https://cpw.cvlcollections.org/files/original/8776bc9bb8bb948eca340df7e9abbc8d.pdf a90cadc969fd23e601abccbccb53ab65 PDF Text Text Dan Prenzlow, Director, Colorado Parks and Wildlife • Parks and Wildlife Commission: Marvin McDaniel, Chair • Carrie Besnette Hauser, Vice-Chair Marie Haskett, Secretary • Taishya Adams • Betsy Blecha • Charles Garcia • Dallas May • Duke Phillips, IV • Luke B. Schafer • James Jay Tutchton • Eden Vardy �CHAPTER 16 Wolf Prey Selection in an Elk-Bison System: Choice or Circumstance? Matthew S. Becker,* Robert A. Garrott,* P. J. White,† Claire N. Gower,* Eric J. Bergman,* and Rosemary Jaffe* *Fish and Wildlife Management Program, Department of Ecology, Montana State University † National Park Service, Yellowstone National Park Contents I. Introduction II. Methods A. Detecting and Identifying Wolf Kills B. Factors Influencing Wolf Prey Selection C. Prey Switching III. Results A. Factors Influencing Wolf Prey Selection B. Wolf Prey Switching with Murdoch’s Equation IV. Discussion V. Summary VI. References Appendix 306 309 309 311 315 316 318 322 324 330 331 335 Theme What a predator eats when given choices, and the subsequent effects of this behavior on ecosystem stability, has long been a topic of interest for ecologists. Prey selection is influenced by the absolute and relative abundances of prey types, the life history characteristics of predators and prey, and the attributes of the environment in which these interactions occur. Strong preference by a predator for a particular prey type can lead to ecosystem instability, while prey switching can lessen predation effects on the less abundant prey and enhance system stability. Evaluating prey selection in large mammal systems is difficult due to the broad spatial and temporal scales at which these predatory interactions occur, and investigations, particularly with wolf-ungulate systems, typically involve only the primary prey. Multiple prey species characterize most large mammal predator-prey systems, therefore research into predator-multiple prey dynamics has the potential to yield important ecological insights. We studied winter prey selection during 1996–1997 through 2006–2007 in a newly established wolfelk-bison system where prey differed substantially in their vulnerability to wolf (Canis lupus) predation and wolves preyed primarily on elk (Cervus elaphus) but also used bison (Bison bison) to varying degrees within and The Ecology of Large Mammals in Central Yellowstone R. Garrott, P. J. White and F. Watson ISSN 1936-7961, DOI: 10.1016/S1936-7961(08)00216-9 Copyright # 2009, Elsevier Inc. All rights reserved. �306 Becker et al. among winters and packs. We analyzed the relative influences of prey abundance, predator abundance, and environmental variables on the selection of prey species and age classes and evaluated whether wolves exhibited prey switching from elk to bison. I. INTRODUCTION Predator-prey dynamics can be broadly classified by whether they are single prey or multiple prey systems. Predator diets can be relatively simple when only one prey type is present and few options exist. However, what a predator eats when given choices is a fundamental question germane to the multiple prey assemblages characteristic of most natural predator-prey systems. During the predatory sequence of encountering, attacking, and killing prey, most predators are assumed to select prey types based on abundance, thereby typically relying on encounter rates (Holling 1959). Once encountered, prey are frequently selected by sex, age, size, condition and behavior when individuals differ in their vulnerability to predation (Errington 1946, Morse 1980, Pastorok 1981, Greene 1986, Stephens and Krebs 1986, Quinn and Cresswell 2004). Selection across prey species can also differ based on their absolute or relative abundances and the life history characteristics of both predator and prey as manifested in morphology, defenses, and behavior. Thus, all of these variables have the potential to dramatically influence the dynamics of multiple-prey systems (Murdoch 1969, Fitzgibbon and Lazarus 1995, Moran et al. 1996, Denno and Peterson 2000, Denno et al. 2002, Rosenheim et al. 2004). While the physical vulnerability of a species or individual is of considerable importance in predatory interactions, environmental attributes can also influence vulnerability. Variables such as heterogeneity in climate, habitat structure, and landscape attributes can act alone or in concert with physical vulnerability to influence a predator’s diet (Smuts 1978, Peckarsky and Penton 1989, Hunter and Price 1992, Langellotto and Denno 2004, Hopcraft et al. 2005, Chapter 24 by Garrott et al., this volume). Prey selection by predators in multiple-prey systems can have fundamental positive or negative effects on community stability and prey diversity (Oaten and Murdoch 1975, Murdoch and Bence 1987, Holt and Lawton 1994, Bonsall and Hassell 1997, Synder and Ives 2001). When a predator consumes a prey item disproportionately to its abundance it is said to exhibit a preference (Begon et al. 1996). Many predators have strong preferences for a certain prey type regardless of its abundance (i.e., specialist), and this strong preference is typically viewed as destabilizing to a predator-prey system (Andersson and Erlinge 1977, Hanski et al. 1991, Turchin and Hanski 1997, Eubanks and Denno 2000). Conversely, other predators consume a wide variety of prey, with changes in prey availability strongly affecting their patterns of selection (i.e., generalist). Prey switching behavior is typically associated with generalist predators and occurs when attacks are disproportionately frequent when a prey species is abundant and disproportionately infrequent when a prey species is rare (Murdoch 1969). Switching behavior by predators is generally viewed as stabilizing to a system because a predator can have a regulating influence through density-dependent predation on both prey species (Oaten and Murdoch 1975, Fryxell and Lundberg 1994). Thus, the patterns of prey selection a predator exhibits can have dramatically different ecological consequences (Paine 1966, Holt 1977, Caswell 1978, Hanski et al. 1991, Fryxell and Lundberg 1994, Krivan and Eisner 2003). Studies of predation and its effects on ecosystem stability are difficult because, even in experimental settings, disentangling the myriad factors influencing prey selection is quite complicated. Nevertheless, decades of investigations have increased our understanding of these processes (Murdoch 1969, Oaten and Murdoch 1975, Post et al. 2000, Krebs et al. 2001, van Balaan et al. 2001, Prugh 2005). Relative to investigations of smaller taxa, intensive long-term studies of prey selection and stability in large mammal systems are hindered by the logistic and financial constraints imposed by the broad spatial and temporal scale of investigations. However, large mammal systems have the potential to yield significant insights because the life history characteristics of top predators and large herbivores with �Chapter 16 . Wolf Prey Selection: Choice or Circumstance? 307 strong ecological influences differ substantially from those of smaller taxa typically used in prey selection studies (McNaughton 1985, Temple 1987, Frank and McNaughton 1992, Hobbs 1996, Terborgh et al. 2001, Garrott et al. 2007). Unlike systems of smaller taxa where prey typically rely on avoiding detection, large herbivore prey species are formidable and diverse in their array of defenses and behaviors they can employ once encountered and attacked. These defenses preclude large predators from killing all types of prey with equal effort and subject the predator to constant risk of severe injury and even death (Makacha and Schaller 1969, Mech 1970, Kruuk 1972, Schaller 1972, Carbyn and Trottier 1987, Creel and Creel 2002, Mech and Peterson 2003, Smith et al. 2003, MacNulty et al. 2007). As a result, substandard or vulnerable individuals are frequently selected. This vulnerability often depends on attributes of the prey (e.g., age, size, physiological condition, behavior), environmental attributes, prey density, encounter rates, and the availability of alternative prey (see Mech and Peterson 2003 for review). An impressive body of work has been compiled on wolf prey selection during the last several decades. Wolves are typically considered consummate generalists—opportunistic coursing predators taking advantage of whatever vulnerable prey are available within their territories (Mech 1970, Mech and Peterson 2003). However, virtually all wolf-ungulate investigations in multiple-prey systems have also demonstrated a clear selection for a particular prey species relative to other species in an assemblage (Carbyn 1983, Huggard 1993a, Dale et al. 1995, Je˛drzejewski et al. 2000, Hebblewhite et al. 2003, Smith et al. 2004). Also, within a prey species wolves generally select certain age classes such as young-of-the-year (Mech 1970, Mech et al. 1995, Jaffe 2001, Smith et al. 2004) that could be considered different prey types due to differences in vulnerability. The dynamics of multiple-prey systems, and the mechanisms and conditions whereby wolf prey selection affects community stability and diversity have not been extensively investigated (Dale et al. 1995) and should yield insights into predation processes and the dynamics of large-mammal predator-prey systems. Furthermore there is a need to formally evaluate prey switching for wolves (Dale et al. 1994), because switching has often been incorporated into models of wolf-ungulate dynamics and used to describe simple changes in predator diet composition rather than density-dependent predation (Garrott et al. 2007). Wolves exhibit many of the attributes common to predators that switch prey, including typically hunting by sight, cueing into the different areas where each prey species can be found (Bergman et al. 2006), and testing and evaluating individual prey (Murie 1944, Carbyn et al. 1993, MacNulty et al. 2007). Thus, it is feasible wolves could exhibit prey switching under some conditions. However, if prey preference for a particular species is strong, perhaps due to differences in vulnerability, then