 4,650 600 4,650 0 (40*$430) 0 (250*$6) 354,684 17,200 1,500 177,150 4,300 1,500 189,500 0 1,500 243,715 0 1,500 185,856 Two snowmobiles and 1 trailer are already available for the project. 85 �APPENDIX I Sex Determination of Mountain Lions Bayed in Trees With little effort the sex can be determined for a mountain lion bayed in a tree. Refer to the photos of the different lions, 4 males (A―D) 2 females (E, F), attached to these tips. Male adult and subadult lions have a conspicuous black spot of hair, about 1 inch diameter, surrounding the opening to the penis sheath behind the hind legs and about 4 to 5 inches below the anus. In between the black spot and the anus is the scrotum, which is usually covered with silver, light brown, and white hair. Therefore, look for the black spot and scrotum. The anus is usually hidden below the base of the tail. Female adult and subadult lions do not have the black spot or scrotum behind the hind legs and below the base of the tail. There is just white hair there. The anus is directly below the base of the tail, and the vulva is directly below the anus. The anus and vulva are usually hidden by the base of the tail. Teats of females are usually inconspicuous, even of mothers with weaned cubs or mothers that have just finished nursing cubs. So teats are usually not a good indicator of sex in treed lions. Sometimes sex determination of lions can be done with the naked eye. But use a pair of binoculars to make sexing lions easier. If a lion’s position in a tree obscures your view, you can get the lion to move around for a better look. Pick up a baseball-bat-size branch and bang on the trunk of the tree. If there is snow on the ground, throw a few snow balls at the lion. You can even climb the tree toward the lion. These actions usually get the lion to move. When it does, be ready to sex the lion. Also, sometimes the lion urinates when bayed by dogs or when a person climbs the tree toward it. Look for the origin of the urine stream. If the urine stream comes from behind the hind legs about 4 to 5 inches below the anus, then the lion is probably a male. If the urine stream comes from under the base of the tail, then it’s probably a female. Tracks may also be indicative of sex. Adult and large subadult male lions usually have hind foot plantar (“heel”) pad widths that exceed 2 1/16 inches (52 mm). Adult and subadult female lions usually have hind foot plantar pad widths less than or equal to 2 1/16 inches. Carry a small ruler or wind-up metal tape in your pocket to make measurements 86 �Male Mountain Lions (A―D) Penis Spot, Scrotum, Anus. Penis (black) spot ~1 inch dia. is ~4-5 inches below anus. A B K. Logan photo K. Logan photo D K. Logan photo Female Mountain Lions (E, F) Vulva directly below anus, both usually hidden by base of tail. No “black spot” 4-5 inches below anus C E K. Logan photo K. Logan photo F 87 K. Logan photo � https://cpw.cvlcollections.org/files/original/0f0af4b0f2b27135880519535d4be4ba.pdf 3326bfe975ebf1787bb8f268b2de5016 PDF Text Text 239 Colorado Division of Wildlife Wildlife Research Report July 2001 and July 2002 PROGRESS REPORT State of_________C=ol=o=rad=o,___ _ __ Division of Wildlife - Mammals Research Work Package No._ _ _ _. ::;3.. :a.0=03:::___ _ __ Predatory Mammals Conservation Task No_ - - - - - - - - - - ' 3 " - - - - - - - Peiod Covered: Pilot Study - Evaluation of GPS Technology in Measuring Chronic Wasting Disease Prevalence Among Deer Preyed upon by Puma January 1, 2001 - December 31, 2001 Author: C. E. Krumm, T.D.I. Beck, M. W. Miller. Personnel: C. E. Krumm, T.D.I. Beck, M. W. Miller Interim Report - Preliminary Results This work continues, and precise analysis of data has yet to be accomplished. A1anipulation or interpretation of these data beyond that contained in this report should be labeled as such and is discouraged. ABSTRACT A prospectus for a pilot study to ascertain the efficacy and feasibility of using Global Position Systems (GPS) technology to measure chronic wasting disease prevalence among puma prey, as well as in other studies of puma, was developed. Objectives of the pilot study are to: 1) Evaluate the potential utility of Televilt Positioning GPS collars in studies of selective predation in puma under field conditions; and 2) Develop and assess the adequacy of field sampling techniques for studying selective predation on CWD-infected mule deer. Two adult puma are to be captured and fitted with GPS collars for the pilot study. ��241 PILOT STUDY Evaluation of new GPS technology in measuring chronic wasting disease prevalence among deer preyed upon by mountain lions C. E. Krumm, T. D. I. Beck, and M. W. Miller Background As a pilot study to test a new technology in Global Positioning Systems (GPS) and its application to studies of predator-prey relationships, we plan to capture and collar two free-ranging puma (Puma concolor) in the foothills west of Ft. Collins in early April of 2001. Our pilot study will evaluate new GPS technology, as well as the potential utility of data collected with this system in testing hypotheses about selective predation; specifically, we will evaluate the ability to compare chronic wasting disease prevalence among puma-killed deer to prevalence among harvested deer. Chronic wasting disease (CWD) is a naturally occurring spongiform encephalopathy of captive and freeranging deer and elk. CWD has become a concern in managing deer herds in northeastern Colorado. Studies conducted the past several years have provided important data on prevalence of CWD (Miller et al. 2000) and the potential effects of selective population control on affected populations (Gross and Miller 2001). It follows that processes fostering selective removal of affected individuals, like test-andslaughter or predation, should be closely evaluated in the context of disease management. New technology in GPS tracking of animals by Televilt Positioning (Lindesberg, Sweden) allows location data to be downloaded remotely without retrieval of collars. Testing the effectiveness and accuracy of these collars will benefit a suite of studies that are being planned across Colorado to examine the selectivity of puma for prey animals (specifically mule deer) of varying condition. These studies will help to answer a fundamental ecological question: Do puma selectively prey on debilitated or compromised animals rather than healthy ones? Objectives Our specific objectives are to: 1. evaluate the potential utility of Televilt Positioning GPS collars in studies of selective predation in puma under field conditions; and 2. develop and assess the adequacy of field sampling techniques for studying selective predation on CWD-infected mule deer. Study Design Because this is a pilot study, we will capture and collar only two adult puma to evaluate equipment and sampling techniques. We regard two individuals as the fewest needed to adequately assess all aspects of equipment use and performance, sampling techniques, and other logistical facets of larger prospective studies. Capture Methods and Handling We plan to capture adult puma for this study using methods described in Shaw ( 1979). Briefly, a tracker with experience in tracking and handling mountain lions will be hired to facilitate capture and will use trained dogs to track and tree or bay each mountain lion. Field anesthesia will be supervised by an attending veterinarian. Anesthetic drugs will be administered intramuscularly via projectile syringe using a gas-powered projector. For capture, puma will be anesthetized with ketamine (10-1 lmg/kg) and xylazine HCl (1.8-2mg/kg) orketamine (2 mg/kg) and medetomidine (0.075 mg/kg) (Shaw 1979, Kreeger 1996). We will observe darted puma for signs of sedation (salivation, unsteadiness of head and body, and a wide-eyed expression). If the puma is treed, then people and dogs will be removed from the immediate area to give the animal a chance to descend before becoming completely anesthetized. If the puma remains in the tree until almost completely anesthetized, then someone wearing climbing gear will climb to the puma and attach either a chest harness (preferred) or hind leg noose and quickly lower the animal �242 before it falls; others will hold a taut net below to break the puma's fall should it slip before a harness or rope can be secured. If signs of anesthesia are inapparent after 15 minutes, then a second full injection will be given. Upon first approach of an apparently anesthetized puma, a 4-5 foot stick will be used to gently prod the paws and muzzle of the animal; if there is no response (i.e. snarling or biting), then we will assume anesthesia is sufficient for handling. Once anesthetized, we will apply eye ointment and a blindfold to reduce visual stimuli, place gauze pads in the puma's ears to reduce auditory stimuli, and restrain its legs with nylon belts or hobbles. A GPS-Simplex collar (Televilt Positioning; maximum weight 600 g) will be fastened around the puma's neck. The leg restraints will be quickly removed, and the puma will be allowed to recover from the sedation either naturally or with the aid of an antagonist; when prescribed, yohimbine HCl (0.125 mg/kg IV) will be used to antagonize xylazine sedation and atipamezole (0.3 mg/kg) will be used to antagonize medetomidine sedation. Postcapture Monitoring According to the manufacturer, the locations of collared animals can be retrieved and plotted several times a day without removing the collars. Up to 2000 satellite positions can be stored in the memory, allowing us to closely monitor the puma's movement on a daily basis. If a puma remains in one location for several hours, we will assume that it has made a kill. Based on data from studies elsewhere (e.g., Homocker 1970, C. Anderson, personal communication), we anticipate that each collared animal will make an ungulate kill every 7 to 11 days on average. We will locate the prospective kill site using the GPS-Simplex system. We will evaluate whether using this system allows us to locate kill sites quickly enough to retrieve a suitable tissue sample to test for CWD. If the animal killed is a deer, the presence of suitable diagnostic samples (brain stem and tonsil tissues) and overall carcass condition will be noted, and tissues will be taken to test for CWD when available. To evaluate the effect of carcass sampling activities on puma behavior, we will alternate taking the entire head of the kill with sampling only the necessary tissues in the field to compare the effect on the puma's return to the kill. The animals will be monitored closely after the kill has been sampled to ensure our handling does not interfere with their return to the kill site. Generally, researchers' presence at and inspection of a kill site does not dissuade a puma from returning to that site (T. Beck, unpublished data). However, if it becomes apparent that one technique is more disruptive than the other, then we will adopt the least disruptive sampling technique for the remainder of the study. Both puma will remain collared for a period of no less than one month unless the collars appear to be adversely affecting them. We will monitor each animal for changes in behavior like decreased kill rates or mobility that may be attributed to the collars. If the collars seem to have no adverse effects on the puma, then they will remain in place until the batteries must be replaced (about 3-4 mo, depending on final programming configuration). If the collars need to be removed for any reason, the same capture and handling methods as described above will be used for recapture. • Data from this pilot study will be used in designing more comprehensive studies ofpuma~eer relationships in Colorado, and may be of use in other studies of predator-prey ecology. Literature Cited Gross, J. E., and M. W. Miller. 2001. Chronic wasting disease in mule deer: A model of disease dynamics, control options, and population consequences. J. Wildl. Manage. In press. Homocker, M. G. 1970. An analysis of mountain lion predation upon mule deer and elk in the Idaho Primative Area. Wildlife Monographs 21. Kreeger, T. J. 1996. Handbook of wildlife chemical immobilization. International Wildlife Veterinary Sciences, Inc. Laramie, Wyoming, USA. Miller, M. W., E. S. Williams, C. W. McCarty, T. R. Spraker, T. J. Kreeger, C. T. Larsen, and E.T. Thome. 2000. Epizootiology of chronic wasting disease in free-ranging cervids in Colorado and Wyoming. Journal of Wildlife Diseases 38:676-690. Shaw, H.G. 1979. Mountain lion field guide. Fourth edition. Arizona Game and Fish, Phoenix, Arizona, USA. � https://cpw.cvlcollections.org/files/original/2fe1f3f8bb3bc1b59bdade9a25849384.pdf 03f83b1b0aa6b61ba3c15936da81fc1c PDF Text Text Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Project No. Work Package No. Task Colorado Federal Aid Project: N/A 3740 : : : : Cost Center 3430 Mammals Research Wildlife Diseases Pilot evaluation of GPS technology in chronic wasting disease prevalence and management at artificial feeding sites in urban areas. : Period Covered: April 1 2003 through July 31, 2004 Author: Eric J. Bergman, Michael W. Miller and L. L. Wolfe Personnel: M. Sirochman All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT A pilot study for assessing the utility of GPS technology in the evaluation of CWD prevalence and management in urban areas was designed is being implemented. Objectives of this pilot study are to: 1) Evaluate the utility of GPS radio collar technology in identifying artificial feed sites in urban settings, 2) Evaluate if there is evidence that artificial feed sites reduce the size of deer home ranges, 3) Evaluate if deer density is elevated at artificial feed sites, and 4) Evaluate if CWD prevalence is higher at artificial feed sites . 119 �JOB PROGRESS REPORT PILOT EVALUATION OF GPS TECHNOLOGY IN CHRONIC WASTING DISEASE PREVALENCE AND MANAGEMENT AT ARTIFICIAL FEEDING SITES IN URBAN AREAS Eric J. Bergman, Michael W. Miller and L. L. Wolfe INTRODUCTION Analyses of data from recent field studies and from culling have revealed areas of relatively high CWD prevalence associated with urban areas along the northern Front Range (Wolfe et al. 2002, 2004; Conner and Miller 2004; Farnsworth et al. 2004). Within these, artificial and illegal feeding sites may be particularly important because they appear to congregate deer in one location, thereby increasing local deer density and exposure to contaminated environments (Miller et al. 2004). Although the nature of the relationship between disease prevalence and mule deer density has not been definitively identified, it seems likely (Barlow 1996) that CWD prevalence is being indirectly elevated through artificial feeding. The development of global positioning system (GPS) technology and its incorporation into radio collars for wildlife research presents a tool for better understanding CWD in urban areas. We have initiated a pilot field study to: 1) evaluate the effectiveness of different GPS collars in identifying illegal feed sites in urban settings, and 2) develop and evaluate a strategy for utilizing GPS technology in studying and managing CWD in urban mule deer populations. METHODS The study area for this work is located within two subdivisions in Estes Park, Colorado. The subdivisions, separated by approximately 1.6 km, were identified as treatment and control sites based on the presence and absence of known feeding sites (Fig.1, Wolfe et al. 2004). Between five and eight adult (>1 yr old) female deer from each subdivision were captured and collared with one of two different brands of GPS collars (HABIT Research, British Columbia, Canada and LOTEK Wireless, Ontario, Canada). Collars from each company will be evenly distributed between sites. Capture will occur as part of an ongoing "test and cull" research project (Wolfe et al. 2004) during April 2004 and from August to October of 2004 as needed. Deer will be recaptured and collars will be removed prior to battery failure (~220 days service) in order to retrieve GPS data. No specific hypotheses are being tested in this pilot study; rather, we are attempting to determine if GPS radio collar technology is adequate for use as a tool in refining CWD epidemiology and management. We will record and report on the performance of GPS collars, and calculate costs (mean, range per animal tested) associated with our artificial feed site identification strategy as implemented in this pilot study. However, we will compare home range sizes of deer from each site to determine if artificial feeding reduces home range size of deer. We will also incorporate ground survey data (Wolfe et al. 2004) to estimate and compare mule deer density and ultimately CWD prevalence from sampled deer at each site. CWD prevalence will be compared between sites as well as to previous estimates from the greater Estes Park area (Wolfe et al. 2004) to explore future research potential. 120 �RESULTS AND DISCUSSION GPS Collar Comparison A total of 16 GPS collars (10 LOTEK, 6 HABIT) were available for testing in this study. Prior to initiation of this study no HABIT collars were on hand for deployment, rather, all 6 had to be built to specification and delivered. GPS collars from HABIT Research, ~$1,800/unit, were programmed to: collect GPS locations every 2 hours, to transmit GPS data (via VHF signal) over two day intervals every two weeks and to transmit the most recent GPS location (via VHF signal) at the start of each minute. Due to delays in the manufacturing process, no HABIT collars were received in time for spring deployment (≥2 weeks pre-fawning). Additionally, due to programming errors, 0 of 6 HABIT collars were ready for deployment after initial testing. Upon servicing by HABIT Research (~3.5 weeks), 3 of 6 collars appear to be ready for deployment in late summer 2004. The remaining HABIT collars (3 of 6) will be serviced and deployed upon satisfactory performance. All LOTEK collars were on hand prior to initiation of this study. Eight of 10 collars were deployed in spring of 2004, with 1 of 10 needing service. GPS collars from LOTEK Wireless, ~$3,500/unit, were also programmed to collect GPS locations every 2 hours, but did not offer remote download capabilities. All GPS locations collected by LOTEK collars will be acquired upon retrieval of the collar. GPS Collar Performance Data from LOTEK GPS collars continues to be collected and HABIT GPS collars will be deployed between August-September 2004. LITERATURE CITED Barlow, N.D. 1996. The ecology of wildlife disease control: simple models revisited. Journal of Applied Ecology 33:303-314. Conner, M.M., and M.W. Miller. 2004. Spatial epidemiology in natural populations: a case study of movement and prion disease prevalence relationships among mule deer population units. Ecological Applications (in press). Farnsworth, M.L., L.L. Wolfe, N.T. Hobbs, K.P. Burnham, D.M. Theobald, and M.W. Miller. 2004. Human land use influences chronic wasting disease prevalence in mule deer. Ecological Applications: in review. Miller, M.W., E.S. Williams, N.T. Hobbs, and L.L. Wolfe. 2004. Environmental sources of prion transmission in mule deer. Emerging Infectious Diseases: in press. Wolfe, L.L., M.M. Conner, T.H. Baker, V.J. Dreitz, K.P. Burnham, E.S. Williams, N.T. Hobbs, and M.W. Miller. 2002. Evaluation of antemortem sampling to estimate chronic wasting disease prevalence in free-ranging mule deer. Journal of Wildlife Mangement 66:564-573. _________, M.W. Miller, and E.S. Williams. 2004. Feasibility of "'test-and-cull" for managing chronic wasting disease in urban deer. Wildlife Society Bulletin 32:500-505. Prepared by ____________________________ Eric J. Bergman, Wildlife Researcher 121 � https://cpw.cvlcollections.org/files/original/2376790d1e8298352a11238991361be7.pdf 0536b4699cd92aed373d8e6d7594a909 PDF Text Text Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Project No. Work Package No. Task Federal Aid Project Colorado 3740 N/A : : : : Cost Center 3430 Mammals Research Mammals Support Services Veterinary Services – Medical Support : Period Covered: July 1 2003 through June 30, 2004 Author: L. L. Wolfe Personnel: M. W. Miller, L. A. Baeten, M. M. Conner, K. Cramer, T. R. Davis, K. Griffin, D. O. Hunter, J. E. Jewell, E. Knox, C. E. Krumm, C. T. Larsen, J. Rhyan, M. Sirochman, T. Sirochman, E. S. Williams, D. Wroe All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. 123 �JOB PROGRESS REPORT VETERINARY SERVICES – MEDICAL SUPPORT L.L. Wolfe INTRODUCTION Veterinary services are provided as support for a variety of wildlife research projects, transplants and reintroductions conducted by the Colorado Division of Wildlife (CDOW) and its collaborators throughout the year. The following overviews and summarizes key wildlife veterinary medical support services provided during 2003−2004. VETERINARY MEDICAL SUPPORT Location of services & primary investigator CDOW Foothills Wildlife Research Facility (FWRF), Tracy Davis and researchers Rocky Mountain Arsenal, Sherry Skipper Species mule deer, white-tailed deer, elk, bighorn sheep, pronghorn, puma, others mule deer, white-tailed deer Uncompahgre Plateau, Chad Bishop mule deer Pinion Canyon maneuver site, Elizabeth Joyce swift fox Colorado Springs, Brian Dreher mule deer CDOW FWRF, Department of Defense contract in cooperation with Elizabeth Williams mule deer, white-tailed deer Type of medical support Preventive, routine, and emergency medical care for all research animals housed at FWRF for use in ongoing CDOW and research. Chemical immobilization of adult does for survival study and CWD surveillance. Does were ultrasounded, tonsil biopsied, blood was collected, and vaginal implant transmitters (VITs) were inserted. Medical care of injured animals, assisted with ultrasound, VIT, and blood collection for viral serosurvey and thyroid study. Swift fox kits were anesthetized and abodminal radiotransmitters were surgically inserted and blood was collection. Adult deer were captured and radiocollared for CWD surveillance. We tonsil biopsied deer, collected blood, and provided area training for future efforts Provided medical care for hand raised deer fawns, including diarrhea outbreak management and treatment of injured fawns. TRAINING Capture and Sampling A capture and handling training class was provided for the district wildlife manager trainees. A second class was held for researchers and biologists. Capture classes included lectures on drug use regulations and recordkeeping, pharmacology of select capture drugs, dosing, safety and types of equipment. These classes also included “hands on” capture and handling of animals at FWRF. This last year, we also devised and administered a written and practical exam for the DWM trainees. As needed, spaces in these classes also provide opportunities for graduate students and technicians to learn sample collection techniques. 124 �TONSIL BIOPSY Tonsil biopsy training sessions were provided for staff from Wyoming Game and Fish Department and the Wisconsin Department of Natural Resources. In addition, CDOW personnel from the Colorado Springs area were trained on-site in capture and sampling techniques. Tonsil biopsy training sessions included lectures on sampling techniques and recognizing signs of CWD. The training also involves “hands on” time in the necropsy lab for sampling techniques. We also provide hands on training, as scheduling allows, on research animals. TARGETED CWD SURVEILLANCE A training module was developed and used to instruct USFWS and tribal biologists in recognizing signs of CWD as a tool in developing targeted surveillance programs on national wildlife refuges and tribal lands. This training included a lecture and PowerPoint tutorial illustrating clinical signs of CWD, as well as first-hand observation of captive mule deer showing signs of CWD. The tutorial file was subsequently modified and made available to other state and federal agencies as a self-teaching tool to aid in respective CWD management programs. TRANSLOCATION During transplant and translocation operations, we provide emergency medical care or humane euthanasia to injured animals. We also provide blood sampling, health exams, health certificates, vaccinations, anthelmintics, and antibiotics as needed to assure safe transport and improve survival in translocated wildlife. Species BHS swift fox black-footed ferrets Lynx Services Provided Vaccination, anthelmintics, antibiotics, pharyngeal swabs Health exams and health certificates Health exams and health certificates Entry and release exams, medical care for capture injuries Comments Transplanted within Colorado Released in South Dakota Released in Utah Reintroduction project* *Lynx reintroduction: Thirty eight lynx were received for the 2004 release. Most of the lynx were in good condition on arrival at the lynx holding facility in Del Norte. Three lynx required digit amputation due to trapping injuries, but all three healed without complications. One female was euthanized due to a compound fracture of the radius-ulna. In 2003/2004, the anesthetic protocol for transported lynx was changed from 2.0-2.5 mg/kg Telazol delivered intramuscularly (IM) to 20-40 mg (0.02 -0.05 mg/kg) ketamine and 0.6-0.8 mg (0.050.11 mg/kg) medetomidine given IM. No adverse anesthetic reactions were seen. Lynx were given the ketamine/medetomidine by IM injection while held in a squeeze cage. Induction time averaged 5.1 minutes (S.E. 0.4). Anesthetic time (induction to reversal) averaged 24.5 minutes (S.E. 0.9). Lynx were given atipamazole (0.25−0.55 mg/kg) in equal volume to medetomidine by IM injection. Lynx were recovered with minimal stimulation in their den boxes. Recovery time (time from reversal to standing with coordination) averaged 48.5 minutes (S.E. 2.5). There were no poor recoveries or anesthetic reactions. This new drug combination offered substantial reduction in processing and recovery times for lynx being handled at different points in the reintroduction process. ® DRUG DISTRIBUTION Since 2002, there has been extensive reorganization of drug distribution procedures and recordkeeping for chemical capture of wildlife. Overall, there has been a dramatic improvement in drug tracking and accountability. This has resulted in a reduction in wasted expired drugs and improvement in field logs. 125 �Telazol summary 800 700 □ Total bottles Telazol bottles 600 ■ bottles prescribed 500 - - 400 - - 300 200 100 0 □ bottles "out" unknown ~ - - ~ - - □ field use logged - ■ unreported field logs - - - - □ bottles returned expired 1- l r- - 2000 2001 2002 2003 ■ bottles in FW RF safe 2004 year CLINICAL TRIALS The following table summarizes the clinical trials from 2003. These trials were designed and conducted to improve veterinary medical care associated with various research and management programs conducted by CDOW. More complete reports of these trials are in the appendices. Clinical Trial Investigators Clostridium perfringens type A vaccine trial Plague vaccine in Canada lynx Wolfe, Miller, Davis, Ellis BHS, MD A Wolfe, Shenk, Baeten, Miller, Roke Wolfe, Miller lynx B mountain lions, MD, WTD C Wolfe, Miller, Lance MD Published Baker, Wolfe, MD Wolfe, Ryan, Miller fallow deer See progress report for Dan Baker D Chemical immobilization field trial with medetomidine and ketamine combination Chemical immobilization with A-3080 in mule deer Chemosterilization with GnRH toxin in mule deer Comparison of dart injection quality between 2 brands of collared and uncollared darts in fallow deer 126 Species Appendix �APPENDIX A EXPERIMENTAL EVALUATION OF A VACCINE FOR CLOSTRIDIUM PERFRENGENS TYPE A IN CAPTIVE BIGHORN SHEEP (Ovis canadensis) AND CAPTIVE MULE DEER (Odocoileus hemionus) L. L. Wolfe, R. P. Ellis, K. Fox, T. Davis, and M. W. Miller INTRODUCTION Clostridium perfringens is found naturally in the intestines of animals and in the environment. This bacterium possesses the ability to produce heat-resistant endospores and potent extracellular toxins. Isolates of C. perfringens can be subdivided into types based on the production of these exotoxins. The four major toxins implicated in disease are α, β, ε, and ι. Of these four major toxins, type A produces α toxin only, type B produces α, β, and ε, type C produces α and β, type D produces α and ε, and type E produces α and ι toxins. Other minor toxins also exist within the five types of C. perfringens, although they are not used to identify the specific type due to overlap between types. These toxins include δ (found in types B and C), θ (found in all five types), κ (found in all five types), λ (found in types B, D, and E), µ (found in types A, B, C, and D) ν (found in all five types), and neuraminidase or sialidase (found in all five types). In addition, C. perfringens enterotoxin (CPE) is often produced. CPE is most often found occurring with type A, although it has also been documented with all five types of C. perfringens. Many toxins produced by C. perfringens organisms are hydrolytic enzymes, necessary for life as a saprobe found naturally in the soil. Type A, the focus of this study, also possesses enzymes with hydrolytic properties, including phospholipase C and sphingomyelinase activities (from the α toxin). (Petit et al. 1999) Clostridium perfingens type A has recently been implicated as a cause of enterotoxemia in a variety of species including lambs and goats. Tympany, hemorrhagic enteritis and abomasitis, and abomasal ulceration in calves characterize the disease. Lesions include necrotic enteritis in domestic chickens; necrotizing enterocolitis and villous atrophy in suckling and feeder pigs; and hemorrhagic gastroenteritis in dogs. (Bueschel et al. 1998). Other reports of enteric disease associated with C. perfringens type A include enterotoxemia in minks, muskrats, and racing camels, acute toxemia in water buffaloes (Songer, 1996), gastroenteritis in black-footed ferrets (Schulman et al., 1993) and dairy cattle (Dennison et al., 2002). The α toxin in type A C. perfringens acts by way of phospholipase C activity and sphingomyelinase activity, breaking down phosphatidylcholine and sphingomyelin found in the membranes of erythrocytes, platelets, leukocytes and endothelial and muscle cells. By way of this action, α toxin is thought to be responsible for the cytotoxicity, necrosis, and hemolysis observed with type A C perfringens. There is evidence suggesting that minor differences in the amino acid sequence of α toxins exist, creating two strains with different pathways of infection. One strain has an increased resistance to chymotrypsin, allowing survival and multiplication in the gut, followed by entry into circulation. This strain is believed to be the primary cause of type A related enterotoxemia. The other strain, lacking a resistance to chymotrypsin, is believed to have a higher affinity for invasion of muscle tissue, and perhaps the cause of type A related gas gangrene (Songer, 1996). Clinical signs of animals suffering from type A C. perfringens vary from species to species, but consistently include depression, anorexia, diarrhea, bloating in non-avian species, and death. Postmortem findings from these cases varies from species to species and between specific cases, but consistently tend to include gram positive bacilli surrounding necrotic tissue, necrosis, particularly in the small intestine, 127 �and hemorrhage and ulceration, again in the small intestine; in ruminants, abomasitis, tympany, and abomasal hemorrhage and ulceration are also common findings. Infection by type A C. perfringens is believed to occur in a variety of ways. One theory, especially in neonatal ruminant cases, is that engorgement on milk or esophageal groove dysfunction allows milk to spill over into the rumen, providing a substrate for growth of the bacterium, as well as an anaerobic environment in which to proliferate. Another suggested scenario, as found in cattle herds in Nebraska and Wyoming is that bacterial infection is secondary to copper deficiency. The findings of this study indicated that low copper concentrations may have weakened the abomasal mucosa and compromised immune function (Roeder et al., 1988). Environmental contamination may play a role in the acquisition of C. perfringens type A because these toxins are known to exist in the soil and many ruminant species ingest soil in attempts to acquire essential minerals. In addition, α toxin can be detected in the feces of birds with necrotic enteritis (Bueschel et al., 1998), and thus avian vectors may provide an additional method of toxin movement. The presence of C. perfringens type A at the Foothills Wildlife Research Facility (FWRF) in Ft. Collins, Colorado appears to be relatively recent, with the first case identified in 1997. Since then, the number of cases has increased exponentially, approximately doubling each year. A total of 29 cases had been attributed to C. perfringens type A since the first case was diagnosed in 1997. At the FWRF, the disease has affected primarily bighorn sheep and mule deer; these 2 species account for 25 of the 29 cases. In adult animals, bighorn sheep have been the primary species affected (6 out of 10 cases), and sudden death has been common. These animals often exhibited bloating and diarrhea shortly before death, and showed signs of hemorrhage such as bleeding from the mouth or anus. Necropsies of these animals consistently included large amounts of rod-shaped bacteria, especially in the small intestines. Other lesions included necrosis and hemorrhage, particularly in the heart and small intestine as well as intestinal and abomasal ulcers. In neonatal and juvenile animals (<1 yr old), mule deer have been the primary species affected (14 of 19 cases), and chronic symptoms have been most common. These animals consistently exhibited chronic bloating, emaciation, depression, and soft brown diarrhea that was sometimes chronic, usually present early on in the animal’s life, and often never alleviated despite various treatment attempts. Therapies included a barrage of antibiotics (benzathine penicillin and florfenicol appeared most effective), subcutaneous fluids, transfaunation, kaolin pectin with lactobacillus granules, probios powder, electrolytes, and medicated “Deccox” feed distributed by Ranchway Feeds. Despite therapy, most of these cases ended in death -- those that survived exhibited symptoms that were short-lived and often only exhibited a single case of bloating and/or diarrhea. Necropsies of affected animals consistently included: abomasitis; hemorrhage and ulcers in the intestine, abomasum, and lungs; fluid and gas throughout the intestines; watery to frothy green fluid in the rumen, and sometimes extending into other stomachs; necrosis; and rod-shaped bacteria in the abomasum and/or small intestine. Of the 19 neonatal cases, 5 occurred in bighorn sheep, and it is noteworthy that these cases were primarily in lambs born late in the spring, after the majority of lambing had already occurred. These cases were similar to those occurring in mule deer neonates. Mortality caused by C. perfringens type A is a growing impact on FWRF operations and ongoing research: it is the leading cause of death in captive bighorns and second only to chronic wasting disease (CWD) as cause of death in mule deer. Moreover, because infections occur primarily in juvenile animals, many long-term studies (e.g., CWD and fertility control) have been hampered by lack of available animals for planed experiments. Here, we proposed to develop and evaluate efficacy of a vaccine to prevent C. perfringens type A morbidity and mortality in captive bighorn sheep and mule deer. 128 �METHODS Initial vaccine development was pursued in Dr. Robert Ellis’ laboratory in the Department of Microbiology, Immunology, and Pathology at Colorado State University. We used captive Rocky Mountain bighorn sheep (O. canadensis canadensis) and mule deer (Odocoileus hemionus) in this experiment. All animals were housed at the CDOW's Foothills Wildlife Research Facility (FWRF) throughout the study and resided in 3-7 ha pastures. In addition to natural forage, grass/alfalfa hay mix and a pelleted high-energy supplement was provided as prescribed under FWRF feeding protocols for bighorn sheep and mule deer in respective age/sex classes throughout the study; fresh water and mineralized salt blocks was provided ad libitum. The general health of all animals was evaluated immediately after vaccination, as well as daily thereafter, and observations recorded throughout our experiment. Injection sites were also examined weekly for 4 weeks after vaccine administration to assess local reactions to vaccine. Bighorn sheep (n = 19) and Mule deer (n = 10) were randomly assigned to vaccinated or unvaccinated groups. The vaccinated group was injected IM with the C. perfringens vaccine in the right hind leg on day 0 and in the left hind leg 4 weeks later (booster). Blood was collected prevaccination, at the booster injection and 4 weeks after the final booster. The control group was weighed and blood was drawn at the time of the vaccine group’s booster and 4 weeks after the final booster. Serum was separated and stored frozen until it was submitted to Colorado State Diagnostic Lab for antibody titer using enzyme-linked immunosorbent assay (ELISA). As necessary deer were sedated with xylazine HCl (5-20 mg IV or 25-100 mg IM) or immobilized with a cocktail of thiafentanil HCl (8-10 mg), or ketamine HCl (100 mg), and xylazine HCl (20 mg), delivered IM by projectile syringe, to facilitate collections; narcotic effects were be reversed with naltrexone HCl (150 mg SC + 50 mg IV). RESULTS There were no vaccine site reactions or adverse effects from vaccination observed. No serum neutralizing antibody titers to C. perfringens were seen in either the BHS or MD. On follow up evaluation of the vaccine by Colorado State Diagnostic Lab, there was no type A antigen in the vaccine. DISCUSSION The vaccine in this study failed due to lack of quality control by the manufacturer, however, we anticipate that a safe and effective vaccine can be readily developed, and that its incorporation into FWRF’s preventive animal health program will reduce morbidity and mortality associated with C. perfringens type A infection. Managing clostridial enteritis is essential to improving success of preventative health programs at the FWRF and minimizing impacts on planned and ongoing research. 129 �LITERATURE CITED Bueschel, D., R. Walker, L. Woods, J. Kokai-Kun, B. McClane, J. G. Songer. 1998. Enterotoxigenic Clostridium perfringens type A necrotic enteritis in a foal. J. Am. Vet. Med. Assoc. 213(9) 1305—1307. Dennison, A. C., D. C. VanMetre, R. J. Callan, P. Dinsmore, G. L. Mason, R. P. Ellis. 2002. Hemorrhagic bowel syndrome in dairy cattle: 22 cases (1997—2000). J. Am. Vet. Med. Assoc. 221 (5) 686—689. L. Petit, M. Gibert and M. R. Popoff. 1999. Clostridium perfinrgens: toxinotype and genotype. Trends in Microbiology. 7 (3) 104—110. Roeder, M. M. Chengappa, T. G. Nagaraha, T. B. Avery, G. A. Kennedy. 1988. Experimental induction of adominal typany, abomasitis, and abomasal ulceration of intraruminal inoculation of Clostridium perfringens type A in neonatal calves. Am. J. Vet. Res. 49 (2) 201—207. Songer, J. Glenn. 1996. Clostridial enteric diseases of domestic animals. Clinical Microbiology Reviews. 9 (2) 216—234. Tillotson, K., J. Traub-Dargatz, C. E. Dickinson, R. P. Ellis, P. S. Morley, D. R. Hyatt, R. J. Magnuson, W. T. Riddle, M. D. Salman. 2002. Population-based study of fecal shedding of Clostridium perfringens in broodmares and foals. J. Am. Vet. Med. Assoc. 220 (3) 342—348. 130 �APPENDIX B SAFETY AND EFFICACY OF RECOMBINANT F1-V FUSION PROTEIN VACCINE TO PROTECT LYNX FROM PLAGUE L. L. Wolfe1, T. E. Rocke2, S. M. Dieterich3, T. M. Shenk1, A. M. Friedlander4, and M. W. Miller1 1 Colorado Division of Wildlife, Wildlife Research Center, 317 West Prospect Road, Fort Collins, Colorado 80526-2097, USA; 2U.S. Geological Survey, Biological Resources Division, National Wildlife Health Laboratory, 6006 Schroeder Road, Madison, Wisconsin 53711, USA; 3Frisco Creek Wildlife Rehabilitation Center, POB 488, Del Norte, Colorado 81132-0002, USA; 4U.S. Army Medical Research Institute of Infectious Diseases, Bacteriology Division, Fort Detrick, Frederick, Maryland 21702, USA. INTRODUCTION Plague, caused by Yersinia pestis, was introduced into the North American continent in the early 1900s, and its impacts on some native wildlife species since that time have been substantial (Cully 1993, Wuerthner 1997, Gasper and Watson 2001). Epidemics in prairie ecosystems have been well documented, and probably contributed to the marked declines observed in both prairie dogs (Cynomys spp.) and black-footed ferrets (Mustela nigripes) over the last century (Cully 1993). Although less extensively studied, it seems likely that sylvatic plague has impacted other wildlife species as well (Gasper and Watson 2001). Canada lynx (Lynx lynx) resided in Colorado historically (Fitzgerald et al. 1994), but apparently were extirpated by the late 1970s. Whether plague played any role in the disappearance of lynx from Colorado is not known. Regardless of plague’s role in the historical decline, this disease now appears to be an obstacle to ongoing efforts to reestablish lynx in southwestern Colorado. To date, Y. pestis infections have been confirmed in 6 Colorado lynx. Plague was the primary cause of death in 27% (4/15) of the known natural deaths and possibly contributed to 1 of the 6 known hit-by-vehicle deaths in adult lynx released in Colorado since 1999 (Wild 2000, Shenk 2003; T. M. Shenk, Colorado Division of Wildlife, unpublished data). Plague also killed at least 1 kitten born in the wild during the first year of documented natural reproduction in Colorado’s reintroduced lynx population (T. M. Shenk, Colorado Division of Wildlife, unpublished data). Practical tools for preventing plague in reintroduced lynx could benefit species recovery efforts in Colorado and perhaps elsewhere. Effective vaccines for preventing plague in mammalian species, including felids, have been developed only recently (Heath et al. 1998, Gasper and Watson 2001, Creekmore et al. 2002). Of these, a recombinant capsular F1-V fusion protein vaccine (Heath et al. 1998) has shown a promising combination of safety and efficacy in black-footed ferrets (Rocke et al. in press), and could be useful in lynx restoration as well. Here, we propose to (1) evaluate F1-V vaccine in captive lynx being held in southwestern Colorado prior to release as part of an ongoing restoration program and (2) compare number of lynx mortalities caused or complicated by plague in vaccinated and unvaccinated lynx released in Colorado. METHODS Our study was conducted in conjunction with the 2004 release program. All lynx were captured, transported, held, cared for, and handled as described in established protocols for Colorado’s restoration program (Wild 2000). Lynx were held at the Frisco Creek Wildlife Rehabilitation Center (FCWRC) prior to and throughout the study until release. Whenever possible, vaccination and sampling was done in 131 �conjunction with other handling activities to minimize stress that could arise from repeated handling of captive lynx. We initially evaluated safety and efficacy of F1-V vaccine (U.S. Army Medical Research Institute of Infectious Diseases, Fort Detrick, Frederick, MD) in 10 adult lynx; 10 age- and originmatched lynx will remain unvaccinated as controls. Blocks will consist of age and origin: age will be either ≤ 5 years old or ≥ 6 years old; origin will be either British Columbia (where prior exposure to plague is possible) or Manitoba/Quebec (where prior exposure is unlikely). We estimated ages based on tooth wear; animals ≤1 year old were excluded. Within each block (age and origin) of lynx, half were selected at random to receive the vaccine while the remainder will serve as controls. Vaccine was be diluted and combined with Alhydrogel adjuvant (United Vaccines, Madison, WI) as described by Rocke et al. (in press). We administered vaccine via subcutaneous (SQ) injection in the hindquarter on day 0 and a second dose was given 21 days later. Initial vaccine doses were delivered by hand-held syringe when lynx are examined upon entry into FCWRC; booster doses were delivered via hand-held syringe while the lynx was restrained in a squeeze cage. Vaccinated lynx were observed immediately after vaccination, immediately upon recovery from anesthesia (when applicable), and daily thereafter for evidence of adverse effects. To evaluate serological responses of vaccinated lynx as an index of efficacy, we will collected blood (~6 ml) from all captive lynx at each handling during the 2004 season regardless of vaccination status. For the 10 principal vaccinates and controls, at minimum blood will be collected on day 0 and again 42 days later (21 days after the booster vaccination). Serum was harvested and stored frozen until assayed. We will measure antibody titers against F1 and V antigens with phytohemagglutinaiton assay (PHA) at the Center for Disease Control and an enzyme-linked immunosorbent assay (ELISA) using methods of Rocke et al. (in press). For the 10 principal vaccinates and controls, we compared changes in log10 anti-F1 and anti-V antibody titers-1 .. Mortality of vaccinated lynx due to or complicated by plague will be compared to mortality due to or complicated by plague of unvaccinated lynx from previous releases. A suite of models developed a priori will be evaluated through AICc model selection (Burnham and Anderson 2002) to investigate the possible effects of vaccination status, age, location of birth, and time to death on mortality of lynx due to or complicated by plague. RESULTS All PHA prevaccination titers were negative. All vaccinated lynx showed seroconversion on the PHA assay at the after the first and second booster (figure 1.). ELISA results are pending. There were no vaccine site reactions and no adverse side effects were seen. 132 �Vaccine Response 10000 PHA antibody titer □ pre 1000 100 ~ ~ - post 1st dose □ post 2nd dose - - - • - ~ ... 