switching is unlikely to occur (Murdoch 1969, Murdoch and Marks 1973). In addition to the ecological complexity inherent in studies of predation, comparative investigations are further complicated by the frequent use of terminology without consistent and explicit definitions and distinctions. Specifically, the concepts of prey selection, vulnerability, prey preference and prey switching are ubiquitously employed in predator-prey literature, but typically without concise definitions or differentiation. Prey selection is often considered to be what a predator eats when given choices, with no reference to the abundances of the various prey types available. Fundamental to this choice is the concept of differential vulnerability, or the factors that make an individual more susceptible to predation than other animals in a system. Vulnerability is considered to be of overriding importance in many predator-prey systems, especially those involving large mammals (Ivlev 1961, Mech 1970, Menge 1972, Power et al. 1992, Sinclair and Arcese 1995). However, precise definitions of vulnerability are infrequent and highly variable, ranging from the product of encounter rates and attack probabilities (Greene 1986, Pastorok 1981), ‘‘a combination of capture efficiency and profitability relative to risk’’ (Mech and Peterson 2003:140), to comprising ‘‘all of the behaviors that a prey can adopt to modify its risk of being targeted and caught when attacked’’ (Lind and Cresswell 2005:946). Disparities in definitions owe to the fact that quantifying vulnerability in most natural systems is extremely difficult because it is influenced by physical, behavioral, and environmental factors, and can vary among individuals, populations, species, and landscapes (Chapter 24 by Garrott et al., this volume). �308 Becker et al. Consequently, defining and quantifying prey preference is equally fraught with difficulties and most ecologists employ what Taylor (1984) terms the ‘‘black box’’ definition of preference, defined as when a predator selects a prey type disproportionately to its occurrence in the environment. This interpretation is frequently employed in analyses because it serves as an umbrella for data that are rarely available in natural systems, encapsulating every decision a predator makes based on the myriad physical, behavioral, and environmental factors acting upon all stages of the predatory encounter, attack, and capture (Taylor 1984). Lastly, the definition of prey switching comes from Murdoch (1969) where ‘‘the number of attacks on a species is disproportionately large when the species is abundant relative to the other prey, and disproportionately small when the species is relatively rare.’’ This indicates a preference that is not constant across all levels of abundances of the prey types but changes across a gradient of relative prey abundances, with a predator having a preference for the most abundant prey. Evaluations of dynamics in multiple-prey systems suffer from a lack of consistency in using the term ‘‘prey switching,’’ with some investigators employing it to indicate a density-dependent change in predator preference (Murdoch 1969), while others use it to simply describe changes in predator diet composition. We evaluated prey selection by wolves in a newly-colonized, bison-elk prey system in the Madison headwaters area of Yellowstone National Park during the winters of 1996–1997 through 2006–2007 (Figure 16.1). Wolf numbers varied between 2–50 wolves in 1–5 packs after they were reintroduced and colonized the area beginning in 1995–1996 (Chapter 15 by Smith et al., this volume). Elk were resident throughout the year, but their numbers decreased from approximately 600 to 174 following wolf establishment (Chapters 11 and 23 by Garrott et al., this volume). In contrast, bison were seasonally migratory with numbers increasing through each winter (200–1500) until they exceeded elk numbers by several orders of magnitude in late winter (Chapter 12 by Bruggeman et al., this volume). The dramatic contrasts in life history characteristics, movements, and abundance between these two prey species, coupled with variations in snow pack and wolf abundance, offered a unique opportunity to evaluate prey selection. Elk are smaller in size and prone to flight as an anti-predator behavior, while bison are larger and tend to employ sophisticated group defenses (Carbyn 1974, Carbyn and Trottier 1987, Carbyn et al. 1993, MacNulty et al. 2007). Our objectives were to: (1) characterize FIGURE 16.1 A member of the Hayden pack feeds on a freshly-killed bull elk in the Madison Canyon. Wolves exhibited a preference for elk but selected bison to varying degrees within and among winters (Photo by Shana Dunkley). �Chapter 16 . Wolf Prey Selection: Choice or Circumstance? 309 wolf prey selection over time and among packs, (2) evaluate the drivers of wolf prey selection at the species- and age-class levels, and (3) evaluate if wolves switched from elk to bison in this multiple-prey system. II. METHODS A. Detecting and Identifying Wolf Kills We conducted intensive predation investigations in the primary winter ranges of bison and elk in the Madison headwaters area (31,000 ha), with concurrent investigations of these prey species allowing collection of wolf predation data in a tractable area with a well-described ungulate prey base. We documented prey selection by wolves during 15 November through 30 April each winter from 1996–1997 through 2006–2007. Our sampling unit was radio-collared wolf packs that used the study area as part of their territory. Wolves were aerially darted from helicopters by National Park Service biologists and equipped with VHF telemetry collars. A total of 37 wolves from four packs were collared during the course of the study (Chapter 15 by Smith et al., this volume). The number and sizes of wolf packs using the study area were dynamic within and among winters. Thus, we used ground observations, snow-tracking, and aerial counts during tracking flights by park biologists to estimate the wolf population. We defined two metrics, wolf days and pack days, as one wolf or one pack in the study area for one day, respectively (Chapter 15 by Smith et al., this volume). We also defined multiple pack days as the number of days when more than one pack was present in the study area. We used roads traversing each river drainage in the study area (Chapter 2 by Newman et al., this volume) to sample for wolf presence daily throughout the winter. Sampling began at dawn with ground crews of 3–4 people covering all roads by snowmobile or vehicle, and using strategic high points in the landscape to facilitate telemetry triangulations (White and Garrott 1990) and observations of wolves (Figure 16.2). When possible, multiple locations were obtained in early morning and evening each day. We also recorded any uncollared wolves detected opportunistically via tracks or FIGURE 16.2 Using telemetry to detect, locate, and monitor radio-collared wolf packs. High points on the landscape were utilized for signal detection, scanning for avian scavengers, and observations (Photo by Shana Dunkley). �310 Becker et al. observations to aid in the estimation of the wolf population using the study area. In addition, biologists studying elk and bison routinely covered backcountry areas to assist with wolf detection. When wolves were located, we used visual scans and monitoring of avian scavengers in the vicinity to detect kills. Ravens preferentially associate with wolves in winter, and an average of 28.6 ravens (Corvus corax) were present at fresh wolf kills in the northern range of Yellowstone (Stahler et al. 2002), with slightly lower averages in the Madison headwaters area (D. Stahler, National Park Service, personal communication). This association facilitated the detection of kills. We also conducted extensive snowtracking after wolves departed the area to further facilitate kill detection (Huggard 1993a, Dale et al. 1995, Je˛drzejewski et al. 2000, Jaffe 2001, Hebblewhite et al. 2003). We necropsied ungulate carcasses to determine cause of death, species, sex, age, and condition (Figure 16.3). Wolf kills were inferred from collective evidence of subcutaneous hemorrhaging indicative of injuries sustained before death, signs of struggle or chase at the kill site, blood trails, signs of predator presence, and our knowledge of wolf movements and activities. We documented frequent spring grizzly bear (Ursus arctos) predation on bison during the latter years of the study. Thus, when both bears and wolves were present on a kill, we classified it based on the patterns of injury and subcutaneous hemorrhaging. Bears typically attacked the head and spine, while wolves attacked the hindquarters and flanks. Similarly, mountain lion (Puma concolor) kills of elk were determined based on characteristics of the kill site and patterns of injury. Kills were sexed using the presence of genitalia, horns, antlers, or pedicels, and aged based on FIGURE 16.3 Performing a necropsy on a wolf-killed calf elk in the Madison canyon. Extensive examinations were performed on every ungulate carcass to determine whether the animal was killed by predators or died of other causes, as well as to assess the condition of the animal and the attributes of the site in which it was killed. Wolves preferred elk calves over other prey types and the abundance of calves strongly influenced variation in wolf prey selection (Photo by Kevin Pietrzak). �Chapter 16 . Wolf Prey Selection: Choice or Circumstance? 311 size and patterns of tooth eruption and replacement (Fuller 1959, Hudson et al. 2002). When available, an incisor or canine was removed from adult ungulates and aged using cementum annuli (Moffitt 1998, Hamlin et al. 2000). Marrow fat from the femur or humerus was assessed visually based on color and consistency, and classified as: (1) solid and white, (2) 50–75% solid with red spots, (3) 25–50% solid, reddish, and (4) 0–25% solid, gelatinous and red (Cheatum 1949). We also classified the extent of fluoride toxicosis and necrosis (0 ¼ none, 1 ¼ mild, 2 ¼ moderate, 3 ¼ severe) in jaws from adult animals because these ailments were relatively common due to the strong geothermal influence (Shupe et al. 1984, Garrott et al. 2002, Chapter 10 by Garrott et al., this volume). Observed patterns of wolf predation can be biased by differing rates of detection for various prey types, with smaller prey such as calves consumed faster and, thus, potentially detected less frequently (Fuller and Keith 1980, Fuller 1989, Hebblewhite et al. 2003). While this bias is more likely in aircraftbased studies or studies that do not use snow-tracking (Fuller 1989, Dale et al. 1995), recent studies have considered kill detection efficiency in ground-tracking (Je˛drzejewski et al. 2000, Jaffe 2001, Hebblewhite et al. 2003, Smith et al. 2004). We empirically evaluated our efficiency in detecting kills and concluded our methods provided accurate data on wolf prey selection patterns (Jaffe 2001). B. Factors Influencing Wolf Prey Selection The probability of a prey animal being consumed by a predator is the product of the probability of being encountered by the predator, the probability of the predator attacking the prey once encountered, and the probability that the attack is successful (Endler 1991). We did not assess encounter rates and attack rates, but recognized that bison and elk calves could be considered separate prey items given their dramatic differences in vulnerability compared to adults (Figure 16.4, Mech 1970, Carbyn and Trottier 1987, Mech et al. 1995). Thus, we identified four main prey types available to wolves (i.e., elk calves, elk adults, bison calves, bison adults) in the Madison headwaters area. FIGURE 16.4 An elk calf and bull bison feeding along the banks of the Firehole river. Elk calves and bull bison represent the extremes in prey sizes and defenses confronting wolves (Photo by Kevin Pietrzak). �312 Becker et al. Studies of prey selection in natural systems where encounter and attack rate data are unavailable typically employ selection indices (Lechowicz 1982), whereby the occurrence of a prey type in the predator’s diet is compared to its abundance in the system. Thus, we calculated selection indices for the four prey types using Chesson’s (1978) alpha method across early-, middle-, and late-winter periods from the establishment of resident packs in winter 1998–1999 through 2006–2007 to determine whether wolves selected any prey types disproportionately to their abundance. However, selection indices are limited to prey abundance questions. Therefore, we employed a multinomial logit analysis (Menard 2002) to evaluate the relative importance of prey abundance, predator abundance, and environmental variables on selection across prey species and age classes. We modeled four response categories corresponding to the four main prey types available to resident wolf packs. Each kill comprised an observation and we used elk calves as the base model because wolves tended to select them when available (Jaffe 2001, Smith et al. 2004). Three logits were modeled as La ðxÞ ¼ log ½pa ðxÞ=p0 ðxÞ�ða ¼ 1; 2; 3Þ, where p0(x), p1(x), p2(x), and p3(x) were the probabilities of a calf elk, adult elk, calf bison, and adult bison response, respectively. p0(x) was the denominator (i.e., baseline response) of each odds and x ¼ (x1, x2, . . ., xp) was a vector of model covariates. We developed a suite of covariates corresponding to prey, predator, and snow pack variables to evaluate factors influencing prey selection by wolves. These covariates were chosen based on our knowledge of the system and variables reported to significantly influence prey vulnerability and selection by wolves in other systems. We estimated covariates for the date that each wolf kill occurred. Wolf population structure consisted of territorial packs, each occupying a particular territory (Chapter 15 by Smith et al., this volume). Prey abundance also varied temporally and spatially, with bison prone to frequent large-scale movements between drainages (Chapter 12 by Bruggeman et al., this volume). Thus, all prey in the study area were not equally available to all wolf packs, and such disproportionate abundance could affect wolf diets. Therefore, we estimated prey abundance at a scale contained within the river drainages (Madison, Gibbon, Firehole) of each pack’s respective territory. We determined the drainages used by each pack each winter by constructing 95% fixed kernel territories from ground locations collected from 15 November through 30 April each winter (Chapter 15 by Smith et al., this volume). We excluded temporary probes by packs into other drainages from these calculations. We estimated abundance covariates (ELKcalf, ELKadt, BISONcalf, and BISONadt) for the four prey types within each drainage by decomposing the 167-day winter field season into three approximately 8-week periods corresponding to early, middle, and late winter. The non-migratory elk population (Craighead et al. 1973, Garrott et al. 2003) was not subject to the dramatic fluctuations in abundance characteristic of migratory populations, and typically only experienced decreases across winter due to starvation and predation, particularly of calves (Chapter 23 by Garrott et al., this volume). Consequently, we estimated the abundance of adult elk and calves during early, middle, and late winter using mark-resight techniques and age composition data (Chapters 11 and 23 by Garrott et al., this volume). We conducted multiple mark-resight surveys in late winter (n ¼ 10–33) when elk were concentrated in meadow complexes. We also estimated calf:cow ratios during early and late winter using the respective first and last 100 random elk groups obtained from telemetry sampling (Chapters 11 and 23 by Garrott et al., this volume). We then estimated adult elk abundance (ELKadt) for the early- winter period by multiplying the previous spring’s mark-resight population estimate by a pooled summer survival rate of 0.95 derived from telemetry data (Chapter 23 by Garrott et al., this volume). We assumed 85% of the adult population was females (Chapter 11 and 23 by Garrott et al., this volume) and multiplied the early-winter adult female estimate by the early-winter calf:cow ratio to obtain an early-winter calf (ELKcalf ) estimate. Similarly, we multiplied the adult cow estimate by the late-winter calf:cow ratio to obtain the number of elk calves remaining in the late-winter period. A late-winter adult elk estimate was then calculated by subtracting the late-winter calf estimate from the total late-winter mark-resight population estimate. We averaged the respective means of the early- and late-winter estimates to approximate the abundance of both adult and calf elk for the mid-winter period. While elk distribution �Chapter 16 . Wolf Prey Selection: Choice or Circumstance? 313 among the three drainages varied among winters (Chapter 21 by White et al., this volume), there was little elk movement between drainages within winters (Chapter 18 by Gower et al., this volume). Thus, we multiplied our estimates by the proportion of the elk population observed within each drainage during the spring mark-resight surveys to estimate the abundance of both prey types within each of the three drainages. Lastly, we estimated a covariate for the total abundance of elk (ELK) by summing the adult and calf estimates. We estimated bison adult and calf abundance (BISONadt and BISONcalf, respectively) by conducting ground counts through the winter range every 10–16 days, with observers recording the number, location, sex, and age class of all observed bison. Each drainage was subdivided into discrete survey units and bison totals for each drainage were calculated by summing the respective unit totals. Because substantial changes in bison abundance could occur between surveys, we interpolated between estimates to derive the bison abundance estimates for the date of each wolf kill. We also estimated a total bison abundance covariate (BISON) by summing the adult and calf estimates. In addition, we calculated the ratio of bison to elk abundance (BISON.ELK) by dividing the bison estimate by the elk estimate. The high density ungulate winter range of the Madison headwaters experienced considerable wolf use despite comprising a relatively small area. Following the establishment of multiple packs in the system there was a substantial increase in wolf abundance, spatial and temporal territory overlap, and inter-pack strife (Chapter 15 by Smith et al., this volume). These dynamics, coupled with increases in elk anti-predator responses to increasing wolf numbers (Chapters 18, 19, and 20 by Gower et al., Chapter 21 by White et al., this volume) and decreasing elk numbers (Chapter 23 by Garrott et al., this volume), were negatively related to kill rates for wolves using the system (Chapter 17 by Becker et al., this volume) and likely affected prey selection. Thus we developed three covariates to index the strength of these competitive interactions: the wolf:ungulate ratio (WOLF:UNG); the wolf:elk ratio (WOLF: ELK); and multiple pack days (PACKmult). We estimated each of these as population level indices for early-, middle-, and late-winter periods across the entire study area because wolf territories often overlapped extensively, pack territories often included more than one drainage, and adjacent packs likely influenced each others’ movements and behaviors (Chapter 15 by Smith et al., this volume). We calculated the wolf:ungulate and wolf:elk ratios by dividing the total wolf days estimated for the time period by the number of days in the period, and then dividing by the mean elk and bison estimates. We estimated multiple pack days by summing the total number of days in a period during which more than one pack was detected in the study area. Snow pack substantially decreases ungulate mobility and increases their vulnerability to wolf predation (Peterson 1977, Parker et al. 1984, Nelson and Mech 1986, Huggard 1993b). Snow depth, density, and crusting can impede escape for ungulates that employ flight as an anti-predator tactic. Snow depth is not an accurate integrator of snow pack attributes due to differences in density, crust conditions, and layers. Thus, we described the temporal and spatial dynamics of snow pack using a validated