10 1 ~ ~ - - ~ ~ ,- - - ,- ~ ... - QF2 QF3 QF4 QF6 QF7 BF1 animal id BF2 BF3 BF4 BF6 Figure 1. PHA antibody titer for individual lynx. All pre titers were negative. All lynx showed seroconversion following vaccination. DISCUSSION Lynx were examined on entry and 5 females from Quebec and 5 females from British Columbia were randomly chosen for vaccination with F1-V fusion protein plague vaccine. All pre vaccine PHA titers were negative. All vaccinates showed seroconversion but the quantitative titer assays are still pending. The vaccine appears to be safe in lynx; there were no vaccine site reactions or adverse systemic reactions. LITERATURE CITED Burnham, K. P. and D. R. Anderson. 2002. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. Second edition. Springer-Verlag. New York. Cully, J. F. 1993. Plague, prairie dogs, and black-footed ferrets. In Management of prairie dog complexes for the reintroduction of the black-footed ferret, J. L. Oldemeyer, D. E. Biggins, B. J. Miller, and R. Crete (eds.). U.S. Fish and Wildlife Service, Biological Report 13, Washington, D.C., pp. 38– 49. Creekmore, T. E., T. E. Rocke, and J. Hurley. 2002. A baiting system for delivery of an oral plague vaccine to black-tailed prairie dogs. Journal of Wildlife Diseases 38: 32–39. Fitzgerald, J. P., C. A. Meaney, and D. M. Armstrong. 1994. Mammals of Colorado. Denver Museum of Natural History and University Press of Colorado, Denver, Colorado, pp. 368–371. Gasper, P. W., and R. P. Watson. 2001. Plague and yersiniosis. In Infectious diseases of wild mammals, 3rd edition, E. S. Williams and I. K. Barker (eds.). Iowa State University Press, Ames, Iowa, pp. 313–329. Heath, D. G., G. W. Anderson, Jr., J. M. Mauro, S. L. Welkos, G. P. Andrews, J. Adamovicz, and A. M. Friedlander. 1998. Protection against experimental bubonic and pneumonic plague by a recombinant capsular F1-V antigen fusion protein vaccine. Vaccine 16: 1131–1137. 133 �Rocke, T. E., J. Mencher, S. R. Smith, A. M. Friedlander, G. P. Andrews, and L. A. Baeten. Recombinant F1-V fusion protein vaccine protects black-footed ferrets (Mustela nigripes) avainst virulent Yersinia pestis infection. Journal of Zoo and Wildlife Medicine, in press.Shenk, T. 2003. Species conservation: Colorado’s lynx. Lynx Home Page, Colorado Division of Wildlife, Denver, Colorado. Accessed 23 October 2003 at http://wildlife.state.co.us/species_cons/lynx.asp. Wild, M. A. 2000. Lynx veterinary services and diagnostics. Federal Aid: Wildlife Research Report for the Colorado Division of Wildlife, pp 47-62. Wuerthner, G. 1997. Viewpoint: The black-tailed prairie dog ⎯ headed for extinction? Journal of Range Management 50: 459–466 APPENDIX C EFFICACY OF KETAMINE MEDETOMIDINE COMBINATION IN MOUTAIN LIONS (Puma concolor) , MULE DEER (Odocoileus hemionus) AND WHITE-TAILED DEER (Odocoileus verginianus) FOR CHEMICAL IMMOBILIZATION IN THE FIELD L. L. Wolfe, W. R. Lance and M. W. Miller In cooperation with Wildlife Pharmaceuticals, Inc. (Fort Collins, CO) we are using ketamine (200mg/ml) and medetomidine (20 mg/ml) compounded at a higher concentration than commercially available. By concentrating the drugs we are able to use an effective dose in a 1 cc dart for mountain lions and a 2 cc dart for deer. To date over 200 deer have been captured and 5 mountain lions using this combination. No adverse side effects have been seen. Evaluation of this drug combination for field capture is ongoing. 134 �APPENDIX D EVALUATION OF COLLARED AND UNCOLLARED DANINJECT AND PNEUDARTS L. L. Wolfe, D. M. Okeson, W. R. Lance, J. Rhyan, and M. W. Miller INTRODUCTION Remote delivery systems, powered by compressed CO2 or blank charge, are an important tool for wildlife immobilization. Darts used with these remote delivery systems are barbed, collared or have uncollared needles. The purpose of the barbs and collars are to hold darts in place long enough to ensure complete drug delivery. However, barbed darts can only be used to deliver anesthetics, thus allowing for dart retrieval. Collared darts and uncollared darts will fall out on their own, but drug delivery may be incomplete. Drugs are delivered from the darts with a powder charge (Pneu-dartTM, Williamsport, PA) or pressurized air (DaninjectTM, Wildlife Pharmaceuticals, Fort Collins, CO). Some level of trauma is inherent, and varies greatly with the type of dart used (Valenburg et al. 1999, Kreeger 2002). Powder charged darts deliver drug rapidly, but are potentially more traumatic than the air pressurized darts; consequently, induction times vary when these darts are used to deliver anesthetic drugs. In this study we compared drug delivery between collared and uncollared darts. We compared dart trauma between collared and uncollared darts and we comopared Daninject darts with Pneu-darts. METHODS This study was conducted in conjunction with a previously-approved terminal study testing fallow deer susceptibility to chronic wasting disease (CWD) (CDOW ACUC 12-2000). Because these animals were already slated for euthanasia, we opportunistically evaluate and compare trauma and drug delivery associated with the respective dart types immediately prior to euthanasia. Deeply anesthetized fallow deer were placed on a stand to facilitate darting the hindquarters. Each animal was darted in the hindquarter with a collared Daninject dart and Pneu-dart dart on one side and uncollared Dainject dart and Pneu-dart dart on the opposite side. Darts were loaded with 2 cc India ink. The animals were then be euthanized by intravenous injection of Euthansol (8.8 mg/kg). Each dart site was evaluated for amount of India ink leakage at dart site, degree of trauma (recorded on a scale of 03. (0= none, 3 = extensive hemorrhage and tissue disruption) and ink injection pattern. All fallow deer were be captured with thiafentanil oxalate (0.1 mg/kg) delivered intramuscularly (IM) via projectile syringe using an adjustable air-powered rifle and xylazine hydrochloride. Anesthetic drugs were delivered to the shoulder and neck to avoid confounding subsequent assessment of dart effects. RESULTS On necropsy the ink pattern at the injection site was evaluated. A common ink pattern noted at necropsy was a “T”shaped pattern. The superficial area of ink (top of T) averaged 20-25 mm in diameter (recorded on line #4- spread of ink in muscle). Note that this refers only to the most superficial spread (horizontally) of the ink in the muscle; ink forming the top part of the T was typically only 2 mm thick. Ink then typically followed the needle track 20-30 mm into muscle. The pattern of “focal” was often recorded, but would probably be more correctly stated as “along needle track” as these 20-30 mm deep tracks were typically only 2 mm wide (roughly the diameter of a needle). 135 �Most injection sites had minimal hemorrhage or trauma this was recorded on a scale of 0-3. (0= none, 3 = extensive hemorrhage and tissue disruption) Overall, most dart sites were rated a “0” (12 of 32 darts) or a “1” (7 of 32 darts). Uncollared PneuDarts Eight of 8 uncollared PneuDarts bounced off the animal immediately after impact. However, ink was delivered into the animals, in all but one case. In 5 cases darts bounced, but no ink was noted on the hair of the animal or was observed spraying from the bounced dart. At necropsy, ink was noted in muscle in all of these cases, indicating that the bounced darts did deliver ink prior to ejecting from animal. One dart that bounced was noted to have sprayed a large amount of ink. However on necropsy, ink was found in muscle, indicating that dart did deliver at least some of the ink into the muscle. One dart bounced was noted as “did inject”, but some ink was noted on hair around injection site. At necropsy, ink was found in muscle, indicating that dart did deliver at least some of the ink into the muscle. One dart bounced was noted as “did inject”, but there was ink running down from the injection site. At necropsy, it was hard to distinguish any ink from bruising, so this dart may have not injected any ink into the muscle. There were 2 of 8 animals scored as “3” (extensive hemorrhage and tissue disruption). However there were also 3 of 8 animals scored as a “0” (no hemorrhage and tissue disruption.). In addition 2 of 8 were NR (not recorded). It is difficult to draw a conclusion as to whether or not there is a tendency for this type of dart to cause more tissue damage. Fallow deer dart study 4/16/03 - Results for smooth Pneudart (symbol +) Deer # Dart stayed Ink on in muscle? hair 1102 N 1 202 N NR 302 N 3 2002 N 1 1002 N 1 Depth of ink in muscle 5 mm Pattern of ink in muscle Superficial spread of ink on muscle Hemorrhage/ trauma 25 mm NR ? 2 unknown 0 ? NR Unable to distinguish ink from hemorrhage/bruising. Did dart inject? 20 mm located 20 mm deep 5 mm deep, focal 20 mm 15 mm focal spot located 20 mm deep dissecting on tendon sheaths 1802 N 0 2502 N 0 < 2 mm only superficially delivered 1202 N 1 20 mm NR spot located 20 mm deep focal, but no needle track 5 mm 60 mm superficial spot plus spot located 20 mm deep into muscle with a 20 mm diameter 20 mm Fits typical T pattern of ink spread? 3 N 0 N 3 Y 0 N NR ? Collared PneuDarts All 8 collared PneuDarts stayed in the animal after impact. In 3 of 8 cases the darts delivered ink only very superficially (not deep in muscle). Overall the collared darts caused very little trauma. There were 2 of 8 animals with a rating of “0”, and 2 of 8 with a rating of “1”. Only 1 of 8 animals had a rating of either “2” or “3”. Not recorded = 2 of 8. 136 �Fallow deer dart study 4/16/03 - Results for collared Pneudarts (symbol #) Deer # Dart stayed Ink on in muscle? hair Depth of ink in muscle Pattern of ink in muscle Superficial spread of ink on muscle Hemorrhage/ trauma only superficially delivered 20 mm 2 5 mm 3 Not sure of pattern 0 N; Ink only superficially delivered with spread along superficial fascial planes 1102 Y 0 "very superficial" (?<2mm) 202 Y 0 20 mm deep, focal 70 mm Fits typical T pattern of ink spread? N; Ink only superficially delivered, minimal muscle penetration of ink. 302 Y 0 < 2mm 70 mm superficial "splotch", dissects along fascial planes 2002 Y 0 20 mm focal 20 mm 1 Y 1002 Y 0 30 mm focal 5 mm NR Y 1802 Y 0 35 mm focal 5 mm 1 Y 2502 Y 0 < 2mm only superficially delivered 60 mm 0 N; Ink only superficially delivered, minimal muscle penetration of ink. 1202 Y 0 15 mm NR 20 mm NR ? Uncollared Dan-Inject Darts Seven of 8 uncollared Dan-Inject darts stayed in the animal’s muscle after impact. The result of one uncollared Dan-Inject dart was not recorded. Overall, the uncollared Dan-inject darts caused very little hemorrhage or trauma. Four of 8 animals had a rating of “0” hemorrage/trauma rating and 2 scored 1 and 1 scored 2. Fallow deer dart study 4/16/03 - Results for smooth DanInject darts (symbol *) Deer # Dart stayed Ink on Depth of ink Pattern of ink in muscle? hair in muscle in muscle 1102 Y 0 < 2mm 202 Y 0 0 Superficial spread of ink on muscle very superficial no ink injected into muscle Hemorrhage/ Fits typical T pattern trauma of ink spread? 15 mm 1 2 smooth Daninjects hit animal. 1st "penetrated leg, injection out back side of leg". 2nd dart also recorded as Yes stayed in muscle & 0 ink on hair; but not sure if depth, pattern, and spread info is for 1st or 2nd dart. ??? 0 0 N; dart went deep into limb but injected ink out medial aspect 302 Y 0 30 mm deep, focal 20 mm 1 Y 2002 Y 0 50 mm deep, focal 5 mm 0 Y 1002 NR 0 15 mm focal 5 mm 0 Y 1802 Y 0 50 mm NR (not recorded) 20 mm 0 Y? 2502 Y 3 75 mm 1202 Y 0 0 NR 10 mm 2 1st smooth Daninject dart went through leg, injected some out caudal aspect (3 for ink on hair refers to ink on back of limb from 1st dart). 2nd dart recorded as - stayed in muscle, no ink on hair; but not sure if 75 mm & 10 mm is for 2nd dart or 1st dart??? only superficially delivered 20 mm NR N; Ink only superficially delivered, no muscle penetration of ink. Collared Dan-Inject Darts Seven of 8 collared Dan-Inject darts stayed in the animal’s muscle after impact. The result of one collared Dan-Inject dart was not recorded. These darts shows a tendency to cause little to no tissue damage. There were 6 of 8 animals with a rating of either “0” or “1”. These darts show a tendency to cause little to no tissue damage. There were 3 of 8 animals with a rating of “0”, and 3 of 8 with a rating of “1”. Only 1 of 8 animals had a rating of “3”. 137 �Fallow deer dart study 4/16/03 - Results for DanInject collared darts (symbol @) Deer # Dart stayed Ink on in muscle? hair Depth of ink Pattern of ink in muscle in muscle Superficial spread of ink on muscle Hemorrhage/ Fits typical T pattern trauma of ink spread? 1102 Y 0 5 mm focal 15 mm 1 Y 202 Y 0 2 mm deep, focal 10 mm 0 N, basically round superficial, spot 302 Y 0 20 mm diffuse ~15 mm 10 mm "deep" (?) 1 ? Not sure of pattern 2002 Y 0 5 mm 3 ? Not sure of pattern 1002 NR 1 25 mm deep, focal 25 mm 1 ? Not sure of pattern 1802 Y 0 40 mm deep, focal 5 mm 0 Y? may be 5 mm wide all the way down 2502 Y 0 10 mm deep, focal 25 mm 0 Y 1202 Y 0 35 mm NR (= not recorded) 35 mm NR ?; Note 0.5 mL of ink left in dart deep 15 mm dissects btw muscle masses DISCUSSION All of the collared darts stayed in the muscle (did not bounce). All of the Daninject uncollared darts also stayed in the muscle. Only the uncollared pneudarts bounced out of the muscle. There was no inject sprayed on the hair of animals darted with collared pneudarts. Only one animal in each group of animals darted with daninjects had ink sprayed on the hair. Five of the animals darted with uncollared pneudarts had ink sprayed on the hair. LITERATURE CITED Kreeger, T. J. 2002. Analyses of immobilizing dart characteristics. Wildlife Society Bulletin. 30(3) 968−970. Valkenburg, P., R. W. Tobey, and D. Kirk. 1999. Velocity of tranquilizer darts and capture mortality of caribou calves. Wildlife Society Bulletin. 27(4) 894−896. Prepared by _______________________________ Lisa L. Wolfe, Veterinarian 138 � https://cpw.cvlcollections.org/files/original/67b18b9a50ae62c96bbae585d027c7e0.pdf 5f4009aa19fc000085191acb4d186423 PDF Text Text 133 JOB PROGRESS REPORT State of -------=C'--"o""'lo=r=a=do=------- : Division of Wildlife - Mammals Research Work Package No. ----=3-'-7--'-4-=-0_ _ _ _ _ : Chronic Wasting Disease and Other Wildlife Disease Management Task _ _ _ _ _ _ _ _ _1_ _ _ _ _ _ _ : Chronic Wasting Disease in Mule Deer Monitoring & Management Period Covered: July 1 2002 through June 30, 2003 Author: Michael W. Miller and L. L. Wolfe Personnel: L.A. Baeten, T. H. Baker, M. M. Conner, K. Cramer, T. R Davis, V. Dreitz, C. P. Hibler, N. T. Hobbs, E. Hoover, D. 0. Hunter, E. Knox, C. E. Krumm, C. T. Larsen, N. Mier, B. E. Powers, J. Rhyan, C. J. Sigurdson, T. R Spraker, K. Taurman, E. S. Williams, D. Wroe Interim Report- Preliminary Results This work continues, and precise analysis ofdata has yet to be accomplished. Manipulation or interpretation of these data beyond that contained in this report should be labeled as such and is discouraged. ABSTRACT We continued conducting research on various aspects of chronic wasting disease (CWD) epidemiology and management. Here, we report progress in ongoing and recently-completed work. Studies focused on improving and expanding surveillance in free-ranging populations, understanding and modeling transmission mechanisms, identifying ecological and anthropogenic factors that may influence epidemic dynamics, and evaluating and applying alternative diagnostic and control strategies. In addition to preliminary findings reported here, eight original studies, as well as one review article, were published publication during this segment; citations are appended to the report. INTRODUCTION We continued conducting research on various aspects of chronic wasting disease (CWD) epidemiology and management. Some parts of this work were conducted in collaboration with investigators at Colorado State University, the University of Wyoming, and elsewhere. Specific projects were supported with various combinations of funds from the Colorado Division of Wildlife (CDOW), Federal Aid in Wildlife Restoration Project W-153-R, the U.S. Department of Agriculture, and National Science Foundation/National Institutes of Health Grant DEB-0091961. �134 METHODS Our work on CWD is both multidisciplinary and multifaceted, but broadly falls under the topics of "epidemiology and management" or "pathogenesis and diagnosis". For simplicity, we describe progress under those headings below. STUDIES OF CWD EPIDEMIOLOGY & MANAGEMENT We continued or initiated studies related to surveillance, transmission mechanisms, epidemic trend forecasting, potential host range and strain variation, risk factors, and management tools and feasibility as aids to understanding and controlling CWD in free-ranging deer and elk in Colorado. Statewide surveillance: The discovery of CWD in northwestern Colorado in January 2002 created a sudden demand for both more widespread surveillance and more rapid turnaround on laboratory results. Consequently, the CDOW's CWD surveillance program was overhauled and its capacity greatly expanded over the summer of 2002 in order to meet anticipated demands for surveillance data, as well as to meet policy-based decisions to provide carcass quality assurance information for individual hunters. The most notable changes were the addition of three regional submission laboratories, streamlining of tissue sampling methods, and incorporation of a rapid screening test for CWD diagnosis. Details of overall programmatic features and changes were described on a new CWD-oriented CDOW web page (http://wildlife.state.co.us/CWD/index.asp): details of the evaluation of modified sampling and testing procedures are described below. Transmission mechanisms: We summarized findings on empirical evidence of animal-animal transmission of CWD and the relative importance of this mechanism in epidemic dynamics. We also completed an experiment comparing the relative contributions of live animals, contaminated environments, and infected carcasses to CWD transmission. In this study, 34 free-ranging mule deer from two sources distant to known endemic foci of chronic wasting disease (Rocky Mountain Arsenal National Wildlife Refuge, US Air Force Academy) were captured for use as experimental subjects during MarchMay 2002. We transported these deer to the Colorado Division of Wildlife's Foothills Wildlife Research Facility (FWRF), where they were placed in paddocks (n = 3 replicates/exposure route; n = 3 deer/paddock). Exposure treatments were: confinement in paddocks housing naturally-infected deer ( l infected deer/paddock), confinement in paddocks where infected deer previously resided, and confinement in paddocks where carcasses from CWD-infected deer have decomposed in situ ( l carcass/paddock); unexposed control paddocks are also being maintained. Entire paddock groups will be sacrificed and examined at the first sign of CWD in any subject deer within a paddock. We compared infection rates within and among treatments to examine which of these may contribute to perpetuation of CWD epidemics. Modeling epidemic dynamics in captive mule deer: Developing detailed, temporally dynamic models of CWD in wild populations is a pressing management need, but available field data are presently insufficient to clearly reveal natural trends in ongoing epidemic dynamics. Moreover, there are several plausible ways to model CWD transmission mechanisms, yet field data will likely not provide sufficient resolution for discerning the most appropriate representation. To begin understanding how to best model CWD transmission, we have undertaken a model selection exercise using a time series of data on prevalence on CWD in captive mule deer. We assembled 26 years of data (1974-2000) from CDOW's Foothills Wildlife Research Facility. These data are being used to evaluate strength of evidence for a set of candidate models involving indirect and direct transmission, as well as with and without latency �135 periods. Estimates of transmission rates derived from these models will provide an upper bound on what could be expected in wild populations and will guide construction of candidate sets for modeling those populations. Host range and strain variation: We continued a series of experimental studies in cattle, fallow deer, and mountain lions to explore potential host range of CWD after intense but natural exposure; these experiments compliment ongoing surveillance for evidence of infection in species not known to be natural hosts of CWD, including moose, mountain lions, and cattle. We also continued work looking for evidence of strain variation in CWD agent from various deer sources using domestic ferrets as a laboratory model. Effects of land use on prevalence: Because land-use changes are likely to shape the spatial and temporal dynamics of CWD, as well as options for its management, we have been working to understand the effect ofland use on patterns of CWD prevalence in free-ranging mule deer. We conducted a study to determine whether CWD prevalence in urban areas is higher than prevalence in non-urban areas. We categorized two land use types: urban areas contained==:, 1 housing unit/IO acres and non-urban areas (e.g., ranch, state, and federal lands) contained< 1 housing unit/10 acres. We compared CWD prevalence between land use types in 3 study areas in northern Colorado (Glacier View Meadows [GVM], Horsetooth [HT], Estes Park [EP]) in which urban and non-urban areas were juxtaposed. In each study area, we delineated urban areas surrounded by a 1-2 km buffer and non-urban areas concentric to the buffer. Deer were sampled in approximately equal numbers from the two land use categories. We used a combination of data collected from mule deer sampled via postmortem (Miller et al., 2000, J. Wildl. Dis. 36:676-690; Miller & Williams, 2002, Vet. Rec. 151:610-612) and antemortem (Wolfe et al., 2002, J. Wildl. Manage. 66:564-573) methods described previously; our target was 210 samples for each land-use category, which provided the ability to detect I 0% differences in prevalence between categories with 90% probability at the 0.05 confidence level. We assumed sampling was normally distributed and tried to balance sampling equally among study areas. Selective predation upon infected mule deer: To test for evidence of selective predation, we began a study to compare prevalence of CWD among puma-killed mule deer to prevalence among mule deer harvested or randomly culled by humans within home ranges of collared mountain lions. Sample size calculations were based on the number of deer samples needed to detect two-fold differences in CWD prevalence: we assume that if the mountain lions are showing selectivity for deer with CWD, then the prevalence in the deer killed by mountain lions will be at least twice the prevalence of CWD in the local deer population. Telemetry-marked mountain lions are being monitored and, when available, brainstem (medulla oblongata at the level of the obex), retropharyngeal lymph nodes, and tonsils are collected from pumakilled mule deer carcasses; where none of these tissues are available, we will try to locate and sample other lymphoid tissues (e.g., submandibular or mesenteric lymph nodes, Peyers patches, etc.). Representative subsamples of collected tissues will be fixed in 10% neutral buffered formalin, and the remainder stored frozen. Tissues will be evaluated for presence of PrPewo accumulations using established irnmunohistochemistry (IHC) techniques; IHC-positive cases will be further evaluated with hematoxylin and eosin staining to stage the duration of CWD infection. We will compare CWD prevalence among puma-killed deer to prevalence among deer harvested by hunters in the same area. Using cumulative location data from each collared puma, home range will be estimated. Data from mule deer harvested and sampled within each home range will be extracted from our harvest survey database, preferentially using data collected during the period of predation sampling where sufficient harvest data are available for that time period. To assess differences between predation- and harvest-associated prevalence, we will calculate the CI on the difference as described above; if the CI does not include 0, then we will conclude that these rates differ. �136 Influence of trace minerals on susceptibility: To investigate the potential influence of trace minerals on CWD susceptibility, we began two independent studies. In a retrospective study, we will use archived tissues to compare tissue levels of copper (Cu), molybdenum (Mo), and manganese (Mn) in mule deer infected with CWD to levels in apparently uninfected deer from the same geographic area. We also started an experiment to examine the effect of Cu supplementation on CWD susceptibility in white-tailed deer, wherein we will compare the natural infection rate and course of CWD in captive deer receiving a sustained-release oral Cu supplement to the rate and course in unsupplemented controls residing in the same paddock. Vaccination as a preventive tool: We collaborated with investigators from Colorado State University to conduct a pilot study evaluating safety and serologic responses of mule deer to an anti-PrP vaccine. Four captive deer (2 vaccinates and 2 controls) were monitored and sampled over a 4-month period for evidence of vaccine effects on health and serum antibody levels. Evaluation of an urban CWD management strategy: Recognizing the need for alternatives to traditional strategies for controlling CWD, we initiated a pilot study to evaluate "test and cull" as an approach for managing CWD in urban habitats. Previously, models exploring probable consequences of various management strategies identified selective removal of infected individuals as a potentially effective method for reducing CWD prevalence in mule deer populations, provided that infected deer were detected early and a large (>50%) proportion of the population could be sampled annually (Gross and Miller, 2001, J. Wildl. Manage. 65:205-215). During November-December 2002, 113 free-ranging mule deer were captured, tested, and marked with timed-release radiocollars in urban areas throughout Estes Park to assess the feasibility of such a management approach. This sampling effort represented testing of about 25% of the adult mule deer residing Estes Park. In January 2003, biopsy-positive deer were culled. Dropped radiocollars were recovered in March-April 2003 for reuse in a second round of sampling planned for April-May 2003. In addition to the primary goal of assessing feasibility, data gathered in the course of this study will also be useful in improving our understanding and modeling of the influences of urban landscapes on CWD epidemiology. STUDIES OF CWD PATHOGENESIS & DIAGNOSIS We continued or initiated studies related to rapid screening test evaluation, pathogenesis in natural hosts, and live-animal diagnostic test refinement and evaluation as aids to improving approaches for CWD surveillance and diagnosis in free-ranging deer and elk in Colorado. Evaluation of a rapid screening test: In conjunction with expanded CWD surveillance in Colorado during Sep-Dec 2002, tissue samples (n = 25,050 total) from 23,256 mule deer, white-tailed deer, and Rocky Mountain elk collected statewide were examined using an ELISA developed by Bio-Rad Laboratories, Inc. (brELISA) in a two-phase study. In the validation phase of this study, a total of 4,175 retropharyngeal lymph node (RLN) or obex (OB) tissue samples were examined independently by brELISA and immunohistochemistry (IHC). There were 137 IHC positive samples and 4,038 IHC negative samples. Optical density (OD) values from brELISA were classified as "not detected" or "suspect" based on recommended cut-off values during the validation phase. Based on the validation phase data, only RLN samples were collected for the field application phase of this study and only samples with brELISA OD values> 0.1 were examined by IHC. We estimated assay performance parameters (sensitivity, specificity, agreement) for brELISA to determine the utility of this rapid screening assay in CWD surveillance programs. �137 Pathogenesis in natural host species: We continued our work studying the pathogenesis of CWD in whitetailed deer after oral inoculation with infectious, conspecific brain tissue. This study will complement studies documenting CWD pathogenesis in mule deer and elk that already have been completed. Evaluation of antemortem diagnostic techniques: In order to better study and manage CWD across landscapes where hunting and culling are not feasible sources of diagnostic samples, we continued working to refine and evaluate techniques for sampling live animals. Previously, we conducted a field study to evaluate tonsil biopsy immunohistochemistry (IHC) as a tool for diagnosing CWD in live, freeranging mule deer and estimating prevalence. Based on our initial success, we have applied these techniques to gather data for new studies related to effects of land use patterns on CWD prevalence and its management, as described elsewhere in this report. We also initiated a study to evaluate a prospective rapid blood test for diagnosing CWD in live deer. A total of 37 samples from 21 different captive mule deer, some infected with CWD, were submitted to a private testing laboratory (GeneThera, Denver, CO) for evaluation using collection materials and instructions provided by the laboratory. In order to objectively assess reliability and repeatability of the candidate assay, the testing laboratory was blinded to the infection status and animal identification for individual samples that we submitted. RESULTS AND DISCUSSION STUDIES OF CWD EPIDEMIOLOGY & MANAGEMENT Statewide CWD surveillance: The CDOW sampled over 26,000 deer and elk harvested or culled in northeastern Colorado and other select locations. Survey results were posted on the Division's CWD web page. Prevalence data also will be used to augment an existing database that is the foundation for ongoing analyses and modeling of temporal and spatial aspects of CWD epidemiology, as well as for evaluating responses to management. This year's data will be particularly useful in further exploring local patterns of disease prevalence related to deer movement, density, and land use patterns. Moreover, the surveillance strategy and methods first devised and implemented in Colorado recently served as a model for developing national recommendations on CWD surveillance in free-ranging populations. Transmission mechanisms: A manuscript describing our findings on the relative importance of animal-animal transmission of CWD, as compared to maternal transmission, was accepted for publication and should appear this fall. Our experiment comparing the relative contributions of live animals, contaminated environments, and infected carcasses to CWD transmission revealed that CWD can be transmitted indirectly, from environments contaminated by excreta or decomposed carcasses to susceptible animals. Under experimental conditions, mule deer became infected in 2 of 3 paddocks containing naturally infected deer, in 2 of 3 paddocks where infected deer carcasses had decomposed in situ ~ 1.8 years earlier, and in 1 of 3 paddocks where infected deer had last resided 2.2 years earlier. Our data suggest that indirect transmission and environmental persistence of infectious prions will complicate efforts to control CWD, and perhaps other animal prion diseases. Modeling epidemic dynamics in captive mule deer: Preliminary analyses suggest that indirect transmission models best represent epidemic data; moreover, our model selection results align well with independent empirical findings on CWD transmission mechanisms. We will continue refining candidate �138 models before making final comparisons and parameter estimations. Findings should be of use in refining epidemic models of CWD in free-ranging mule deer populations. Host range and strain variation: Cattle (n = 11) living in paddocks with naturally-infected mule deer remained healthy through 6 years of exposure; in contrast, only I of 12 mule deer introduced into these same paddocks in 1997 is still alive. Our results are consistent with data from cell-free conversion (Raymond et al., 2000, EMBO 19:4425-4430) and intracerebral (IC) challenge (Hamir et al., 2001, J. Vet. Diag. Invest. 13:91-96) studies that suggest the probability of natural susceptibility to CWD in cattle is extremely low. Similarly, neither signs nor postmortem evidence of infection have been observed in fallow deer (n = 24) exposed to infected mule deer for ~2.5 years, and mountain lions (n = 3) consuming carcasses of CWD-infected deer and elk for > I year also have remained healthy. No evidence of infection has been observed in moose, mountain lions, or cattle examined via ongoing surveillance. Clinical signs and postmortem findings consistent with CWD in ferrets were observed in four of five ICinoculated with tissue from infected deer, but have not been observed in the free-ranging white-tailed deer or control groups. Incidence and incubation periods were consistent among affected groups. Preliminary assessment of Western blots (WB) revealed no apparent differences in glycosylation patterns among WBpositive ferrets. In the absence of changes in status in the unaffected groups, we will terminate this study in the next 6 months and summarize our findings. Effects of land use on prevalence: Preliminary analyses revealed that CWD prevalence was higher among deer sampled from urban areas (12.5%, CI=8.4-16.8%, n=243) than among deer from juxtaposed nonurban areas (7 .3%, CI=4.3-l 0.3%, n=288) (Fisher's exact test, P=0.04). The magnitude of difference between CWD prevalence rates associated with urban and non-urban land use (5.3%, CI=2.4-8.2%) further emphasized the apparent effect of urban land use on CWD prevalence. Although CWD prevalence varied somewhat among study sites, it did not differ (Fisher's exact test, P=0.088). Areaspecific differences may reflect greater risk or exposure among subpopulations. However, the trend of higher CWD prevalence in areas of urban land use was consistent across all three sites. Our findings suggest that urbanization is playing an undesirable role in CWD epidemic dynamics in northcentral Colorado's mule deer populations. The underlying cause of this influence on CWD prevalence remains unclear. Urban landscapes may attract or artificially congregate wild cervids. Supplemental feeding, although illegal in Colorado, occurred throughout urban areas in all 3 of our study sites. Urban areas also may provide refuge from predation. Mountain lions are likely the main predator of deer in this area, but they are reclusive and seldom hunt in urban lands. Deer may become more sedentary in urban areas - in extreme cases, urban development may even promote elimination or modification of seasonal migration patterns made by resident deer. Regardless of the reason(s), urban landscapes clearly cannot be ignored in attempts to manage CWD and perhaps other important wildlife disease problems. Selective predation upon infected mule deer: Our work continues from a pilot study conducted to evaluate available global positioning system (GPS)-based telemetry collars for use in this sampling application. We are now sampling mule deer carcasses to test for evidence of CWD infection by monitoring collared mountain lions 1-3 times/week and locating prospective kill sites using a remotely downloadable GPS telemetry system (Lotek, Inc.; model GPS4000). We will continue refining our monitoring approach to ensure that we find kill sites quickly enough to retrieve a suitable tissue sample to test for CWD. Whether target sample sizes can be attained in the time planned for this work remains to be determined. Influence of trace minerals on susceptibility: Both studies are underway. �139 Vaccination as a preventive tool: We observed no adverse effects of vaccination on captive mule deer; serology results are pending. Evaluation of an urban CWD management strategy: Data from our December pilot trial indicate that testing and culling mule deer appears to be a viable approach for managing CWD in Estes Park. Based on the success of the first round of pilot testing, the CDOW has committed to a 5-year management experiment to evaluate the efficacy of test and cull in lowering CWD prevalence in an urban mule deer population. A manuscript describing the results of our feasibility study is in preparation. STUDIES OF CWD PATHOGENESIS & DIAGNOSIS Evaluation of a rapid screening test: In the validation phase, using II-IC-positive cases as known CWDinfected individuals and assuming II-IC-negative cases were uninfected, the relative sensitivity of brELISA depending on species ranged from 98.3-100% for RLN samples and 92.1-93.3% for OB samples; the relative specificity of brELISA depending on species ranged from 99. 9- l 00% for RLN samples and was 100% for OB samples. Overall agreement between brELISA and IHC was ~97.6% in RLN samples and ~95.7% in OB samples of all species where values could be calculated; moreover, mean brELISA OD values were ~46x higher in II-IC-positive samples than in II-IC-negative samples. Discrepancies were observed only in early-stage cases of CWD. Among 20,875 RLN samples screened with brELISA during the field application phase, 155 of 8,877 mule deer, 33 of l l,73 l elk, and 9 of 267 white-tailed deer samples ( 197 total) had OD values > 0 .1 and were further evaluated by IHC to confirm evidence of CWD infection. Of cases flagged for IHC follow-up, 143 of 155 mule deer, 29 of 33 elk, and all 9 white-tailed deer were confirmed positive. Mean (± SE) OD values for II-IC-positive cases detected during the field application phase were comparable to those measured in RLN tissues during the validation phase. Based on these data, brELISA was determined to be an excellent rapid test for screening large numbers of samples in surveys designed to detect CWD infections in deer and elk populations. Pathogenesis in natural host species: Although our study of CWD pathogenesis in white-tailed deer is ongoing, some white-tailed deer inoculated orally with about 2.5 g of brain tissue homogenate (containing about 15 µg PrPcwo) already developed clinical CWD and were euthanized in end-stage disease 16-30 mo postinoculation (Pl). The clinical course in inoculated white-tailed deer was similar to that previously observed in mule deer inoculated with about 15 µg PrPcwo from infected mule deer. Laboratory evaluations of tissues from both our white-tailed deer and mule deer pathogenesis studies are pending. Evaluation of antemortem diagnostic techniques: Tonsil biopsy is a useful tool for estimating CWD prevalence in nonhunted mule deer populations. In addition to applications in the two field studies described here, the techniqu_es we developed are being used in at least four other field studies of CWD ·epidemiology (WY, NM, WI, SD). Thus far, we have been unable to assess the reliability or repeatability of the "GeneThera test". Over 6 mo have passed since blind samples were submitted, but we have been unable to obtain any test results despite repeated attempts to contact the laboratory. Until such evaluations can be completed, we cannot recommend incorporation of this candidate test into any of our ongoing CWD research or management programs. �140 APPENDIX Publications arising from ongoing CWD work: Gould, D. H., J. L. Voss, M. W. Miller, A. M. Bachand, B. A. Cummings, and A. A. Frank. 2003. Survey of cattle in northeast Colorado for evidence of chronic wasting disease: Geographical and high risk targeted sample. Journal of Veterinary Diagnostic Investigation 15: 274-277. Hibler, C. P., K. L. Wilson, T. R Spraker, M. W. Miller, RR. Zink, L. L. DeBuse, E. Andersen, D. Schweitzer, J. A. Kennedy, L.A. Baeten, J. F. Smeltzer, M. D. Salman, and B. E. Powers. 2003. Field validation and assessment of an enzyme-linked immunosorbent assay for detecting chronic wasting disease in mule deer (Odocoileus hemionus), white-tailed deer (Odocoileus virginianus), and Rocky Mountain elk (Cervus elaphus nelsoni). Journal of Veterinary Diagnostic Investigation 15: 311-319. Race, R. E., A. Raines, T. G. M. Baron, M. W. Miller, A. Jenny, and E. S. Williams. 2002. Comparison of abnormal prion protein glycoform patterns from transmissible spongiform encephalopathy agentinfected deer, elk, sheep, and cattle. Journal of Virology 76(23): 12365-12368. Samuel, M. D., D. 0. Joly, M.A. Wild, S. D. Wright, D. L. Otis, R. W. Werge, and M. W. Miller. 2003. Surveillance strategies for detecting chronic wasting disease in free-ranging deer and elk. Results of a CWD surveillance workshop. USGS, BRD, National Wildlife Health Center, Madison, Wisconsin. Sigurdson, C. J., C. Barillas-Mury, M. W. Miller, B. Oesch, L. J. van Keulen, J.P. Langeveld, and E. A. Hoover. 2002. PrP(CWD) lymphoid cell targets in early and advanced chronic wasting disease of mule deer. Journal of General Virology 83: 2617-2628. Spraker, T. R., K. I. O'Rourke, A. Balachandran, R. R. Zink, B. A. Cummings, M. W. Miller, and B. E. Powers. 