snow model (Chapter 6 by Watson et al., this volume) to estimate mean daily snow-water equivalents (i.e., the amount of water in a column of snow; SWEmean), for the Firehole, Gibbon, and Madison drainages during 1 October through 30 April. When a pack’s territory encompassed more than one drainage, we calculated mean snow pack metrics across the two drainages. SWEmean was estimated for the date of each wolf kill to provide an indirect measure of prey escape ability. The nutrition and condition of ungulates in mid- to high-latitude systems decrease through the winter because most forage is senescent and animals must forage and travel through snow (Chapter 9 by White et al., this volume). Consequently, the accumulation and duration of snow pack can have a long-term weakening influence on ungulate physiological condition that can ultimately be lethal in severe winters (Murie 1944, Severinghaus 1947, Je˛drzejewski et al. 1992, Garrott et al. 2003). We estimated the sum of daily snow-water equivalent values beginning on 1 October each winter (SWEacc; Garrott et al. 2003) for the date of each kill to provide an indirect measure of ungulate physiological condition. �314 Becker et al. We developed and evaluated a priori hypotheses to estimate the relative influence of prey abundance, snow pack, and wolf competition on wolf prey selection of ungulate species and age classes. Our a priori hypotheses were expressed in four main model structures incorporating prey abundance and other potential covariates of snow pack and wolf competition (Appendix). For each covariate, we then identified the metrics we believed were appropriate for estimation. We developed 84 candidate models in the form of multinomial logit equations to evaluate our hypotheses. Three logit equations were generated for each candidate model, describing: (1) the probability of an elk adult kill compared to an elk calf kill, (2) the probability of a bison calf kill compared to an elk calf kill, and (3) the probability of a bison adult kill compared to an elk calf kill. For comparison of coefficient estimates, we scaled and centered each covariate prior to analysis by subtracting the dataset’s midpoint from each covariate value and dividing them by the dataset’s midrange. This restricts each covariate’s values to fall within �1 and 1, inclusively. We assessed potential colinearity between covariates using variance inflation factors and did not use covariates with values >6 in the same model (Neter et al. 1996). Covariates that were not used in the same model due to strong colinearity were BISONadt and BISONcalf, SWEmean and SWEacc, and combinations of MULTPK, WOLF:UNG, or WOLF:ELK. We fitted all models in R version 2.4.1 using the function multinom in the nnet package (R Development Core Team 2006). Models were compared using Akaike’s Information Criterion corrected for small samples (AICc; Burnham and Anderson 2002). We calculated Akaike weights and evaluated the importance of each covariate by its predictor weight (wp), which we calculated by summing the Akaike weights for all models containing the covariate in the final model suite (Burnham and Anderson 2002). Our model selection followed a stepwise procedure within each suite, whereby we first fit all candidate models, calculated AICc and model weights and then determined for a given model structure which metric best estimated a given covariate. For example, if the three top models had identical structure and differed only by their inclusion of a different metric of wolf competition (WOLF:UNG, WOLF:ELK, MULTPK), then we determined which model was best-supported and removed the other two models from among suite comparisons. Among suites, we then recalculated model weights for the reduced set of models once we had determined the best metric for a given model structure. Because elk are considerably more vulnerable to wolf predation than bison (MacNulty 2002), we predicted that covariates of bison abundance (BISONadt, BISONcalf, BISON, BISON.ELK) would have no effect on the probability of wolves eating adult versus calf elk. We also predicted that elk abundance covariates (ELKadt, ELKcalf, ELK) would be negatively related to the probability of wolves eating a bison adult or bison calf given that low elk abundance (absolute and relative to bison) would likely compel wolves to kill bison with increasing frequency. In addition, we predicted the probability that wolves would kill bison compared to elk calves would increase with the relative and absolute numbers of bison because the number of vulnerable individuals in the bison population would likely increase with the influx of migrating animals during winter. We predicted that winters with more severe snow pack, as indexed by SWEacc, would be positively correlated with wolves killing both bison age classes and adult elk because the larger and less vulnerable species and age classes would become weakened and relatively more vulnerable. Given that elk and bison differ in their responses to wolves, with elk typically employing flight and bison resorting to group defense (Carbyn 1974, Carbyn and Trottier 1987, MacNulty 2002), the two snow pack metrics could have different effects on vulnerability across species. Specifically SWEmean could be more influential in predation of elk, as increasing values of SWEmean equate to decreased mobility, while SWEacc could be more influential in predation of bison by weakening their ability to defend themselves against attack. In addition, we predicted that competition imposed by multiple wolf packs (indexed by MULTPK, WOLF:UNG, and WOLF:ELK) would have a positive influence on the probability of wolves taking bison because packs competing for limited and decreasing elk resources would likely need to pursue other prey species to persist. Similarly, we predicted increasing wolf competition would result in increased selection for adult elk as wolves expanded through the study area. �Chapter 16 . Wolf Prey Selection: Choice or Circumstance? 315 Cooperative-hunting large carnivores often exhibit a positive relationship between group size and prey size (Rosenzweig 1966, Gittleman 1989, Creel and Creel 1995). However, the relationship between pack size and prey size for wolves is unclear (Mech and Boitani 2003). We did not expect pack size to be positively correlated with prey size (Chapter 17 by Becker et al., this volume), but added a pack size covariate (WOLFpk) to the best-supported a priori models to determine if the covariate improved model fit. C. Prey Switching Evaluating if wolves are capable of prey switching, or have a strong preference for elk regardless of bison abundance, cannot be determined by examining diet composition alone (Garrott et al. 2007). Thus, we evaluated wolf preference and potential prey-switching by relating the relative availability of bison and elk in the study system with the ratio of the two prey species in the wolves’ diet. Murdoch (1969) provided the classic equation that relates the ratio of two prey types eaten by a predator (g1/g2) to the ratio of the prey types available to the predator (N1/N2). We evaluated the existence and extent of wolf prey switching by regressing the ratio of bison and elk in wolf diets to the ratio of bison and elk available in the population and evaluating the subsequent form (i.e., linear or nonlinear) of the relationship (Murdoch 1969, Garrott et al. 2007). We used Murdoch’s (1969) selection coefficient where the ratio of the two prey types eaten is denoted by: gbison Nbison ¼c gelk Nelk ð16:1Þ The left-hand side of Equation 1 is the ratio of bison to elk in wolf diets and Nbison/Nelk is the ratio of bison to elk in the population. The proportionality constant, c, measures ‘‘the bias in the predator’s diet to one prey species’’ and relates the ratio of prey eaten to their relative abundance (Murdoch 1969:337). If wolves exhibit a high plasticity in their diet, then prey selection would likely change depending on the relative availability of the two prey types as determined by their abundance, vulnerability, and actual predator preference (Garrott et al. 2007). This dynamic nature of c can be incorporated by modifying the equation to allow changes in diet with changes in relative availability of elk and bison: � � gbison Nbison b ¼ c ð16:2Þ gelk Nelk The variable b is a measure of the extent of prey switching, with values greater than one denoting switching (Greenwood and Elton 1979, Elliot 2004). If wolves preferred elk proportional to relative abundance ratios of the two prey, then we would expect the relationship between diet and abundance ratios to be linear (Murdoch 1969, Garrott et al. 2007) because wolves would continue to prefer elk over all ranges of relative abundance ratios, even when elk were rare relative to bison (Figure 16.5). However, if wolves exhibit prey-switching then the relationship should be curvilinear and indicate a diet switch to the more abundant prey with increasing bison:elk ratios (Murdoch 1969, Garrott et al. 2007). Obtaining sufficient data on the ratio of bison to elk in wolf diets required a time-scale of three winter periods of approximately eight weeks each, during which time bison abundance varied substantially among drainages (Chapter 12 by Bruggeman et al., this volume). To account for this, we estimated prey abundance for the entire study system each winter period and calculated the ratio of the two prey species in wolf diets by pooling all wolf-killed elk and bison detected during the respective periods and deriving a ratio of bison to elk (gbison/gelk). Relative abundance ratios of prey (Nbison/Nelk) were estimated by calculating mean bison population estimates from surveys conducted during the early-, middle-, and late-winter periods and dividing by elk population estimates for these periods. �316 Becker et al. gbison/gelk (ratio in diet) 6 b=1 b=2 4 2 0 0 10 20 30 40 Nbison/Nelk (ratio available) 50 60 FIGURE 16.5 Theoretical relationship between the ratio of bison to elk in wolf diets versus the ratio of bison to elk available in the population. Curves for the scenarios for no prey switching (b ¼ 1) and prey switching (b ¼ 2) are depicted (Murdoch 1969, Garrott et al. 2007). Elk estimates by winter period were calculated similarly to the multinomial model. Twenty-seven data points were generated corresponding to nine years of three winter periods each. To determine the form of the relationship between the wolf diet ratio and the ratio of prey abundance, we fit Equation 2 to these data and estimated parameter coefficients using the nls function from the nlme package in R version 2.4.1 (R Core Development Team 2006). III. RESULTS Wolves were detected in the study area on 1306 days of the 1837 day study period, comprising a total of 16,801 wolf days, 1872 pack days, and 437 multiple pack days. We obtained 1369 telemetry locations, 534 visual locations, and 4175 km of backtracking. Approximately 6600 person days were spent in the field, and an estimated 368,000 km were logged on snowmobiles and vehicles. Wolf presence during the 167-day winter field season ranged from 60–3964 wolf days, with pack days and multiple pack days ranging from 19–383 and 0–128, respectively. Established packs ranged in size from 2 to 22 wolves (mean ¼ 9.6; 95% CI ¼ 9.4, 9.8), and the percentage of days wolves were detected during the field season ranged from 19% to 96%. Ten different wolf packs used the Madison headwaters area to varying degrees over the course of the study, with wolves first detected during the winter of 1996–1997 when several itinerant wolves used the area and the Nez Perce pack was soft released into the Firehole drainage (Chapter 15 by Smith et al., this volume). The Nez Perce pack became established in the study area during 1998–1999, and was the only resident pack until the winter of 2002–2003 when the Cougar pack and another uncollared pack in the Gibbon drainage used portions of the study area. Two more packs became established in the study area during 2003–2004, and wolf presence peaked in 2004–2005 with up to five packs totaling approximately 45 wolves using the study area (Chapter 15 by Smith et al., this volume). The wolf population decreased precipitously during winter 2005–2006 to primarily one pack, before increasingly to primarily three packs and an estimated 21 wolves the following winter (Chapter 15 by Smith et al., this volume). Elk population estimates for the study area ranged from 290–664 in autumn to 174–577 in spring, with the population decreasing 5–42% during winter. A progressive decrease in the elk population �Chapter 16 . Wolf Prey Selection: Choice or Circumstance? 317 began in 2003–2004 and continued through spring 2007 when the population was estimated at 174 animals (Chapter 23 by Garrott et al., this volume). Elk were equally distributed among the three drainages until the winter of 2000–2001 when the proportion of animals in the Madison drainage abruptly increased, accompanied by the virtual elimination of elk in the Gibbon drainage, and a gradual decrease in the Firehole drainage to a low of 28 animals (16% of the population) by the end of the study (Chapter 18 by Gower et al., and Chapter 21 by White et al., this volume). The proportion of elk in the Madison drainage gradually increased to 84% of the population in late-winter 2006–2007. Calf abundance also decreased 48–98% during each winter (Chapter 23 by Garrott et al., this volume). Elk abundance estimates within pack territories ranged from 28–331 total animals, 0–84 calves, and 28–271 adults. We conducted 114 ground distribution surveys of bison from 1997–1998 to 2006–2007, with abundance ranging from 205–1538 animals using the study area. Bison abundance generally increased as winter progressed and animals migrated into the study area from the Hayden and Pelican Valleys (Chapter 12 by Bruggeman et al., this volume). Estimated bison abundance in pack territories ranged from 20–1108 animals, with 3–271 calves and 17–952 adults. With fluctuating populations of both prey and predator within and across seasons, we also documented considerable variation in the ratios of bison:elk, wolf:ungulate, and wolf:elk. Estimates ranged from 0.10–29.57 for bison:elk ratios, 0.003–0.038 for wolf:ungulate ratios, and 0.006–0.100 for wolf:elk ratios, respectively. Snow pack accumulation in the study area typically began in late October (Chapter 6 by Watson et al., this volume) and increased until late March when spring melt began, particularly in the lowerelevation meadows and drainages. Snow pack during the course of the study was below historical averages, with annual maximum SWEacc values ranging from 1023–3612 cm days and averaging 2044 cm days (95% CI ¼ 2038, 2050). Maximum SWEmean values ranged from 9.8–31.9 cm and averaged 18.8 cm (95% CI ¼ 14.0, 23.6). A total of 759 wolf-killed ungulates and 21 canids were detected during the study period. Ungulate kills were comprised of elk (79.8%, n ¼ 606), bison (19.9%, n ¼ 151), moose (0.1%, n ¼ 1), and mule deer (Odocoileus hemionus; 0.1%, n ¼ 1), while wolf-killed canids consisted of coyotes (71.4%, n ¼ 15), wolves (23.8%, n ¼ 5), and red fox (4.8%, n ¼ 1). Detected kills varied from 14–106 among winters, with a mean of 81.0 kills (sd ¼ 16.8) following wolf establishment in 1998–1999. Elk were the primary prey species for wolves, with calves and adult females comprising 38% (n ¼ 292) and 32% (n ¼ 241) of total kills, respectively (Table 16.1). The percentage of bison in the pooled diets of resident wolf packs increased from zero soon after wolf recolonization to 53% (n ¼ 29) in winter 2005–2006, when bison TABLE 16.1 Numbers of bison and elk killed by wolves and detected in the Madison headwaters area of Yellowstone National Park during 1996–1997 through 2006–2007 Winter Total elk Total bison Elk calves Elk cows Elk bulls Bison calves Bison cows Bison bulls 1996–97 1997–98 1998–99 1999–00 2000–01 2001–02 2002–03 2003–04 2004–05 2005–06 2006–07 Totals 13 15 51 49 71 75 61 82 61 47 81 606 1 0 12 3 2 16 14 24 33 37 9 151 5 11 31 28 39 32 30 30 31 22 33 292 5 4 11 19 30 33 23 41 19 21 35 241 0 0 8 2 1 10 5 11 11 4 12 64 1 0 12 1 2 9 5 12 16 26 3 87 0 0 0 1 0 5 6 9 7 11 4 43 0 0 0 0 0 2 3 3 9 0 2 19 Age class and sex totals do not include kills that could not be categorized by age and sex. �318 Becker et al. 0.6 Bison Elk Proportion bison 70 0.5 60 0.4 Kills 50 0.3 40 30 0.2 20 Proportion of bison in diet 80 0.1 10 0 0 19981999 19992000 20002001 20012002 20022003 20032004 20042005 20052006 20062007 FIGURE 16.6 Wolf-killed elk and bison from resident packs in the Madison headwaters area of Yellowstone National Park during 1998–1999 through 2006–2007 (n ¼ 566). Bison steadily increased in the diet until the mild winter of 2006–2007 when considerably more elk were killed despite a substantial decrease in elk abundance. comprised the primary prey species (Figure 16.6). However, the proportion of bison in wolf diets decreased to 15% during winter 2006–2007 (n ¼ 9). Prey selection by wolves was predictably variable within winters, with elk calves primarily killed in early-middle winter, bison killed during middle-late winter, and adult elk killed throughout winter (Figure 16.7). Selection indices calculated for early-, middle-, and late-winter periods from 1998–1999 through 2006–2007 demonstrated a strong preference only for elk calves (mean ¼ 0.82, sd ¼ 0.12) throughout every winter, with selection by winter period summarized in Table 16.2. Mean ages of adult elk (n ¼ 280) and bison (n ¼ 44) killed by wolves were 8.3 years (95% CI ¼ 7.8, 8.8) and 10.0 years (95% CI ¼ 8.4, 11.6), respectively. Adult female elk killed by wolves were older than males, with a mean female age of 9.1 years (n ¼ 220; 95% CI ¼ 8.6, 9.7) and a mean male age of 5.6 years (n ¼ 58; 95% CI ¼ 4.6, 6.5). Ages of adult bison killed by wolves were older for females compared to males, with a mean age of 11.0 years (n ¼ 31; 95% CI ¼ 9.3, 12.6) and 7.7 years (n ¼ 13; 95% CI ¼ 4.3, 11.1) respectively, with considerably more variability for males. Wolves killed all age classes of adult bison and elk, but the highest proportion of kills was in the older age classes (Figures 16.8 and 16.9). The proportion of elk kills in the older age classes was higher during winters 2002–2003 through 2006– 2007 when the number of resident packs and wolves increased compared to winters 1998–1999 through 2002–2003 when Nez Perce was the primary resident pack (Figure 16.9). Marrow samples from 121 bison (52 adults, 69 calves) and 481 elk (275 adults, 206 calves) indicated condition decreased from early- to late-winter periods (Figure 16.10). Of the 120 elk jaws that were rated for necrosis, 50% (n ¼ 60) showed no signs of necrosis, 18% (n ¼ 21) were mild, 16% (n ¼ 19) moderate, and 17% (n ¼ 20) severe. Eleven bison jaws were rated, with 55% (n ¼ 6) showing no signs of necrosis, 36% (n ¼ 4) moderate, and 9% (n ¼ 1) severe. A. Factors Influencing Wolf Prey Selection We fitted 84 models from three a priori suites to data from 564 kills (216 elk calves, 223 elk adults, 69 bison calves, 56 bison adults) by resident wolf packs during 1998–1999 through 2006–2007. Model selection results supported two top models with Akaike model weights (wk) of 0.70 and 0.28, �Chapter 16 . Wolf Prey Selection: Choice or Circumstance? A 319 1.00 Proportion of wolf diet 0.80 0.60 0.40 0.20 0.00 Early winter Mid winter Bison calf Elk adult B Late winter Bison adult Elk calf 1.00 Proportion of wolf diet 0.80 0.60 0.40 0.20 0.00 Early winter Mid winter Bison calf Elk adult Late winter Bison adult Elk calf FIGURE 16.7 Winter trends in pooled diet composition from resident wolf packs in the Madison headwaters area of Yellowstone National Park during (A) 1998–1999 through 2002–2003 (n ¼ 262) and (B) 2003–2004 through 2006–2007 (n ¼ 302). Kills were classified by species, age class, and winter period (early ¼ 15 November—10 January; middle ¼ 11 January—6 March; late ¼ 7 March—30 April). �320 Becker et al. TABLE 16.2 Selection indices (Chesson 1978) for four prey types from resident wolf pack kills during winters 1998–1999 through 2006–2007 Index of Selectivity Prey Type Elk Calf Elk Adult Bison Calf Bison Adult Early Winter 0.752 0.176 0.059 0.013 Middle Winter 0.827 0.094 0.073 0.006 Late Winter 0.816 0.098 0.070 0.016 Kills were classified by species, age class, and winter period (early ¼ November 15–January 10; middle ¼ January 11–March 6; late ¼ March 7–April 30). For m prey types a value greater than 1/m (i.e., 0.25) indicates a preference. 