2002a. Validation of monoclonal antibody F99/97.6. l for immunohistochemical staining of brain and tonsil in mule deer (Odocoileus hemionus) with chronic wasting disease. Journal of Veterinary Diagnostic Investigation 14:3-7. Spraker, T. R., R.R. Zink, B. A. Cummings, M.A. Wild, M. W. Miller, and K. I. O'Rourke. 2002b. Comparison of histological lesions and immunohistochemical staining of proteinase resistant prion protein in a naturally-occurring spongiform encephalopathy of free-ranging mule deer (Odocoileus hemionus) with those of chronic wasting disease of captive mule deer. Veterinary Pathology 39: 110119. Wild, M. A., T. R. Spraker, C. J. Sigurdson, K. I. O'Rourke, and M. W. Miller. 2002. Preclinical diagnosis of chronic wasting disease in captive mule deer (Odocoileus hemionus) and white-tailed deer (Odocoileus virginianus) using tonsillar biopsy. Journal of General Virology 83: 2629-2634. Williams, E. S., and M. W. Miller. 2003. Transmissible spongiform encephalopathies in non-domestic animals: origin, transmission, and risk factors. In Risk analysis of prion diseases in animals. C. I. Lasmezas and D. B. Adams, (eds.). Revue scientifique et technique Office international des Epizooties 22: 145-156. �Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Project No. Work Package No. Task 1 Colorado Federal Aid Project: N/A : : : : 3740 Cost Center 3430 Mammals Research Wildlife Diseases Chronic Wasting Disease in Mule Deer Research and Development : Period Covered: July 1 2003 through June 30, 2004 Author: Michael W. Miller and L. L. Wolfe Personnel: L. A. Baeten, S. Bender, M. M. Conner, K. Cramer, T. R. Davis, M. Farnsworth, K. Griffin, C. P. Hibler, N. T. Hobbs, D. O. Hunter, J. E. Jewell, E. Knox, C. E. Krumm, C. T. Larsen, B. E. Powers, J. Rhyan, M. Sirochman, T. Sirochman, T. R. Spraker, M. K. Watry, E. S. Williams, D. Wroe All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT We continued conducting research on various aspects of chronic wasting disease (CWD) epidemiology and management. Here, we report progress in ongoing and recently-completed work. Studies focused on improving and expanding surveillance in free-ranging populations, understanding and modeling transmission mechanisms, identifying ecological and anthropogenic factors that may influence epidemic dynamics, and evaluating and applying alternative diagnostic and control strategies. In addition to preliminary findings reported here, 12 original studies and review articles were published during this segment; citations are appended to the report. 103 �JOB PROGRESS REPORT CHRONIC WASTING DISEASE IN MULE DEER RESEARCH AND DEVELOPMENT Michael W. Miller and L. L. Wolfe INTRODUCTION We continued conducting research on various aspects of chronic wasting disease (CWD) epidemiology and management. Some parts of this work were conducted in collaboration with investigators at Colorado State University, the University of Wyoming, and elsewhere. Specific projects were supported with various combinations of funds from the Colorado Division of Wildlife (CDOW), Federal Aid in Wildlife Restoration Project W-153-R, the U.S. Department of Agriculture (APHIS/VS, the U.S. Department of Interior (USGS/BRD), and National Science Foundation/National Institutes of Health (NIH) Grant DEB-0091961. METHODS Our work on CWD is both multidisciplinary and multifaceted, but broadly falls under the topics of “epidemiology and management” or “pathogenesis and diagnosis”. For simplicity, we describe progress under those headings below. STUDIES OF CWD EPIDEMIOLOGY & MANAGEMENT We continued or initiated studies related to surveillance, transmission mechanisms, epidemic trend forecasting, potential host range and strain variation, risk factors, and management tools and feasibility as aids to understanding and controlling CWD in free-ranging deer and elk in Colorado. Statewide surveillance: Surveillance for CWD continued throughout Colorado to determine the extent of distribution, to estimate prevalence in affected areas, and to monitor prevalence trends. Surveillance methods were as described elsewhere (Miller et al., 2000, J. Wildl. Dis. 36:676–690; Miller & Williams, 2002, Vet. Rec. 151:610–612; Hibler et al., 2003, J. Vet. Diag. Invest. 15:311−319). In addition to reporting of annual survey findings, we also analyzed cumulative surveillance data to examine the potential influences of demographic, spatial, and temporal factors on observed prevalence patterns. We also began exploring ways of improving the efficiency of our CWD surveillance program. Since 1996, tissue samples have been collected from deer killed in vehicle collisions throughout Colorado as part of our monitoring program for detecting CWD in free-ranging populations. We estimated CWD prevalence among vehicle-killed mule deer statewide and compared this to estimated CWD prevalence among mule deer sampled in the vicinity of these collision sites to determine if CWD-infected individuals were more vulnerable to vehicle collisions than otherwise healthy deer. Transmission mechanisms: We summarized findings on empirical evidence of direct and indirect CWD transmission and the relative importance of these mechanisms in epidemic dynamics. Modeling epidemic dynamics in captive mule deer: We continued analyses of 26 years of data (1974−2000) from CWD epidemics at CDOW’s Foothills Wildlife Research Facility to evaluate strength of evidence for a set of candidate models involving indirect and/or direct transmission, with and without latency periods. Estimates of transmission rates derived from these models will provide an upper bound 104 �on what could be expected in wild populations and will guide construction of candidate sets for modeling those populations. Host range and strain variation: We continued a series of experimental studies in cattle, fallow deer, and mountain lions to explore potential host range of CWD after intense but natural exposure; these experiments compliment ongoing surveillance for evidence of infection in species not known to be natural hosts of CWD, including moose and mountain lions. We also completed work looking for evidence of strain variation in CWD agent from various deer sources using domestic ferrets as a laboratory model. Effects of land use on prevalence: We summarized findings on the apparent effects of urban vs. nonurban land use patterns on CWD prevalence in mule deer. Selective predation upon infected mule deer: We continued a study comparing prevalence of CWD among puma-killed mule deer to prevalence among mule deer harvested or randomly culled by humans within home ranges of collared mountain lions to assess whether predation is selective for CWDinfected mule deer. Methods were as described previously (Miller and Wolfe, 2003, Work Package 3430, Task 7410, Progress Report, Colorado Division of Wildlife, Ft. Collins). A total of eight adult mountain lions have been collared, resulting in 39 collared cat months between February 2001 and May 2004. Sampling of predator-killed deer is ongoing. Influence of trace minerals on susceptibility: We continued two independent studies to investigate the potential influence of trace minerals on CWD susceptibility. In a retrospective study, we completed analyses of archived tissues to compare tissue levels of copper (Cu), molybdenum (Mo), and manganese (Mn) in mule deer infected with CWD to levels in apparently uninfected deer from the same geographic area. We also continued an experiment to examine the effect of Cu supplementation on CWD susceptibility in white-tailed deer, wherein we are comparing the natural infection rate and course of CWD in captive deer receiving a sustained-release oral Cu supplement to the rate and course in unsupplemented controls residing in the same paddock. Genetic influences on susceptibility: We continued collaborating with investigators from the University of Wyoming (UWYO) in studies of genetic influence on CWD susceptibility in mule deer. The main objective of ongoing UWYO research has been to search the DNA sequence of the PrP encoding region in exon 3 of the Prnp gene of mule deer for genetic variations that may influence occurrence of naturally acquired CWD. Recent analyses included samples from 529 free-ranging mule deer from four Colorado DAUs (326 from D-10, 63 from D-19, 71 from D-9, and 69 from D-7). Total genomic DNA was extracted from each sample and the PrP coding region from each deer genome was amplified by polymerase chain reaction (PCR). Genotyping was done using commercial sequencing or a simple restriction enzyme digestion (J. E. Jewell, unpublished data) of the PCR amplified PrP gene. Preventive therapies: We collaborated with investigators from the Rocky Mountain Laboratories, NIH, to conduct a pilot study evaluating safety and efficacy of three prospective therapies for preventing CWD in mule deer. Twenty hand-raised mule deer were randomly divided into groups of 5 and assigned to receive a candidate therapy (coded PP, TA, or TC) or no therapy (control). We administered therapies continuously or 3× daily depending on the drug used; administration began 14 days before inoculation, and continued for 14 days after challenge. All groups received pelleted feed and alfalfa hay from the 105 �same source. We used a novel oral inoculation method (M. W. Miller & L. L. Wolfe, unpublished data) for experimental challenge. We collected tonsil biopsies (Wolfe et al., 2002, J. Wildl. Manage. 66:564– 573) from controls about 4 mo post inoculation (PI) and principals about 5 mo PI to assess efficacy of respective therapies in preventing CWD in mule deer. Evaluation of an urban CWD management strategy: We completed an assessment of the feasibility of “test and cull” as an approach for managing CWD in urban habitats, and continued a 5-year study to evaluate the efficacy of this approach in reducing CWD prevalence among urban mule deer. During October 2003−May 2004, we again captured and tested free-ranging mule deer, and marked them with timed-release radiocollars in urban areas throughout Estes Park; our work was complimented by a parallel, coordinated effort by the US National Park Service (NPS) to capture and test deer inside Rocky Mountain National Park (RMNP). The collective annual goal was to test ≥50% of the adult mule deer residing the Estes Park population unit (Conner and Miller, 2004, Ecol. App. in press); target sample sizes (52 adult males and 153 adult females) were estimated based on a mark-resight inventory conducted in December 2003. Field methods were as previously described (Wolfe et al., 2004, Wildl. Soc. Bull. in press). In addition to the primary goal of assessing the efficacy of test and cull as a management strategy, data gathered in the course of this study will also be useful in improving our understanding and modeling of the influences of urban landscapes on CWD epidemiology. STUDIES OF CWD PATHOGENESIS & DIAGNOSIS We continued or initiated studies related to pathogenesis in natural hosts and live-animal diagnostic test refinement and evaluation as aids to improving approaches for CWD surveillance and diagnosis in free-ranging deer and elk in Colorado. Pathogenesis in natural host species: We completed our work studying the pathogenesis of CWD in white-tailed deer after oral inoculation with infectious, conspecific brain tissue. This study will complement studies documenting CWD pathogenesis in mule deer and elk that already have been completed. Evaluation of antemortem diagnostic techniques: We continued working to refine and evaluate tonsil biopsy techniques for diagnosing CWD in live animals. In light of our continued success in applying established techniques (Wolfe et al., 2002, J. Wildl. Manage. 66:564–573), we have continued using tonsil biopsy to gather data for field studies and epidemiological investigations. We also began using tonsil biopsy IHC as diagnostic benchmark for evaluating other candidate tests for diagnosing CWD in live animals. In conjunction with ongoing studies on CWD transmission, we evaluated a candidate rapid test developed by Prion Developmental Laboratories, Inc. (PDL), modified for potential use under field laboratory conditions (J. E. Jewell, unpubl. data). Initial evaluation of this test on biopsy-sized pieces of tonsil tissue collected postmortem from culled mule deer revealed that sensitivity (about 80%) was near the lower limit of acceptability for field use. Modifications to improve sensitivity were made, and subsequently we evaluated sensitivity of the PDL test under conditions simulating those anticipated in field applications using tonsil biopsies collected from captive mule deer naturally infected with CWD. Tissue samples collected via tonsilar biopsy (Wolfe et al., 2002, J. Wildl. Manage. 66:564–573) were examined within 10 min of collection via the candidate PDL test; details of laboratory techniques were proprietary. Laboratory equipment and conditions simulated those that we anticipated would be available at a field site. Paired biopsies were collected from infected and uninfected deer (n = 16); we randomly assigned one of each pair to PDL and the other to immunohistochemistry (IHC) evaluation. Biopsies were processed, test reactions evaluated, and deer categorized as CWD-positive or -negative based on observed reactions; laboratory personnel were unaware of the infection status of sampled deer. Time 106 �from sample collection to reporting of test result was recorded for each biopsied deer. Sensitivity (estimate, 95% CI) of the PDL test was calculated, using IHC as the reference standard. We compared the proportion of positive deer detected by the PDL test to results from IHC using a one-sided Fishers exact test; we used α = 0.1 for all analyses. In addition, the mean reporting time and range of reporting times was calculated for use in assessing the utility of the PDL test under anticipated field conditions. In conjunction with ongoing studies on CWD prevention, we also reevaluated nictitating membrane (also called the “third eyelid”) biopsy as an approach for detecting CWD in live mule deer. We used a modified technique devised for domestic sheep (S. Bender, unpublished data) to identify lymphoid tissue on the nictitating membrane and adjacent conjunctiva, then collected biopsies using established techniques (O’Rourke et al., 1998, Vet. Rec. 142:489-491). We sampled both eyes of 11 mule deer experimentally infected with CWD and known to be tonsil biopsy positive. Nictitating membrane biopsies were evaluated by IHC using published methods (O’Rourke et al., 1998, Vet. Rec. 142:489-491; O’Rourke et al., 2002, Clin. Diag. Lab. Immunol. 9:966-971). We calculated the proportion (± 95% CI) of usable nictitating membrane biopsies, as well as the sensitivity (± 95% CI) of nictitating membrane biopsy IHC for CWD diagnosis using tonsil biopsy IHC as the reference; criteria for regarding this nictitating membrane biopsy technique as potentially useful in diagnosing CWD in mule deer were ≥ 90% of samples containing usable lymphoid tissue and estimated sensitivity ≥ 95%. We also collaborated in a second study to evaluate a prospective rapid blood test (GeneThera, Denver, CO) for diagnosing CWD in live deer. A total of 10 blood samples from tonsil biopsy-positive, captive mule deer were collected by GeneThera representatives using collection materials and protocols provided by the laboratory; samples were immediately taken to their laboratory for evaluation. By previous agreement, the status of sampled animals was known to GeneThera personnel prior to blood collections. RESULTS AND DISCUSSION STUDIES OF CWD EPIDEMIOLOGY & MANAGEMENT Statewide CWD surveillance: The CDOW sampled over 15,000 deer and elk harvested or culled in northern Colorado and other select locations, as well as smaller numbers of deer and elk submitted as clinical suspects. Surveillance revealed two previously undetected CWD foci in mule deer, one on the Grand Mesa (DAU D-51) and the other in Colorado Springs (DAU D-16). Survey results will be posted on the Division’s CWD web page (). Surveillance data also will be used to augment an existing database that is the foundation for ongoing analyses and modeling of temporal and spatial aspects of CWD epidemiology, as well as for evaluating responses to management. In addition to reaffirming the spatial heterogeneity among wintering mule deer subpopulations observed previously (Miller et al., 2000, J. Wildl. Dis., 36:676–690; Conner & Miller, 2004, Ecol. App., in press), our analyses revealed marked differences in CWD prevalence by sex and age groups, as well as clear local trends of increasing prevalence over a 7-yr period. CWD prevalence differed (P < 0.0001) by age (yearling vs. adult), sex, and geographic area at two different spatial scales (game management unit [GMU] or population unit winter range), and increased over time at both geographic scales (GMU: β = 0.064, 95% CI = 0.009−0.119, P = 0.0219; population unit: β = 0.263, 95% CI = 0.134−0.399, P < 0.0001). Disease status (positive or negative) was not independent of age for males (n = 947, df = 3, χ2 = 459, P < 0.0001) or females (n = 549, df = 4, χ2 = 71, P < 0.0001). For both sexes, prevalence peaked in the 4−6-yr old age class, with the largest increase occurring between the 2−3-yr-old and 4−6-yr-old age classes. This differential was larger for males: prevalence rose from 5.9% (95% CI = 4.9−6.8) among 2−3-yr-olds to 19.4% (95% CI =12.1−26.7) among 4−6-yr-olds (P = 0.0002); for the 4−6 yr age class, prevalence among males (19.4%) was 2.7× greater (P = 0.0006) than among females (7.2%). 107 �Demographic, spatial, and temporal factors all appear to contribute to the marked heterogeneity in CWD prevalence in endemic portions of northcentral Colorado. These factors likely combine in various ways to influence epidemic dynamics on both local and broad geographic scales. A manuscript describing our findings is in review for publication in the Journal of Wildlife Diseases. Sampling of vehicle-killed mule deer may be exploited in increasing the efficiency of surveillance programs designed to detect new foci of CWD infection and direct management actions; however, this differential vulnerability also may bias prevalence estimates in natural populations when data from vehicle-killed deer are included in calculating such estimates. Overall CWD prevalence was 1.66× higher in vehicle-killed deer; prevalence among vehicle-killed deer was 0.101 (95% confidence interval [CI] = 0.064−0.139) compared to 0.061 (95% CI = 0.051−0.072) prevalence among mule deer harvested, culled, or biopsied within 3 km of collision sites. The probability of detecting a CWDinfected, vehicle-killed deer, given that at least one other CWD-infected deer had been detected within a 3 km radius of the vehicle-kill site, was 16.7%. Our data suggest increased susceptibility of CWD-infected individuals to vehicle collisions. Evidence of increased susceptibility to vehicle collisions also may aid in understanding vulnerability of CWD-infected individuals to other forms of death, particularly predation. A manuscript describing our findings is in review for publication in the Journal of Wildlife Diseases. Transmission mechanisms: Manuscripts describing our findings on the relative importance of animal−animal transmission of CWD and on the relative contributions of live animals, contaminated environments, and infected carcasses to CWD transmission were accepted for publication and subsequently published in peer-reviewed journals (see Appendix for citations). Modeling epidemic dynamics in captive mule deer: Preliminary analyses suggest that indirect transmission models best represent epidemic data; moreover, our model selection results align well with independent empirical findings on CWD transmission mechanisms (Miller et al., 2004, Emerg. Inf. Dis. 10:1003−1006). We will continue refining candidate models before making final comparisons and parameter estimations. Findings should be of use in refining epidemic models of CWD in free-ranging mule deer populations. Host range and strain variation: Cattle (n = 11) living in paddocks with naturally-infected mule deer remained healthy through 7 years of exposure; in contrast, only 1 of 12 mule deer introduced into these same paddocks in 1997 is still alive. Our results are consistent with data from cell-free conversion (Raymond et al., 2000, EMBO 19:4425-4430) and intracerebral (IC) challenge (Hamir et al., 2001, J. Vet. Diag. Invest. 13:91–96) studies that suggest the probability of natural susceptibility to CWD in cattle is extremely low. Similarly, neither signs nor postmortem evidence of infection have been observed in fallow deer (n = 24) exposed to infected mule deer for ≤3.5 years, and mountain lions (n = 3) consuming carcasses of CWD-infected deer and elk for >2 years also have remained healthy. No evidence of infection has been observed in moose, mountain lions, or cattle examined via ongoing surveillance. Clinical signs and postmortem findings consistent with CWD in ferrets were observed in four of five IC-inoculated with tissue from infected deer, but were not observed in the free-ranging white-tailed deer or control groups. Incidence and incubation periods were consistent among affected groups. Preliminary assessment of Western blots (WB) revealed no apparent differences in glycosylation patterns among WB-positive ferrets, and no evidence of infection in the unaffected white-tailed deer or control groups. Effects of land use on prevalence: Urban land use appears to affect CWD prevalence: rates were higher in developed areas and among male mule deer, suggesting anthropogenic influences on the 108 �occurrence of CWD. We also observed relatively high variation in prevalence across three study sites (Estes Park, Horsetooth Mountain, Glacier View Meadows), suggesting that spatial patterns may be influenced by other factors operating at a broader, landscape scale. Our results suggest that multiple factors, including changes in land use, differences in exposure risk between sexes, and landscape-scaled heterogeneity, are associated with CWD prevalence in north-central Colorado. A manuscript describing these findings in currently “in press” (see Appendix for citation). Selective predation upon infected mule deer: Our work continues from a pilot study conducted to evaluate available global positioning system (GPS)-based telemetry collars for use in this sampling application. Three collar styles have been deployed, and we are continuing to test and evaluate this new technology; aside from our main objective of data gathering related to CWD ecology, evaluation of this technology should be a substantial contribution to future studies of predator-prey relationships. We have detected and examined over 85 kill sites from radio-collared mountain lions and successfully sampled tissues from 28 sites where adult mule deer carcasses were present; we also have collected 17 samples opportunistically from mule deer killed by mountain lions that were not radio-collared. We will continue capturing mountain lions to reach the objective of six to nine collared cat years, and will continue sampling carcasses of lion-killed mule deer to reach our target sample size (n = 157). We also will continue refining our monitoring approach to ensure that we find kill sites quickly enough to retrieve a suitable tissue sample to test for CWD. Whether target sample sizes can be attained in the time planned for this work remains to be determined. Influence of trace minerals on susceptibility: Both studies are well underway. Laboratory analyses of retrospective samples are complete, and data analysis is underway. Experimentally- treated and control deer are being sampled on a regular schedule, but laboratory analyses are incomplete. Genetic influences on susceptibility: Only four codons in the open reading frame of the Pnrp gene exhibit variation in mule deer, and only one of the four results in a change in the final version of PrP (Brayton et al., 2004, Gene 326:167−173; J. E. Jewell, unpublished data) -- this is a change from the amino acid serine (S), the high frequency allele, to phenylalanine (F) at codon 225. Preliminary results showed that estimated frequency of F allele occurrence in gene pools was similar in Colorado DAUs with (D-10: 0.095; n=652) and without endemic CWD (D-19: 0.111; n=126). However, F225 was not detected in the genomes of CWD-infected deer from D-10 (n = 50), and the F225 gene frequency was lower than in uninfected D-10 deer (0.1; χ c2 = 4.6, P < 0.05). We observed a similar pattern of low F225 gene frequency among mule deer infected with CWD after experimental exposure via direct and indirect routes (Miller et al., 2004, Emerg. Inf. Dis. 10:1003−1006). Whether F225 affects truly affects CWD susceptibility or transmission in mule deer remains to be determined, and is the subject of continued investigation. Preventive therapies: All 4 control deer that survived to 4 mo PI showed evidence of PrPCWD accumulation in tonsil biopsies collected ~4 mo PI. Unfortunately, all but 2 of the 15 treated deer also showed PrPCWD accumulation in tonsil biopsies collected ~5 mo PI; the 2 apparently uninfected deer were both from the same treatment group (PP), but overall infection rate did not differ (P = 0.28) from control. We will continue following these deer to examine potential differences in post-exposure survival that could be attributable to therapies, and to further document the outcomes of the alternative inoculation method used. We also plan to continue this work if other candidate therapies become available. Evaluation of an urban CWD management strategy: Data from the 2002−2003 field season indicated that testing and culling mule deer in Estes Park could be done at rates needed to evaluate the efficacy of this approach in reducing CWD prevalence. A manuscript describing the results of our feasibility study is in preparation is “in press” for publication in the Wildlife Society Bulletin. 109 �0.3 Prevalence Because we were successful in reaching objectives for population-level testing, the 2002−2003 field season became year 1 of a 5-year study to evaluate the efficacy of “test and cull” as a CWD control strategy. In year 2 (2003−2004 field season), we captured and tested 44 adult (≥1.3 yr old) male and 119 adult female mule deer in Estes Park. CWD prevalence was about 13.6% among males and 5% among females tested in --- Males --------------- □ Females 0.2 0.1 0 2002 2003 Year Estes Park (Fig. 1); although no clear evidence of a Figure 2. Chronic wasting disease (CWD) prevalence among male (teal bar) and female (plum treatment effect has emerged (Fig. 1), it is probably bar) mule deer tested in Estes Park, Colorado, unrealistic to expect measurable changes in prevalence after 2002−2004. Prevalence between years did not only 1 year of test and cull management. The combined differ (Fisher exact test P≥0.4). Vertical lines are upper 95% confidence limits on estimated efforts of CDOW and RMNP programs resulted in an prevalence. overall testing rate of 63% of the deer wintering in the Estes Park vicinity, including about 90% of the estimated 103 male and 55% of the estimated 306 female deer in this population unit. STUDIES OF CWD PATHOGENESIS & DIAGNOSIS Pathogenesis in natural host species: White-tailed deer inoculated orally with about 2.5 g of brain tissue homogenate (containing about 15 µg PrPCWD) developed clinical CWD and were euthanized in endstage disease 16−30 mo postinoculation (PI). The clinical course in inoculated white-tailed deer was similar to that previously observed in mule deer inoculated with about 15 µg PrPCWD from infected mule deer. Laboratory evaluations of tissues from both our white-tailed deer and mule deer pathogenesis studies are pending. Evaluation of antemortem diagnostic techniques: Tonsil biopsy is a useful tool for estimating CWD prevalence in nonhunted mule deer populations. In addition to applications in the two field studies described here, the techniques we developed are being used in at least six other field studies of CWD epidemiology (WY, NM, WI, SD, NE, CO). Although the PDL test showed considerable promise as a potential field test, assay performance will need to be improved before it can be incorporated into ongoing CWD research or management programs. We observed good assay sensitivity (1.0; 6/6), but relatively low specificity (0.7; 7/10); overall agreement with IHC was 0.64 (95% CI = 0.29−0.98). There appeared to be an unacceptably high number of “false positive” tests -- application in a low prevalence population (e.g., Estes Park) would likely lead to unnecessary culling of numerous healthy deer, and could erode public support for our field study. Consequently, the PDL test was not incorporated into the 2003−2004 field study in Estes Park. The nictitating membrane biopsy technique provided a high proportion of usable samples: all 22 samples contained at least 1 lymphoid follicle and 12−16/22 (55−73%) samples contained ≥ 9 follicles. Unfortunately, IHC of nictitating membrane biopsies detected PrPCWD accumulation in only 2/22 biopsies, both from the same deer. Because estimated sensitivity (0.09; 95% CI 0.01−0.29) is inadequate, we cannot recommend incorporation of nictitating membrane biopsy IHC into any of our ongoing CWD research or management programs. We remain unable to assess the reliability or repeatability of the “GeneThera test”. No test results were provided on the 10 blood samples from positive mule deer; instead, a company representative indicated that extractions from samples were unsuccessful, and that consequently tests could not be run. 110 �This is our second unsuccessful attempt to obtain results from blood samples submitted to GeneThera for CWD testing. Until an evaluation of their test can be completed, we cannot recommend its incorporation into any of our ongoing CWD research or management programs. APPENDIX Publications arising from ongoing CWD work: Belay, E. D., R. A. Maddox, E. S. Williams, M. W. Miller, P. Gambetti, and L. B. Schonberger. 2004. Chronic wasting disease and potential transmission to humans. Emerging Infectious Diseases 10:977−984. Brayton, K. A., K. I. O’Rourke, A. K. Lyda, M. W. Miller, and D. P. Knowles, Jr. 2004. A processed pseudogene contributes to apparent mule deer prion gene heterogeneity. Gene 326:167−173. Miller, M. W. and M. A. Wild. 2004. Epidemiology of chronic wasting disease in captive white-tailed and mule deer. Journal of Wildlife Diseases 40: 320−327. Miller, M. W., and E. S. Williams. 2003. Horizontal prion transmission in mule deer. Nature 425: 35−36. Miller, M. W., and E. S. Williams. 2003. Chronic wasting disease of cervids. In Mad cow disease and related spongiform encephalopathies. D. A. Harris, (Ed.). Current Topics in Microbiology 284:193−214. Miller, M. W., E. S. Williams, N. T. Hobbs, and L. L. Wolfe. 2004. Environmental sources of prion transmission in mule deer. Emerging Infectious Diseases 10: 1003−1006. Miller, M. W., E. S. Williams, B. E. Powers, L. A. Baeten, L. L. Wolfe, and K. L. Green. 2004. Epidemiology and management of chronic wasting disease in free-ranging cervids. In Proceedings of the One Hundred and Seventh Annual Meeting of the United States Animal Health Association, pp. 60−63. O’Rourke, K. I., D. Zhuang, A. Lyda, G. Gomez, E. S. Williams, W. Tuo, and M. W. Miller. 2003. Abundant PrPCWD in tonsil from mule deer with preclinical chronic wasting disease. Journal of Veterinary Diagnostic Investigation 15: 320−323. Powers, B. E., C. P. Hibler, T. R. Spraker, and M. W. Miller. 2004. Large-scale surveillance for chronic wasting disease: The Colorado laboratory experience. In Proceedings of the One Hundred and Seventh Annual Meeting of the United States Animal Health Association, pp. 64. Sigurdson, C. J., and M. W. Miller. 2003. Other animal prion diseases. In Prions for physicians. C. Weissmann, A. Aguzzi, D. Dormont, and N. Hunter, (Eds.). British Medical Bulletin 66: 199−212. Williams, E., and M. Miller. 2003. Prions in the wild: CWD in deer and elk. Microbiology Today 30: 172−173. Wolfe, L. L., W. R. Lance, and M. W. Miller. 2004. Immobilization of mule deer with thiafentanil (A3080) or thiafentanil plus xylazine. Journal of Wildlife Diseases 40: 282−287. Prepared by ____________________________ Michael W. Miller, Veterinarian 111 � https://cpw.cvlcollections.org/files/original/b5591f5db4ca458a29bccdc7cb2f03a3.pdf e4cdf6bbe2a98490cb52044f13029007 PDF Text Text 141 JOB PROGRESS REPORT State of ______C=ol=o=ra=d=o_ _ _ __ Division of Wildlife - Mammals Research Work Package No. ---=3-'-74-'-0=-------- Chronic Wasting Disease and Other Wildlife Disease Management Chronic Wasting Dise~e Surveillance and Laboratory Support T~k 2 --------~------- Period Covered: July 1 2002 through June 30, 2003 Author: L. A. Baeten Personnel: K. Cramer, K. Green, E. Knox, C. T. Larsen, N. Mier, M. W. Miller, K. Taurman, and L. L. Wolfe Interim Report - Preliminary Results This work continues, and precise analysis ofdata has yet to be accomplished. Manipulation or interpretation of these data beyond that contained in this report should be labeled as such and is discouraged. ABSTRACT We established and staffed a Wildlife Health Laboratory (WHL) to facilitate expanded needs for chronic wasting disease (CWD) surveillance throughout Colorado. WHL activities supported CWD epidemiology and management work, as well as various new and ongoing CWD research projects. INTRODUCTION We established and staffed a Wildlife Health Laboratory (WHL) to facilitate expanded needs for chronic w~ting disease (CWD) surveillance throughout Colorado. WHL activities supported CWD epidemiology and management work, as well~ various new and ongoing CWD research projects. Key contributions are described herein. METHODS Statewide CWD surveillance: The discovery of CWD in northwestern Colorado in January 2002 created a sudden demand for both more widespread surveillance and more rapid turnaround on laboratory results. Consequently, the CDOW's CWD surveillance program was overhauled and its capacity greatly expanded over the summer of 2002 in order to meet anticipated demands for surveillance data, ~ well ~ to meet policy-based decisions to provide carc~s quality ~surance information for individual hunters. The most notable changes were the addition of three regional submission laboratories, streamlining of tissue sampling methods, and incorporation of a rapid screening test for CWD diagnosis. Details of overall programmatic features and changes were described on a new CWD-oriented CDOW web page (http://wildlife.state.co.us/CWD/index.asp); details of the evaluation of modified sampling and testing procedures are described below. Evaluation of a rapid screening test: In conjunction with expanded CWD surveillance in Colorado during Sep-Dec 2002, tissue samples (n = 25,050 total) from 23,256 mule deer, white-tailed deer, and Rocky Mountain elk collected statewide were examined using an ELISA developed by Bio-Rad Laboratories, �142 Inc. (brELISA) in a two-phase study. In the validation phase of this study, a total of 4,175 retropharyngeal lymph node (RLN) or obex (OB) tissue samples were examined independently by brELISA and immunohistochemistry (IHC). There were 137 IHC positive samples and 4,038 IHC negative samples. Optical density (OD) values from brELISA were classified as "not detected" or "suspect" based on recommended cut-off values during the validation phase. Based on the validation phase data, only RLN samples were collected for the field application phase ofthis study and only samples with brELISA OD values> 0.1 were examined by IHC. We estimated assay performance parameters (sensitivity, specificity, agreement) for brELISA to determine the utility of this rapid screening assay in CWD surveillance programs. RESULTS AND DISCUSSION Statewide CWD surveillance: The CDOW sampled over 26,000 deer and elk harvested or culled in northeastern Colorado and other select locations. Survey results were posted on the Division's CWD web page. Prevalence data also will be used to augment an existing database that is the foundation for ongoing analyses and modeling of temporal and spatial aspects of CWD epidemiology, as well as for evaluating responses to management. This year's data will be particularly useful in further exploring local patterns of disease prevalence related to deer movement, density, and land use patterns. Moreover, the surveillance strategy and methods first devised and implemented in Colorado recently served as a model for developing national recommendations on CWD surveillance in free-ranging populations. Evaluation of a rapid screening test: In the validation phase, using II-IC-positive cases as known CWDinfected individuals and assuming II-IC-negative cases were uninfected, the relative sensitivity of brELISA depending on species ranged from 98.3-100% for RLN samples and 92.1-93.3% for OB samples; the relative specificity ofbrELISA depending on species ranged from 99.9-100% for RLN samples and was 100% for OB samples. Overall agreement between brELISA and IHC was ~97.6% in RLN samples and ~95.7% in OB samples of all species where values could be calculated; moreover, mean brELISA OD values were ~46x higher in II-IC-positive samples than in II-IC-negative samples. Discrepancies were observed only in early-stage cases of CWD. Among 20,875 RLN samples screened with brELISA during the field application phase, 155 of8,877 mule deer, 33 of 11,731 elk, and 9 of267 white-tailed deer samples (197 total) had OD values > 0.1 and were further evaluated by IHC to confirm evidence ofCWD infection. Of cases flagged for IHC follow-up, 143 of 155 mule deer, 29 of33 elk, and all 9 white-tailed deer were confirmed positive. Mean (± SE) OD values for II-IC-positive cases detected during the field application phase were comparable to those measured in RLN tissues during the validation phase. Based on these data, brELISA was determined to be an excellent rapid test for screening large numbers of samples in surveys designed to detect CWD infections in deer and elk populations. Publications: Hibler, CP, Wilson, KL, Spraker, TR, Miller, MW, Zink, RR, DeBuse, LL, Andersen, E, Shcweitzer, D, Kennedy, JA, Baeten, LA, Smeltzer, JF, Salman, MD, Powers, BE Field Validation and assessment of an enzyme-linked immunosorbent assay for detecting chronic wasting disease in mule deer (Odocoileus hemionus), white-tailed deer (Odocoileus virginianus), and Rocky Mountain elk (Cervus elaphus nelsoni). 2003 J. Vet Diagn Invest 15:311-319. �143 APPENDIX Publications arising from WHL contributions to ongoing CWD work: Hibler, C. P., K. L. Wilson, T. R. Spraker, M. W. Miller, R.R. Zink, L. L. DeBuse, E. Andersen, D. Schweitzer, J. A. Kennedy, L. A. Baeten, J. F. Smeltzer, M. D. Salman, and B. E. Powers. 2003. Field validation and assessment of an enzyme-linked immunosorbent assay for detecting chronic wasting disease in mule deer (Odocoileus hemionus), white-tailed deer (Odocoileus virginianus), and Rocky Mountain elk (Cervus elaphus nelsoni). Journal of Veterinary Diagnostic Investigation 15: 311-319. Samuel, M. D., D. 0. Joly, M.A. Wild, S. D. Wright, D. L. Otis, R. W. Werge, and M. W. Miller. 2003. Surveillance strategies for detecting chronic wasting disease in free-ranging deer and elk. Results of a CWD surveillance workshop. USGS, BRD, National Wildlife Health Center, Madison, Wisconsin. �Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Project No. Work Package No. Task Colorado : : : : 3740 Federal Aid Project Cost Center 3440 Wildlife Health Program Wildlife Diseases Wildlife Disease Surveillance Technical and Laboratory Support : Period Covered: July 1 2003 through June 30, 2004 Author: L. A. Baeten Personnel: K. Cramer, K. Green, K.A. Griffin, E. Knox, C. T. Larsen, M. W. Miller, L. L. Wolfe and D. Wroe All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT The Wildlife Health Laboratory (WHL) was initially created in 2002 to meet expanded needs for chronic wasting disease (CWD) surveillance throughout Colorado. WHL activities supported CWD epidemiology and management work, as well as various new and ongoing CWD research projects. In addition, the WHL has been able to meet demands for diagnostic and laboratory services related to other wildlife diseases that have come to the forefront of concern in the management of Colorado’s wildlife resources. 113 �JOB PROGRESS REPORT WILDLIFE DISEASE SURVEILLANCE TECHNICAL AND LABORATORY SUPPORT L. A. BAETEN INTRODUCTION The Wildlife Health Laboratory (WHL) was created in 2002 in response to the H-1.1 objective of the Division’s Strategic Plan. The purpose delineated in this objective is to “aggressively research, identify, detect, contain and eliminate, where possible, diseases in free-ranging wildlife and captive wildlife that could negatively impact wildlife populations”. The WHL was developed to meet the expanded needs for chronic wasting disease (CWD) surveillance throughout Colorado. WHL activities supported CWD epidemiology, harvest testing and management work, as well as various new and ongoing CWD research projects. In addition, the WHL has been able to meet demands for monitoring, detection, and diagnostic laboratory services related to other wildlife diseases that have come to the forefront of concern in the management of Colorado’s wildlife resources (i.e. West Nile virus, plague, Pasteurellosis, etc.). SUMMARY Statewide CWD surveillance: The discovery of CWD in northwestern Colorado in January 2002 created a sudden demand for both more widespread surveillance and more rapid turnaround on laboratory results. The CDOW’s CWD surveillance program was modified in 2002-2003 to decrease turnaround time (from initial submission to acquisition and posting of results), improve data collection and quality control. The notable changes in 2003-2004 were the addition of an electronic data collection system that was used statewide for collection of field and laboratory data. The WHL staff was instrumental in helping to delineate system mechanics, provide testing and troubleshooting capabilities and assist with training efforts. Details of overall programmatic features and changes were described on the CWDoriented CDOW web page (http://wildlife.state.co.us/CWD/index.asp); details of the efficiencies in the sampling and testing procedures are described below. Numerous state agencies have request demonstrations of this new system for possible implementation in their CWD surveillance programs. During 2003-2004, the CDOW sampled 17,268 deer, elk and moose harvested or culled in northeastern Colorado and other select locations. Survey results were posted on the Division’s CWD web page (http://wildlife.state.co.us/CWD/index.asp). The data generated provided annual CWD survey results. These data were added to the cumulative surveillance data that is the foundation for ongoing analysis and modeling of temporal and spatial aspects of CWD epidemiology, examining potential influences of demographics, as well as evaluating responses to management. In an effort to improve surveillance efficiencies, tissue samples were collected from deer killed from vehicle collisions throughout the state. Prevalence data from this group were analyzed to determine if CWD-infected individuals were more vulnerable than otherwise healthy animals. Moreover, the surveillance strategy and methods first devised and implemented in Colorado continue to serve as a model for developing national recommendations on CWD surveillance in freeranging populations. 