0.35 Proportion of kills 0.30 0.25 0.20 0.15 0.10 0.05 0.00 Yearling Y2–5 (n = 6) (n = 5) Y6–9 (n = 7) Y10–13 Y14–17 Y18–20 (n = 11) (n = 14) (n = 1) FIGURE 16.8 Age distribution of adult bison killed by wolves in the Madison headwaters area of Yellowstone National Park during 1996–1997 through 2006–2007 (n ¼ 41). 0.50 Proportion of kills 0.40 1996–1997 to 2002–2003 (n = 133) 2003–2004 to 2006–2007 (n = 147) All winters (n = 280) 0.30 0.20 0.10 0.00 Yearling Y2-Y5 Y6-Y9 Y10-Y13 Y14-Y16 FIGURE 16.9 Age distribution of adult elk killed by wolves in the Madison headwaters area of Yellowstone National Park during 1996–1997 through 2002–2003 and 2003–2004 through 2006–2007. Reductions in younger age classes during the latter time period likely reflect lack of recruitment. �Chapter 16 . Wolf Prey Selection: Choice or Circumstance? 321 1.00 Proportion of kills 0.80 0.60 0.40 0.20 0.00 Early Middle 4 rating 3 rating Late 2 rating 1 rating FIGURE 16.10 Marrow classifications for ungulates killed by wolves in the Madison headwaters area of Yellowstone National Park during early, middle, and late winter, 1996–1997 through 2006–2007 (n ¼ 602). respectively (Table 16.3). The covariates ELKcalf, BISONcalf, WOLF:UNG, and SWEacc were included in each of these models, with predictor weights (wp) of 0.99, 0.99, 0.99, and 1.00, respectively. The structure of the two models differed only in the inclusion of ELKadt in the top model, with wp ¼ 0.71. All other models had DAICc>8.5 and differed in structure from the best-supported models by their lack of inclusion of the WOLF:UNG covariate. An exploratory analysis adding wolf pack size (WOLFpk) to the two best-supported models did not improve the top model, but improved the second best model’s AICc to 1259.56, with a resultant DAICc of 0.31. Elk abundance was negatively related to the probability that wolves would kill bison of both age classes relative to elk calves and, in particular, the abundance of elk calves was strongly negatively correlated with predation of adult bison (Table 16.4). The abundance of bison, as estimated by the covariate BISONcalf, was positively related to the probability of wolves killing a bison calf relative to an elk calf. This relationship was similar for bison adults, though confidence intervals for the coefficient estimates spanned zero. In contrast to our hypotheses, bison abundance was negatively correlated with the probability that adult elk would be killed relative to calf elk. The wolf:ungulate ratio was positively correlated with the predation probability of all other prey types relative to elk calves, but was strongest for the bison adult logit (Table 16.4). As predicted, there was not a significant positive relationship between pack size and prey size, and there was a negative correlation with the probability of predation of bison calves relative to elk calves. The effect of SWEacc was strongly positive for all logit equations, indicating that increasing snow pack resulted in increased probability of predation for all prey types relative to elk calves (Table 16.4). There were significant increases in the odds of elk adult, bison calf, and bison adult kills with increases in bison calf abundance, wolf:ungulate ratios, and SWEacc (Table 16.4). The odds of predation by wolves for adult elk were 0.57 lower for every 134 animal increase in bison calf abundance, 1.7 times greater for every 0.018 increase in wolf:ungulate ratios, and 6.5–7.6 times higher for each 1800 cm days �322 Becker et al. TABLE 16.3 A priori model structure and results from top models within and among suites for multinomial logit analyses of winter prey selection by resident wolf packs in a bison-elk system in the Madison headwaters area of Yellowstone National Park during 1998–99 through 2006–07. Covariate codes are the abundance of elk adults (ELKadt, ELKcalf), bison calves (BISONcalf), accumulated snow pack (SWEacc), wolf:elk ratio (WOLF:ELK), wolf:ungulate ratio (WOLF:UNG), and multiple pack days (MULTPK) Within Suite Model Structure Prey Suite ELKcalf þ BISONcalf ELKcalf þ ELKadt þ BISONcalf ELKcalf Prey þ Wolf Competition Suite ELKcalf þ BISONcalf þ WOLF:ELK ELKcalf þ BISONcalf þ MULTPK ELKcalf þ ELKadt þ BISONcalf þ WOLF:ELK ELKcalf þ BISONcalf þ WOLF:UNG ELKcalf þ ELKadt þ BISONcalf þ WOLF:UNG ELKcalf þ ELKadt þ BISONcalf þ MULTPK ELKcalf þ WOLF:UNG Prey þ Snow Pack Suite ELKcalf þ ELKadt þ BISONcalf þ SWEacc ELKcalf þ BISONcalf þ SWEacc ELKcalf þ ELKadt þ SWEacc ELKcalf þ SWEacc Prey þ Wolf Competition þ Snow Pack Suite ELKcalf þ ELKadt þ BISONcalf þ WOLF:UNG þ SWEacc ELKcalf þ BISONcalf þ WOLF:UNG þ SWEacc ELKcalf þ ELKadt þ BISONcalf þ WOLF:ELK þ SWEacc ELKcalf þ BISONcalf þ WOLF:ELK þ SWEacc ELKcalf þ ELKadt þ BISONcalf þ MULTPK þ SWEacc ELKcalf þ BISONcalf þ MULTPK þ SWEacc ELKcalf þ ELKadt þ WOLF:UNG þ SWEacc a Among Suites K AICc DAICc wk DAICc wk 3 4 2 1284.6 1286.8 1298.9 0.00 2.14 14.29 0.74 0.26 0.00 25.35 27.49 39.65 0.00 0.00 0.00 4 4 5 4 5 5 3 1283.4 1284.9 1285.7 1286.0 1287.2 1287.6 1300.2 0.00 1.51 2.28 2.57 3.84 4.16 16.77 0.43 0.20 0.14 0.12 0.06 0.05 0.00 24.15 a 26.43 a a a 40.91 0.00 0.00 5 4 4 3 1267.8 1269.1 1277.5 1279.5 0.00 1.33 9.67 11.70 0.65 0.34 0.01 0.00 8.54 9.87 18.21 20.24 0.01 0.01 0.00 0.00 6 5 6 5 6 5 5 1259.3 1261.1 1263.1 1263.7 1265.0 1265.0 1269.5 0.00 1.86 3.86 4.48 5.71 5.76 10.20 0.56 0.22 0.08 0.06 0.03 0.03 0.00 0.00 1.86 a a a a 10.20 0.70 0.28 0.00 0.00 Values not included in among suite evaluations due to identical model structure differing only in wolf competition metric. increase in SWEacc. The odds of predation for bison calves decreased 0.32 and 0.38 times with every 42 animal and 122 animal increase in elk calf and elk adult abundance, respectively. The odds of predation for bison calves increased 2.21–2.67 times per 134 animal increase in bison calf abundance and 4.6–11.8 times per 1800 cm days increase in SWEacc. The odds of predation for adult bison were 6.6–6.9 times greater for each 0.018 unit increase in wolf:ungulate ratio, 17.7–23.9 times greater for each 1800 cm days increase in SWEacc, and 0.03–0.05 times greater for every 42 animal increase in elk calf abundance. B. Wolf Prey Switching with Murdoch’s Equation There was a positive relationship between the ratios of bison to elk in wolf diets and the population (Figure 16.11). Fitting a non-linear model of Murdoch’s equation to the data indicated a curvilinear relationship, with c and b values of 0.229 (95% CI ¼ 0.203, 0.254) and 2.091 (95% CI ¼ 1.175, 3.007), �TABLE 16.4 Coefficient values (Bi), lower and upper 95% confidence intervals (in parentheses), and odds ratios for the three best approximating models identified through AIC model comparison techniques for prey selection by resident wolf packs on bison and elk in the Madison headwaters area of Yellowstone National Park during 1998–1999 through 2006–2007 Model Elk adult PSW15 PSW41 PSW41w Bison calf PSW15 PSW41 PSW41w Bison adult PSW15 PSW41 PSW41w ELKcalf wp ¼ 0.99 0.21 (�0.38, 0.80) 1.23 0.06 (�0.44, 0.56) 1.07 0.07 (�0.43, 0.57) 1.07 �0.31 (�1.49, 0.87) 0.73 �1.13 (�2.20, �0.06) 0.32 �1.06 (�2.13, 0.02) 0.35 �3.00 (�5.32, �0.69) 0.05 �3.46 (�5.53, �1.39) 0.03 �3.34 (�5.38, �1.30) 0.04 ELKadt wp ¼ 0.71 BISONcalf wp ¼ 0.99 SWEacc wp ¼ 1.00 WOLF:UNG wp ¼ 0.99 �0.21 (�0.68, 0.26) 0.81 �0.56 (�1.10, �0.02) 0.57 �0.51 (�1.04, 0.02) 0.60 �0.52 (�1.08, 0.05) 0.60 0.70 (�0.01, 1.42) 2.02 0.79 (0.08, 1.50) 2.21 0.98 (0.24, 1.73) 2.67 0.37 (�0.43, 1.17) 1.45 0.40 (�0.39, 1.19) 1.49 0.36 (�0.47, 1.19) 1.43 2.03 (1.13, 2.93) 7.60 1.87 (1.03, 2.72) 6.51 1.89 (1.01, 2.77) 6.62 2.47 (1.25, 3.68) 11.76 1.89 (0.75, 3.03) 6.60 1.52 (0.35, 2.70) 4.59 3.17 (1.83, 4.52) 23.90 2.87 (1.64, 4.11) 17.70 3.05 (1.71, 4.38) 21.07 0.51 (0.01, 1.02) 1.67 0.48 (�0.02, 0.98) 1.61 0.49 (�0.02, 0.99) 1.62 0.55 (�0.41, 1.51) 1.73 0.52 (�0.40, 1.44) 1.68 0.55 (�0.36, 1.45) 1.73 1.93 (0.94, 2.93) 6.91 1.89 (.90, 2.89) 6.65 1.89 (0.90, 2.88) 6.61 �0.96 (�1.64, �0.27) 0.38 �0.41 (�1.14, 0.32) 0.67 WOLFpack 0.03 (�0.46, 0.51) 1.03 �0.85 (�1.57, �0.12) 0.43 0.29 (�0.62, 1.20) 1.34 Covariate codes are the abundance of elk calves and adults (ELKcalf, ELKadt), bison calves (BISONcalf ), accumulated snow pack (SWEacc), wolf:ungulate ratio (WOLF:UNG), and wolf pack size (WOLFpack). Values in bold indicates confidence intervals do not span zero. �324 Becker et al. respectively. The majority of this relationship was supported by data collected during the winters of 2001–2002 through 2006–2007, in particular two points corresponding to the late-winter periods of 2004–2005 and 2005–2006 appeared to have considerable influence on the shape of the relationship (Figure 16.11). To evaluate the effect of these points we removed them and refit the models. The estimated values of c and b decreased to 0.15 (95% CI ¼ 0.08, 0.22) and 1.27 (95% CI ¼ 0.60, 1.94) respectively, while the standard error of each estimate increased, and thus the confidence interval of b included values less than 1 (Table 16.5). IV. DISCUSSION We contributed new insights into wolf prey selection by investigating a temporally and spatially dynamic system that was unaffected by human harvests and contained two prey species, each with age classes differing in their size and defenses relative to adults. While the complexity of interacting 2.50 gbison/gelk (ratio in wolf diet) 2.00 Predicted Observed 1.50 1.00 0.50 0.00 0 1 4 2 3 Nbison/Nelk (ratio available) 5 6 FIGURE 16.11 Observed versus predicted relationships between the ratio of bison:elk wintering in the Madison headwaters area of Yellowstone National Park and the ratio of bison:elk in wolf diets during 1998–1999 through 2006–2007. Predicted coefficients for fitted line are c ¼ 0.23 and b ¼ 2.09. TABLE 16.5 Coefficient values and lower and upper 95% confidence intervals from analyses of prey switching by wolves in a bison-elk system in the Madison headwaters area of Yellowstone National Park during 1998–1999 through 2006–2007 Parameter Estimates Model Structure (c*BISON:ELK)b Outliers Removed c 0.23 (0.20, 0.25) 0.15 (0.08, 0.22) b 2.09 (1.18, 3.01) 1.27 (0.60, 1.94) Values indicate confidence intervals do not span zero. The constant c measures the bias in a predator’s diet (Murdoch 1969), with values less than one indicating a preference for elk. The constant b measures the extent of prey-switching, with values greater than one indicating switching. The covariate code BISON:ELK measures the relative abundance of bison and elk in the system. �Chapter 16 . Wolf Prey Selection: Choice or Circumstance? 