114 �CWD tissue handling and disposal: The WHL staff prepared documents summarizing published literature on appropriate disposal methods for CWD infected tissues (incineration and chemical). These documents were used extensively for public information during the review process for the incinerator proposal for Wellington. The WHL supervisor worked with EPA and national veterinary diagnostic laboratory representatives to develop “Best Management Practices” when handling CWD infected materials. Research projects The WHL lab staff provided technical and diagnostic support for the ongoing DOW research projects listed below. Major accomplishments and contributions for the WHL this fiscal year include: completion of the experimental phase of the “Molecular epidemiology of strain variations in CWD”; evaluation of antemortem diagnostics (described below); the addition of DNA extractions to the list of diagnostic capabilities (supporting the “Genetic influences study and several species conservation projects); initiation of the pilot study on biosolids and wastewater; preliminary results from studies looking at prion inactivation are ready for presentation; and extensive sample collections for the studies of prions in biological excretions and environmental samples. 1. Molecular epidemiology of strain variations in chronic wasting disease (CWD) 2. CWD host range studies 3. Selective predation upon CWD-infected mule deer 4. Trace mineral influences on CWD susceptibility 5. Genetic influences on CWD susceptibility 6. Evaluation of preventative therapies for CWD 7. Evaluation of antemortem diagnostic techniques for CWD 8. Detection of prions in environmental samples 9. Detection of prions in biosolids and waste water 10. Chemical inactivation of prions 11. Detection of prions in biological excretions 12. West Nile virus in black-tailed prairie dogs 13. Prevalence of CWD in ungulates killed via vehicle collisions 14. Evaluation of a recombinant plague vaccine in lynx 15. Invertebrate role in CWD transmission 16. Evaluation of FWRF diarrhea outbreaks 17. Uncompaghre fawn mortality Evaluation of antemortem diagnostic technique for CWD: In a pilot study, the WHL evaluated the use of a lateral flow strip test (Prion Developmental Laboratories, Inc.) to determine its applicability for “live animal testing” in the field. Approximately 40 tonsil samples were used to assess the applicability of this diagnostic test under field conditions. The test kit procedures were manipulated to determine if modifications to the lymph node procedures could be used to accommodate tonsil tissue (and tonsil biopsy sized tissues). It was determined that the assay could be modified to work under field 115 �conditions (i.e. roving lab). However, despite efforts to modify the test parameters, the sensitivity was not acceptable to pursue further field trials with this new diagnostic test system. In addition, the WHL staff provide technical assistance to collaborators interested in archived tissues for ongoing research projects listed in the table below. The WHL staff accomplished this via additional sample collections from hunter harvest, culls and other DOW submissions. These samples are archived then aliquoted and shipped to the collaborators according to specific tissue requests. Collaborative Agreements IND/HPF/USCF USDA/ARS NYSIBR PDL CSU USDA/ARS NIH/RML CWRU NYSIBR CSU IDEXX USU RMNP GeneThera Brain Brain Brain Lymph nodes Multiple tissues Eyelids, blood Multiple tissues Brain, lymph nodes Urine Brain Lymph nodes DNA extracts (blood) DNA extracts (blood) Blood Transgenic mouse development/ host range studies Strain typing, comparison to other TSE strains Transgenic mouse strains Lateral flow strip test Experimental transmission (ante and post mortem collection CWD assay evaluations Evaluate strain variations Transgenic mouse host strain study, cellular prion transport CWD assay evaluations Effects of composting on prion inactivation Validation of diagnostic assay Epidemiology studies Epidemiology studies Antemortem assay evaluation Wildlife Disease Surveillance: The WHL performed necropsies to assist state wildlife managers and biologists in determining the cause of death for wildlife species including: deer, elk, bighorn sheep, mountain goats, bear, various avian species and rodents (See Table 1). This necropsy effort included support for two species conservation projects. The WHL provided technical and diagnostic support for the field projects listed below. This effort included biological sample collection, data collection, sample processing, diagnostic testing, archiving and/or distribution of samples. 1. Evaluation of diagnostic techniques for avian translocations 2. Disease surveillance for Prairie grouse restoration 3. Disease surveillance for Turkey translocation 3. Bighorn sheep translocations: Identification of Pasturella spp. strains 4. Identification of Johne’s disease in BHS and RMG 5. Identification of lungworm larvae in BHS feces 6. Black-footed ferret restoration: carnivore sampling 7. Lynx restoration 8. Winter deer capture: mule deer survival monitoring 9. Elk Fertility control 10. Test and cull evaluation 11. CWD management culling 12. Foot hills wildlife research facility 116 �West Nile Virus: The WHL established in-house testing for West Nile virus (WNV). Fifty-four carcasses were submitted as suspects for necropsy and testing during this fiscal year. Thirteen positives were identified. In conjunction with the DOW WNV testing performed at the WHL, tissue samples collected during necropsies were provided to CDC (Komar) for their use in experimental trials developing a new post mortem assay for WNV. Avian Translocations: The WHL investigated alternative diagnostic testing for avian translocation projects. An in-house assay for Mycoplasma (synoviae, gallisepticum, meleagridis) was determined to be optimal for testing individual birds being translocated. The use of the ELISA assay for these serological tests minimized the cross reactivity effects that were experienced with diagnostic tests used previously. With the use of the ELISA, next-day releases were possible, therefore, decreasing individual stress levels and increasing survivability for birds moved in translocation efforts. The WHL staff provided technical and diagnostic support for three avian translocation projects during this project year (sharp-tailed grouse, turkey and ring-necked pheasant). In combination with the diagnostic necropsy support, the WHL established a database to allow electronic review of these data over time. All historical diagnostic records were incorporated into the database during this fiscal year. To date, this database contains approximately 400 records. From the diagnostic reports database, the WHL prepared wildlife disease summaries for statewide distribution. The wildlife disease summaries for years 2002 and 2003 delineate animal mortality data by species, quarter and region. This data will assist wildlife veterinarians, managers and biologists in future wildlife disease events. During this fiscal year, the WHL established an archive database which includes all of the historical samples collected since the initial establishment of the WHL in the 1990’s. This database allows WHL staff to determine what tissues are available for use in research projects, delineates physical locations where various tissue samples can be found, tracks distribution of tissue samples and contains appropriate animal identification and specifications. At the end of the fiscal year, there were a total of 4,850 entries with approximately 300 of those added for the year. Table 1: Diagnostic Support Species Necropsies Diagnostic Samples (collected, processed, archived) Carnivore 3 40 Deer 30 385 Elk 10 4 Lynx 0 106 Other avian species (WNV) 36 36 Other ungulate 18 107 Prairie grouse 7 22 Small game 10 50 Small mammals 15 35 Total 126 745 117 �Training sessions: The WHL has provided multiple training session for CWD sample collection. The attendees included CDOW employees as well as federal employees from the Rocky Mountain region. In addition, the WHL staff assisted with training sessions for DWM trainees in necropsy techniques, darting and sample collections. Presentations: The WHL staff made various presentations on wildlife disease to various groups including the Wildlife Society, USFWS, CDOW staff, black-footed ferret subcommittee, Colorado Wildlife Federation. The titles of the presentations were: 1. West Nile Surveillance in Colorado 2003 2. The impact of West Nile virus on wildlife populations 3. The significance of West Nile virus in prairie dogs 4. Common diseases in wildlife populations of Colorado APPENDIX Publications arising from WHL contributions to ongoing CWD work: Brayton KA, O’Rourke KI, Lyda AK, Miller MW, Knowles Jr. DP. A processed pseudogene contributes to apparent mule deer gene heterogeneity. Gene 326: 167-173. Miller MW; Williams ES. Horizontal prion transmission in mule deer. Nature 2003 425: 35-36 _________: __________, Hobbs NT; Wolfe LL. Environmental Sources of Prion Transmission in Mule Deer. Emerging Infectious Diseases 2004 10(6): 1003-1007 _________; Wild MA. Epidemiology of Chronic Wasting Disease in Captive White-tailed and Mule deer. J Wildlife Dis 2004 40(2): 320-327 O’Rourke KI; Zhuang D; Lyda A; Gomez G; Williams ES; Tuo W; Miller MW. Abundant PrPCWD in tonsil from mule deer with preclinical chronic wasting disease. J Vet Diagn Invest 2003 13: 320-323 Powers BE, Hibler CP, Spraker TR, Miller MW. Large scale surveillance for chronic wasting disease: The Colorado laboratory experience. Annual proceedings USAHA 2004 pg 64. Prepared by _________________________ Laurie A. Baeten, Veterinarian 118 � https://cpw.cvlcollections.org/files/original/fbee6b213ac31095475b46b4f8398be5.pdf 9da843fa860f7ec35dd3faf7381af65f PDF Text Text 145 JOB PROGRESS REPORT Stateof_ _~C~o~lo~r~ad~o~-----~ Division of Wildlife - Mammals Research Work Package No. 8160 3740 Chronic Wasting Disease and other Willife Disease Management Animal and Pen Support Facilities Task No. -~3_ _ _ _ _ _ _ __ Period Covered: January 1, 2001 - June 30, 2003. Author: T.R. Davis Personnel: 2001: H. Barr, C. Budler, N. Dryer, D. Finley, J. Foster, M. Foster, J. Habel, M. Hanusack, L. Ho, B. Hotchmuth, E. Jones, S. Liss, M. Lowe, M. Miers, A. Mitchell,, T. Petersburg, T. Terrell, C. Weagley, 2001/2002: B. Bates, D. Biggins, E. Berrill, K. Downing, D. Finely, J. Foster, M. Foster, J. Grigg, J. Habel, M. Hanusack, J. Hatch, C. Hernandez, L. Ho, E. Jones, M. Lowe, A. Mitchell, N. Miers, A. Ray, L. Reimer, R. Rhyan, K. Sparks, T. Stout, T. Terrell, R. Thompson, M. Thonhoff, C. Weagley, D. Weaver, B. Williams, T. Zeaman, 2002/2003: M. Anderson, B. Bates, K. Beamer, L. Dahl, J. Fenwick, D. Fox, K. Fox, J. Habel, J. Hatch, T. Halasinski, M. Hanusack, G. Harvey, L. Ho, E. Jones, G. Kyriacou, M. Lowe, N. Miers, A. Mitchell, A. Northrup, R. Rutledge, K. Steffen, T. Stout, D. Thompson, R. Thompson, D. Weaver, ABSTRACT The Colorado Division of Wildlife's Foothills Wildlife Research Facility (FWRF) maintained captive animals (2000/200 l annual total: 262, 2001/2002 annual total: 320, 2002/2003 annual total: 312) and facilities in support of twenty-one captive wildlife research projects. Chronic wasting disease (CWD) pathology, and etiology, in deer and potential transmission to other species was the primary focus of research during this period, however FWRF supported a number of other significant research projects including contraception and reproductive effects, pathogen immunization, foraging behavior, drug delivery systems, and evaluation of wildlife capture pharmaceuticals. Three new species; fallow deer, domestic ferrets, and mountain lions were added to support CWD research as well as additional numbers of mule deer and white-tailed deer. Chronic wasting disease was again a significant source of mortality in mule deer and white-tailed deer and is reflected by the number of CWD research projects conducted at FWRF during this period. The CWD Management Protocol was updated to incorporate new information and early detection techniques, while maintaining the philosophy of managing the disease for research purposes under heightened bio-safety guidelines and intensive herd management. Additionally, a number of other protocols were revised, and new SO P's developed to accommodate the new species, facility improvements, and expanded research. An expanded database, a 5 year facility capitol construction plan, and a draft facility fee schedule were also implemented. The quality of animal care and facility maintenance provided by temporary, work-study, personal service, intern and volunteer employees is in part reflected by the finding of compliance under the Animal Welfare Act during the annual USDA APHIS inspections of FWRF. In addition to routine maintenance, the FWRF team made significant facility improvements including new facilities to accommodate expanded CWD mule deer research, �146 partial completion of a mountain lion holding facility, and support for construction of the new Wildlife Health Lab now located within the FWRF perimeter. Animal Maintenance: Routine animal husbandry including feeding, health observations, training, weighing, and clean-up, was performed primarily by well trained temporary employees, work-study students, and volunteers. FWRF was inspected by USDA APHIS for compliance with federal animal welfare regulations on March 8 200 I, April 12 2002, and April 30 2003. Table I summarizes the species totals reported to USDA animal welfare and includes all neonates born at the facility, transfers into and out of the facility, and all animals that died or were humanely euthanised during the respective fiscal year. Ungulate herd levels at any one time averaged approximately 70 percent of the ungulate total and 60-65 percent of the total number of animals housed at the facility. Table I. Species reported to USDA Animal Welfare Species Bighorn Sheep 2000/2001 Total 57 26 2001/2002 Total 52 22 2002/2003 Total 28 25 25 25 36 74 126 139 21 20 21 24 40 39 227 285 288 159 200 202 11 11 11 21 21 10 3 3 3 262 320 312 Elk Fallow Deer Mule Deer Pronghorn Antelope White-tailed Deer Ungulate Total Ungulate Mean Cattle Domestic Ferrets Mountain Lions Facility Total Herd Management: Three new species; domestic ferrets, fallow deer, and mountain lions were added to the facility in FY 2000/200 l and mule deer and white-tailed deer herd levels were expanded in FY 0 l/02, and 02/03 through herd management practices and incoming transfers. Additional adult animals were brought in to support expanding CWD, fertility control, and brucellosis vaccine research and consisted primarily of free �147 ranging and habituated mule deer obtained from various locations around the state. Captive mule deer, white-tailed deer, and pronghorn antelope were also brought in from out of state to supplement FWRF herds. The bighorn sheep herd was reduced in FY 2001/2002 and FY 2002/2003 through natural mortality and an out-going transfer of excess animals. The Fallow deer herd was allowed to expand naturally as per the study protocol in FY 2001/2002, while the cattle elk, and pronghorn herd levels remained relatively constant for the period. Commission approval was granted in 200 I to transfer excess FWRF captive wildlife, and/or orphaned neonates out of state to support collaborative and non-agency wildlife research projects. In 200 I the excess bighorn sheep were transferred to a research facility in Idaho, and in 2001, 2002, and 2003 orphaned mule deer neonates were transferred to a captive facility in Wyoming. It is important to note that the 2002 and 2003 out of state transfers were not of FWRF origin, but habituated orphaned fawns not suitable for release. Other facility transfers include several excess bighorn weanlings that went to a zoo for display, several pronghorn bucks that were borrowed from (and returned to), another captive wildlife research facility, and several free ranging bull elk brought in for breeding purposes. Breeding was planned annually to maintain optimal population sizes of the various species required to support current and future research projects. Depending on research objectives, some of the offspring from FWRF animals were hand-raised, and various species of wild orphaned neonates were accepted for hand rearing. Habituated weanlings and adult animals were also accepted whenever herd levels would allow. Hand rearing protocols for mule deer are described by Parker and Wong (1987), and by Wild and Miller { 1991) for bighorn sheep, elk, pronghorn antelope, and white-tailed deer. The male cattle, domestic ferrets and mountain lions were castrated at an early age, and the male fallow deer were vasectomized in the summer of 02/03 to prevent further breeding. Table 3 summarizes the breeding and rearing practices of ungulate species for the period: �148 Table 3 Ungulate breeding and rearing practices FWRF Breeding Species 2000 2000/2001 Bighorn Sheep Bred Elk Bred 5 Cows Fallow Deer Yearlings, did not breed Bred Mule Deer Pronghorn Antelope Bred White-tailed Deer Bred 3 yearlings 2001/2002 Bighorn sheep FWRF Breeding 2001 Bred Elk Did not breed Fallow Deer Mule Deer Bred Bred Pronghorn Antelope White-tailed Deer Bred 2002/2003 Bighorn Sheep Pronghorn Antelope White-tailed Deer Orphans 2001 Hand raised 2, dam raised 2, 1 stillborn No offspring 1 weanling Hand raised 4, dam raised others Hand raised 4 , 2 still born, others Euthanized as per study protocol Dam raised 0 0 0 0 13 orphans FWRF Neonate Rearine 2002 Hand raised 5, dam raised others No offspring Orphans 2002 0 3 orphans, 9 weanlings 1 weanling Bred Dam raised Hand raised 20, dam raised others Euthanized as per study protocol Dam raised FWRF Breeding 2002 Not bred FWRF Neonate Rearing 2003 No offspring Orphans 2003 Bred 3 cows 0 Not bred 1 hand raised, 2 dam raised No offspring Bred Dam raised 5 weanlings, Bred Hand raised 1 orphan Bred Dam raised 0 Elk Fallow Deer Mule Deer FWRF Neonate Rearine 2001 Dam raised I orphan I weanling 11 orphans, 2 weanlings 0 0 �149 Nutritional Maintenance: Feeding protocols for ungulates previously housed at the facility were reviewed by Wild ( 1997). The fallow deer were maintained on a high quality grass alfalfa mix hay and Regular Ranch-way deer and elk ration. The domestic ferrets were maintained on a commercial ferret chow, and the mountain lion kittens were initially maintained on Kitten Milk Replacer, Nurtural, and commercial kitten chow. The kittens were switched to a ground commercial feline diet at weaning, and were introduced to chunk deer and elk meat for training purposes at four months of age. A commercial carnivore supplement was added to the training meat to enhance dietary levels of calcium, and vitamins A and E, and was offered several times weekly. At five months of age, the kittens were gradually introduced to whole deer and elk carcasses and carcass portions with the GI tract removed, and are currently maintained on carcass portions, and training meat with supplement. Individuals of all species maintained reasonable body condition on available diets with the exception of some mule deer fawns, and CWD infected animals at the clinical stage of the disease. Fawn mortalities may have been associated with general poor body condition of does infected with chronic wasting disease, the presence of other etiological agents, and/or interspecies competition for space and cover in paddocks housing cattle and fallow deer. Pen Enrichment: In an effort to provide cover and subsequently reduce stress, the mule deer in the cattle pens were provided with a refuge area not accessible to the cattle, and artificial refuge areas were constructed in all paddocks housing semi-wild deer and dam raised neonates. Single piece and "L" shaped hide-outs, were constructed on site, and vegetation ex-closures were added in early spring and removed later to provide seasonal natural cover. Additionally, the Fort Collins Water Treatment Plant donated rock, labor and the use of equipment to construct two rock mountains in the bighorn sheep pens to enhance the natural structure in these areas. In addition to pen structure, behavioral enrichment was offered through training. The mountain lions were trained using operant conditioning; a form of training based on a reward system, and used widely in wildlife display facilities. Using this system, the lion kittens were taught to sit, platform, kennel, and stretch up on the fence for physical exams. Hand raised ungulate neonates were treat trained using the same philosophy, and were taught to follow their human trainers and stand on the scale for physical exams and weighing. Passive training was also used to habituate animals to the scale and alley-way by feeding the animals supplement in these areas, and allowing free exploration without human interference. Health Maintenance: Animal health care was provided as required and as mandated by the preventive medicine program (Wild 1995) and chronic wasting disease protocols. Overall, captive wildlife maintained at FWRF remained healthy throughout the period. Chronic wasting disease (CWD) continues to be a significant source of mortality in captive mule deer and white-tailed deer and is reflected by the number of animals dedicated to CWD research.projects throughout this period. Dystocia was a significant source of mortality in adult pronghorn does, and was associated with a failure of the cervix to dilate at the time of parturition. The underlying cause of the pronghorn dystocia is still unknown, and the collaborative USDA RB5 l brucellosis vaccine project was put on hold in FY 2002/2003 due to the resulting reduced number of adult females available. Other significant etiological agents included Epizootic hemorrhagic disease (EHD), bluetongue virus {BTV), and clostridium perfringens. �150 Standard Operating Procedures: Chronic wasting disease The CWD management protocol was again revised in FY 2002/2003 (Attachment 1). Generally, CWD continues to be managed as described by Wild (1997): to maintain CWD and maximize potential exposure for specific research objectives. The revised protocol was prepared to incorporate new information resulting from recent research findings: increasing bio-security, incorporating early detection techniques, and intensive herd management of CWD infected animals (Wild et. al 2002, Wolfe et. al 2002). All animals at FWRF were monitored closely for clinical signs of CWD, and tissues from all mortalities occurring at FWRF were examined for evidence of infection with CWD. Systems development In addition to the CWD protocol, all animal husbandry, facility security, FWRF management protocols, and veterinary supply inventories were reviewed and updated. Protocols were developed to manage the new species, and an Access database was developed to track additional information such as projects and veterinary treatments. The old paradox database and hard copies of vital records, necropsy, clinical pathology, and transfer information was integrated into the new database. Facility and animal maintenance costs were analyzed and incorporated into a draft fee schedule for use of research animals and FWRF facilities by professional collaborators, and a draft 5 year facility capitol construction plan was developed to address long term planning needs. Educational Contributions FWRF functions primarily to support wildlife research, however when possible and relevant, facility tours were provided to school, university, and professional groups. We emphasized the importance of maintaining captive wildlife to perform controlled experiments, and the contributions made by current and historic research projects conducted at FWRF. FWRF animals and facilities were also used occasionally for hands-on training of CDOW employees, collaborators, and other professional groups in sampling techniques and chemical immobilization. Research Projects: Facility operations offered support for research projects conducted by CDOW personnel and other collaborators that were initiated, conducted, or continued using FWRF animals and facilities. A total of twenty one research projects were supported by FWRF for the period: • • • • • • • • • • • Cattle susceptibility to chronic wasting disease. Mechanisms of CWD transmission in mule deer. Evaluation ofprospective preventative therapies for chronic wasting disease in mule deer. Validation ofa potential blood test for chronic wasting disease (GeneThera test). Prion peptide immunization and challenge. Molecular epidemiology of strain variations in chronic wasting disease. Susceptibility ofMountain Lions to chronic wasting disease. Susceptibility offallow deer to chronic wasting disease. Pathogenesis of chronic wasting disease in white-tailed deer. Effects ofGnRH-PAP on reproduction and behavior in female mule deer. Evaluation ofGnRH agonist (leuprolide) as a reversible contraceptive in mule deer. �151 • • • • • • • • • • Evaluation ofGnRH agonist (Lupron) as a potential contraceptive in rocky mountain elk: Effects on pregnancy. Development ofa remote delivery system for GnRH agonist (leuprolide) in female elk Paradoxical immunosuppression in bighorn lambs as a mechanism for depressed recruitment following pastuerellosis epidemics. Biosafety and reproductive effects ofRB51 (brucellosis) vaccine in pronghorn. Evaluation ofdrug delivery and dart trauma using collared and un-collared pneudart and daninject darts. Evaluation ofA3080 (thiafentanil oxalate) and naltrexone HCLfor the immobilization and reversal of mule deer. Evaluation ofA3080 (thiafentanil oxalate) and naltrexone HCLfor the immobilization and reversal ofpronghorn antelope. Effects of 2% DRC-1339 treated brown rice on non-target species. Testing alternative models ofherbivore foraging in heterogeneous environments. Field Immobilization Training. Facility Improvement Projects: A variety of scheduled and unscheduled maintenance and repair activities were necessary to support facility operation and ongoing research programs. Highlights include construction of the new Wildlife Health Lab (WHL) housing a laboratory, office space, a necropsy lab, and walk in freezer/cooler space, now located within the FWRF perimeter. This project was designed, constructed, and funded by the CDOW engineering and capitol construction team, while FWRF personnel provided support services. Additional facility modifications include twelve new paddocks, associated buildings, alleys and an access road to support the CWD transmission study, and an automatic water system for all paddocks on the east side of the facility. Other improvements included five new isolation pens, perimeter fence and gate upgrades, and construction of compost bins to hold animal waste material generated from CWD research paddocks. A new mountain lion facility including a concrete block building containing 4 indoor dens and a work space, a 50 x 60 foot outdoor pen, and shift containment system is currently under construction. A 2000 gallon vault was installed on the east side, a new pasture was also constructed on the west side, and the old house trailer was demolished. Additionally, the Soldier Canyon Filter Plant donated several culverts and constructed a detention pond on the west side of the facility to better manage natural water run-off and scheduled water releases from the plant. Facility maintenance and construction projects were prioritized based on animal welfare concerns and anticipated research needs. Table 3 summarizes the completed, current, and on-going facility construction maintenance projects for the period. �152 • Table 3. Facility Improvement Projects Details and Information Replace roofs on existing pens, and add 5 additional pens with shelters Remove garage door, replace with permanent wall, add window Automatic waters installed in all existing pens and new paddocks for transmission study. CSU provided the electric contractor, FWRF contracted out plumbing Purchase Tough Sheds, level sites, pour concrete pads for Transmission study, purchased 1 shed, other was supplied by WHL Construct 9 new pens, and split 2 pens into 4 for Transmission study, CSU provided contractor for installation Construct Feed-sheds for north and south transmission study pens Completion Year Construct feed shelters for transmission study pens Construct 400 feet of alley, 16 walkthru gates for transmission study 2001/2002 Completed Construct road, culvert donated by Fort Collins Water Treatment Plant 2001/2002 Completed Purchased all gates < 14 foot long, 14 'gates donated by CDOW game damage, installed gates, added horse fence, add gate opening in D3 contracted out electric fence modifications Construction of D 1 pasture contracted out by NWRC 2001/2002 Install automatic water in D 1, and a water shut off valve in the west hub, contracted out plumbing and excavation Construct l feed shelter, and l 2001/2002 Project Status 1. Improvements to west rearing area 2. East lab improvements 3. Add 13 automatic waters, 4 shut off valves to east side Completed 4. Add pellet feed storage shed, and feed-shed to east side 5. Construct 12 new pens on east side Completed 6. Construct 2 additional feed storage sheds on east side 7. Construct 13 feed shelters on east side 8.Construct alley system and gates for new pens on east side 9. Construct access road to new alley system on east side l 0. Replace all wooden drive through gates with 7 foot metal tubing gates, add gate to south side of D3 11. Construct new pasture on west side (Dl) 12. Plumbing upgrades to west hub area Completed 13. Construct 2 Completed Completed Completed Completed Completed Completed Completed Completed 2001/2002 2001/2002 2001/2002 2001/2002 2001/2002 2001/2002 2001/2002 2001/2002 2001/2002 �153 shelters in D 1 14.East fawn rearing area improvements Completed 15. House trailer demolition, and FWRF site clean-up 16. Construct ram pen exclosure around feed area in E3 17. Reconstruct shed on west side of west scale-room, modify scale 18. Water damage repair to El/E2 feedshed 19. Perimeter Fence upgrades Completed 20. Upgrade 2 perimeter and main east and west gates 21. Add Secondary perimeter gate and 8 foot fence on south side of facility Completed 22. Compost animal waste from CWD paddocks Initial start-up is completed, composting is ongoing Completed 23. Replace east side septic tank 24. Rock mountains constructed in upper sheep pens 25. Construct west detentionpond 26. Construct mountain lion holding facility Completed animal shelter Reconstruct roof structures, repair shelters, double fencing on N side, add 1 alley gate Demo old trailer, clean up, organize FWRF construction materials and supplies, remove waste Purchased range panels, installed panels, added horse fence 2001/2002 2001/2002 2002/2003 Completed Reconstruct shed, modify scale to accommodate access from west side 2002/2003 Completed Remove soil on west side, reconstruct wall, re-grade soil 2002/2003 Completed Replace rotten posts, add V-mesh to lower 4 feet of perimeter fence, contracted out labor on V-mesh Replace 4 old drive thru gates with 8 foot chain link gates 2002/2003 Close off FWRF access road between the Ft. Collins Water Treatment Plant and Soldier Canyon Filter Plant, contracted out time and materials Construct compost bins, purchase bacteria, train personnel to mix and monitor, Contracted out initial bin construction, and start-up Replaced rusted metal l 000 gallon tank with a 2000 gallon concrete vault, Contracted out time and materials Rocks, equipt, and time to construct the mountains donated by the Ft. Collins Water Treatment Plant Construct pond to maintain drainage water inside our perimeter fence, time and equipt. to construct pond was donated by the Soldier Canyon Filter Plant Utilities, concrete block building, 50 x 60 foot outdoor pen, shift containment system, and 4 indoor dens, building slab, and alley concrete, concrete block building, 2002/2003 Completed Completed Completed Current project: planning, utilities, and building construction completed. finish: 2002/2003 2002/2003 2002/2003 2002/2003 2002/2003 Project began 2001/2002, scheduled for completion 2004/2005 �154 outdoor pen, shift containment, indoor dens Current project: 5 completed, finish: 7 plumbing, electrical, and engineering contracted out 28. New roofs/repair structure on old feed-sheds and animal shelters. On-going project 29. Add additional animal shelters On-going project 30. Road Maintenance 3 1. Paint old building exteriors On-going project Approx. ¼ of the old structures and roofs on the facility have been replaced in the last 2 years using treated lumber and long lasting roofing materials Construct additional shelters in pens with heavy stocking rates. (36 unroilate pens on the facility) Road grading and upkeep 32. Repair/replace latches, and broken or water damaged alley-way boards 33. Replace walk thru alley gates 34. Replace old visual barrier fencing and utility wire on metal gates On-going project 27. Reconstruct west isolation pens 35. Animal holding fence upgrades, and repairs 36. Construct artificial refuge areas inside pens for neonates and adults On-going project Demolish old pens and shelters, reconstruct with upgraded design and materials Now using CCA treated lumber, or metal siding for repairs & building replacements to reduce the amount of painting necessary in the future. Now using CCA treated lumber for all repairs Project began 2001/2002, scheduled for completion 2004/2005 Began 2000/2001, as needed Began 200 l/2002, as needed As needed Old structures are on a painting schedule every 3-5 years As needed On-going project Replace old gates as necessary On-going project: most of the old material has been replaced, but this project is on-going due to animal and environmental ~e On-going project: rotten posts have been replaced all over the facility, and many double fences have been constructed to comply with CWD protocols On-going project: completed for all new east side paddocks, maintain existing, construct Old snow fence and construction fence replaced and moved to the outside of the paddock fence (except interior fences), utility wire is systematically being replaced with horse-fence Began 200 l/2002, as needed Replace old range fence and Vmesh, as well as electric fencing in pens that house deer, Construct double fences as required by CWD protocols Began 2002/2003, as needed Construct single and L-shaped, refuge areas to provide refuge and shade, construct hog panel seasonal exclosures to promote vegetation growth in the spring Began 2002/2003, as needed As needed �155 37. Add windscreen to west and south facing fence-lines 38. Mowing and weed control 39.WHL maintenance 40. Unscheduled miscellaneous emergency facility repairs new On-going project On-going project On-going project On-going project Provide additional shaded areas for animals, and maintain existing Began 2002/2003, as needed Seasonal mowing and manual, chemical noxious weed control Provide maintenance assistance to WHL, and support for initial lab construction Emergency repairs to structures, animal holding facilities, perimeter fence, automatic waters, utilities, etc ... As needed Began 2002/2003, as needed As Needed �156 Addendum 1. PROTOCOL FOR MANAGING CHRONIC WASTING DISEASE AT FOOTHILLS WILDLIFE RESEARCH FACILITY Draft Rev. 2003 HISTORY Chronic wasting disease (CWD) is a transmissible spongiform encephalopathy (TSE) or prion disease of cervids (deer and elk). Other TSE's include scrapie of sheep, bovine spongiform encephalopathy (BSE), and Crutzfeld-Jacob disease of humans. The disease causes behavioral changes and loss of body condition and is invariably fatal to infected deer and elk. Despite a comprehensive program initiated in 1985 to eradicate CWD from cervids and the environment at Foothills Wildlife Research Facility (FWRF), CWD remains endemic at the facility. After the 1985 clean-up, CWD was first diagnosed in elk in 1989 and in mule deer in 1994. Natural transmission is now common in mule deer at FWRF and sporadic cases continue to occur in elk. Additionally, natural transmission rates are markedly higher and self-sustaining in paddocks housing infected animals being used in ongoing CWD research studies compared to paddock areas housing animals for other research studies. Based on these observations, guidelines established in 1985 (and revised in 1993 and again in 1997) for maintaining a CWD-free facility are largely obsolete. Here, we provide additional revisions to those guidelines that are directed at maintaining the disease for research purposes in captive deer and elk while minimizing the risk to personnel and the potential spread of CWD outside the facility. OBJECTIVES 1. 2. 3. 4. 5. 6. Prevent transmission or exposure of CWD from FWRF to animals or facilities outside FWRF. Minimize potential for exposing FWRF personnel and visitors to pathogens or potential pathogens including CWD. Maintain endemic CWD in deer at FWRF; however, animals showing end stage clinical signs of CWD will be euthanized to avoid undue suffering, unless directed otherwise by research protocol. Minimize potential spread of CWD among species of captive wildlife (deer, elk and noncervid research animals). Minimize cross contamination between CWD infected and non-targeted research animals. Prevent cross contamination between CWD research treatment groups. ASSUMPTIONS 1. 2. 3. 4. CWD is an infectious disease of deer and elk caused by an abnormally shaped protein prion. CWD is not widespread in free-ranging cervids. Where it occurs, the prevalence of disease varies greatly. Mode of transmission for CWD is not known, and may be direct, via animal/animal contact, or indirect, through contact with excreta (saliva, urine, feces); animate and inanimate objects may serve as fomites (vehicles) in transmitting CWD. Non-cervid wildlife and domestic species are not naturally susceptible to CWD. It is possible that non-cervids could be inapparent carriers of CWD; however, no data have been produced to support this possiblity. Based on patterns seen in other TSE's, it seems likely that if CWD is transmitted to a new host species, then the likelihood of further transmission to others within that species is increased. �157 5. There is no evidence that CWD is transmissible to humans; however, it is prudent to minimize human exposure to CWD as well as animal pathogens known to be transmissible to humans (e.g. Salmonella spp., E. Coli, etc.). APPROACH Overview: 1. Follow established guidelines that prevent contact of captive research animals with animals outside FWRF (wild and domestic). 2. Minimize potential spread of infectious material outside FWRF perimeter. 3. Minimize potential transmission of CWD between species of captive animals, between CWD and non-CWD research animals, between research projects, and between experimental treatment groups where necessary. This includes transmission from mule deer/cattle pens, mule deer/fallow deer pens, therapy mule deer pens, white-tailed deer, and mountain lion pens via contaminated materials or potentially contaminated, equipment, or clothing. 4. Maintain each species of animal in isolation from others, unless directed by research protocol (e.g., mule deer with cattle, mule deer with fallow deer). 5. Educate animal caretakers about CWD (hazards, protocols, and clinical signs exhibited by affected animals). Perform daily animal observations and maintain detailed records of animal health as a portion of the FWRF CWD surveillance program. Animals: 1. 2. 3. 4. 5. Exclude wild or captive cervids from CWD established areas from entering the captive herd, unless directed by a research protocol. Established areas will now include: northeastern, northcentral, and northwestern Colorado, Park, Albany, and surrounding counties in Wyoming, and the Denver Zoo. However established areas are dynamic and may change as surveillance for CWD increases. Therefore, please consult the latest CWD update for guidance. Depending on intended use, orphans, and neonates raised outside FWRF, may be accepted from areas that are CWD established, as well as areas that are not CWD established. These animals will be maintained separately to minimize potential CWD transmission to uninfected neonates that come from sources outside the established area. Raise and maintain each animal species in isolation from others, unless directed by a research protocol. To prevent transmission of CWD from FWRF to facilities where CWD is not established, noncervid species from FWRF will be transferred or donated to other facilities only if the following criteria are met: 1) the transfer location is within the CWD established area, 2) animals are scheduled for a specific research project, 3) the destination is a closed facility (no egress of live animals), 4) animals will not be used in "tame animal trials" in non-confined environments. 5) transfer is approved by the mammal's research leader, 6) recipients will be notified of CWD risks associated with accepting animals from FWRF. Transfers of live cervids from FWRF are prohibited. Animal Maintenance: l . House and maintain each species in isolation from other species, unless directed by a research protocol. 2. House and maintain CWD research animals separate from non-CWD research animals. 3. Maintain accurate records for all animals. This information includes (but is not limited to): birth date, origin, body weights on tractable animals, vaccinations, health problems and treatments, research projects, and movements (intra and inter facility). Additionally, a. Tag all animals for easy individual identification. �158 4. 5. 6. b. Train FWRF personnel to recognize clinical signs of CWD. FWRF personnel will maintain daily animal observation records describing animal status and will report abnormal observations to the facility manager. Where feasible, weigh and/or briefly examine every animal at least once monthly. Wild research animals usually cannot be handled for weighing; these will be visually examined and immobilized via a dart injection for closer examination if necessary. Follow a preventative medicine program that includes routine vaccination, anthelmintic treatment, hoof trimming, nutritional evaluation, and other measures to optimize overall health of research animals. Initiate early detection measures by conducting annual tonsil biopsies on all deer (WTD, MD) housed within the facility. CWD positive animals will be removed at the discretion of the lead project researcher from non-CWD research paddocks and 1) added to the CWD research herd, 2) held in isolation, or 3)humanely euthanized. Use of Research Animals Outside FWRF: l. The transport of non-cervid species from FWRF to facilities or locations, outside the CWD established area is prohibited. 2. The transport of non-cervid species to facilities or locations outside FWRF but within areas where CWD occurs is prohibited unless expressly approved by the mammal's research leader. 3. The transport of cervids outside FWRF is prohibited. 4. Procedures for isolating cervids at other CDOW facilities will be the same as those at FWRF. 5. Animals of any species maintained at FWRF will not be released into the wild. 6. The FWRF Manager is responsible for maintaining accurate records of animals transferred into and out ofFWRF. General Facilities and Equipment: 1. Exclude free-ranging wildlife and livestock from the facility or from contact with captive animals using interior and perimeter fencing. A minimum 4 foot corridor must be maintained between interior pasture fencing and the 8 foot tall perimeter fence surrounding FWRF. The perimeter gates will remain closed at all times, the perimeter fence is inspected monthly, and necessary repairs are made top priority for facility maintenance. 