325 biotic and abiotic factors influencing prey selection in natural systems makes evaluations difficult, analyses across species and age classes demonstrated that selection was influenced by the absolute and relative abundance of prey types, the abundance of predators, and the duration of snow pack. Prey abundance, particularly elk calf and bison calf abundance, was important because wolves strongly preferred elk calves relative to all other prey types and elk calf abundance was inversely related to the occurrence of bison calves and adults in wolf diets. While wolves preferred elk to bison, their patterns of selection were also driven by the relative abundance of the two prey species with wolves killing disproportionately more bison at high bison:elk ratios, and the curvilinear form of Murdoch’s equation indicating prey switching. In addition to prey abundance, the influence of predator numbers as reflected in the wolf:ungulate ratio resulted in a broadening of wolf prey selection from elk calves, with increasing probabilities of different prey types in the diet with increasing ratios. Lastly, we evaluated both the movement-inhibiting and weakening influence of snow pack on prey and demonstrated that the probability of predation on both bison age classes and adult elk increased dramatically with increasing snow pack duration and accumulation. The profound influence of snow pack illustrates the important role of environmental variables on prey selection in large mammal systems. Preference for elk calves was strong within and among all years, presumably due to their small size and lack of defenses relative to other prey types. Unlike most small taxa, ungulates pose considerable injury risk to predators, and anti-predator defenses vary among species and age classes (Nelson and Mech 1981, Bergerud et al. 1984, Carbyn and Trottier 1987, Dale et al. 1995). Therefore, the ability of prey individuals to repel an attack can substantially influence large mammal predator-prey dynamics (Garrott et al. 2007). Elk typically employed flight as a primary anti-predator tactic and, as a result, elk calves did not benefit from group protection strategies such as those used by bison (Carbyn and Trottier 1987, Carbyn et al. 1993). Elk adults are more capable than calves of inflicting injury on wolves due to their larger size, strength, experience, and the presence of antlers on bulls. However, wolves also killed elk from all other age classes, with the proportion of older adult elk in wolf diets increasing in the latter years of the study due to the effects of consistently low recruitment on the age structure of the population (Chapter 23 by Garrott et al., this volume). While wolves preferred elk calves over adults they also preferred elk over bison due to differences in vulnerability, and selected both age classes of bison less than expected given their abundance. As in other studies (Carbyn et al. 1993, Smith et al. 2000, MacNulty et al. 2007), bison constituted an extremely formidable and dangerous prey to wolves due to their physical and behavioral defenses. We documented numerous instances of wolves seriously injuring bison and returning to kill and feed on them later (Carbyn et al. 1993). We also frequently witnessed bachelor and cow-calf herds continually defending injured or weak animals under attack by wolves (Carbyn and Trottier 1987, MacNulty et al. 2007). The majority of bison predation occurred in late winter when ungulates are likely in their most substandard physiological condition (Figure 16.12, Chapter 9 by White et al., this volume) and in the latter years of the study when bison were most abundant relative to elk (Chapter 12 by Bruggeman et al., this volume). The smallest bison prey with the fewest defenses, calves, comprised the majority of bison kills, followed by cows and bulls, respectively (Table 16.1). Bison adults of both sexes were likely weakened in late winter, but cows had increased energetic demands because they were at or nearing parturition (Chapter 14 by Geremia et al., this volume). Bison were more abundant than elk during all years of our study, and were predictably found feeding in the open meadow complexes (Chapter 28 by Bruggeman et al., this volume). Therefore encountering bison was unlikely to be a limiting factor influencing wolf predation on bison, even if a group was considered the unit of encounter (Huggard 1993a). Thus, while general models of predator behavior typically focus on encounter rates (Taylor 1984) we support the caution of Dale et al. (1995) against wolf predation models incorporating only encounter rates given the dramatic differences in vulnerability and risk of injury ungulates pose when attacked. While absolute and relative prey abundance and life history characteristics were influential, the physiological effect of snow pack, as indexed by SWEacc, had an overwhelming influence on predation �326 Becker et al. FIGURE 16.12 Winter-starved bull bison (A) and calf bison (B) in the Firehole river drainage. Late winter starvation was the primary source of mortality for both elk and bison prior to wolf recolonization and the weakening influence of snow pack made formidable prey such as bison considerably more vulnerable to wolf predation (Photos by Jeff Henry and Matt Becker). by wolves among the four prey types. Though snow depth has been frequently attributed to increased vulnerability due to its inhibition of movement in a predation event and its longer-term weakening influence on ungulates (Nelson and Mech 1986, Huggard 1993b, Mech and Peterson 2003), the relative importance of each has not been analyzed for specific predator-prey systems and prey species. While both metrics have important and interacting influences, the mean snow pack on the ground at the time of a kill (SWEmean) explained much less variation in prey selection than the duration and accumulation of snow up to that date (SWEacc). Physical condition of wolf-killed ungulates decreased through each �Chapter 16 . Wolf Prey Selection: Choice or Circumstance? 327 winter, consistent with nutritional profiles of the elk population (Chapter 9 by White et al., this volume), and winter starvation mortalities were the predominant source of mortality for both elk and bison prior to wolf recolonization (Chapter 11 by Garrott et al., this volume). The odds of wolf predation on bison increased many orders of magnitude with increasing accumulation and duration of snow pack, presumably weakening bison such that they were less able to defend themselves or their calves. While we did observe bison being killed in deep snow, observations of wolves attacking bison in late winter typically occurred in low snow meadow complexes and defense sometimes lasted several hours, as wolves continually attempted to isolate and injure vulnerable individuals. An animal in a weakened state is likely much less able to sustain such defense in the face of an attack. Snow pack is also highly influential in driving broad-scale movements of bison, such as their winter migrations into the Madison headwaters area and movements among drainages (Chapters 12 and 28 by Bruggeman et al., this volume). Snow pack also increased the vulnerability of adult elk to wolf predation both by weakening their condition and impeding their escape during flight. Because of their habitat selection and anti-predator defenses elk are likely more susceptible to environmental vulnerability in the form of hard habitat edges, structure, and changes in snow pack that can impede flight (Bergman et al. 2006). We frequently found wolf-killed elk that had been either encountered and killed in deep snow and complex forest structure or chased into it and killed. Thus, environmental vulnerability (Chapter 24 by Garrott et al., this volume) can assume considerable importance in large mammal predator-prey interactions given the severe weakening influence of snow pack on prey animals, the potential for snow pack effects to differ among prey species depending on their life history characteristics, and the potentially negative effect of edge and the accompanying differences in habitat structure. In addition to prey and snow pack variables, the ratio of wolves to ungulates was influential in wolf prey selection within and among species, likely due to a combination of wolf competition and elk antipredator behavior. The transition from a one-wolf-pack system to a multiple-wolf-pack system resulted in wolves occupying the entire study area and overlapping extensively, with inter-pack strife the main cause of wolf mortality (Chapter 15 by Smith et al., this volume) and wolf functional responses to elk best described as a Type II ratio-dependent response, indicating significant predator dependence (Chapter 17 by Becker et al., this volume). Although the winter range comprised a relatively small area, packs did not exhibit temporal avoidance in their use of the system and were routinely detected in the same drainages, though they typically avoided direct encounters (Chapter 15 by Smith et al., this volume). However, despite this intense use there was typically one dominant pack, and smaller packs were often displaced to more marginal areas of the study system. For example, the majority of the Biscuit Basin and Nez Perce pack territories overlapped with each other during winters 2003–2004 and 2004–2005; however when both packs were detected in the same drainage Nez Perce appeared to occupy the main hunting areas while Biscuit Basin was often displaced. Similar dynamics were apparent with the Hayden pack and dominant Gibbon pack in 2006–2007. Packs whose territories included large areas of marginal foraging based on the paucity of elk (e.g., Gibbon drainage, Chapter 