2. Maintain each species of animal separately and allow no direct or fence-line contact unless directed by a research protocol. 3. Minimize runoff between pens housing different species through appropriate pen assignment and drainage control, unless directed by a research protocol. 4. Use drainage control to minimize runoff outside the facility in areas where natural and/or man made drainages occur inside CWD paddocks. 5. Minimize common use of equipment between pens housing different species, between CWD and non-CWD paddock areas, and between CWD treatment groups. When it is necessary to use the same equipment (vehicles) a 20 % chlorine, or 5 % LPH solution can be used to disinfect equipment immediately following the use of equipment inside CWD infected paddock areas. 6. All equipment, materials, organic, inorganic, materials that have been exposed to CWD pathogens must either remain on site or follow EPA treatment guidelines prior to leaving FWRF. 7. Feed and handle animals or clean pens using the following traffic pattern: Clean CWD controls (MD,WTD), non-cervids, elk, non-CWD research mule deer, CWD research mule deer, CWD infected white-tailed deer, mule deer with cattle/fallow deer. Additionally, follow specific protocols for traffic patterns between various CWD research treatment groups. 8. Clean animal pens (especially feed areas and waters) weekly. Dispose of waste from pens housing non-CWD research animals, and clean controls in the main dumpster. Waste from all �159 CWD infected paddocks must never leave the facility. 9. Fecal material and non-palatable feed from CWD research paddocks will be reduced through on-site composting and palatable feed will be recycled to the cattle. 10. Isolation pens, digestion cages, and other areas where animals are held for extended periods, will be cleaned of organic matter and disinfected with a 20% chlorine solution, or 5% LPH solution after use. The researcher last using the area will be responsible for cleanup. Cooperative compliance will be made a condition of all study plans using FWRF ungulates and facilities. 11. Different species may be held concurrently in isolation pens if a buffer zone (empty pen) is used. Feed: 1. Hay will not be accepted from areas where domestic sheep have grazed on cultivated pastures. Personnel: 1. Wash hands before and after handling each species of animal, before and after handling nonCWD and CWD research animals. 2. No eating or drinking allowed in animal areas. 3. Dedicate one pair of shoes/boots to FWRF. Change into/out of this pair of shoes when you arrive at work/when you leave. Alternately, shoes can be sprayed liberally, and/or washed thoroughly in 20% chlorine or 5% LPH solution. 4. Coveralls, boots, and gloves, are required when handling animals showing clinical signs of CWD; and face masks and eye protection are available for use if desired. 5. Coveralls and/or boots are a protocol requirement for CWD infected areas, and CWD control groups. Additionally, each set of treatment groups within a research project may require a separate set of boots and/or coveralls depending on research objectives. Please do not enter a paddock unless you know the protocol. 6. Unsupervised access to FWRF will be limited to authorized personnel. Unauthorized persons will not enter animal pens or be permitted direct contact with research animals. The facility will be locked except when attended during normal business hours. 7. Visitors will be informed that FWRF houses CWD infected animals and is within the CWD established area, and will be given the option of wearing rubber overshoes which will remain on site. 8. All researchers and collaborators and their subordinates will comply with this protocol. All personnel working at FWRF will be required to read this protocol and other appropriate literature and to sign the attached sheet of informed consent. Additional Requirements for CWD research Pens: 1. Protective clothing such as designated boots/shoe covers and/or coveralls and must be worn when entering all pens housing CWD infected animals (currently these are: mule deer/cattle pens, mule deer/fallow deer pens, mule deer therapy pens, infected WTD pens, and mountain lion pens), as well as all CWD control pens. 2. Place waste feed and manure from infected mule deer and white tailed deer pens in the storage compost pile at FWRF (NOT in the dumpster, or working compost piles). Compost will be mixed appropriately and put into composting bins by assigned personnel. Finished compost will be incinerated, or used for topsoil in CWD infected paddocks as needed. 3. Waste feed from the mountain lion pens is disposed of through incineration or sent to CSU for chemical digestion. Fecal material from the lion pens is composted along with other CWD pen waste material. 4. Dedicated (separate) equipment (wheelbarrows, rakes, shovels, water brushes, bucket scrapers, etc.) must be used for cleaning CWD infected vs. CWD control and non-CWD research �160 paddocks. Additionally, separate cleaning equipment may be required for each treatment group within specific research projects. Please ask the facility manager if you are not sure of the cleaning protocol. 5. Vehicles must be cleaned after use in all CWD paddocks. Wash organic material from tires, remove all organic material from the truck bed and disinfect with a 20 % chlorine, or 5% LPH solution. 6. Clean-up procedures following depopulation of a CWD infected paddock: disinfect feed bunks and feed pans in 5% LPH solution and rinse thoroughly, disinfect water receptacle with a 20% bleach solution and rinse thoroughly, rake out all fecal material, spray feed shelter and soil under and around shelter with a 50% bleach solution. Allow all to dry thoroughly before repopulation of paddock. Additional clean-up procedures may be required such as removing the top 6 inches soil around a feed area, soaking with bleach solution, and adding road-base. This will depend on the specific research project. 7. Keep gates to pens, hub/working area, and main east and west gates closed at all times except when passing through. 8. Animal carcasses must be enclosed with a protective cover to contain potentially infectious materials during transportation to the Wildlife Health Lab (WHL) on site, or off site to the CSU Vet Teaching Hospital (VTH) or the Wyoming State Veterinary Lab (WSVL). Alternatively, the truck/equipment could be cleaned with a 20% chlorine solution after use if transported to the necropsy lab on site. 9. Cattle will not leave the facility alive unless transferred to a biosecurity level 2 or greater facility and this requirement is part of a written change to the established research protocol. I 0. Report any abnormalities or accidents immediately to facility supervisor. CWD SURVEILLANCE PROGRAM 1. 2. 3. 4. 5. Euthanize any animal showing clinical signs of CWD and examine tissues grossly and histologically. Perform complete postmortem examination and histologically examine brain tissue of any animal that dies at FWRF. Carcass disposition will be by incineration (required for cattle), chemical digestion, or appropriate burial at the Larimer County Landfill. If CWD is diagnosed in any noncervid species at FWRF, this protocol will be immediately revised and biosecurity at FWRF further increased. The attending veterinarian, facility manager, and Research Facility Animal Care Committee (RFAC) will evaluate and amend this program as necessary. The FWRF CWD PROTOCOL WAS FIRST ESTABLISHED IN 1985 AND REVISED: 1993 1997 2003 INFORMED CONSENT I, - - - - - - - - - - ~ have read the Foothills Wildlife Research Facility (FWRF) protocol concerning chronic wasting disease (CWD) and agree to follow the protocol. Although there is no evidence that CWD is transmissible to humans, I realize that I will be working with research animals and in an environment potentially infected with CWD. I understand that this protocol reflects current knowledge on measures for minimizing exposure to and spread of CWD and other potential pathogens at FWRF. Signature Date �161 LITERATURE CITED Parker, K. L., and B. Wong. 1987. Raising black-tailed deer fawns at natural growth rates. Can. J. Zool. 65:20-23. Wild, M.A., and M. W. Miller 1991. Bottle raising wild ruminants in captivity. Colorado Div. Wild!. Outdoor Facts No. 114. Wild, M.A. 1995. Animal and pen support facilities for mammals research. Colorado Div. Wildl. Res. Rep., WP la, Jl, Jul 1994 - Jun 1995, Fort Collins. Wild, M.A. 1997. Animal and pen support facilities for mammals research. Colorado Div. Wild!. Res. Rep., WP la, Jl, Jul 1996 - Jun 1997, Fort Collins. Wild, M. A., T. R. Spraker, C. J. Sigurdson, K. I. O'Rourke, and M. W. Miller. 2002. Preclinical diagnosis of chronic wasting disease in captive mule deer (Odocoileus hemionus) and white-tailed deer (Odocoileus virgineanus) using tonsillar biopsy. J. General Virol. 83:2629-2634. Wolfe, L. L., M. M. Conner, T. H. Baker, V. J. Dreitz, K. P. Burnham, E. S. Williams, N. T. Hobbs, and M. W. Miller. 2002. Evaluation of antemortem sampling to estimate chronic wasting disease prevalence in free-ranging mule deer. J. Wild!. Manage. 66:564-573 �Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Colorado Project No. Work Package No. 3740 Task No. 3 : : : : Cost Center 3430 Mammals Research Mammals Support Services Animal and Pen Support Facilities for Mammals Research Period Covered: July 1, 2003 – June 30, 2004. Author: T.R. Davis Personnel: M. Anderson, K. Beamer, T. Bogardus, E. Crawford, E. Donegan, M. Dupire, K. Fagerstone, J. Faue, E. Featherman, K. Fox, T. Halasinski, L. Ho, M. Hanusack, G. Harvey, E. Jones, K. Kanapeckas, J. Kint, G. Kyriacou, I. Levan, T. McCollum, A. Mitchell, A. Northrup, M. Paulek, A. Phillips, R. Rhyan, T. Sanders, J. Sirochman, T. Sirochman, J. Stout, T. Stout, D. Thompson, R. Thompson, D. Weaver, A. Wilson All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT The Colorado Division of Wildlife's Foothills Wildlife Research Facility (FWRF) maintained captive animals (2003/2004 annual total: 360) and facilities in support of seventeen captive wildlife research projects. The primary focus of research during this period was chronic wasting disease (CWD) pathology, epidemiology, preventative therapies, sources of transmission in deer and potential transmission to other species. FWRF supported a number of other significant research projects including contraception and reproductive effects, pathogen immunization, evaluation of wildlife capture pharmaceuticals and personnel training in ante mortem sampling and field immobilization. The quality of animal care and facility maintenance provided by temporary, work-study, personal service, intern and volunteer employees is in part reflected by the finding of compliance under the Animal Welfare Act during the annual USDA inspection of FWRF. Herd management practices allowed pronghorn antelope and bighorn sheep herd levels to decline through natural mortality and the remaining domestic ferrets were removed as per study protocols. Alternatively, herd levels of mule deer and white-tailed deer were managed for maximum growth to support on going CWD research. Chronic wasting disease was again a significant source of mortality in mule deer and white-tailed deer and is reflected by the number of CWD research projects conducted during this period. We continue to manage CWD with the philosophy of managing the disease for research purposes under heightened bio-safety guidelines and intensive herd management. Fawn mortalities were higher than expected during the 2004 rearing season however a number of disease causing agents and contributing factors were identified through various diagnostic tests and evaluations. Neonate training was intensified for hand raised animals and training SOP’s developed to accommodate CWD epidemiology research. Administrative actions include compiling a summary of all current and historic FWRF published research, an animal husbandry change order request was 139 �implemented, and the FWRF tour policy revised. New SOP’s were implemented for routine equipment maintenance, seasonal winterizing, tree/shrub care, and the construction maintenance work request forms were revised and reinstated. In addition to routine maintenance, the FWRF team made significant facility improvements including new facilities to accommodate CWD epidemiology research, completion of the mountain lion holding facility, and installation of a perimeter fence around the Wildlife Health Lab. The capitol construction team allocated funds for a new hay barn which is scheduled for replacement in the summer of 2005, and engineering assisted in the development of electronic facility site maps. In addition, the FWRF landowner; Colorado State University approved an easement for the Northern Colorado Water Conservation District to install a 68 inch water pipeline through the center of the facility (north to south). The installation process partially disrupted FWRF management activities for a six week period, but resulted in an upgraded road system, and replacement of existing windrows with five gallon potted trees and shrubs. 140 �JOB PROGRESS REPORT ANIMAL AND PEN SUPPORT FACILITIES FOR MAMMALS RESEARCH T.R. Davis Animal Maintenance: Routine animal husbandry including feeding, health observations, training, weighing, and cleanup, was performed primarily by well trained temporary employees, work-study students, and volunteers. FWRF was inspected by USDA APHIS for compliance with federal animal welfare regulations on July 28, 2004. Table 1 summarizes the number of animals by species reported to USDA animal welfare for the period of October 1, 2003 – September 30, 2004. Table 1. Total number of animals by species reported to USDA Animal Welfare Species Bighorn Sheep Number of animals held, not dedicated to research 11 16 Number of animals dedicated to research 2003/2004 Total 16 15 27 31 0 28 28 75 84 159 19 0 19 1 0 1 45 25 70 0 11 11 167 179 346 201 0 10 10 1 0 1 0 3 3 168 192 360 Elk Fallow Deer Mule Deer Pronghorn Antelope Sika Deer White-tailed Deer Cattle Ungulate Total Ungulate Mean Domestic Ferrets Prairie Dog Mountain Lions Facility Total 141 �The number of animals held but not dedicated to research includes all animals being bred, conditioned, or held for use in research, but not yet used for such purposes. This group consists primarily of breeding animals, and young animals. The relatively high number of ungulates not dedicated to research during this period is the result of a large influx of young animals (born at FWRF, and orphaned neonates) dedicated to the CWD epidemiology study scheduled to begin data collections in the fall of 2004. The total number of animals dedicated to research includes all animals used in experiments at any time during the period. Experiments include those involving no pain, distress, or use of pain relieving drugs, and experiments where pain relieving drugs were necessary to minimize stress on the animal. No animals at FWRF were used in experiments involving pain, without the use of anesthetic, analgesic or tranquilizing drugs. The species total includes all adult animals housed at the facility, neonates born at the facility, transfers into and out of the facility, and all animals that died or were humanely euthanized during the respective fiscal year. It is important to note that ungulate herd levels at any one time averaged approximately 60 percent of the ungulate total and 55% percent of the total number of animals housed at the facility for the entire period. Herd Management: One habituated sika deer and one habituated prairie dog were brought into the facility to support division law enforcement efforts. Mule deer, white-tailed deer, and elk herd levels were expanded through herd management practices and incoming transfers to support CWD and fertility control research. Incoming transfers consisted primarily of habituated adult animals and orphaned neonates obtained from various locations around the state, as well as Wyoming, Nebraska, Iowa, and Kansas. The bighorn sheep and pronghorn antelope herds were reduced through natural mortality and out going transfers as the experiments these animals were dedicated to, reached a stage of completion. Eight fallow deer and the remaining domestic ferrets were removed through planned euthanasia as per the study protocols, while mountain lion and cattle numbers remained constant for the period. Commission approval was granted in 2001 to transfer excess FWRF captive wildlife, and/or orphaned neonates out of state to support collaborative and non-agency wildlife research projects. In 2004 eight pronghorn antelope neonates were transferred to the National Wildlife Research Center (NWRC) in Fort Collins. Two of these animals were orphaned neonates, and six were excess animals of FWRF origin. Other facility transfers include a pronghorn buck that was borrowed from, and returned to, the Sybille Wildlife Research Unit in Wyoming. FWRF herd management practices include planned breeding to maintain optimal population sizes of the various species required to support current and future research projects. Depending on research objectives, some of the offspring from FWRF animals are hand-raised, and various species of wild orphaned neonates are accepted for hand rearing. Habituated weanlings and adult animals are also accepted whenever herd levels will allow. Hand rearing protocols for mule deer are described by Parker and Wong (1987), and by Wild and Miller (1991) for bighorn sheep, elk, pronghorn antelope, and whitetailed deer. Table 3 summarizes the breeding and rearing practices of ungulate species for the period: 142 �Table 3. FWRF Ungulate breeding and rearing practices FWRF Breeding FWRF Neonate Orphan/Transferred 2003 Rearing 2004 Neonates 2004 Hand raised 2 Dam Bighorn Sheep Bred 7 Ewes raised 3 0 Species 2003/2004 Elk Bred 3 Cows Mule Deer Pronghorn Antelope White-tailed Deer Bred 17 Does Bred 4 does Bred 9 does Dam raised 3 Hand raised 16 Dam raised 11 Transferred to NWRC 6 Hand raised 6 Dam raised 7 1 32 0 24 Table 3 does not include three orphan mule deer fawns euthanized on arrival due to severe injuries, or very poor body condition, and five animals (3 mule deer, 2 white-tailed deer) which were still born or died shortly after birth due to parturition complications. The mountain lions, domestic ferrets, and fallow deer are also not included in the above table, as the mountain lions and domestic ferrets were neutered at an early age and the male fallow deer were vasectomies prior to the 2003 breeding season and therefore no breeding occurred in these species. Nutritional Maintenance: Feeding protocols for ungulates previously housed at the facility were reviewed by Wild (1997), and feeding protocols for the fallow deer and mountain lions were described by Davis (2003). The sika deer was maintained on a high quality grass alfalfa mix hay and Regular Ranch-way deer and elk ration. The prairie dog was maintained on Mazuri ADF # 25 herbivore diet, grass/alfalfa mix hay, and fresh vegetables. Individuals of all species maintained reasonable body condition on available diets with the exception of some hand raised neonates (primarily mule deer fawns), and CWD infected animals at the clinical stage of the disease. Fawn mortalities may have been associated with general poor body condition of does infected with chronic wasting disease, the presence of other etiological agents identified (see health maintenance below), and/or interspecies competition for space and cover in paddocks housing cattle and fallow deer. Pen Enrichment: In an effort to provide cover and subsequently reduce stress, additional artificial refuge areas were constructed in paddocks housing semi-wild deer and dam raised neonates. “Y” shaped hide-outs, were constructed on site, vegetation ex-closures were added in early spring and removed later to enhance natural cover, and creep areas with natural cover were provided for dam raised fawns. All pen structure enrichments were readily accepted and utilized by the animals. In addition to pen structure, behavioral enrichment was offered through training. Expanding on the operant conditioning system for mountain lions described by Davis (2003) hand raised ungulate neonates were "treat" trained using the same philosophy. Bighorn sheep, mule deer and white-tailed deer were taught to follow their human trainers and stand on the scale for physical exams, injections, treatments and weighing. Additionally, mule deer and white-tailed deer were gradually conditioned to the metabolic cages in preparation for CWD epidemiology sample collections. Passive training was used in conjunction with the above techniques to habituate animals to the scale and alley-way through supplemental feeding to encourage free exploration without human interference, in these areas. 143 �Health Maintenance: Animal health care was provided as required and as mandated by the preventive medicine program (Wild 1995) and chronic wasting disease protocols. Overall, captive wildlife maintained at FWRF remained healthy throughout the period. Chronic wasting disease (CWD) continues to be a significant source of mortality in captive mule deer and white-tailed deer and is reflected by the number of animals dedicated to CWD research projects throughout this period. Mortality of an adult pronghorn doe was attributed to dystocia, and as described in previous years (Davis 2003) was associated with a failure of the cervix to dilate at the time of parturition. Several cases of dystocia with variable presentations were also observed in mule deer (n=2) and white-tailed deer (n=1). Epizootic hemorrhagic disease (EHD) and bluetongue virus (BTV) were not significant etiological agents during this period and may be associated with a management effort to reduce the quantity of free standing water on the facility, coinciding with the time of documented seasonal peaks of the disease. Mortality rates and disease were higher than expected in hand raised mule deer fawns. Hand raised fawn mortalities were primarily associated with two types of illness: 1.) Intestinal disease resulting in diarrhea, bloating, and/or dehydration accompanied by a general lack of appetite and failure to thrive, and 2.) Respiratory disease (acute bacterial pneumonia) resulting in nasal discharge, coughing, labored breathing, and in some cases, no preliminary signs and acute death. Post mortem sampling and fecal isolation, revealed clostridium perfringens, salmonella, Escherichia coli, and rotovirus. Nasal cultures and post mortem sampling of lung tissue revealed mixed bacterial infections including, Alcaligenes species, Pasteurella species, Pseudomonas aeruginosa, however Arcanobacterium pyogenes was consistently diagnosed and is likely responsible for those cases resulting in acute death. In addition to the etiological agents identified, several management and natural conditions may have contributed to fawn mortalities: 1) Inadequate hospital facilities, and clean isolation areas to separate sick animals (facility carrying capacity), 2) Higher than normal precipitation levels contributing to viable pathogens surviving in the soil for longer periods, and greater exposure to damp/cool conditions, 3) immuno-compromised animals to start with, as FWRF born fawns are exposed to a very pathogen rich environment at birth, and, a high percentage of the hand raised animals were orphans who are often in poor body condition and/or ill when they arrive. Due to animal welfare concerns, management recommends construction of adequate animal holding and hospital facilities prior to hand raising mule deer in the future, as well as a review of the neonate nutrition, health maintenance, and fawn rearing programs. Chronic Wasting Disease: Following the recent revision of the CWD protocol (Davis 2003), we continue to manage CWD with the philosophy of managing the disease for research purposes under heightened bio-safety guidelines and intensive herd management. Intensive herd management is accomplished using the early detection techniques described by Wild et. al (2002) and Wolfe et. al (2002). All animals at FWRF were monitored closely for clinical signs of CWD, and tissues from all mortalities occurring at FWRF were examined for evidence of infection with CWD. Systems Development: Administrative actions include compiling a summary of published articles generated from FWRF research. Hard copies of the 100 + articles filed by date of publication are available at FWRF and the research library. Currently, we are compiling an Access database of the articles to facilitate searches by author, subject, date, etc. In addition, an animal husbandry change order request was implemented as suggested by the Mammals research leader. The change order, modeled after the construction/maintenance work request, was designed to track the origin and justification of facility changes in herd stocking levels, species needs, and basic husbandry techniques including animal care, breeding, rearing, and training practices. 144 �Other administrative actions include the development of new standard operating procedures for routine equipment maintenance, seasonal winterizing, and tree/shrub care. The SOP’s are designed to put all FWRF equipment on a routine maintenance and winterizing schedule, and the schedule is specific to what level of maintenance is necessary at each interval. In the same fashion an SOP was developed for soil moisture testing and watering of tree and shrub windrows. Due to the increasing demands for unscheduled (but necessary and often emergency) construction and maintenance needs, the work request forms were revised and reinstated. The forms were designed to assist in prioritizing and assigning tasks, as well as provide a format for information transfer (an accurate description of the need), and to track labor costs associated with specific projects, routine and emergency maintenance. Educational Contributions: The FWRF tour policy was also revised. The revised policy allows for use of FWRF animals and facilities for hands on training of CDOW employees, collaborators, and other professional groups in sampling techniques and chemical immobilization when pre-approved by the Mammals research leader and/or the Animal Care and Use Committee (ACUC). FWRF functions primarily to support wildlife research, but will no longer function secondarily as an educational facility due to the overwhelming demand for this service. Protecting the integrity of the research, facility management, and increasing animal welfare concerns were sufficient justifications for the policy change. Research Projects: Facility operations offered support for research projects conducted by CDOW personnel and other collaborators that were initiated, conducted, or continued using FWRF animals and facilities. A total of twenty one research projects were supported by FWRF for the period: • • • • • • • • • • • • • • • • • Cattle susceptibility to chronic wasting disease. Susceptibility of fallow deer to chronic wasting disease. Susceptibility of Mountain Lions to chronic wasting disease. Mechanisms of CWD transmission in mule deer. Evaluation of prospective preventative therapies for chronic wasting disease in mule deer. Validation of a potential blood test for chronic wasting disease (GeneThera test). Molecular epidemiology of strain variations in chronic wasting disease. Pathogenesis of chronic wasting disease in white-tailed deer. Effect of copper pathogenesis of CWD in white-tailed deer. Epidemiology of chronic wasting disease: detection of PrPres, shedding, and environmental contamination. Evaluation of third eyelid biopsy for detection of chronic wasting disease infection in mule deer. Survey for chronic wasting disease in cottontail rabbit populations. Leuprolide as a contraceptive agent in female elk: determination of effective minimum dose. Evaluation of GnRH-PAP as a chemosterilant in captive mule deer. I. Effects on animal health. Experimental evaluation of a vaccine for clostridium perfringens type A in captive bighorn sheep (Ovis canadensis) and captive mule deer (Odocoileus hemionus). Training personnel for tonsil biopsy for chronic wasting disease in mule deer. Field Immobilization Training. Facility Improvement Projects: A variety of scheduled and unscheduled maintenance and repair activities were necessary to support facility operation and ongoing research programs. Highlights include construction of new animal holding facilities and restoration of the metabolic cages to accommodate CWD epidemiology research. The mountain lion holding facility was also completed with the exception of the scale and squeeze/treatment area, as we are still in the design phase on this portion of the project. Additional 145 �funding assistance from the Wildlife Health Laboratory (WHL) permitted installation of a separate perimeter fence around WHL. The new fence will allow easier access to the laboratory, while controlling traffic into and out of the animal holding facility. Additional facility modifications include allocation of funding from the capitol construction team to replace the main hay barn. The barn is scheduled for replacement in the summer of 2005, and will be relocated to a central site with better access. The engineering office assisted with the development of electronic site maps of the facility. The site maps show exact locations for buildings and animal holding facilities, as well as locations for all known utilities. The maps will be updated periodically as facility construction and modifications occur. In addition, the FWRF landowner; Colorado State University approved an easement for the Northern Colorado Water Conservation District (NCWCD) to install a 68 inch water pipeline through the center of the facility (north to south). The easement was approved by the CDOW legal staff and included stipulations to maintain the perimeter fence, vehicle access, and keep all excavated soil within the perimeter of FWRF. The installation process partially disrupted FWRF management activities for a six week period, but resulted in an upgraded road system, and replacement of existing windrows with 200 five gallon potted trees and shrubs. NCWCD donated three culverts, built up the road system with excavated dirt, and added road-base which will allow for better water run-off, and should reduce road maintenance costs in the future. Facility maintenance and construction projects were prioritized based on animal welfare concerns and anticipated research needs. Table 3 summarizes the completed, current, and on-going facility construction maintenance projects for the period. 146 �Table 3. Facility Improvement Projects Project 1. CWD Therapy Pens 2. Travel Installation Status Completed Trailer Completed 3. WHL and FWRF Completed Parking Area Improvements 4. D3-6 Feed Area Completed Exclosures /catch areas 5. Electrical Upgrades 6. New Tool Equipment Shed Completed and Completed 7. E4 Site Clean-up Completed 8. Lion Facility Walk-in Completed, Freezer /Cooler Engineering request approved 9. Road Improvements Completed 10. FWRF Electronic Site Maps Completed 11. WHL Water Shut- Completed off Valve 12. Office Septic Pipe Completed Repair 13. WHL Perimeter Completed Fence Details Split 2 pens into 4, Construct 3 new shelters, add 1 automatic water (2 others included in east side plumbing upgrades above) Prepare 3 sites, winterize, hook into electric, water, septic, and purchase propane tank for one, electric and water hook-up for the others, electrical and furnace repair for the third, and misc. repairs for housing, office, and lab space Add road base, gravel, landscaping timbers to expand and improve parking areas Re-set poles, replace range wire and snow fence with Hog panels in MD feed areas and catch areas Increased power needed for expanding facilities- East and West sides, West side upgrades provided by WSVL Provided by NCWCD to replace shed demolished for water pipeline construction Remove 6 top inches of soil, saturate with 20% bleach soln., add 2 inches of road-base to lambing area Add a 10 x 10 cooler, and 10 x 10 freezer unit to the mountain lion complex The road system was built-up with extra dirt to enhance water run-off, 4 new culverts, time and equipt. to build up road system and install culverts were donated by NCWCD Generated electronic site maps from an aerial photo, with all utilities, animal holding facilities, and structures Valve was added to allow shut off the pen and necropsy lab water, while still providing water to the lab Completion Year 2003/2004 2003/2004 2003/2004 2003/2004 2003/2004 2003/2004 2003/2004 2003/2004 2003/2004 2003/2004 2003/2004 Emergency repairs to a cracked septic 2003/2004 pipe Construct a new perimeter fence around 2003/2004 the lab to allow access to the lab without compromising the animal holding facility perimeter fence 147 �Project Status 14. Emergency Fawn Completed Rearing Shelters 15. Equipment storage Completed slab 16. DOD study Completed Facilities 17. Trailer Water Shut- Completed Off Valve Replacement 18. Mountain Lion Completed Facility 19. New roofs/repair On-going project structure on old feedsheds and animal shelters. 20. Add additional On-going project animal shelters 21. Road Maintenance On-going project 22. Paint old building On-going project exteriors 23. Repair/replace On-going project latches, and broken or water damaged alleyway boards 24. Replace walk thru On-going project alley gates Details Dairy calf shelters purchased and installed in waterfowl pen, fences modified to accommodate fawns Pour a concrete slab to store tractor and bobcat attachments out of the mud Convert E7 into 4 pens, add 6 automatic waters, construct 2 alleys, repair N. alley and Dig. Cage ramps, add drainage ditch, double fencing, refurbish the metabolic cages Some materials and labor donated by WSVL Replace leaking water shut-off valve to FWRF travel trailer Utilities, concrete block building, 50 x 60 foot outdoor pen, shift containment system, and 4 indoor dens Approx. ¼ of the old structures and roofs on the facility have been replaced in the last 2 years using treated lumber and long lasting roofing materials Construct additional shelters in pens with heavy stocking rates. (36 ungulate pens on the facility) Road grading and upkeep Now using CCA treated lumber or metal siding for repairs & building replacements to reduce the amount of painting necessary in the future. Now using CCA treated lumber for all repairs Replace old gates as necessary Completion Year 2003/2004 2004/2005 2004/2005 2004/2005 2004/2005 Began 2000/2001, as needed Began 2001/2002, as needed As needed Old structures are on a painting schedule every 3-5 years As needed As needed 25. Replace old visual barrier fencing and utility wire on metal gates On-going project: most of the old material has been replaced, but this project is ongoing due to animal and environmental damage Old snow fence and construction fence Began replaced and moved to the outside of 2001/2002, the paddock fence (except interior as needed fences), utility wire is systematically being replaced with horse-fence 26. Animal holding fence upgrades, and repairs On-going project: rotten posts have been replaced, double fences constructed Replace old range fence and V-mesh, as Began well as electric fencing in pens that 2002/2003, house deer, Construct double fences as As needed required by CWD protocols 148 �Project Status 27. Construct artificial On-going project: refuge areas inside pens completed for all for neonates and adults new east side paddocks, maintain existing, construct new 28. Add windscreen to On-going project west and south facing fence-lines 29. Mowing and weed On-going project control 30.WHL maintenance On-going project 31. Unscheduled On-going project miscellaneous emergency facility repairs Completion Details Year Construct single and L-shaped, refuge Began areas to provide refuge and shade, 2002/2003, construct hog panel seasonal exclosures As needed to promote vegetation growth in the spring Provide additional shaded areas for Began animals, and maintain existing 2002/2003, As needed Seasonal mowing and manual, chemical As needed noxious weed control Provide maintenance assistance to Began WHL, and support for initial lab 2002/2003, construction As needed Emergency repairs to structures, animal As Needed holding facilities, perimeter fence, automatic waters, utilities, etc… LITERATURE CITED Davis, T. R., 2003. Animal and pen support facilities for mammals research. Colorado Div. Wildl. Res. Rep., Jan. 2001 – Jun. 2003, Fort Collins. Parker, K. L., and B. Wong. 1987. Raising black-tailed deer fawns at natural growth rates. Can. J. Zool. 65:20-23. Wild, M. A., and M. W. Miller 1991. Bottle raising wild ruminants in captivity. Colorado Div. Wildl. Outdoor Facts No. 114. Wild, M. A. 1997. Animal and pen support facilities for mammals research. Colorado Div. Wildl. Res. Rep., WP1a, J1, Jul 1996 - Jun 1997, Fort Collins. Wild, M. A., T. R. Spraker, C. J. Sigurdson, K. I. O’Rourke, and M. W. Miller. 2002. Preclinical diagnosis of chronic wasting disease in captive mule deer (Odocoileus hemionus) and white-tailed deer (Odocoileus virgineanus) using tonsillar biopsy. J. General Virol. 83:2629-2634. Wolfe, L. L., M. M. Conner, T. H. Baker, V. J. Dreitz, K. P. Burnham, E. S. Williams, N. T. Hobbs, and M. W. Miller. 2002. Evaluation of antemortem sampling to estimate chronic wasting disease prevalence in free-ranging mule deer. J. Wildl. Manage. 66:564-573 Prepared by ________________________________ Tracy R. Davis, Wildlife Technician 149 � https://cpw.cvlcollections.org/files/original/1a38a7d5bf82e1f047949a7b60aa5f49.pdf 1620fa19343093e23848b9b5c70349a9 PDF Text Text 243 Colorado Division of Wildlife Wildlife Research Report July 2001 and July 2002 PROGRESS REPORT State of~------'C=o=l=ora==d=o_ _ _ __ Division of Wildlife - Mammals Research Work Package No._ _ _3~00~4_ _ _ _ __ Other Ungulate Conservation Task No. - - - - - - - - - - - - - - Annual Winter Count of Middle Park Pronghorn Period Covered: July 1, 2000 - June 30, 2001 Authors: Thomas M. Pojar Personnel: CDOW -T.M. Pojar, R. Firth, I.Claassen, M. Crosby, K. Holinka, T. Kroening, R. Thompson, C. Wagner. Others - C. Cesar (BLM). Volunteers- B. Kraft, R. Nutter, D. O'Sullivan, M. Palowoda. MIDDLE PARK PRONGHORN WINTER 2000-01 COUNT Tom Pojar, February 13, 2001 The annual winter count of the Middle Park pronghorn herd was conducted during February 9 through February 12, 2001. The major effort was done on February 9th when several observation crews from Area 9 assisted and the largest pronghorn groups were counted. The smaller, more isolated groups were counted during subsequent days. Division of Wildlife personnel from Area 9 that participated in the count were: Jerry Claassen, Mike Crosby, Kris Holinka, Tom Kroening, Jim Liewer, Bob Thompson, and Chuck Wagner. Chuck Cesar of BLM assisted as well as the following volunteers: Ben Kraft, Ron Nutter, Dan O'Sullivan, and Marie Palowoda. This year's count was made more interesting with the infusion of animals into the population from the Blue Valley Ranch transplant. The purpose of this transplant was to expand the range of the Middle Park population to wintering areas south of the Colorado River. Throughout the years of habitation, the pronghorn have only wintered north of the Colorado River. BVR pronghorn were released from enclosures they were held in during winter 1999-00 around June 1, 2000. At this point they were free to select the summer, and subsequently, the winter range they found desirable. Of 50 BVR pronghorn radioed, 39 survived to the winter of 2000-01. Twenty-six of these are wintering south of the Colorado River; 15 west ofBVR ~eadquarters west of the Blue River, 2 west ofJim Yust's (west of the Blue River), and 9 east of Junction Butte, which is east of the Blue River. Thirteen of the BVR radioed animals are wintering north of the Colorado River with groups of the "native" pronghorn. COLO ;·;;s·;,;;; •c< :·:: IIIIIIIIIHllllllli BDOW016831 �244 Count North of the Colorado River AREA COUNT Back Troublesome 176 Starr Gulch 221 NE Red Mt. 138 Pinto Ranch 4 TOTAL 539 AREA COUNT Junction Butte - east 13 Blue Valley Ranch - west 28 Jim Yust's ranch - west 2 TOTAL 43 Count south of the Colorado River GRAND TOTAL= 582 The conclusion from this count is that there are at least 582 pronghorn wintering in Middle Park during winter 2000-01. A spreadsheet population model has been maintained for this herd since it was first being tracked in 1986. The projected winter population has corresponded with the actual count quite closely through the years with a mean deviate count of 17. This year, the model projects a population of 675 for the largest discrepancy ever encountered - 93 animals. The model was adjusted for the infusion of BVR animals with the fawn to doe ratio applied to all mature females in the entire Middle Park population. There are several possible scenarios to explain the apparent undercount for this year. No new radios have been deployed on the population north of the Colorado River since 1998. Population growth, radio failure, and harvested radioed animals have contributed to dilution of the proportion of radioed animals. This year about 5% of the population north of the river are carrying working radios. As this percentage decreases the probability of having a group of animals without at least one radio in it to allow detection increases. Finding pronghorn groups with winter conditions featuring a mottled background of snow and sagebrush can be very difficult. Radios are crucial to locating pronghorn groups during winter. All of the radios that were presumed to be working were located except 3. Although these radios were located earlier in the fall during the herd structure survey they may have failed since then or it is possible we missed detecting them during the winter count. The Antelope Creek, Antelope Pass, Cow Gulch, Wolford Mountain, Sulphur Gulch, and Corral Creek areas were searched and radio scanned for these 3 radios. The fact that these radios were not found does not eliminate the possibility that a group (or groups) of pronghorn were missed. Winter conditions are mild thus far this year. The winter began with a much colder than normal November and some snow. However, milder conditions prevailed during December, January, and so far in February. Snow depth where the pronghorn are wintering ranges from 4-10 inches with clear south facing slopes and adequate wind-blown ridges. In brief, the pronghorn should come through this winter in very good condition barring any severe late winter weather. � https://cpw.cvlcollections.org/files/original/ddb68ca712960cb10a55f71e5088ea1c.pdf 113ef8fab45e730df971320245f7827e PDF Text Text Colorado Division of Wildlife July 2009 − June 2010 WILDLIFE RESEARCH REPORT State of: Cost Center: Work Package: Task No.: Colorado 3430 3001 4 Federal Aid Project No. W-185-R : : : : : Division of Wildlife Mammals Research Deer Conservation Development of an Automated Device for Collaring and Weighing Mule Deer Fawns Period Covered: July 1, 2009 − June 30, 2010 Authors: C. J. Bishop, D. P. Walsh, M. W. Alldredge, E. J. Bergman, and C. R. Anderson. All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT We designed and produced a trap-like device for mule deer that would automatically attach a radio collar to a ≥6-month-old fawn and record the fawn’s weight and sex, without requiring physical restraint or handling of the animal. Our passive collaring device is designed to allow biologists and researchers to radio-collar, weigh, and identify sex of ≥6-month-old mule deer fawns with minimal expense and labor when compared to traditional mule deer capture techniques. This technique should significantly reduce stress that is typically associated with capture and handling and eliminate capturerelated mortality. We collaborated with students and faculty in the Mechanical Engineering Department at Colorado State University to produce a conceptual model and early prototype. We then worked with professional engineers at Dynamic Group Circuit Design in Fort Collins, Colorado, to produce a fullyfunctional prototype of the device. We will conduct an extensive field evaluation of the device with freeranging mule deer during 2010-11. 93 �WILDLIFE RESEARCH REPORT DEVELOPMENT OF AN AUTOMATED DEVICE FOR COLLARING AND WEIGHING MULE DEER FAWNS CHAD J. BISHOP, DANIEL P. WALSH, MATHEW W. ALLDREDGE, ERIC J. BERGMAN, AND CHUCK R. ANDERSON P. N. OBJECTIVE To develop and evaluate a trap-like device for mule deer that would automatically attach a radio collar to a ≥6-month-old deer fawn and record the fawn’s weight and sex, without requiring physical restraint or handling of the animal. SEGMENT OBJECTIVES 1. Work with a professional engineering firm to produce a fully-functional prototype of an automated collaring device for ≥6-month-old mule deer fawns. INTRODUCTION The Colorado Division of Wildlife (CDOW) captures and radio-marks 6-month-old mule deer (Odocoileus hemionus) fawns each year to support research and management of mule deer. Approximately 240 deer fawns are captured annually to monitor survival among 4 populations distributed across western Colorado and an additional 100−350 deer fawns are captured as part of ongoing research studies. Other state agencies in the western United States capture large numbers of mule deer fawns annually also. Most capture is accomplished with net-guns fired from helicopters (Barrett et al. 1982, van Reenen 1982, Webb et al. 2008), which is becoming increasingly expensive (i.e., >$500 per captured deer). Also, net gunning is inherently dangerous with a small market, which at times limits availability of contractors. Drop nets (Ramsey 1968, Schmidt et al. 1978), clover traps (Clover 1956), drive nets (Beasom et al. 1980), and darting (Wolfe et al. 2004) are used occasionally in the western United States to capture deer, but these techniques can be time consuming and labor intensive. Many biologists lack time and resources given other job requirements to conduct such capture operations for any length of time. The increasing cost of helicopter net-gun capture coupled with increasing demand for capturing and radio-collaring 6-month-old fawns has created a need for another capture alternative. Specifically, there is need for a capture technique that is relatively inexpensive to employ considering both operating and personnel costs. In response to CDOW’s capture needs, we conceived the idea of an automated marking device for ≥6-month-old deer fawns that would attach a radio collar and record weight and sex without physically restraining the animal or requiring handling. The idea of automatically attaching radio transmitters to animals is not new, although to our knowledge, there are no proven methods or devices for use on deer or other ungulates. Even a relatively expensive trap or device (e.g., >$5,000 ea.) would reduce CDOW’s capture costs assuming the device could be reused over time with few maintenance expenses. Such a device would enable seasonal wildlife technicians or graduate students to radio-collar samples of deer fawns independently or with little assistance from researchers and biologists because no animal handling would be required. We want the device to record weight and sex because these variables are useful covariates in survival analyses and are typically measured when fawns are captured and handled. A passive marking device would minimize animal stress associated with capture and should have virtually no potential to cause capture-related mortality. The large-mammal capture techniques described 94 �above place considerable, temporary stress on animals as part of netting and handling. Roughly 2-3% of animals typically die from capture-related injuries or stresses under routine capture conditions. Thus, successful development of a passive marking system would reduce CDOW’s operating expenses and improve animal welfare. Therefore, our objective is to design, produce, and evaluate a fully-functional prototype of an automated collaring device for ≥6-month-old mule deer fawns. STUDY AREA We conducted all evaluations with captive deer at the FWRF in Fort Collins, Colorado. We conducted limited evaluations with free-ranging deer near Fort Collins in north-central Colorado. We plan to conduct extensive field evaluations with free-ranging deer in north-central Colorado and elsewhere in Colorado once a fully-functioning device is produced. METHODS We initially wrote a study plan and identified detailed device specifications to guide development of the automated collaring device. We approached Colorado State University’s Mechanical Engineering Department to discuss their interest in helping design such a device. In result, the collaring device became a senior design project for 6 CSU engineering students during the 2008-09 school year. We met with the students weekly and provided them a materials budget of $10,000 to produce a prototype device. We conducted staged evaluations of device components during the year by working with captive deer at FWRF. We also conducted limited evaluations with free-ranging deer near the end of the year. Field evaluations focused primarily on how deer utilized and interacted with the device to guide subsequent design and development decisions. We documented utilization and interactions using direct observation and motion-sensor digital cameras. We relied exclusively on digital cameras when we were not on-site during an evaluation. Automation of the collaring device was disabled any time we were not present to prevent any potential harm to deer. Following preliminary field evaluations, we refined our design specifications and developed a contract with Dynamic Group Circuit Design (DGCD), located in Fort Collins, Colorado, to produce a fully-functional prototype device. We routinely met with electrical engineers from DGCD, and a mechanical engineer subcontracted by DGCD, during the course of the year. These meetings ensured that our device specifications were being satisfactorily met from both engineering and deer biology perspectives. RESULTS AND DISCUSSION We produced a fully-functional prototype device that met our design specifications as set forth in the contract. The prototype device comprises an aluminum cage attached to a bait compartment. Deer enter the device through an adjustable opening at the front of the cage. The adjustable opening can be used to deter entry of larger animals by adjusting both width and height. The sides of the cage comprise one-way gates that prevent entry into the device but allow an animal to exit the device at any point. The bait compartment is accessed through an opening positioned at the rear of the cage. An expandable radio collar is placed in this opening by extending it around four rectangular, aluminum plates that hold the collar in the fully-expanded position (Fig. 1). Radio collars are made expandable by attaching springs to each end of the transmitter; that is, springs are used in place of belting on standard radio collars. Clear plexiglass separates the cage from the bait compartment to maximize visibility. A deer is able to extend its head and neck through the expanded radio collar positioned in the rear opening to access the bait in the bait compartment, which is the only access point to the bait (i.e., it cannot be reached by an animal outside of the device). The floor of the cage is a scale that continuously records weight and informs device operation. Only animals in a specified weight range can be collared, which allows the user to 95 �target fawns and avoid collaring adult deer. Specifically, the mechanism that releases the collar around a deer’s neck will not trigger when an animal is too heavy or too light. Also, an actuator moves a plexiglass plate into the space between the rear cage opening and the bait pan, preventing animals outside of the weight range from accessing the bait. Shortly after a non-target animal exits the device, the collar release mechanism is once again ready to fire and the actuator lowers the plexiglass plate so that the bait is accessible. To prevent an animal from being collared twice, a loop antenna is placed around the entrance to the cage and connected to a radio frequency identification (RFID) reader. All collars used with the device include a small RFID transponder sewn into the collar material. If a previously-collared fawn enters the cage, the RFID transponder is detected, which in turn prevents the collar from being released and activates the actuator to block access to the bait. If a deer enters the cage that is in the specified weight range and has not been previously collared, the collar will release around the deer’s neck once it accesses the bait. The collar release is triggered when a deer’s head breaks an infrared beam positioned immediately above the bait pan. The collar is released by activating a solenoid, which in turn releases a lever that causes the upper 2 aluminum plates holding the expanded collar in place to collapse (Figs. 2 and 3). The collar is then situated around the deer’s neck. When the collar is released, 2 different cameras are immediately activated to take a series of 3 photographs each. One camera is positioned in the back of the bait compartment and set to take a closeup photo of the top of the deer’s head. The second camera is positioned in the floor of the cage and set to take a photo of the deer’s abdomen and groin. These cameras are activated only when a collar is released and facilitate determination of deer sex. Last, when a collar is released, the device records and stores the weight of the deer. An external computer can be hooked up to the device to change program settings, remotely operate the device, and upload weight data. The device is powered by a 12 volt battery that must be recharged every 2-3 days assuming continuous operation. DGCD prepared a user’s manual that explains device operation and detailed schematics to allow future production. We will evaluate effectiveness of the device in the field during 2010-11. Initially, we will only set the device with a collar when we are present and able to directly observe deer interactions with the device. After collaring 5-10 animals in this manner and troubleshooting any problems with the device, we will set the device to operate remotely without an observer on-site, which is how it is intended to be used. SUMMARY We developed a fully-functional prototype of an automated collaring device for mule deer in collaboration with professional engineers. The automated collaring device is designed to allow biologists and researchers to radio-collar portions of their deer samples with minimal time and expense because no animal handling is required and deer can be collared at any time. Primary time commitments include baiting sites, moving device(s) among sites, and adding collars to the devices. The collaring device should also have distinct benefits for studies in urban environments by providing a non-invasive technique for collaring deer. The collaring device should significantly reduce stress that is typically associated with capture and handling and there should be no capture-related mortality. We also have designed the collaring device so that it should be relatively easy to adjust to target adult deer and other ungulate species. Last, the collaring device should have wide applicability for ungulate researchers and managers beyond Colorado. We will be evaluating the device in the field with free-ranging mule deer during the coming year and making additional modifications as necessary. 96 �LITERATURE CITED Barrett, M. W., J. W. Nolan, and L. D. Roy. 1982. Evaluation of a hand-held net-gun to capture large mammals. Wildlife Society Bulletin 10:108−114. Beasom, S. L., W. Evans, and L. Temple. 1980. The drive net for capturing western big game. Journal of Wildlife Management 44:478−480. Clover, M. R. 1956. Single-gate deer trap. California Fish and Game 42:199−201. Ramsey, C. W. 1968. A drop-net deer trap. Journal of Wildlife Management 32:187−190. Schmidt, R. L., W. H. Rutherford, and F. M. Bodenham. 1978. Colorado bighorn sheep-trapping techniques. Wildlife Society Bulletin 6:159−163. van Reenen, G. 1982. Field experience in the capture of red deer by helicopter in New Zealand with reference to post-capture sequela and management. Pages 408−421 in L. Nielsen, J. C. Haigh, and M. E. Fowler, editors. Chemical immobilization of North American wildlife. Wisconsin Humane Society, Milwaukee, USA. Webb, S. L., J. S. Lewis, D. G. Hewitt, M. W. Hellickson, and F. C. Bryant. 2008. Assessing the helicopter and net gun as a capture technique for white-tailed deer. Journal of Wildlife Management 72:310−314. Wolfe, L. L., M. W. Miller, and E. S. Williams. 2004. Feasibility of “test-and-cull” for managing chronic wasting disease in urban mule deer. Wildlife Society Bulletin 32:500−505. Prepared by _______________________ Chad J. Bishop, Mammals Research Leader 97 �Figure 1. View of the radio collar and bait compartment of an automated collaring device for mule deer. To reach bait, deer must extend their head and neck through the expanded radio collar. 98 �Figure 2. View of the collar release mechanism in an automated collaring device for mule deer. 99 �Figure 3. Female mule deer fawn accessing bait by extending her head through an expanded radiocollar. The prototype device will be evaluated extensively in the field with free-ranging deer during 2010-11. 100 � https://cpw.cvlcollections.org/files/original/0e12b1dabc99edcfae1ddf74c39d7285.pdf 52282c9dd1416e8cac8d68b47a07b7d2 PDF Text Text Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Project No. Work Package No. Task Colorado Federal Aid Project: N/A 3740 : : : : Cost Center 3430 Mammals Research Wildlife Diseases Pilot evaluation of GPS technology in chronic wasting disease prevalence and management at artificial feeding sites in urban areas. : Period Covered: April 1 2003 through July 31, 2004 Author: Eric J. Bergman, Michael W. Miller and L. L. Wolfe Personnel: M. Sirochman All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT A pilot study for assessing the utility of GPS technology in the evaluation of CWD prevalence and management in urban areas was designed is being implemented. Objectives of this pilot study are to: 1) Evaluate the utility of GPS radio collar technology in identifying artificial feed sites in urban settings, 2) Evaluate if there is evidence that artificial feed sites reduce the size of deer home ranges, 3) Evaluate if deer density is elevated at artificial feed sites, and 4) Evaluate if CWD prevalence is higher at artificial feed sites . 119 �JOB PROGRESS REPORT PILOT EVALUATION OF GPS TECHNOLOGY IN CHRONIC WASTING DISEASE PREVALENCE AND MANAGEMENT AT ARTIFICIAL FEEDING SITES IN URBAN AREAS Eric J. Bergman, Michael W. Miller and L. L. Wolfe INTRODUCTION Analyses of data from recent field studies and from culling have revealed areas of relatively high CWD prevalence associated with urban areas along the northern Front Range (Wolfe et al. 2002, 2004; Conner and Miller 2004; Farnsworth et al. 2004). Within these, artificial and illegal feeding sites may be particularly important because they appear to congregate deer in one location, thereby increasing local deer density and exposure to contaminated environments (Miller et al. 2004). Although the nature of the relationship between disease prevalence and mule deer density has not been definitively identified, it seems likely (Barlow 1996) that CWD prevalence is being indirectly elevated through artificial feeding. The development of global positioning system (GPS) technology and its incorporation into radio collars for wildlife research presents a tool for better understanding CWD in urban areas. We have initiated a pilot field study to: 1) evaluate the effectiveness of different GPS collars in identifying illegal feed sites in urban settings, and 2) develop and evaluate a strategy for utilizing GPS technology in studying and managing CWD in urban mule deer populations. METHODS The study area for this work is located within two subdivisions in Estes Park, Colorado. The subdivisions, separated by approximately 1.6 km, were identified as treatment and control sites based on the presence and absence of known feeding sites (Fig.1, Wolfe et al. 2004). Between five and eight adult (>1 yr old) female deer from each subdivision were captured and collared with one of two different brands of GPS collars (HABIT Research, British Columbia, Canada and LOTEK Wireless, Ontario, Canada). Collars from each company will be evenly distributed between sites. Capture will occur as part of an ongoing "test and cull" research project (Wolfe et al. 2004) during April 2004 and from August to October of 2004 as needed. Deer will be recaptured and collars will be removed prior to battery failure (~220 days service) in order to retrieve GPS data. No specific hypotheses are being tested in this pilot study; rather, we are attempting to determine if GPS radio collar technology is adequate for use as a tool in refining CWD epidemiology and management. We will record and report on the performance of GPS collars, and calculate costs (mean, range per animal tested) associated with our artificial feed site identification strategy as implemented in this pilot study. However, we will compare home range sizes of deer from each site to determine if artificial feeding reduces home range size of deer. We will also incorporate ground survey data (Wolfe et al. 2004) to estimate and compare mule deer density and ultimately CWD prevalence from sampled deer at each site. CWD prevalence will be compared between sites as well as to previous estimates from the greater Estes Park area (Wolfe et al. 2004) to explore future research potential. 120 �RESULTS AND DISCUSSION GPS Collar Comparison A total of 16 GPS collars (10 LOTEK, 6 HABIT) were available for testing in this study. Prior to initiation of this study no HABIT collars were on hand for deployment, rather, all 6 had to be built to specification and delivered. GPS collars from HABIT Research, ~$1,800/unit, were programmed to: collect GPS locations every 2 hours, to transmit GPS data (via VHF signal) over two day intervals every two weeks and to transmit the most recent GPS location (via VHF signal) at the start of each minute. Due to delays in the manufacturing process, no HABIT collars were received in time for spring deployment (≥2 weeks pre-fawning). Additionally, due to programming errors, 0 of 6 HABIT collars were ready for deployment after initial testing. Upon servicing by HABIT Research (~3.5 weeks), 3 of 6 collars appear to be ready for deployment in late summer 2004. The remaining HABIT collars (3 of 6) will be serviced and deployed upon satisfactory performance. All LOTEK collars were on hand prior to initiation of this study. Eight of 10 collars were deployed in spring of 2004, with 1 of 10 needing service. GPS collars from LOTEK Wireless, ~$3,500/unit, were also programmed to collect GPS locations every 2 hours, but did not offer remote download capabilities. All GPS locations collected by LOTEK collars will be acquired upon retrieval of the collar. GPS Collar Performance Data from LOTEK GPS collars continues to be collected and HABIT GPS collars will be deployed between August-September 2004. LITERATURE CITED Barlow, N.D. 1996. The ecology of wildlife disease control: simple models revisited. Journal of Applied Ecology 33:303-314. Conner, M.M., and M.W. Miller. 2004. Spatial epidemiology in natural populations: a case study of movement and prion disease prevalence relationships among mule deer population units. Ecological Applications (in press). Farnsworth, M.L., L.L. Wolfe, N.T. Hobbs, K.P. Burnham, D.M. Theobald, and M.W. Miller. 2004. Human land use influences chronic wasting disease prevalence in mule deer. Ecological Applications: in review. Miller, M.W., E.S. Williams, N.T. Hobbs, and L.L. Wolfe. 2004. Environmental sources of prion transmission in mule deer. Emerging Infectious Diseases: in press. Wolfe, L.L., M.M. Conner, T.H. Baker, V.J. Dreitz, K.P. Burnham, E.S. Williams, N.T. Hobbs, and M.W. Miller. 2002. Evaluation of antemortem sampling to estimate chronic wasting disease prevalence in free-ranging mule deer. Journal of Wildlife Mangement 66:564-573. _________, M.W. Miller, and E.S. Williams. 2004. Feasibility of "'test-and-cull" for managing chronic wasting disease in urban deer. Wildlife Society Bulletin 32:500-505. Prepared by ____________________________ Eric J. Bergman, Wildlife Researcher 121 � https://cpw.cvlcollections.org/files/original/2c8cfb764ba7ff42f9d3d5a4cdb42e84.pdf d9c11598d5e99b617aeffaeb13dfb64d PDF Text Text Colorado Division of Wildlife Wildlife Research Report July 2003 – June 2004 JOB PROGRESS REPORT State of Project No. Work Package No. Task 2 Colorado Federal Aid Project: N/A : : : : 3004 Cost Center 3430 Mammals Research Other Ungulate Conservation Potential Research Project Assessment : Period Covered: July 1, 2003 through June 30, 2004 Author: Eric J. Bergman and Gary C. Miller Personnel: D. Freddy, C. Bishop, B. Watkins, J. Madison, J. Broderick All information in this report is preliminary and subject to further evaluation. Information MAY NOT BE PUBLISHED OR QUOTED without permission of the author. Manipulation of these data beyond that contained in this report is discouraged. ABSTRACT As part of the research planning process, general assessments were made for the potential to conduct research projects on 3 main topics identified by Colorado Division of Wildlife field management personnel. These topics were: impacts of mule deer/elk interactions on mule deer population performance; improving success of bighorn sheep reintroductions and translocations; and, impacts of natural gas and oil extraction on mule deer. All 3 topics present challenges to conducting successful research endeavors with deer-elk interaction studies potentially providing the most predictable research and funding situations. 89 �JOB PROGRESS REPORT POTENTIAL RESEARCH PROJECT ASSESSMENT Eric J. Bergman INTRODUCTION The Colorado Division of Wildlife is charged with protecting, preserving and enhancing Colorado's ungulate populations for the use, benefit, and enjoyment of the people. The management principles guiding managers in this mission include: wildlife conservation, use and enjoyment, maintaining healthy, diverse and abundant populations and maintaining/conserving habitat quality/quantity through science-based decision making (Colorado Div. of Wildlife 2002-2007 Strategic Plan:9). The objective of ungulate research is to provide information to facilitate the making of these science-based decisions. METHODS Members of the Division’s Mammals Research staff met with wildlife managers and biologists from the Northwest and Southwest Regions to consult on ungulate management issues and research needs. The following topics were identified as the primary statewide issues of concern: ● Negative impacts of deer/elk interactions on deer population performance. ● Variable success of bighorn sheep reintroduction/translocation efforts. ● Impacts of natural gas and oil extraction on deer. In general, research topics identified by wildlife managers favored issues occurring in their respective areas and regions. Topics identified by biologists emphasized need for broad-scale research efforts to address issues that exist throughout the state. The objective of research project selection is to successfully merge these two viewpoints into a study that: 1) Provides clear results through control/treatment experiments. 2) Addresses issues of local and statewide concern. 3) Allows inference to wide spatial/temporal boundaries. RESULTS AND DISCUSSION Impacts of Deer/Elk Interactions on Deer Population Performance A general long-term trend of mule deer population declines has existed both in Colorado and throughout the Western U.S. since the late 1950s (Unsworth et al. 1999, Gill 2001). While insular deer herds have functioned outside of this trend and others have experienced pulses of population growth nested within this long-term decline, declines of as much as 50% have been reported (Colorado Division of Wildlife, unpublished data). Along with these overall population declines, depressed fawn:doe ratios have been simultaneously measured (Gill 2001, White et al. 2001). In an attempt to address concerns over these declines and to develop working hypotheses as to the underlying causes, CDOW hosted a conference for employees of CDOW, CSU, federal agencies, invited publics and deer experts in 1998. A result of this conference was the identification of five potential sources of mule deer population depression: A) habitat deterioration, B) predation, C) competition with elk, D) disease, and E) weather (Fig. 1). Based on a coupling of the ideas from this conference with existing knowledge of mule deer/coyote predation dynamics (Bartmann et al. 1992), CDOW entered a collaborative mule deer 90 �research program with Idaho Fish and Game. Two studies were simultaneously developed. Idaho began a mule deer predation study that is currently assessing the impacts of predator control (both puma and coyote) on mule deer survival. The Division of Wildlife began a nutrition study to simulate and quantify the demographic impacts of habitat enhancement on mule deer fawn survival (Bishop 2003). Through a treatment/control, cross-over study design, CDOW is providing supplemental forage as a mechanism to test the effects of an immediate, short-term enhancement of mule deer habitat (Bishop 2003). Preliminary results of this ongoing research indicate that by enhancing the nutritional value of mule deer diets, overwinter fawn survival can be increased (Bishop 2003). While the methods employed through this research are not an acceptable management strategy to the Division, the pending conclusions exemplify the need for landscape level treatments to enhance mule deer habitat (discussed below). As such, through a series of reductionist experiments, it has been demonstrated that mule deer population performance is directly related to habitat quality, yet functions somewhat independently of coyote predation. However, a confounding factor in these conclusions stems from the fact that concurrent with mule deer declines in Colorado, there has been a dramatic increase in the distribution, number and density of elk. Accordingly, while enhancing the quality of habitat should improve mule deer survival, the ultimate causes of habitat degredation have not been tested. It remains unclear whether interference/exploitation by elk has further reduced carrying capacity for mule deer in late seral stage pinyon-juniper ecosystems. Based on past research, there is variable evidence that competition between mule deer and elk exists. Both interference and exploitation competition have been reported for deer and elk (Singer and Norland 1994, Kirchoff and Larsen 1998, Stewart et al. 2003). However, due to logistical constraints, these studies have been observational and are highly dependent on the specific conditions at which the data was collected. Thus, strong inference is not possible. In order to accurately quantify the impacts of deer/elk competition, a study would need tight control over ungulate densities. Under ideal conditions, this would be done in a controlled setting (i.e. large enclosures). Due to financial constraints and disease potential, this is not possible in Colorado. Additionally, enclosures with controlled densities of deer and elk remove much of the natural variation that exists in nature. Elk are highly mobile animals with weak site fidelity. Thus, while deer may encounter high elk densities one week, they could very easily encounter no elk the following week. Enclosures with static deer and elk densities would remove this variability and subsequent results would be of limited utility. Elk mobility is also the primary reason that density reduction treatments cannot be applied in field settings. A dramatic harvest treatment to reduce local elk density could be erased as elk immigrate into the treatment area once harvest is finished. Given these constraints, the possibility of studying the interactions between deer and elk still exists. Currently, the potential to do this is presenting itself on the Uncompahgre Plateau. The proposed follow-up to the short-term deer nutrition research is the implementation of large scale deer habitat enhancement treatments. However, there is concern that elk will monopolize treatment areas and will either displace deer or deprive deer of receiving the intended treatment. As such, regardless if elk are being studied, mitigation efforts will have to be employed to insure the delivery of treatments to deer. Due to the baseline knowledge on deer in the Uncompahgre system (C. Bishop, unpublished data, B. Watkins, unpublished data), the existing capital investment and given the fact that elk treatments would be an integral part of any deer habitat enhancement study, many of the financial and logistic hurdles to starting a complimentary elk research program would be minimized. 91 �Declining Mule Deer Habitat Quality Bishop Phase I deer/elk research --Collaborative ----- -,- --- --- Enhanced nutrition study showing increased fawn survival l--, NO Are late seral stage pinon/juniper communities conducive to quality deer production? Deer/Elk Interactions I I I enhance nutrition through , Can - -we-habitat - -treatments? -I ........... . ----.l.•······· ........ . =········'·~ Problems: -Will elk displace deer in treatment areas? -Can we successfully apply treatments on large temporal & spatial scales? Has elk population expansion: -displaced deer? -put deer at an exploitation disadvantage? I I I_ - - - - - - - - , . I I No Evidence To Date Is disease depressing deer populations statewide? Have drought conditions negatively impacted deer? NO No clear Evidence Disease Bishop study, 5 statewide study areas Measure species/sex/age class of puma kills using deer & elk radio collars (K. Logan) Do coyotes force poor population performance? Miller, Wolfe, Baeten, Etc. Monitoring of deer mortalities for disease Puma Research & Mgmt. )--------- MULE DEER ISSUES -Low Populations -Poor population parameters Ongoing CWD Research Do puma force poor population performance? Compensatory mortality in a Colorado mule deer population Weather Predation Bartmann et al. (1992) Figure 1. Conceptual model concerning mule deer population declines in Colorado with an emphasis on the relationships between sources of population depression (light gray boxes), current and past research (dark gray boxes) and potential research (dashed lines). A final benefit to conducting a study on the interactions between elk and mule deer on the Uncompahgre Plateau pertains to ecosystem level understanding. As the likelihood of puma research being conducted on the Uncompahgre Plateau increases, the addition of elk research would allow for integrated, broad scale understanding of deer, elk and puma management. Over the past 25 years, wildlife ecologists have strived to understand ecosystem level interactions (Sinclair and Norton-Griffiths 1979, Houston 1982, Jedrzejewska and Jedrzejewski 1998, Krebs et al. 2001). However, a fundamental feature of these studies is that they were conducted in systems that were managed for species conservation (i.e. no hunting), limiting the utility and applicability in systems where species are managed as a consumptive resource. All elements will likely be in place to do ecosystem level research in a system that is managed with a multi-use strategy. The level of interest within the CDOW and other western state agencies in studying deer/elk interactions is high (D. Freddy, unpublished study plan, 1998). Additionally, based on high priority achievements H-1.2 and H-1.3 of the Colorado Division of Wildlife Strategic Plan (2002-2007), research on these issues are highlighted as guiding principles of Division activities. This interest is being echoed by external conservation groups. Over the past decade, the Uncompahgre Project of southwest Colorado has actively pursued habitat enhancement on the Uncompahgre Plateau. This group has expressed interest in coordinating their proposed habitat manipulations with wildlife research, potentially relieving the Division of the financial burden of conducting landscape level treatments (R. Sherman, personal communication). Additionally, it is believed that groups such as the Mule Deer Foundation and the Rocky Mountain Elk Foundation would be willing to invest in research that further explains the impacts of deer/elk interactions as well as the effectiveness of habitat enhancement for these species. 92 �Variable success of Bighorn Sheep Reintroduction/Translocation Efforts During the late 1800's and early 1900's, the rocky-mountain west experienced dramatic declines in bighorn sheep populations (Singer et al. 2000a). Due to the role that bighorns play in ecosystems and their value as both a watched and hunted species, recovery is a high priority. Approximately 55%-58% of the existing populations are a result of reintroduction efforts (Singer et al. 2000b). However, reintroduction efforts have been variable in success (success rates fall between 40% and 58%) (Singer et al. 2000a, Singer et al. 2000b). While these trends and statistics are for the west as a whole, most are indicative of bighorn management efforts within Colorado (Bailey 1990). There are many parameters that influence the success of bighorn sheep reintroduction efforts. Among these, survival, density, metapopulation characteristics, habitat quality/quantity, disease and predation have all been identified as potential limiting factors (Fig. 2). However, due to the lack of experimental manipulation, the relative importance of any single parameter is largely unknown. Bighorn population characteristics are poorly understood, yet are likely important for planning and implementing successful recovery efforts. Many recovery efforts appear to be typified by brief periods of very high population growth, followed shortly by stabilization and population crash (Singer et al. 2000c). In the absence of disease, adult survival is typically high. However, lamb survival can be impacted by density as well as weather (McCarty and Miller 1998, Holl et al. 2004). In the absence of dispersal corridors, and due to potential growth rates as high as λ = 1.30 (Singer et al. 2000c), many recovering bighorn populations suffer from sedentariness. Because bighorns can quickly reach the local carrying capacity of a patch, and due to high adult survival, the instigation of senescence at approximately 7-9 years of age, and low dispersal, it is possible that recovery efforts fail simply because no new individuals are entering the population. A manipulative, research approach to addressing this question is to emulate dispersal through the selective removal of senescent individuals, thereby reducing density and forage pressure such that lamb survival can increase. Bighorn sheep are thought to have traditionally functioned as a metapopulation, with a single population being characterized as a discrete group with limited movement of individuals between groups. While limited in overall quantity, the role of immigration and emigration is vital in metapopulation stability. Dispersal typically occurs in the form of animals leaving one population to join another existing population (i.e. into contiguous, occupied habitat) (Singer et al. 2000a). While providing relief for dense populations, dispersal also provides other populations with new genetic stock. Unfortunately, most recolonization efforts have focused on filling insular habitat patches with a single population and dispersal is not considered. By taking a multi-patch view of the landscape and populating several patches with animals, a metapopulation could be established. Bighorn sheep habitat is typified by open vegetation structures with high visibility and rough terrain to avoid predation (Singer et al. 2000a). Additionally, typical bighorn sheep habitat is composed of climax (late-seral stage) plant communities. These habitats occur naturally (as well as through human influence) in a fragmented fashion across the landscape, further explaining the metapopulation structure. Movement between habitat patches is easily arrested by physical barriers such as rivers and roads (Singer et al. 2000b). Absence of these key habitat characteristics and presence of barriers are all factors that potentially have a negative impact on recovery efforts. Unfortunately, experimental research has not been conducted to determine the actual importance of any single parameter. Disease has also been identified as a potential arresting factor in recolonizing bighorn populations. Epizootic breakouts can reduce bighorn sheep populations by >20%/year (McCarty and Miller 1998). Typically epizootic breakouts are in the form of Pasturella haemolytica, however, parainfluenza-3 and other Mycoplasma outbreaks have also been documented (Singer et al. 2000c). In Colorado, lungworm (Protostrongylus spp.) has also been prevalent. Diseases such as bacterial 93 �pneumonia are often introduced to bighorn populations via exposure to domestic sheep (Goodson 1982). In a massive reintroduction program throughout the west during the 1990’s, a 16km buffer between wild and domestic sheep populations was deemed necessary for habitat to be classified as suitable for bighorn reintroductions (Singer et al. 2000a). Precautionary vaccination programs have been instituted, but reported results have not indicated that such approaches greatly enhance bighorn population recovery (Miller et al. 2000). - - - - - - -, Inoculation Experiment: Harvest I I Experiment: --- I I Can population growth I I be slowed via Can reintroduced animals I be immunized to reduce susceptibility? limited harvest? 1 - - - - - - -1 ~------' ~ Density Does initial population growth exceed carrying capacity at release sites, instituting density dependent limitations prior to initiation of dispersal events? Habitat -------, Livestock Removal Experiment: -- - Disease - Does disease limit growth? - Can disease be eliminated? -Is habitat quality/quantity at reintroduction sites conducive to desired population growth? r----------- Habitat Experiment: 1 -Can we ----I improve habitat by reducing I fragmentation and providing dispersal I corridors? I - - - - - - - - - - _1 Bighorn Issues -Variable success of reintroduction efforts Puma Mgmt. ~-------··► (See Logan’s Program Narrative) ----------, Predator Control Experiment: Metapopulation -Do reintroduction efforts establish metapopulations that function outside of immigration and are thus extremely susceptible to devastational stochastic events -· Can disease vectors be removed? r-----------1 -Does blanket lion removal facilitate re-establishment of bighorn populations? - Is selective removal of offending cats less effective than blanket removal? - - I Reintroduction Experiment: I -- - -long-term --- I -Does a multi-stage, release strategy that emulates immigration increase reintroduction success? I I Predation - Do lions limit population growth in reintroduced populations? Figure 2. Conceptual model concerning the variable success of bighorn sheep reintroduction efforts with emphasis on the relationships between limiting factors (gray boxes) and research possibilities (dashed lines). The role of puma predation on bighorn sheep populations is heavily debated in the literature, but in some circumstances is thought to be the limiting factor in bighorn population growth (Wehausen 1996, Hayes et al. 2000, Holl et al. 2004). In Colorado, the impacts of puma predation on desert bighorn sheep are an issue of concern. For instance, it is believed that recent desert bighorn recovery efforts in the Dolores canyon were unsuccessful due to puma predation. Of the 12 radio-collared animals in this population, 11 have died. Nearly 100% of these deaths were due to predation (those not classified as predation were classified as unknown) (B. Watkins, unpublished data). Due to this preliminary evidence, an experimental research project that addresses the impact of puma predation on desert bighorns would likely be beneficial to future desert bighorn recovery efforts. In summary, there are many management-driven experiments that could be designed to elucidate the effectiveness of different management approaches to the recolonization of bighorn sheep. In fact, many of the potential treatments have been conducted in a non-experimental fashion (see Bailey 1990 for a review). On the western slope of Colorado, the number of potential study sites for bighorn sheep research is high. The Dolores Canyon (west of Durango) has had several failed reintroduction efforts, 94 �despite being classified as viable bighorn habitat. The Escalante, Dominguez and Roubideau canyons east of the Uncompahgre Plateau all have bighorn sheep. However, this population is currently suffering from a Pasturella outbreak (B. Watkins, personal communication). Finally, Colorado National Monument and Debeque Canyon are potential study sites. Because many of these issues take place at the population level, any study would be long-term and broad in scale. Although not explicitly detailed, conservation of bighorn sheep (especially desert bighorn sheep) is loosely prioritized in section S-2 of the CDOW strategic plan (conservation of native species). External funding for this type of research would be difficult to secure. While private groups are interested in bighorn restoration, the level of support needed to address these questions is likely greater than that which they can provide. In terms of logistical support, based on conversations with Division employees from local through the regional levels, interest and support for this research is high. Impacts of Natural Gas and Oil Extraction on Deer and Elk. The impact of natural gas and oil development on deer populations in Colorado appears to be a cyclical issue driven by economics, political policy and current world affairs. Of the impacts that resource extraction can have on deer populations, the two of most immediate concern are space use patterns and population performance (Fig. 3). In terms of space use, deer behavior can shift on broad scales, fine scales or both. For instance, a broad scale shift might manifest itself in the form of animals vacating any area where development is occurring. A fine scale shift might manifest itself in the form of avoiding habitat types, but not abandoning areas of development. Little information for this type of impact has been published. In face of the sparse literature pertaining to the impacts of development on deer and elk, there is information available that pertains to the impacts of development on caribou. Behavioral adaptations similar to those mentioned above have been documented for caribou in response to resource extraction in the arctic (Cameron et al. 1992, Nellemann and Cameron 1998, Dyer et al. 2001, Nellemann et al. 2003). Despite documented shifts in caribou distribution and density, there has been no documentation of negative population level impacts caused by resource extraction. In fact, the number of caribou in the Central Arctic caribou herd increased from 5,000 to 20,000 during oil-field development (between 19751997, Cronin et al. 2000). Similarly, while calving caribou and cow/calf pairs have been observed to avoid roads, a depression of calf survival was not reported (Nelleman and Cameron 1998). While space use behaviors are important, the DOW is primarily concerned about mule deer population performance. Of the population parameters that could potentially be impacted, fawn survival is the most sensitive to disturbance and is subsequently the key parameter for monitoring. Of additional note, measuring changes in the proximate physiological factors that lead to depressed fawn survival may be an avenue for exploring the impacts of resource extraction if monitoring fawn survival is unrealistic. Within the history of Colorado, the impacts of natural resource extraction on wildlife is a subject that has not been ignored. During the 1980's, the U.S. Department of Energy funded research on deer survival on the CA and CB tracts of the Naval Oil Shale Reserve in northwestern Colorado. This research was to be extended to include the impacts of oil-shale development. However, due to financial limitations in the extraction process, development never progressed beyond the initial phase (G. White, personal communication). A pilot study addressing the impacts of natural gas drilling on deer space-use is currently underway in southern Colorado (S. Wait, personal communication). Based on what is known about the current distribution of development, there are essentially two potential study areas in Colorado. There is interest in studying these impacts along the I-70 corridor where gas pad density approaches 1 per 20 acres. Development for natural gas extraction is also either occurring or proposed for the Roan plateau, as well as the Mamm and Divide Creek basins. Unfortunately, the current density of extraction platforms in the Mamm and Divide Creek basins is too high for a controlled experiment. The feasibility of conducting research on the Roan Plateau is unknown. Based on public sentiment reflected in the news media, development on the Roan Plateau is hotly contested. In order for this study to be accomplished, 95 �the Division would likely be placed at odds with this public sentiment due to the need for intense development as a treatment effect. Away from the I-70 corridor, development is also underway in southwestern Colorado (east of Durango, see above). This area likely offers the greatest possibility for advanced research on this issue due to the ongoing dialogue between management agencies, and the advanced stages of research activity. Additionally, because development is progressing at a slower rate in this region, and because much of the proposed development is on the Southern Ute Indian Reservation, there is potential for conducting experimental research. Fawn Survival Is fawn survival negatively impacted by development? -----------:------- ------- I \ _____________ \ Adult Survival :' ' ', ',, ___ Is adult survival negatively impacted by development? • ••. ',,t: ',,_ ', ' \, [ Population Performance Pregnancy Rates What are the population level impacts of development/? Will pregnancy rates be reduced due to stress and/or nutrition? Impacts of Natural Gas and Oil Development on deer and elk populations Temporal Response '' ,, Are the impacts of development permanent, or will deer and elk become accustomed to the change and adapt accordingly? Spatial Response ·········+··················:;:1 _ _ Do deer and elk respond to development by changing space-use patterns? • • ,------ I Resource Selection Do deer and elk respond to development by changing selection patterns? Figure 3. Conceptual model concerning the impacts of oil and gas development on deer and elk populations, with emphasis on potential impacts (gray boxes) and research possibilities (dashed lines). From a greatly simplified viewpoint, natural resource extraction can be broken into two phases. Phase one is primarily composed of the building of infrastructure (i.e. road building, pipeline building, drill pad leveling, pump installation, etc.). The subsequent phase two is composed of steady state resource extraction (Fig. 4). In terms of impacts on deer, phase one is likely to have a strong, negative impact that is relatively short-lived. Infrastructure construction is typically high intensity and may be accompanied by a shift in space use behavior of deer (the longevity of this shift is open to debate). Phase two is typified by the physical extraction of the resource, a highly mechanized process that could require very little human presence on a daily basis. Due to the longevity of phase two, it is likely to have the longest lasting impact on deer (though the impact itself may be more subtle). However, the above described progression of development is an oversimplified scenario of how events could take place. Numerous uncertainties affect the rate of development, many of which are tightly linked to current governing policy and the overall state of the economy. For instance, it is possible that due to economic hardship or due to inefficiencies in the resource refining process, development would be aborted before the completion of phase one (Fig. 5). This sequence of events are similar to those that took place during the 1980's on the CA and CB tracts of the Naval Oil Shale Reserve (G. White, personal communication). 96 �Conversely, it is also possible that regulatory policy could be relaxed during phase one of development. Thus, instead of progressing into phase two, phase one would be closely followed by a period of even more intense development (Fig. 6). The most likely scenario for the rate of development over the next 10-15 years, however, is a merger of both of these last two possibilities (i.e. development that is marked by peaks and valleys driven by the policy and economic events of the time, Fig. 7). As such, any study would have to accommodate a highly unpredictable and sporadic development pattern (Fig. 8). As mentioned above, published experimental research on the impacts of resource extraction has been scarce. The reasons for this can largely be condensed to three primary problems: 1) the lack of experimental control, 2) the necessity for long-term commitment, and 3) cost and logistics. For experimental research to occur, the needed approach would be a treatment/control study design. Due to the fact that the natural process variation of mule deer fawn survival ranges between 0.04 and 0.81 (Unsworth et al. 1999) and to the longevity of this study, a pre/post experimental design would not provide meaningful results. However, through a treatment/control design, the confounding effects of this annual variation would be eliminated. Thus, the issue turns to maintaining clean control/treatment study areas and clean treatment affects. The criteria for a control study area include: 1) quality deer winter range similar to the treatment area (i.e. similar geographic, topographic and botanical composition), 2) close proximity to the treatment study area, such that annual weather patterns are shared between the treatment and control areas, 3) being located far enough from the treatment study area that development activities do not influence deer on the control site, 4) having no pre-existing development, and 5) remaining free of on site and nearby resource extraction/exploration during the 10-15 years covering the experiment. The criteria for an adequate treatment study area are equally complex. In addition to the underlying criteria for a control site, a treatment study area would need: 1) an absence of any pre-existing development, 2) a "phase one" development schedule that is not subject to change (regardless of political, social or economic factors), and 3) a "phase two" period of steady-state extraction and maintenance that is also void of further mining and exploration. Despite great efforts, these criteria would be difficult to meet. Mining companies cannot afford to slow the extraction process if the allowable rate increases, and likewise, they cannot afford to continue pumping if it isn't economically feasible. These economic and social factors that drive production are outside the control of DOW, DNR, high level political officials and the mining companies themselves. Long term commitment is a factor that plagues all long term research programs. However, a failed commitment to see an experiment on oil/gas development to completion would provide very little in terms of knowledge. For many long term research projects the delivery of a treatment occurs in the early stages, only to be monitored in the later years. The treatment in a study concerning oil/gas development would be two-fold, i.e. the treatment (or lack thereof) would be applied every year for 10-15 years. Willingness to maintain this program would need to persist in light of the political and economic changes that will inherently occur. Policy regulating natural resource extraction could become more stringent or more relaxed (discussed above). If the economic potential of development is not realized, long term commitment is not feasible. 97 �Phase Two Phase One HIGH I LEVEL OF DISTURBANCE / . ___..---·-· r-- -·-·- I - ·-· -·- I ·-· -·-·-·- I I I NONE J Figure 4. A conceptual diagram showing the ideal separation between phases of natural resource extraction and subsequent development. Phase one is typified by high intensity, infrastructure construction. Phase two is typified by lower intensity, higher longevity, extraction processes. Phase Two Phase One HIGH I LEVEL OF DISTURBANCE / I .--- --·- ....... \ \ I \ I \ I '--. 1 NONE 1,'_ -..l Figure 5. A conceptual diagram showing potential separation between phases of natural resource extraction and subsequent development. As opposed to ideal conditions, phase one could be cut short due to changes in the economy or changes in regulatory policy. Phase Two Phase One LEVEL OF DISTURBANCE HIGH I I I I NONE Figure 6. A conceptual diagram showing potential separation between phases of natural resource extraction and subsequent development. As opposed to ideal conditions, phase one could be followed immediately by further development, a result of economic incentives or relaxation in regulatory policy. 98 �Phase Two Phase One HIGH LEVEL OF DISTURBANCE / I .---·-·-·-·-·--l I \ I "J I _,..., .T·\ " I. I \ . \ I\ \ \ I \ .. I I\ ·J \: I '· i \\/I \," I NONE V Figure 7. A conceptual diagram showing likely separation between phases of natural resource extraction and subsequent development. As opposed to ideal conditions, phase one will probably be followed by highly fluctuating stages of further development and extraction. The peaks and valleys of this scenario would be driven by normal economic variation and changes in regulatory policy. Phase Two Phase One LEVEL OF DISTURBANCE HIGH NONE Figure 8. A diagram showing the potential development patterns that could be encountered during experimental research on the impacts of resource extraction. The variability shown in phase two is an uncontrollable product of changing economic patterns and changing regulatory policy. The lack of control highlighted by this figure demonstrates the underlying reason that experimental research on this issue is impossible. Cost, the third factor, is also an obstacle encountered during any study. Based on power calculations from other mule deer fawn survival studies, a sample size of 40 marked fawns per study area was deemed necessary to detect a 15% change in fawn survival (Bishop 2000). However, in this example, experimental design increased the power of detecting this difference in excess of 80% because annual results could be pooled over consecutive years. A compiling of consecutive years is not possible when studying the impacts of resource extraction because a carry over effect is present and the impact of development compounds each year. Thus, reductions in fawn survival would need to be captured on a yearly basis. A preliminary power analysis (α=.05, β=.20) using estimates of fawn survival of µ=.444 and SD=.217 (Unsworth et al. 1999) indicated that 60 marked fawns per study area would be needed to have an 80% chance of detecting a 15% change in survival. However, a 15% change in survival is unrealistic in the face of published literature. A more likely expectation is in the realm of a 5%-10% decrease. Thus, more appropriate sample size estimates are between 125 (to detect a 10% change) and 480 (to detect a 5% change) for each of 2 study areas (annual sample sizes would thus range from 250-960 individuals). 99 �Due to a lack of dialogue with personnel from natural gas and oil companies, the DOE, the Colorado Oil and Gas Conservation Commission and the Southern Ute Indian Reservation, the possibility of external funding for this type of research is unknown. Based on past efforts of these entities, it is believed that some level of financial support is possible. The issue of development is addressed by high priority achievement H-1.3 of the CDOW Strategic Plan (2002-2007). While the impacts of natural gas and oil extraction development, per se, are not addressed, they do qualify as a developmental issue of concern. SUMMARY Mule deer/elk interactions, bighorn sheep translocation, and impacts of natural gas and oil extraction are all topics that present suitable and needed research investigations. At this time, moving forward on mule deer/elk interactions appears to be the most reasonable course of action. Cooperative commitments between industry and the State of Colorado appear needed before research could be initiated to meaningfully assess the impacts of natural gas extraction on mule deer populations LITURATURE CITED Bailey, J.A. 1990. Mangement of rocky mountain bighorn sheep herds in Colorado. Colorado Division of Wildlife: Special Report No. 66. Bartmann, R.M., White, G.C. and L.H. Carpenter. 1992. Compensatory mortality in a Colorado mule deer population. Wildlife Monographs No. 121. Bishop, C.J. 2000. Effects of habitat enrichment on mule deer recruitment and survival rates. Program Narrative, Colorado Division of Wildlife. Bishop, C.J. 2003. Effect of nutrition and habitat enhancements on mule deer recruitment and survival rates. Colorado Division of Wildlife: Job Progress Report. Cameron, R.D., Reed, D.J., Dau, J.R., and W.T. Smith. 1992. Redistribution of calving caribou in response to oil field development on the arctic slope of Alaska. Arctic, 45: 338-342. Cronin, M.A., Whitlaw, H.A., and W.B. Ballard. 2000. Northern Alaska oil fields and caribou. Wildlife Society Bulletin, 28: 919-922. Dyer, S.J., O'Neill, J.P., Wasel, S.M. and S. Boutin. 2001. Avoidance of industrial development by woodland caribou. Journal of Wildlife Management, 65: 531-542. Gill, R.B. 2001. Declining mule deer populations in Colorado: Reasons and Responses. Colorado Division of Wildlife: Special Report No. 77. Goodson, N.J. 1982. Effects of domestic sheep grazing on bighorn sheep populations: a review. Proceedings of the biennial symposium of the Northern Wild Sheep and Goat Council 3:287-313. Hayes, C.L., Rubin, E.S., Jorgensen, M.C., Botta, R.A. and W.M. Boyce. 2000. Mountain lion predation of bighorn sheep in the Peninsular Ranges, California. Journal of Wildlife Management 64:954959. Holl, S.A., Bleich, V.C. and S.G. Torres. 2004. Population dynamics of bighorn sheep in the San Gabriel Mountains, California, 1967-2002. Wildlife Society Bulletin 32:412-426. Houston, D.B. 1982. The northern Yellowstone elk herd: ecology and management. Macmillan Publishing Co., New York, New York, USA. Jedrzejewska, B. and W. Jedrzejeswki. 1998. Predation in vertebrate communities: the Bialowieza Primeval Forest as a case study. Springer-Verlag, New York, New York, USA. Kirchoff, M.D. and D.N. Larsen. 1998. Dietary overlap between native Sitka black-tailed deer and introduced elk in southeast Alaska. Journal of Wildlife Management 62:236-242. Krebs, C.J., S. Boutin and R. Boonstra. 2001. Ecosystem dynamics of the boreal forest: the Kluane project. Oxford University Press, Oxford, England. McCarty, C.W. and M.W. Miller. 1998. Modeling the population dynamics of bighorn sheep: a synthesis of literature. Colorado Division of Wildlife: Special Report No. 73. 100 �Miller, M.W., J.E. Vayhinger, D.C. Bowden, S.P. Roush, T.E. Verry, A.N. Torres and V.D. Jurgens. 2000. Drug treatment for lungworm in bighorn sheep: Reevaluation of a 20-year-old management prescription. Journal of Wildlife Management 64:505-512. Nellemann, C. and R.D. Cameron. 1998. Cumulative impacts of an evolving oil-field complex on the distribution of calving caribou. Canadian Journal of Zoology, 76: 1425-1430. Nellemann, C., Vistnes, I., Jordhøy, P., Strand, O., and A. Newton. 2003. Progressive impact of piecemeal infrastructure development on wild reindeer. Biological Conservation, 113: 307-317. Sinclair, A.R.E., and M. Norton-Griffiths. 1979. Serengeti: dynamics of an ecosystem. University of Chicago Press, Chicago, IL, USA. Singer, F.J., Bleich, V.C. and M.A. Gudorf. 2000a. Restoration of bighorn sheep metapopulations in and near western national parks. Restoration Ecology 8(4 special supplement):14-24. Singer, F.J., Moses, M.E., Bellew, S. and W. Sloan. 2000b. Correlates to colonizations of new patches by translocated populations of bighorn sheep. Restoration Ecology 8(4 special supplement):66-74. Singer, F.J. and J.E. Norland. 1994. Niche relationships within a guild of ungulate species in Yellowstone National Park, Wyoming, following release from artificial controls. Canadian Journal of Zoology 72:1383-1394. Singer, F.J., Williams, E., M.W. Miller and L.C. Zeigenfull. 2000c. Population growth, fecundity, and survivorship in recovering populations of bighorn sheep. Restoration Ecology 8(4 special supplement):75-84. Stewart, K.M, Bowyer, R.T., Kie, J.G., Dick, B.L. and M. Ben-David. 2003. Niche partitioning among mule deer, elk, and cattle: Do stable isotopes reflect dietary niche? Ecoscience 10:297-302. Unsworth, J.W., Pac, D.F., White, G.C. and R.M. Bartmann. 1999. Mule deer survival in Colorado, Idaho and Montana. Journal of Wildlife Management 63:315-326. Wehausen, J.D. 1996. Effects of mountain lion predation on bighorn sheep in the Sierra Nevada and Granite Mountains of California. Wildlife Society Bulletin 24:471-479. White, G.C., D.J. Freddy, R.B. Gill and J.H. Ellenberger. 2001. Effect of adult sex ratio on mule deer and elk productivity in Colorado. Journal of Wildlife Management 65:543-551. Prepared by ______________________ Eric J. Bergman, Wildlife Researcher 101 � https://cpw.cvlcollections.org/files/original/dd6011b6a5d8b5f4d9d7bae2220575de.pdf 7fc4c9456bf7e6bb238d96c0e5d4e3a2 PDF Text Text 163 JOB PROGRESS REPORT Stateof_ _ _ _ _--=C=o=lo=r=ad=o=------ Division of Wildlife - Mammals Research Work Package No. _ _ _ _ _ _ _ _ _ __ Multispecies Investigations Task No. _ _ _ _ _ _ _ _ _ _ _ _ __ Prairie Dog Research and Wildlife Extension Period Covered: July 1, 2002 - June 30, 2003 Author: W. F. Andelt, Colorado State University, Dept. Fishery and Wildlife Biology Personnel: M. Christopher, J. Dennis, L. Gepfert, E. Hollowed, BLM, S. M. Quinlivan, P. M. Schnurr, A. Seglund, Utah Division of Wildlife Resources, G. C. White, D. Younkin PRAIRIE DOG AND PREDATOR-GROUSE RESEARCH, AND WILDLIFE EXTENSION W. F. Andelt OBJECTIVES I. Objectively assess and document the current scientific knowledge base about Gunnison's prairie dogs by I September 2002 via a technical review draft publication, submitted to the CDOW research peer review process. 2. Conduct on-the-ground surveys, and collect measurements of key elements of Gunnison's prairie dog colonies at 50 sites in western Colorado by September 30, 2002, and provide a report, including data summaries, by October 30, 2002, to CDOW's project leader. By October 30, 2002, provide a data set that can be used by other investigators to develop a defensible, quantified Gunnison's prairie dog inventory technique. 3. Provide information specifically directed toward chronic wasting disease from DOW/DNR to the public through CSU's Extension network of 57 county Extension offices and provide intensive training to at least 4 offices and I 00 employees/volunteers in key western slope counties by April 30, 2003. 4. Provide general wildlife information and information regarding human-wildlife conflicts from DOW/DNR to the public through the CSU's Extension network of 57 county Extension offices and provide intensive training to at least 4 offices and 100 employees/volunteers by April 30, 2003. 5. Provide analyses of data on the possible role of predators in the sage grouse decline in northwest Colorado. �164 STATUS OF GUNNISON'S PRAIRIE DOGS IN COLORADO W. F. Andelt The Gunnison's prairie dog (Cynomys gunnisoni) occurs in Colorado, Arizona, New Mexico, and Utah. Their geographical range probably has not changed much during the past century (Knowles 2002). However, acreage of Gunnison's prairie dogs within their range likely has contracted during the past century. The extent of decline is unknown because there were no accurate accounts of the abundance of prairie dogs prior to settlement (Clark 1973, Anderson et al. 1986, Knowles 2002), and the abundance of Gunnison's prairie dogs today also is not well known. Approximately 22% of the range of Gunnison's prairie dog occurs in Colorado (Knowles 2002), where it is distributed primarily across the southwestern quarter of the state at elevations of 6,000 to 12,000 feet (Fitzgerald et al. 1994). The Gunnison's prairie dog consists of 2 subspecies (C. g. gunnisoni and C. g. zuniensis). In Colorado C. g. gunnisoni occurs in the Gunnison River drainage, the upper Arkansas and South Platte drainages, and in the San Luis Valley (Fitzgerald et al. 1994). In Colorado, C. g. zuniensis occurs at lower elevations in Montezuma, La Plata, Dolores, San Miguel, and Montrose counties (Fitzgerald et al. 1994). Densities of Gunnison's prairie dogs range from 5 to 10 per acre (Knowles 2002). The primary threat to Gunnison's prairie dogs is plague (Yersinia pestis), whereas poisoning, recreational shooting, agricultural land conversion, and urbanization are of secondary importance (Knowles 2002). Plague became apparent in Gunnison's prairie dog colonies during the late 1940s (Lechleitner et al. 1968, Cully 1993). Plague often kills >99% of Gunnison' s prairie dogs (Lechleitner et al. 1968). South Park, Colorado apparently contained 913,000 acres of Gunnison's prairie dogs in 1941, but an epizootic of sylvatic plague entered this area in 1947, and by 1949 plague reduced the acreage of prairie dogs by 95% (Ecke and Johnson 1952, Fitzgerald 1969, Armstrong 1972). Plague has continued in this area during the 1950s and 1960s (Lechleitner et al. 1962, Fitzgerald and Lechleitner 1974). During the first half of the 20 th century, Gunnison's prairie dogs were mostly eliminated from the major valleys in Colorado (Burnett and McCampbell 1926, Longhurst 1944) due to plague or poisoning (Knowles 2002). Recently, most wildlife biologists interviewed by Knowles (2002) felt that plague was the dominant controlling factor of prairie dogs. Recover of Gunnison's prairie dogs from plague appears to range from no recovery to a pattern where colonies are regularly lost, but new colonies appear and grow in other areas (Knowles 2002). Gunnison' s prairie dogs were subject to poisoning in the higher valleys of Colorado during the 1950s (Lechleitner et al. 1968). Control of Gunnison' s prairie dog continues on private land, but control of prairie dogs on Federal lands currently does not appear to be a conservation issue (Knowles 2002). The current abundance of Gunnison's prairie dog in Colorado is not well known. Some biologists (Fitzgerald 1991), environmental proponents, and other individuals have expressed concern that populations of Gunnison' s prairie dogs have been reduced by epizootics of plague (Lechleitner et al. 1962, 1968; Fitzgerald 1969, 1978, 1993; Rayor 1985), and control of prairie dogs (Fitzgerald 1991) in Colorado. Speculation exists that the Gunnison's prairie dog might be petitioned for listing as threatened or endangered. Decisions to list the Gunnison's prairie dog should be based upon the most accurate and most current data. In this report, I summarize information from various sources about the status of Gunnison's prairie dog in Colorado. �165 Colorado Agricultural Statistics Service (1990) Survey Colorado Agricultural Statistics Service (1990) surveyed 9,046 farmers and ranchers and obtained nearly 3,000 surveys to estimate that 1,553,000 acres were occupied by prairie dogs in Colorado during 1989. This survey estimated acres occupied by prairie dogs in each county, but it did not differentiate between acres occupied by Gunnison's prairie dogs, black-tailed prairie dogs (Cynomys ludovicianus), and white-tailed prairie dogs (Cynomys leucurus). Thus, I used distribution maps in Fitzgerald et al. (1994) to ascertain which counties were occupied by the 3 species of prairie dogs. In counties where Gunnison's prairie dogs overlapped with 1 of the other species of prairie dogs, I estimated the relative proportion of the county that was occupied by Gunnison's prairie dogs. I multiplied that proportion by the acreage reported occupied by all prairie dogs in a county to obtain an estimate of the acreage occupied by Gunnison's prairie dogs for that county. I summed the acres ofreported Gunnison's prairie dogs in each county and obtained an estimated 445,500 acres ofreported Gunnison's prairie dogs in Colorado during 1989 (Table 1). Jim Fitzgerald (1991) letter to Galen Buterbaugh, U.S. Fish and Wildlife Service Fitzgerald (1991) expressed concern about the status of the gunnisoni subspecies of the Gunnison's prairie dog. He indicated that plague and poisoning have eliminated almost all populations in South Park. He also indicated populations appear to be in poor condition in the San Luis Valley, they appear to be gone from the extreme upper Arkansas River valley, and populations appear to be small and patchy in other parts of its historic range in Colorado. He believed Gunnison's prairie dogs are gone from Jefferson, Douglas, and Lake Counties in Colorado. He noted that a large complex exists on the Curecante National Recreation Area west of Gunnison, Colorado. Fitzgerald (1991) sent inquiries to all Colorado Division of Wildlife District Wildlife Managers and Wildlife Biologists and reported that a disappointing number of colonies were identified. He indicated that the low number of reports of colonies sent to him by the Colorado Division of Wildlife and his low estimates are in direct contrast to acreage of Gunnison' s prairie dogs reported by Colorado Agricultural Statistics Service ( 1990). Robert Finley (1991) Survey of Distribution and Status of Gunnison's Prairie Dogs in Colorado Finley ( 1991) conducted a broad reconnaissance survey of the distribution of Gunnison' s prairie dogs by driving some highways and roads and recording observations of prairie dogs. He observed 74 Gunnison's prairie dog colonies, of which 42 were active. He recorded colonies in 10 counties. He reported the largest active colonies were in the Gunnison drainage. He reported that South Park was almost devoid of prairie dogs, but he found a medium sized colony near Hartsel and a few on the periphery. He indicated that some mammalogists suspect that the spread of Wyoming ground squirrels southward through Colorado, after prairie dogs die out from plague, may be preventing prairie dogs from repopulating their former towns east of the Continental Divide and north of the Arkansas River. Finley (1991) concluded that populations of Gunnison's prairie dogs "seem to be far below those reported in the years prior to plague epizootics", "but I do not feel that the present situation is serious enough to warrant protection by Threatened status." �166 Mike Threlkeld, Chief of Rodent Control, Colorado Department of Agriculture (Personal Communication, 11 June 2002) Mike Threlkeld indicated that there are large acreages of Gunnison's prairie dogs around Cortez (perhaps 7,000 acres), Dolores, Montrose (perhaps 7,000 acres), Blue Mesa Reservoir, between Dove Creek and Nucla/Naturita, west of Canyon City, north of Salida, and on the Ute Mountain Indian Reservation (perhaps over 7,000 acres). Colorado Division of Wildlife (2002) Report on Acreage of Gunnison's Prairie dogs Field personnel from the Colorado Division of Wildlife, Forest Service, and the Bureau of Land Management placed Gunnison's prairie dog colonies on 1:50,000 US Geological Survey County sheets during July and August, 2002 (Colorado Division of Wildlife 2002). The colonies were assigned as active (prairie dogs know to be present in the last 3 years) or unknown status (prairie dogs have been active but current presence in the area is unknown and requires field verification). From this exercise, the Colorado Division of Wildlife (2002) reported 85, 795 acres of active and 194,777 of unknown acres of Gunnison's prairie dogs in Colorado (fable 1). In addition, 53,832 acres of active prairie dogs were identified in Delta County, where it was not know if these acres represented Gunnison's or white-tailed prairie dogs. These acreages are considered preliminary minimum estimates of the number of acres occupied by Gunnsion's prairie dogs. Craig Knowles (2002) Report on Status of White-tailed and Gunnison's Prairie Dogs Knowles (2002) primarily summarized Colorado Division of Wildlife (2002) for his assessment of the current status of Gunnison's prairie dogs in Colorado. He criticized the Colorado Agricultural Statistics Service (1990) report of acreage of prairie dogs in Colorado by stating " ... these estimates clearly greatly inflate the acreage at least in some counties." However, it is worth noting that Knowles (1998) reported that there was only 44,000 acres of black-tailed prairie dogs in Colorado during 1998, whereas the Colorado Agricultural Statistics Service (1990) estimated about 930,000 (calculated from their report). Recent aerial surveys by the Colorado Division of Wildlife (following Sidle et al. 2001) indicate that there are about 631,000 acres occupied by black-tailed prairie dogs in Colorado (F. Pusaterie, personal communication). Thus, the estimates provided by Colorado Agricultural Statistics Service (1990) were much closer than Knowles (1998) to the acreage reported by the Colorado Division of Wildlife. Knowles (2002) indicated that Gunnison's prairie dog populations in Colorado were greatly reduced by plague and poisoning during the 1900s and this decline may be continuing, or at best, the populations may be stable. Synthesis of Reports on Abundance of Gunnison's Prairie Dogs in Colorado Abundance of Gunnison's prairie dogs likely has declined in Colorado, particularly starting during the 1940s when plague became endemic. Our best estimates of the acreage ofGunnison's prairie dogs in Colorado seem to be provided by Colorado Division of Wildlife (2002) and Colorado Agricultural Statistics Service (1990). The Colorado Division of Wildlife (2002) reports a preliminary minimum of 85,700 acres of active Gunnison's prairie dogs, another 194,800 acres ofGunnison's prairie dogs where their status is unknown, and another 53,800 acres of prairie dogs in Delta County which are either Gunnison's or white-tailed prairie dogs. The Colorado Agricultural Statistics Service (1990) survey of acreage of prairie dogs in Colorado during 1989, from which I derived 445,500 acres of reported Gunnison's prairie dogs, has been criticized as biased by Knowles (1998, 2002). However, Colorado Agricultural Statistics Service (1990) and Colorado Division of Wildlife (2002) seem to concur at least to some extent. The Colorado Division of Wildlife is assessing the feasibility of aerial surveys for estimating acreage of Gunnison's prairie dogs in Colorado. Pending feasibility, these surveys are needed to provide better estimates of the acreage of Gunnison's prairie dogs in Colorado. �167 LITERATURE CITED Anderson, E. A., S. C. Forrest, T. W. Clark, and L. Richardson. 1986. Paleobiology, biogeography, and systematics of the black-footed ferret, Mustela nigripes (Audubon and Bachman), 1851. Great Basin Naturalist Memoirs 8: 11-62. Armstrong, D. M. 1972. Distribution of mammals in Colorado. Museum of Natural History, University of Kansas Monograph 3. 415pp. Burnett, W. L., and S. C. McCampbell. 1926. The Zuni prairie dog in Montezuma County, Colorado. Office of State Entomologist, Colorado Agricultural College, Fort Collins, Colorado. Circular 49, 16pp. Clark, T. W. 1973. Prairie dogs and black-footed ferrets in Wyoming. Pages 88-101 in Linder, R. L., and C. N. Hillman, editors. Proceedings of the black-footed ferret and prairie dog workshop. South Dakota State University, Brookings, South Dakota. Colorado Agricultural Statistics Service. 1990. Vertebrate rodent infestation survey. Colorado Department of Agriculture, Lakewood, Colorado. Colorado Division of Wildlife. 2002 (September). Report of acreages of active colonies for Gunnison's prairie dogs (Cynomys gunnisoni) and white-tailed prairie dogs (Cynomys leucurus). 2pp. Cully, J. F., Jr. 1993. Plague, prairie dogs, and black-footed ferrets. Pages 38-49 in J. L. Oldemeyer, D. E. Biggins, B. J. Miller, and R. Crete, editors. Proceedings of the symposium on the management of prairie dog complexes for the reintroduction of the black-footed ferret. U.S. Fish and Wildlife Service Biological Report No. 13. Ecke, D. H., and C. W. Johnson. 1952. Plague in Colorado. Part I. Plague in Colorado and Texas. U.S. Public Health Service, Public Health Monograph 6:1-54. Finley, R. B., Jr. 1991. Survey of present distribution and status of Cynomys gunnisoni gunnisoni in Colorado. Unpublished manuscript. 9pp. Fitzgerald, J.P. 1969. Sylvatic plague in Gunnison's prairie dog (Cynomys gunnisoni) and associated mammals in South Park, Colorado. Journal of the Colorado Wyoming Academy of Science 7:45. Fitzgerald, J. P. 1978. Plague (Yersinia pestis) epizootics in introduced Gunnison's prairie dogs: implications for prairie dog management. New Mexico Academy of Science Bulletin 18:40. Fitzgerald, J. P. 1991. Letter to Galen Buterbaugh, Regional Director, U.S. Fish and Wildlife Service. Dated 1 April 1991. Fitzgerald, J. P. 1993. The ecology of plague in Gunnison's prairie dogs and suggestions for the recovery of black-footed ferrets. Pages 50-59 in J. L. Oldemeyer, D. E. Biggins, B. J. Miller, and R. Crete, editors. Proceedings of the symposium on the management of prairie dog complexes for the reintroduction of the black-footed ferret. U.S. Fish and Wildlife Service Biological Report No. 13. Fitzgerald, J.P., and R.R. Lechleitner. 1974. Observations on the biology ofGunnison's prairie dog in central Colorado. American Midland Naturalist 92:146-163. �168 Fitzgerald, J.P., C. A. Meaney, and D. M. Armstrong. 1994. Mammals of Colorado. University Press of Colorado, Niwot, Colorado. 488pp. Knowles, C. J. 1998. Status of the black-tailed prairie dog. Unpublished manuscript prepared for U.S. Fish and Wildlife Service, Pierre, South Dakota. 12pp. Knowles, C. 2002. Status of white-tailed and Gunnison's prairie dogs. National Wildlife Federation, Missoula, Montana and Environmental Defense, Washington, DC. 30pp. Lechleitner, R.R., L. Kartman, M. I. Goldenberg, and B. W. Hudson. 1968. An epizootic of plague in Gunnison's prairie dogs (Cynomys gunnisoni) in south-central Colorado. Ecology 49:734-743. Lechleitner, R. R, J. V. Tileston, and L. Kartman. 1962. Die-off of a Gunnison 's prairie dog colony in central Colorado, I. Ecological observations and description of the epizootic. Zoonoses Research 1:185-199. Longhurst, W. 1944. Observation on the ecology of the Gunnison prairie dog in Colorado. Journal of Mammalogy 25:24-36. Rayor, L. S. 1985. Dynamics of a plague outbreak in Gunnison's prairie dogs. Journal ofMammalogy 66:194-196. Sidle, J. G., D. H. Johnson, and B. R. Euliss. 2001. Estimated areal extent of colonies of black-tailed prairie dogs in the northern great plains. Journal of Mammalogy 82:928-936. �169 Table I. Acres of Gunnison's prairie dogs reported and estimated from the Colorado Agricultural Statistics Service (1990) survey during l 989 and estimated by Colorado Division of Wildlife (2002) during 2002 in Colorado. Colorado Agricultural Statistics Service survey Acres of all County Qrairie dogs Alamosa 6,200 48,900 Archuleta 3,200 Chaffee 20,500 Conejos Costilla 1,600 Custer 5,900 Delta 52,500 Dolores 56,000 Douglas 12,600 16,700 El Paso Fremont 15,300 5,800 Gunnison 300 Hinsdale Huerfano 6,400 Jefferson 1,700 Lake 900 La Plata 80,000 Las Animas 18,500 Mineral 200 Montezuma 92,000 Montrose 52,100 Ouray 7,400 Park 5,100 Rio Grande 14,300 Saguache 13,200 San Juan 13,400 San Miguel Teller 5,200 TOTAL: 555,900 Proportion acres• occupied by Gunnison's Q.dogs 1.0 1.0 1.0 1.0 1.0 1.0 0.12 1.0 0.25 0.05 1.0 1.0 1.0 0.63 0.31 1.0 1.0 0.2 1.0 1.0 0.73 0.5 1.0 1.0 1.0 1.0 1.0 1.0 Acres of Colorado Division Gunnison's of Wildlife Qrairie dogs Active acres Unknown acres 6,200 2 12,220 48,900 18,226 15,978 3,200 2,467 0 20,500 4,707 67,218 1,600 14,948 25,439 5,900 6,300 56,000 3,363 2,549 3,150 58 0 835 15,300 5,800 611 221 300 4,032 527 900 80,000 6,816 619 3,700 200 449 1,221 92,000 12,223 0 38,033 6,482 0 3,700 647 0 5,100 42 3,150 14,300 12,263 2,094 13,200 2,659 58,891 13,400 5,200 448,277 2,017 ____fil_ 85,795 2,927 - -0 194,777 •obtained by estimating the proportion of a county (from Fitzgerald et al. 1994) that was occupied by Gunnison's prairie dogs, white-tailed prairie dogs, and black-tailed prairie dogs, and then dividing the proportion for Gunnison's prairie dogs by the sum of proportions for all 3 species. �170 EVALUATION OF AERIAL SURVEYS FOR ESTIMATING ACREAGE OF GUNNISON'S AND WHITE-TAILED PRAIRIE DOGS IN COLORADO AND UTAH W. F. Andelt, P. M. Schnurr, and A. Seglund During November 2002, we (Andelt and Schnurr 2002) reported our assessment of 3 survey techniques, including ground surveys, interpretation of satellite imagery (Sidle et al. 2002), and aerial surveys (Sidle et al. 2001 ), for obtaining a valid estimate of the distribution and acreage of Gunnison' s prairie dogs (Cynomys gunnisoni) in Colorado. We concluded that ground surveys likely would be very difficult, if not impossible to implement for obtaining a valid scie~tific estimate of acreage of Gunnison' s prairie dogs in Colorado. However, we recognized that ground surveys could be used to provide an estimate of the minimum acreage of Gunnison's prairie dogs in Colorado. We concluded that satellite imagery is very expensive ($2,000 per 36 mi2 or $2,880 per 100 mi 2 of digital imagery [John Norman, Natural Resources Ecology Lab, CSU; personal communication]), the imagery would need to be interpreted and verified, activity of prairie dog towns would need to be ascertained on the ground, and it is unknown if the technology would be suitable in rolling terrain. Aerial surveys, using line intercept methodology, have been used to estimate area occupied by black-tailed prairie dogs (Cynomys ludovicianus) (Sidle et al. 2001, J. Dennis and F. Pusaterie, Colorado Division of Wildlife; personal communication). We concluded that the technique held promise for estimating acreage of Gunnison' s prairie dogs in Colorado. In this paper, we report on our current progress in evaluating aerial surveys for estimating acreage of Gunnison's and white-tailed prairie dogs (Cynomys leucurus) in Colorado and Utah. Initially, on 13 June 2002, William Andelt accompanied Jim Dennis and Dave Younkin on an aerial survey of black-tailed prairie dogs to gain additional familiarity with the technique. On 24 June 2002, William Andelt and Larry Gepfert, CDOW, flew over the 32 active Gunnison's prairie dog colonies reported by Joe Cappodice. With the aid of a GPS unit, all colonies were located, although some of the smaller colonies were somewhat difficult to observe. We ascertained that aerial surveys appear to have potential for establishing distribution of Gunnison' s prairie dogs and that further investigation of the technique was merited. However, because of some difficulty in observing some colonies, we, in collaboration with Gary White, decided that future test flights should also obtain photos of prairie dog colonies; classify colonies as being located in grassland, short shrubs, tall shrubs, or agriculture; rank the colonies as barely detectable, detectable, or highly detectible; and classify colonies as active, inactive, or unknown. Our plans were to use these data to estimate detection probabilities for the various categories of colonies. We then planned to use the detection probabilities to correct acreages of prairie dog colonies observed from the air (White 2002). Subsequently, during summer 2002, Pam Schnurr and Gary White met with Amy Seglund and Bill Bates, biologists with the Utah Division of Wildlife Resources (UDWR). Both states agreed to coordinate and cooperate to further ascertain the feasibility of aerial surveys to estimate acreage of Gunnison's and white-tailed prairie dogs, and to develop detection probabilities for both species. METHODS We entered the boundaries of known Gunnison's and white-tailed prairie dog colonies in both Colorado and Utah into GIS Arc/Info. We established 31, 17, 19, and 11 transects across these Gunnison's and white-tailed prairie dog colonies in Colorado and Utah, respectively. These transects were established across known colonies in both states along with a number of control transects (i.e. transects over areas without colonies). Beginning and ending UTM coordinates were ascertained for each transect and placed in a spreadsheet. We hired and trained a ground crew that verified the distribution of all white-tailed prairie dog colonies on the transects in Colorado. �171 Jim Dennis and Dave Younkin, CDOW, and Brad Crompton and Craig Hunt, from the Utah Division of Wildlife Resources flew all 4 sets of transects and obtained GPS coordinates for the beginning and end of prairie dog colonies on the transects. The crew from Colorado had extensive experience surveying black-tailed prairie dogs, whereas the crew from Utah had extensive experience with aerial surveys of wildlife, other than prairie dogs. The Utah and Colorado survey teams flew the transects in opposite directions. We plotted the endpoints of the prairie dog colonies that were ascertained by both aerial crews on all transects in GIS Arc/Info. We used Arc/Info to determine the lengths of each colony on each transect and then entered these data in a spreadsheet. We summed the lengths of colonies ascertained on the ground and from the air on each transect. We analyzed these data in SAS using Proc GLM to determine the effect of aerial team, rating of colony visibility, and rating of habitat type on the proportion of colonies observed on aerial versus ground surveys. We censored transects without prairie dogs known on ground surveys, and then used Spearman Correlation (Proc CORR) analyses to ascertain correlations for proportion of colonies observed, ratings of visibility, and ratings of habitat types between the 2 aerial crews. We also used Spearman Correlation analyses to ascertain correlations between ratings of visibility of colonies and proportion of colonies detected, and ratings of habitat types and proportion of colonies detected. RESULTS The Colorado and Utah teams overestimated lengths of Gunnison's prairie dog colonies on transects in Colorado and Utah (Table l ). Both teams also overestimated lengths of white-tailed prairie dog colonies on the white-tailed site in Utah. In contrast, the Colorado team underestimated lengths of colonies on the white-tailed site in Colorado. Although the Utah team closely estimated the overall average lengths of colonies on this site, we found considerable variation between total lengths of colonies on transects observed by this team versus those known on the ground. The Utah aerial team (x = 5.3; S.E. = 1. ll), compared to the Colorado team (x = 2.3; S.E. = 0.36), observ'ed a greater proportion oflengths of colonies on transects (Tables 1, 2), however both teams significantly overestimated the lengths of e<;>lonies compared to the lengths ascertained on the ground. The proportion of length of prairie dog colonies observed from the air compared to the lengths ascertained from the ground were not related to ratings of visibility nor to ratings of habitat types observed from the air (Table 2). Proportion of lengths of prairie dog colonies detected by aerial crews from Colorado and Utah were weakly correlated (Table 3). However, ratings of visibility of colonies and ratings of type of habitat found on transects of colonies were not correlated between the Colorado and Utah aerial crews. The 2 crews did not consistently report finding prairie dogs in the same areas along the same transect. This may partially explain the differences between the 2 crews in their ratings of visibility of colonies and rating of habitat types on transects. Proportions of lengths of colonies detected by aerial crews were not correlated with rating of visibility of colonies on transects (Table 4 ). The greatest proportions of lengths of colonies were detected by aerial crews on transects described as grasslands followed by transects described as short shrubs and then followed by transects described as tall shrubs (Table 4). The Colorado team rated prairie dogs on 76% of 51 transects as active, 12% as unknown, and 12% as a combination of active and unknown. The Utah team rated prairie dogs on 28% of 63 transects as active, 2% as inactive, 57% as unknown, and 25% as a combination of active, inactive, and unknown. �172 DISCUSSIONS AND RECOMMENDATIONS We recognize a number of goals when inventorying prairie dogs. We believe the most important goal is to obtain accurate and repeatable estimates (i.e. low variation within and among survey crews) of the acreage of Gunnison' s and white-tailed prairie dogs. Low variation among survey crews is necessary so that differences between estimates of acreage are actually related to increases or decreases in acreage of prairie dogs rather than differences between crews. Another goal for inventorying prairie dogs is to establish minimum acreages of prairie dogs which we can relate to their status and decisions about listing them as threatened or endangered. Our goal has been to ascertain the feasibility of aerial surveys for estimating acreage of Gunnison's and white-tailed prairie dogs in Colorado and Utah. We envisioned this as a multi-step process. We first flew over known Gunnison's prairie dog colonies and noted that many of the colonies were visible from the air. Next, we arranged aerial surveys by crews from Colorado and Utah to estimate . the length of colonies on transects where the distribution of prairie dogs were known to us, but unknown to the crews. Accuracy of aerial surveys was not sufficient to estimate detection probabilities. We found significant variation between the 2 aerial teams in estimates of lengths of prairie dog colonies on transects, however these estimates were weakly correlated between the 2 teams. Shortly after completing the aerial flights and before data were compiled, Jim Dennis noted that his team likely could have more accurately estimated lengths of prairie dog colonies by conducting some flights followed by ground reconnaissance of the same transects to verify what they were observing from the air (see Appendix 1). We anticipate this training would enhance accuracy of estimates. We recommend that training, or other methods to improve estimates between teams, are needed before broad scale surveys are conducted. The large variation between teams in our study indicate that, without improving accuracy and consistency between teams, it would be difficult to ascertain even moderate changes in acreages of prairie dogs. The Colorado and Utah teams surveyed the Colorado white-tailed prairie dog site on 20 September and 28 August 2002, respectively. The Colorado team rated 10 of the transects as active and 2 as unknown. The Utah team rated 4 transects as active, l as inactive, 5 as unknown, and 4 as activeinactive or active-unknown. We surveyed part of the Colorado white-tailed site from the ground on 23 September 2002 and found very little sign of activity by prairie dogs. Thus, we recommend that ground crews verify ratings of activity on a random sample of future transects. If aerial crews are unable to accurately determine activity, a ground crew will need to verify activity on a random portion of transects on future surveys. We reviewed potential causes for why estimates oflengths of prairie dog colonies varied between ground surveys and aerial surveys, and between the 2 aerial crews. We closely surveyed the distribution of prairie dogs on the white-tailed sites in Colorado and Utah, but additional verification on the ground is needed for the 2 Gunnison's prairie dog sites to insure that accuracy of ground surveys is not a cause of error. Coordinates of prairie dog colonies were recorded on the ground and by the Utah team in the NAD27 datum. The Colorado team used the WGS84 datum when they flew the transects. The use of the WGS84 resulted in the Colorado team being 38 to 219 m off the actual transect, depending on the study area and direction of flight (east-west versus north-south). Although we initially suspected that the 38 to 219 m away from transects resulted in some errors, our review of the data suggested that accuracy appeared similar when the airplane was on the transect versus away from the transect. The Utah team strayed over 1,000 m from portions of 4 transects which likely attributed to some errors. �173 We recognize 2 general approaches (ground vs. aerial surveys) for continuing surveys of Gunnison's and white-tailed prairie dogs. To continue aerial surveys, we recommend that the distribution of prairie dogs is more accurately verified on the ground on the 2 Gunnison's prairie dog sites. If distributions are different than what is currently known, the distribution of prairie dogs on aerial and ground surveys should be compared again. Then, we recommend training aerial crews by conducting flights over short transects over some colonies and then surveying the colonies from the ground so that they can better ascertain what they are observing from the air. After this training, we recommend reflying the previous transects to ascertain if accuracy can be improved. If accuracy cannot be improved, we recommend discontinuing aerial surveys. An alternative to surveying prairie dogs from the air would be to continue Pam Schnurr's earlier work of meeting with biologists to plot known distribution of Gunnison' s and white-tailed prairie dogs on maps. A ground crew should then verify a random portion of these distributions. Although this alternative likely would cost less than aerial surveys, it likely would underestimate acreage of prairie dogs and would not provide an adequate and repeatable sample for future comparisons. However, this methodology might be sufficient for considerations of listing prairie dogs as threatened or endangered. LITURATURE CITED Andelt, W. F., and P. Schnurr. 2002. Progress report: inventorying Gunnison's prairie dogs in Colorado. Unpublished Report submitted to the Colorado Division of Wildlife, Denver, Colorado. Sidle, J. G., D. H. Johnson, and B. R. Euliss. 2001. Estimated areal extent of colonies of black-tailed prairie dogs in the northern great plains. Journal of Mammalogy 82:928-936. Sidle, J. G., D. H. Johnson, B. R. Euliss, and M. Tooze. 2002. Monitoring black-tailed prairie dog colonies with high-resolution satellite imagery. Wildlife Society Bulletin 30:405-411. White, G. C. 2002. Memorandum to Pam Schnurr, Bill Bates, and Amy Seglund, Colorado Division of Wildlife and Utah Division of Wildlife Resources. Dated July I, 2002. �174 Table l. Average length (m) of Gunnison's and white-tailed prairie dog colonies, observed from the ground and reported by aerial survey crews from the Colorado Division of Wildlife and the Utah Division of Wildlife Resources, on transects surveyed in Colorado and Utah during August, September, and November 2002. Transects Avg. Avg. length of colonies/transect• Proportion of colony length observedb Team survey N lenlrth Colo 9/19-20 31 8,671 Utah 10/l 31 8,671 Colo 9/20 17 5,446 Utah 8/28 17 5,446 Colo 9/23 19 10,660 19 10,660 Utah 8/28 9/24 Colo 11 40,403 11 40,403 Utah 8/26 Ground Aerial 264 723 246 1,5 l l 1,955 1,202 1,984 1,955 424 1,770 424 3,406 2,912 9,714 2,912 5,418 N 18 18 14 14 11 11 8 8 X 2.6 8.4 0.7 1.8 3.5 7.5 2.7 1.7 S.E. 0.65 2.57 0.16 0.80 0.97 2.09 0.85 0.44 51 51 2.3 5.3 0.36 1.11 Date of Area Colorado Colorado Colorado Colorado Utah Utah Utah Utah S12ecies Gunnison's Gunnison's White-tailed White-tailed Gunnison's Gunnison's White-tailed White-tailed TOTAL: Colo Utah 78 78 12,928 12,928 1,045 1,045 2,350 2,626 aRepresents average length of colonies known primarily from ground reconnaissance, and estimated from aerial surveys on transects with and without prairie dog colonies. bRepresents proportion of length of prairie dog colonies observed from aerial surveys divided by lengths ascertained from ground reconnaissance on transects with prairie dog colonies. Table 2. Effects of aerial teams 3, ratings of visibility of colonies\ and ratings of habitat typesc on proportions oflength ofGunnison's and white-tailed prairie dog colonies observed on aerial transects during August, September, and November 2002. Independent variable Aerial teams Rating of visibility Rating of habitat type 3 df 1 4 5 F 6.79 0.57 0.48 p 0.011 0.684 0.793 Aerial team from Colorado Division of Wildlife and from Utah Division of Wildlife Resources. 113arely detectible, barely detectible-detectible, detectible, detectible-highly detectible, highly detectible. cGrassland, grassland-short shrub, short shrub, short shrub-tall shrub, tall shrub, agricultural. �175 Table 3. Correlations between aerial crews from the Colorado Division of Wildlife and the Utah Division of Wildlife Resources for proportions of lengths of prairie dog colonies detected, ratings of visibility", and ratings of habitat typesb on aerial transects of Gunnison' s and white-tailed prairie dogs observed during August, September, and November 2002. • Colorado team Variable Proportion of colony length detected Rating of visibility of colony Rating of habitat type on colony X N 51 30 22 Utah team X N 51 30 22 S.E. 0.36 0.11 0.12 2.3 2.4 2.2 S.E. 1.11 0.10 0.08 5.3 2.5 1.4 r0.301 -0.020 -0.066 p 0.032 0.916 0.769 •1 = barely detectible, 1.5 = barely detectible-detectible, 2 = detectible, 2.5 = detectible-highly detectible, 3 = highly detectible. bl = grassland, 1.5 = grassland-short shrub, 2 = short shrub, 2.5 = short shrub-tall shrub, 3 = tall shrub. Table 4. Correlations between ratings of visibility" and proportions of prairie dog colony lengths detected, and ratings of habitat typesb and proportions of prairie dog colony lengths detected on transects of Gunnison's and white-tailed prairie dogs combined by aerial crews from the Colorado Division of Wildlife and the Utah Division of Wildlife Resources combined during August, September, and November 2002. Visibility/Habitat Variable Visibility versus proportion of colony length detected Habitat versus proportion of colony length detected Proportion of colony detected N X S.E. N X 77 2.4 0.07 77 4.5 65 1.7 0.07 65 4.3 r p 0.76 0.038 0.742 0.88 -0.246 0.048 S.E. •1 = barely detectible, 1.5 = barely detectible-detectible, 2 = detectible, 2.5 = detectible-highly detectible, 3 = highly detectible. bl = grassland, 1.5 = grassland-short shrub, 2 = short shrub, 2.5 = short shrub-tall shrub, 3 = tall shrub. �176 Appendix 1. Suggestions for Aerial Surveys (from Andelt and Schnurr 2002). Based upon our flight with Larry Gepfert and suggestions from Jim Dennis and Dave Younkin we have developed a number of suggestions for aerial surveys of Gunnison' s prairie dogs and white-tailed prairie dogs: • Elevation and overall range distributions (Armstrong 1972, Fitzgerald et al. 1994) should be ascertained before aerial surveys are conducted to minimize the area that needs to be surveyed. • Flight crews should spend at least I day on the ground in Gunnison's prairie dog and white-tailed prairie dog towns to become more familiar with the towns before they fly transects. The crews should also gain experience by flying over known colonies. After flying over known colonies, the crew should spend some time on the ground in a colony to better ascertain what they have seen from the air. • Transects should be constructed along drainages, instead of across drainages, to minimize changes in elevation while conducting surveys. Further, transects should be flown down the drainage, instead ofup drainages, to maximize aircraft maneuverability while minimizing danger. RECOMMENDED PLANS FOR FUTURE • Complete ground surveys to establish the remaining "known" boundaries for white-tailed prairie dog colony transects already flown in Colorado. Compare known and aerial estimates of the locations of prairie dog colonies to ascertain accuracy of aerial surveys. • Ascertain if a correction for detection probabilities will need to be employed. This will be primarily needed if the aerial crews were unable to observe a significant proportion of the "known" colonies. • Determine strata boundaries utilizing recent WRIS mapped activity areas and elevation limits for prairie dogs to minimize the extent of surveys. • Establish transect lines along drainages and within strata. • Determine who will conduct aerial surveys in Colorado. We suspect that we will need to contract with a commercial company. • Ascertain if prairie dog colony activity can be determined from the air. If colony activity cannot be determined from the air, a subset ground sampling technique will need to be established to determine activity. During September field trips to the white-tailed colony in Colorado, we were unable to ascertain activity of many colonies because many prairie dogs apparently entered hibernation early this year due to the drought (Dean Biggins, personal communication). �177 PRELIMINARY EVALUATION OF SURVEYS OF PLOTS FOR ESTIMATING OCCURRENCE OF WHITE-TAILED PRAIRIE DOGS IN COLORADO AND UTAH W. F. Andelt, G. C. White, P. M. Schnurr, and A. Seglund Our research (see above) indicated that aerial line-intercept surveys likely will not work for reliably estimating acreage of Gunnison's and white-tailed prairie dogs. Thus, during Spring 2003, we established a pilot project and surveyed 19 500 by 500 m plots from the ground and air to ascertain if surveys of plots can be used to ascertain trends in occurrence of white-tailed prairie dogs in Colorado and Utah. We focused on white-tailed prairie dogs because they have been petitioned for listing as a threatened or endangered species, but we also plan to expand this methodology for Gunnison's prairie dogs. METHODS We overlaid 7.5 minute topo maps (NAD27 datum) in GIS with 500 by 500 m grid lines on each of 4 study areas (Wolf Creek and Grand Valley, Colorado, Coyote Basin and Cisco, Utah) where locations of prairie dogs were identified. After reviewing the maps and visiting with colleagues familiar with distributions of prairie dogs in each of the 4 study areas, we visited grids (plots) in the field and choose 6 plots in Wolf Creek and 6 plots Coyote Basin such that 2 had low, 2 had medium, and 2 had relatively high abundance of prairie dogs. Also within this classification, 1 of each of the low, medium, and high abundance grids had low visibility and the other high visibility. We also established 4 plots in the Grand Valley, near Grand Junction, and 3 plots near Cisco, Utah in areas with relatively low abundance of white-tailed prairie dogs. During June 2003, we visited the 4 comers of most study plots 3 times each to establish detection probabilities, with 1 visit during 0700-1100, another visit during 1100-1500, and another visit during 1500-1900 hrs. For each study plot, we recorded the investigator's name, date, time, lITM Zone, GPS coordinates for the lower left (SW) comer of the plot, percent cloud cover, and soil type (from a soil survey map), approximate precipitation during last 24 hours, and approximate precipitation during last 30 minutes. For the 4 comers of each plot, we recorded temperature, wind direction, approximate wind speed, percent of plot that was visible, percent of plot in sunshine, rating of visibility, rating of elevation, number of mounds observed, and groups of prairie dogs observed. On 12 June, William Andelt flew over each study plot to ascertain if prairie dogs could be reliably detected in plots from aircraft. We also hired a commercial company to photograph, with high resolution, 9 by 9 inch, color infrared film, 21 study plots to ascertain its feasibility for establishing occurrence of prairie dogs in plots. RESULTS AND DISCUSSION We are currently analyzing data from our pilot observations of white-tailed prairie dogs within 19 study plots. Initial results indicate that we should be able to reliably monitor occurrence and detect changes in occurrence of white-tailed prairie dogs by visiting plots from the ground. We plan to establish about 300 (based upon computer simulations) random plots within the range of white-tailed prairie dogs in Colorado. We plan to hire a field crew and visit these plots to ascertain occurrence of prairie dogs during spring and summer 2004. Our flight over 21 study plots indicate that an airplane might be used to establish occurrence of prairie dogs in high density plots, especially on warm days with snow cover during spring. We will evaluate aerial photographs after they are developed. We also plan to conduct a pilot study, during spring 2004, of the above methodology for ascertaining occurrence of Gunnison's prairie dogs in Colorado. We are currently writing a proposal which will detail our subsequent work. �178 CHRONIC WASTING DISEASE W. F. Andelt We established links, on my web site {http://www.coopext.colostate.edu/wildlife/, then, go to Diseases), to 9 sites that contain information on chronic wasting disease. I informed all extension personnel, including all county extension agents, in Colorado about the availability of this information on my web site. I also informed 153 Cooperative Extension volunteers at 3 training sessions in Colorado, that information on chronic wasting disease was available on my web site. My web page on Diseases was accessed 701 times during January-June, 2003. EXTENSION'INFORMATION ON RESOLVING HUMAN-WILDLIFE CONFLICTS W. F. Andelt My Cooperative Extension activities included: Refereed Publications: Yoder, C. A., W. F. Andelt, L.A. Miller, J. J. Johnston, and M. J. Goodall. 2003. Effectiveness of twenty, twenty-five diazacholesterol, avian gonadotropin releasing hormone, and chicken riboflavin carrier protein for inhibiting reproduction in Coturnix quail. (Submitted to Poultry Science). Refereed Publications In Preparation: Schwartz, A. M., and W. F. Andelt. Effects of castration on reproduction and social structure in the black-tailed prairie dog (Cynomys ludovicianus). (Manuscript is 95% completed, will be submitted to the Journal of Wildlife Management). Schwartz, A. M., and W. F. Andelt. Effects of castration on body mass and survival in the blacktailed prairie dog (Cynomys ludovicianus). (Manuscript is 95% completed, will be submitted to the Journal of Wildlife Management). Heffernan, D. J., W. F. Andelt, and J. A. Shivik. Coyote exploratory behavior following removal of novel stimuli. (Manuscript is 95% completed, will be submitted to the Journal of Wildlife Management. Book Chapters: Lamb, B. L., R. P. Reading, and W. F. Andelt. 2003. Public attitudes and perceptions toward black-tailed prairie dogs. Pages _to_ in J. L. Hoogland, editor. Conservation and management of prairie dogs. Island Press, Washington, D.C. (Submitted 2nd draft). Andelt, W. F. 2003. Methods and economics of managing prairie dogs. Pages_ to_ in J. L. Hoogland, editor. Conservation and management of prairie dogs. Island Press, Washington, D.C. (Submitted 3rd draft). �179 Extension Publications: Andelt, W. F. 2002. Impacts of drought on wildlife. lpp. (Published at http://drought.colostate.edu/). Andelt, W. F., S.N. Hopper, and M. Cerato. 2002 (revised). Preventing woodpecker damage. Cooperative Extension Bulletin, Colorado State University, Fort Collins. Spp. (Published at http://www.ext.colostate.edu/PUBS/NATRES/pubnatr.html). Andelt, W. F. 2003. Preventing woodpecker damage to trees. The Green Scene (July, In press). Cerato, M., and W. F. Andelt. 2003 (revised). Coping with skunks. Cooperative Extension Bulletin, Colorado State University, Fort Collins. 5pp. (In press; will be published at http://www.ext.colostate.edu/PUBS/NATRES/pubnatr.html). Cerato, M., and W. F. Andelt. 2003 (revised). Coping with snakes. Cooperative Extension Bulletin, Colorado State University, Fort Collins. 6pp. (Published at http://www.ext.colostate.edu/PUBS/NATRES/pubnatr.html). Progress Reports: Andelt, W. F., and P. Schnurr. 2002. Progress report: inventorying Gunnison's prairie dogs in Colorado. Progress report submitted to Gary Miller, Colorado Division of Wildlife, 7 November 2002. 7pp. Andelt, W. F. 2003. Status of Gunnison's prairie dogs in Colorado. Progress report submitted to Gary Miller, Colorado Division ofWildlife, 13 January 2003. !Opp. Andelt, W. F., P. Schnurr, and A. Seglund. 2003. Evaluation of aerial surveys for estimating acreage of Gunnison's and white-tailed prairie dogs in Colorado and Utah. Progress report submitted to Gary Miller, Colorado Division of Wildlife, 24 February 2003. 13pp. Papers Presentation at National. Regional. and State Meetings: Andelt, W. F. 2003. Alternatives to toxicants for managing conflicts with black-tailed prairie dogs. Colorado Prairie Dog Technical Conference, Fort Collins, Colorado (Invited paper). Andelt, W. F. 2003. Behavioral modification of coyotes to reduce predation on livestock. Department of Fisheries and Wildlife, Utah State University (Invited paper). Andelt, W. F. 2003. Evaluation of aerial surveys for estimating acreage ofGunnison's and white-tailed prairie dogs in Colorado and Utah. Colorado Prairie Dog Technical Conference, Fort Collins, Colorado. Andelt, W. F. 2003. Incorporating experimental design in education on managing humanwildlife conflicts at Colorado State University. Tenth Wildlife Damage Management Conference, Hots Springs, Arkansas (Invited paper). Andelt, W. F. 2003. Managing conflicts with coyotes: aversive stimuli, novel stimuli, and livestock guarding dogs. Wyoming Student Chapter of The Wildlife Society, Laramie, Wyoming (Invited paper). �180 Andelt, W. F. 2003. Non-lethal methods for managing conflicts with prairie dogs. Colorado Prairie Dog Technical Conference, Fort Collins, Colorado (Invited paper). Jozwiak, E. A., T. N. Bailey, and W. F. Andelt. 2003. Response of wolves to changing harvest levels on the Kenai NWR, Alaska. The World Wolf Congress 2003 - Bridging Science and Community, The Banff Centre, Banff, Canada (submitted). Analyzed about 200 predator scats to help assess the role of various predators in the decline of sage grouse in northwe~tem Colorado. • Obtained $1,200 from the Renewable Resources Extension Act to revise Cooperative Extension fact sheets on managing conflicts with wildlife. Submitted a research proposal to study Ecology of coyotes and coyote predation on bighorn sheep in Rocky Mountain National Park, Colorado. Project was not funded. Co-coordinator and instructor at 3 2-4 hour workshops for 153 extension volunteers and 3 Colorado Division of Wildlife employees. Speaker at 3 Cooperative Extension meetings with 80 participants. Provided training for 55 biologists and other professionals including wildlife commissioners at l workshop. Presented 3 guest lectures to I 07 students in Colorado State University courses on managing conflicts with wildlife. Advised an M.S. candidate that conducted research on resolving conflicts with prairie dogs, and a Ph.D. candidate that was conducted research on coyotes. Served on 2 M.S. and I PhD. Committees. Evaluated 27 posters for the Cooperative Extension Poster Session at the February 2003 InService training. Served on the Jefferson County Cooperative Extension Natural Resources Agent Search Committee. Served as a Mentor for Thomas Mason, Jefferson County Cooperative Extension Natural Resources County Agent. Served on the Colorado Department of Agriculture Pesticide Review Committee. Commented on impacts of pesticides on wildlife .. Provided extensive reviews of the efficacy data for the Rodex 4000 (an explosive device for killing rodents), and efficacy of 2 repellents (Deer Stopper, Deer Stopper Ready to Use) for deterring deer. Served on the Colorado State University Cooperative Extension, College of Natural Resources, Renewable Resources Extension Act Committee. Served on the Rodent Program Review Panel for the National Wildlife Research Center. �181 Updated my web site on Managing Conflicts with Wildlife at (http://www.coopext.colostate.edu/wildlife/}. Various pages of the web site have been accessed 227 to 3,381 times each during January-June 2003. Provided interviews for 5 newspaper reporters at United Press International, Rocky Mountain News, Denver Post, and others. Provided interviews for 2 radio stations. Wrote 1 news release for CSU Cooperative Extension Agents. Reviewed 2 manuscripts for scientific journals and 1 manuscript for a colleague. Participated in about 75· meetings. Wrote about 50 e-mail messages about conflicts with wildlife. Answered about 50 telephone inquiries about managing conflicts with wildlife. �182 POSSIBLE ROLE OF PREDATORS IN THE SAGE GROUSE DECLINE W. F. Andelt Approximately $5,200 was received from the Moffat County Department of Natural Resources to conduct preliminary research on the possible role of predators in the sage grouse (Centrocercus urophasianus) decline in Moffat County. Red fox (Vulpes vulpes; Flinders [1999]) have been reported as one of the primary mammalian predators of sage grouse, whereas coyotes (Canis la trans; Presnall and Wood [1953]), bobcats (Fe/is rufus; Hartzler [1974], mink (Mustella vison; Hartzler [1974]), badgers (Taxidea taxus; Gill [1965]), and ground squirrels (Spermophilus spp.; Schroeder and Baydack [2001]) also prey on adults or nests of sage grouse. Thus, we obtained data on relative abundance of mammalian carnivores on 2 study areas (immediately northwest of Craig ["Craig"] and north of Maybell ("Maybell"). Sage grouse are scarce on the Craig study area which is fragmented habitat (sagebrush-grassland interspersed with CRP, alfalfa, and wheat), whereas they are moderately abundant on the Maybell study area which is primarily contiguous habitat (mostly sagebrush-grassland). Golden eagles (Aquila chrysaetos; Hartzler [1974]) appear to be the primary avian predator of sage grouse, particularly on leks, whereas prairie falcons (Falco mexicanus; Hartzler [1974]), red-tailed hawks (Buteo jamaicensis), Swainson's hawks (B. swainsoni), ferruginous hawks (B. regalis), northern harriers (Circus cyaneus; references in Schroeder and Baydack 2001) may occasionally kill some sage grouse. Common ravens (Corvus corax; Allred [1942], Autenrieth [1981], Alstatt [1988]) appear to be the primary avian predators of sage grouse nests or simulated nests, whereas black-billed magpies (Pica pica; Autenrieth [1981]) may prey on some nests. Consequently, we collected data on relative abundance of avian predators, and collected carnivore scats on the Craig study area, the Maybell study area, and in the Axial basin (Appendix 1), which consists primarily of contiguous habitat (mostly sagebrushgrassland) where sage grouse are moderately abundant. I also provided information to Dr. Tony Apa and colleagues on identifying which predators killed sage grouse or depredated their nests. Relative Abundance of Mammalian Predators During 5 to IO June 2001, we (Dr. Andelt, 1 graduate student, and 1 technician) set 92 scent stations on the Craig study area and 92 scent stations on the Maybell study area to gain an assessment of general abundance of carnivores on the 2 sites. Scent stations are 1-yard diameter circles of sifted earth with an attractant (fatty acid scent, small traffic cone or both) placed in the center. The scent stations were set in groups of 4 with each station 0.2 miles apart. Each group of 4 scent stations were set at least 2 miles apart to minimize visits to different groups of stations by individual carnivores. The locations for stations were mostly randomly selected from BLM maps. The stations were checked 1 day after they were set. A few of the stations were rendered inoperable by light to moderate rain. I used chi-square tests (PROC FREQ, SAS Inst. Inc. 1988) to analyze the data. Red fox visited more scent stations (X/ = 5 .465; P = 0.0194) and more groups of stations (X/ = 5. 199; P = 0.0226) on the area with few sage grouse compared to the area where grouse were fairly abundant (Table 1). We need to interpret these data with caution. First, we do not know exactly how scent station visitation rates relate to relative abundance of red fox on the 2 sites, but these data do suggest that red fox likely are more abundant on the site where grouse are rare. These data also do not indicate that red fox caused the decline of sage grouse. Surprisingly, we did not positively identify coyote tracks at any of the stations although it is possible that l or 2 of the stations could have been visited by coyotes. �183 Table 1. Visits by red fox to scent stations set in Moffat County, Colorado during 5 to 10 June 2001. Few grouse Grouse moderately abundant (Craig study area) (Maybell study area) 92 92 Number scent stations 78 87 Operable stations 19 21 Operable groups of stations 12 4 Stations visited by red fox 9 3 Groups of stations visited by red fox Stations visited by coyotes 0 0 Raptor Surveys We established 10 I -mile long survey routes on roads on the Craig study area, 10 I -mile long survey routes on the Maybell study area, and 10 I-mile long survey routes in the Axial Basin. We counted raptors (hawks, eagles, magpies), at all distances, along these transects once per month from August 2001 through June 2002 to ascertain if the abundance of raptors differs among the 3 areas. I compared abundance of various raptors on transects with ANOVA (PROC GLM, SAS Inst. Inc. 1988). Abundance of none of the raptors varied among study areas (F2,216 = 0 .15-2. 82; P = 0. 865-0. 062, Table 2). In general, black-billed magpies were the most abundant raptor followed by American crows (Corvus brachyrhynchos; Table 2). Table 2. Average number of raptors observed per month• on the Craig, Maybell, and Axial Basin study areas in Moffat County, Colorado from August 2001 through June 2002. Few grouse Grouse moderately abundant (Craig study area) Maybell study area) Axial Basin Golden eagle 0.7 2.2 1.6 Common raven 0.6 0.9 0.0 Black-billed magpie 7.8 7.8 6.4 Prairie falcon 0.0 0.0 0.0 Red-tailed hawk 0.3 0.5 0.8 Ferruginous hawk 0.0 0.1 0.0 Northern harrier 0.2 0.0 0.1 Bald eagle (Haliaeetus leucocephalus) 0.0 0.2 0.1 American crow 3.7 6.2 9.1 American kestrel (Falco sparverius) 0.1 0.2 0.2 Turkey vulture (Cathartes aura) 0.8 0.0 0.0 Other (unidentified) 0.0 0.2 0.0 aren I-mile long transects were driven once per month and all raptors observed from the vehicle were recorded from August 2001 through June 2002, except observations were not made during January and observations also were not made on the Maybell study area during February due to difficulty traversing roads. �184 Carnivore Food Habits We established 10 I -mile long survey routes on roads on the Craig study area, 10 I -mile long survey routes on roads on the Maybell study area, and 10 I -mile long survey routes on roads in Axial Basin (Appendix 1). We collected carnivore (primarily coyote and red fox) scats along these survey routes once per month from August 2001 through June 2002, except for January when travel was hindered by snow. We measured the diameters of scats with calipers and weighed them on an electronic balance. Green and Flinders (1981) reported that only 5% ofred fox scats are 2:18 mm in maximum diameter and that only 4% of coyote scats were <16 mm in maximum diameter. I extrapolated data from Weaver and Fritts (1979) and Danner and Dodd (1982) which indicated that only about 8 and 11 % of coyote scats are <16 mm in maximum diameter. Thus, I classified scats <16 mm in diameter as red fox and those 2:18 mm in diameter as coyote. Scats that consisted of short segments were classified as bobcat (Murie 1954). We placed these scats in fine-mesh nylon bags and washed and dried them. We visually inspected the scats to determine if they contained sage grouse feathers or egg shells to help ascertain if red fox or coyotes preyed on sage grouse. When feathers were found, we ascertained if they were from sage grouse according to overall size of the feathers, presence and size of quills, presence of aftershafts, and general structure of the feather. Only birds in the order Galliformes, which includes sage grouse, have aftershafts (Elbroch and Marks 2001:235) on their feathers. A total of 224 scats were collected and analyzed (Table 3). Based upon diameter and segmentation of scats, we ascertained that 26 scats were from red fox, 141 from coyotes, 4 from bobcats, and 53 scats could not be assigned to species. Although we collected scats on 10 miles of roads in each study area, the greatest numbers of scats were found on the Maybell and Axial Basin study areas, whereas the fewest scats were found on the Craig study area. Roads on the Craig study area are traveled more frequently by automobiles and are graded more frequently than roads on the other 2 study areas. These activities obliterate scats, thus relative abundance of scats likely is a poor indicator of relative abundance of carnivores on the 3 study areas. We found feathers in only 5 of 224 scats and none of the feathers appeared to be from sage grouse (Table 3). Table 3. Number carnivore scats and presence of feathers in scats found on transects on the Craig, Maybell, and Axial Basin study areas in Moffat County, Colorado from August 2001 through June 2002. Few grouse Grouse moderately abundant (Craig study area) Maybell study area) Axial Basin Total scats 18 101 105 Red fox scats 4 13 9 Red fox scats with feathers 1 1 0 Coyote scats 9 66 66 Coyote scats with feathers 0 1 1 1 3 Bobcat scats 0 Bobcat scats with feathers 0 0 0 Unknown scatsa 5 21 27 Unknown scats with feathers 0 1 0 •Based upon diameter and weight, we could not assign these scats to red fox, coyote, or bobcat. �185 Assistance with Determining which Predators are Responsible for Depredating Sage Grouse and their Nests: I provided Tony Apa and colleagues with information on how to determine which predators killed grouse or depredated their nests. SYNTHESIS OF RESULTS AND DISCUSSION Results of our scent station surveys suggest that red fox are more abundant on the Craig study area, where few sage grouse were present, than on the Maybell study area, where grouse were moderately abundant. The absence of sage grouse feathers in 141 scats, ascertained to be from coyotes, suggests that coyotes perhaps may not be substantial predators of sage grouse. We also did not find grouse feathers in 26 scats ascertained to be from red fox, and 4 scats ascertained to be from bobcats, however these small sample sizes do not allow for strong inferences regarding predation by red fox and bobcats on sage grouse. Even if feathers would have been found in coyote, red fox, or bobcat scats, it would still be difficult to ascertain the impact of either species on sage grouse without knowing densities of these 3 carnivores, densities of sage grouse, carnivore digestion and defecation rates, etc. However, I analyzed carnivore scats in a preliminary attempt to ascertain if either species might be frequently preying on sage grouse. Prior research has indicated that golden eagles and common ravens are the primary avian predators of sage grouse and their nests, respectively. Our raptor surveys indicated that both species were fairly common on most of our study areas. Initially, we expected that we might find more golden eagles and common ravens on the Craig study area, where sage grouse are scarce, if they are having an impact on sage grouse. However, predators are opportunists which often frequent areas of highest prey abundance. Due to these factors, and due to no significant differences in abundance of golden eagles and common raven among the 3 study areas, it is difficult to draw solid inferences from this study about the impact of these species on sage grouse. Ultimately, the best way to ascertain impacts of various predators on adult sage grouse, sage grouse chicks, and sage grouse nests is to monitor survival and causes of mortality for these life stages of sage grouse. ACKNOWLEDGMENTS I thank numerous individuals that assisted with this project. J. Comstock provided continuous encouragement and financial support for the study. G. Miller and the Colorado Division of Wildlife provided financial support while W. Andelt conducted the study. J. Shivik and the National Wildlife Research Center provide salary support for D. Heffernan and D. Martin while they assisted with the study. T. Apa and R. Hoffman provided suggestions for the study. D. Heffernan and D. Martin assisted with scent station surveys. A. Martin and V. Dobrich conducted surveys of raptors and collected carnivore scats. C. Simpson analyzed carnivore scats to determine presence of feathers and egg shells. R. Ryder and R. Hoffman assisted with ascertaining if feathers in carnivore scats were from sage grouse. �186 LITERATURE CITED Allred, W. J. 1942. Predation and the sage grouse. Wyoming Wild Life 71(1):3-4. Alstatt, A. 1988. Sage grouse production and mortality studies. Job performance report, Nevada Department of Wildlife, Reno, Nevada, USA Autenrieth, R. E. 1981. Sage-grouse management in Idaho. Idaho Department of Fish and Game, Wildlife Bulletin 9, Boise, Idaho, USA. Danner, D. A., and N. Dodd. 1982. Comparison of coyote and gray fox scat diameters. Journal of Wildlife Management 46:240-241. Elbroch, M., and E. Marks. 2001. Bird tracks & sign: a guide to North American species. Stackpole Books, Mechanicsburg, Pennsylvania, USA Flinders, J. T. 1999. Restoration of sage-grouse in Strawberry Valley, Utah, 1998-99. Utah Reclamation, Mitigation and Conservation Commission, Progress Report, Brigham Young University, Provo, Utah, USA Gill, R. B. 1965. Distribution and abundance of a population of sage grouse in North Park, Colorado. Thesis, Colorado State University, Fort Collins, Colorado, USA Green, J. S., and J. T. Flinders. 1981. Diameter and pH comparisons of coyote and red fox scats. Journal of Wildlife Management 45:765-767. Hartzler, J.E. 1974. Predation and the daily timing of sage grouse leks. The Auk 91:532-536. Murie, 0. J. 1954. A field guide to animal tracks. The Peterson Field Guide Series. Houghton Mifflin Company, Boston, Massachusetts, USA Presnall, C. C., and A. Wood. 1953. Coyote predation on sage grouse. Journal ofMammalogy 34:127. SAS Institute Inc. 1988. SAS/STAT User's guide, release 6.03 edition, SAS Institute Inc., Cary, North Carolina, USA Schroeder, M. A., and R. K. Baydack. 2001. Predation and the management of prairie grouse. Wildlife Society Bulletin 29:24-32. Weaver, J. L., and S. W. Fritts. 1979. Comparison of coyote and wolf scat diameters. Journal of Wildlife Management 43:786-788. �187 Appendix l. GPS coordinates for transects where carnivore scats were collected and raptors were observed (datum= WGS 84). ----------------------Scat---------------------Start of transect End of transect -------------------Raptor------------------End of transect Start of transect y y X X CRAIG STUDY AREA- FRAGMENTED HABITAT X y X y 4495138 4504934 4498993 4497952 4499870 4503788 4505463 4496358 4493667 4490952 286359 286753 289265 282630 279994 276290 271595 278798 281903 274742 4496685 4503651 4499443 4499345 4500954 4504491 4505789 4496016 4492847 4491859 286417 286993 287842 281182 278429 274892 270687 279930 281766 273755 4498162 4504934 4499021 4499731 4500997 4505050 4504525 4495509 4491265 4492689 Transect# 1 2 3 4 5 6 7 8 9 10 286321 286993 290602 283274 280846 277689 272894 277304 281065 275680 286359 287551 289265 282630 279994 276290 271595 278798 281903 274742 4496685 4506266 4499443 4499345 4500954 4504491 4505789 4496016 4492847 4491859 MAYBELL AREA- UNFRAGMENTED HABITAT 11 12 13 14 15 16 17 18 19 20 747333 747215 749179 745930 742374 739699 745369 743327 749357 248318 4497815 4502458 4508451 4508307 4510170 4510025 4514231 4521876 4520229 4517684 748923 745844 750735 744658 742575 738301 746090 744832 750726 249268 4497867 4501925 4508465 4507713 4511677 4510610 4515477 4522160 4519826 4516416 748923 745844 750735 744658 742575 738301 746090 744832 750726 249268 4497867 4501925 4508465 4507713 4511677 4510610 4515477 4522160 4519826 4516416 750446 744495 752004 744178 741688 737962 ? 746305 752031 249486 4498098 4501667 4509245 4508932 4511365 4511658 ? 4521754 4519026 4514902 4480100 4474342 4476843 4472106 4470006 4472271 4467748 4465355 4470302 4478294 253671 254317 257398 255153 249317 246720 252917 257542 260734 249580 4478589 4474661 4475313 4471305 4471575 4473285 4466698 4464349 4470308 4477016 253671 254317 257398 255153 249317 246720 252917 259268 260734 249580 4478589 4474661 4475313 4471305 4471575 4473285 4466698 4466465 4470308 4477016 253641 255142 257138 254046 249559 247862 251914 258706 260138 248906 4477122 4475609 4473907 4470304 4473119 4474390 4465499 4465355 4468956 4475569 AXIAL BASIN 21 22 23 24 25 26 27 28 29 30 253432 252975 257571 256457 249225 245513 254021 258706 262032 250505 � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Mammals research pdfs Description An account of the resource Pdfs of mammals research projects by title . Pdfs are linked from this page. Includes federal aid progress reports and non federal aid progress reports. If reports cover more than one year they are combined into one pdf. Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Mammals Research Projects pdfs (2001-2010) Rights Information about rights held in and over the resource <a href="http://rightsstatements.org/vocab/NoC-NC/1.0/">No Copyright - Non-Commercial Use Only</a>