21 by White et al., this volume) also preyed on adult elk and bison considerably more than packs occupying areas with relatively abundant elk, and the saturation of the system with wolves likely resulted in fewer places for adult elk to avoid encounters, such as bulls that often resided in minor drainages away from the main meadow complexes. This competition likely intensified as the elk population and the per-capita availability of calves decreased concurrent with a substantial distribution change within and among the three drainages (Chapter 18 by Gower et al., and Chapter 21 by White et al., this volume). While wolves killed elk calves throughout all winters, high wolf numbers in the system resulted in nearly all elk calves being killed by late winter in several of the latter years (Chapter 23 by Garrott et al., this volume). While wolf pack size did not appear to affect selection, we are not aware of other studies that demonstrated the influence of wolf population numbers on prey selection. This may be because most of these analyses were confined to a single prey species or prey item, established wolf-ungulate systems do not typically undergo the dramatic changes in predator and prey �328 Becker et al. populations that accompany newly-established systems, or the spatial and temporal dynamics of wolf territoriality in this system are unusual due to the high prey density of the Madison headwaters relative to the surrounding areas. The significance of the wolf:ungulate ratio was also likely related to elk anti-predator responses. The elk population experienced a substantial decrease and re-distribution following wolf restoration, with 84% of the population residing in the Madison drainage by the end of the study period compared to approximately equal proportions distributed among the three major drainages before wolves (Chapter 21 by White et al., this volume). In addition, elk increased the variability of their group sizes, changed their movements, and intensified their selection for habitats with high snow heterogeneity and possible escape terrain in response to increasing wolf numbers (Chapters 18, 19, and 20 by Gower et al., Chapter 21 by White et al., this volume). Under the behavioral resource depression hypothesis (Charnov et al. 1976), increasing predator presence should decrease the ability of individual predators, in this case packs, to capture prey due to increased wariness (and therefore decreased vulnerability) of prey (Chapter 17 by Becker et al., this volume). Prey-switching is typically thought to result in prey persistence because the relative rarity of the primary prey results in lower encounter and kill rates and, as a result, the predator switches to the more abundant prey (Murdoch 1969). However, rarity alone is likely insufficient for prey persistence and has not been well-demonstrated (Matter and Mannan 2005). In systems where large predators are being restored, changes in prey selection and the potential for prey-switching may be driven in part by a shift in prey behaviors as species adopt more effective anti-predator strategies to reduce their vulnerability rather than changes in predator preference. Virtually all studies of predation in large mammal multiple-prey systems report strong selection for certain prey species (Carbyn 1974, Potvin et al. 1988, Dale et al. 1995, Karanth and Sunquist 1995, Kunkel et al. 1999, Je˛drzejewski et al. 2000, Creel and Creel 2002, Sinclair et al. 2003, Hayward et al. 2006). However, this selection is not consistent among studies or species assemblages. For example, wolves primarily select caribou in some systems and not in others (Dale et al. 1995, Wittmer et al. 2005). Likewise, our documentation of elk as the primary prey of wolves was consistent with some investigations in other multiple-prey systems containing elk (Carbyn 1974, 1983, Huggard 1993a,b, Weaver 1994, Hebblewhite et al. 2003, Husseman et al. 2003, Smith et al. 2004) and contrary to others (Kunkel et al. 1999). These apparent contrasts have led some investigators to conclude that use of the term ‘‘preference’’ to describe wolf prey selection is inappropriate because wolves select individuals of whatever species are most profitable with the least risk (i.e., the most vulnerable; Mech and Peterson 2003). Defining preference as what a predator eats when all prey types are equally abundant and available confines investigations to the sophisticated cafeteria feeding trial experiments used on smaller taxa (Rodgers 1990). Thus, it is admittedly infeasible to determine if wolves have an inherent preference for a particular prey species. However, investigations of prey selection patterns in the context of natural multiple-prey systems, where preference is defined relative to the prey species assemblage available and influenced by the backdrop of landscape and climate variables upon which these interactions occur, have the potential to significantly advance our understanding of wolf-prey dynamics and help explain the prey selection contrasts observed among different systems (Garrott et al. 2007). We used data on wolf diet composition and relative abundance ratios to evaluate whether wolves switched from preying primarily on elk to bison. Murdoch (1969) demonstrated that values of c < 1 are indicative of preference, while values of b > 1 indicate switching. Based on this equation our analyses indicate that wolves in the Madison headwaters area had a strong preference for elk relative to bison (c ¼ 0.229), but switched to bison at high bison:elk ratios (b ¼ 2.091). Garrott et al. (2007) estimated Murdoch’s selection coefficient by decomposing c into vulnerability (v), preference (s), and biomass (m) to account for the profound differences in morphology and anti-predator defenses across ungulate prey species. Based on attack rate and attack success data on elk and bison in the northern portion of Yellowstone and the Pelican Valley (MacNulty 2002), Garrott et al. (2007) estimated the product of svm �Chapter 16 . Wolf Prey Selection: Choice or Circumstance? 329 as 0.04, considerably lower than what we estimated by fitting Murdoch’s equation to our data. This discrepancy may illustrate the difficulty of obtaining sufficient data on attack rates and success when decomposing c into svm, as well as the potential differences among wolf-elk-bison systems. Elk vastly outnumber bison on the northern range of Yellowstone, while the Pelican Valley prey base is primarily one bison herd of �150 animals (MacNulty 2002). Thus, it is reasonable to assume that differences in attack rates on the two prey species might differ as well among the three systems. While the utility of decomposing c into svm may primarily be confined to systems where such data are more readily collected (e.g., Scheel 1993, Creel and Creel 2002, MacNulty et al. 2007), simply estimating c and b for a given system can be accomplished with data on wolf diet composition and relative prey abundance (Garrott et al. 2007). The significance of distinguishing between Murdoch’s (1969) definition of prey switching and using prey switching to simply describe changes in diet ultimately concerns the possible regulatory effects of a predator. By having a preference for the more abundant prey and thereby presumably lessoning predation on the less abundant prey at the same time, the predator can exert strong stabilizing density-dependent effects on the system (Murdoch and Oaten 1975, Oaten and Murdoch 1975). However, much of the experimental work on switching assumes constant predator and prey abundance (Murdoch 1969, Messier 1995). Thus theoretical and empirical treatments of switching that do not consider a numerical response (Messier 1995) or that assume a constant handling time do not address situations where predators respond numerically or can decrease their handling time under certain circumstances when prey become more available. When these equations are applied to natural systems with these characteristics a curvilinear relationship can be derived, yet a predator can take disproportionate amounts of the relatively more abundant prey without diminishing their take of the relatively less abundant prey. In this situation, predators will not exert a stabilizing influence on the less abundant and preferred prey species and consequently the switching evaluations recommended by Garrott et al. (2007) require further refinement to account for potentially common scenarios in natural settings. Though Murdoch’s (1969) equation suggested wolves in the Madison headwaters switched to bison at high bison:elk ratios, we did not detect a concurrent switch away from their preferred prey, elk, and we did not have constant abundances for either wolves or ungulates. Variations in wolf kill rates on elk were not negatively related to bison abundance and the effect of increasing bison abundance and increasing snow pack duration and accumulation was simply to increase the total wolf kill rate and the wolf kill rate on bison rather than reduce the kill rate on elk (Chapter 17 by Becker et al., this volume). Furthermore, carcass consumption was negatively related to total kill rates and bison kill rate variation was best explained by snow pack or bison calf abundance (Chapter 17 by Becker et al., this volume). The curvilinear relationship indicating switching was also heavily leveraged by two points comprising the late winter periods of 2004–2005 and 2005–2006 respectively. Disproportionate bison selection by wolves in late winter 2004–2005 occurred during the peak of wolf abundance in the study area, with calf elk abundance decreasing to an estimated six animals by winter’s end. Peak snow pack accumulation was below average for the study period, but substantial numbers of bison relative to elk and high wolf abundance likely resulted in increased selection for bison. In contrast the winter of 2005–2006 followed a dramatic decrease in wolf abundance coupled with an above-average snow pack accumulation and high numbers of bison relative to elk, resulting in the highest proportion of bison killed during the study period. Nevertheless, elk continued to be preferred during the final winter of the study (2006– 2007), when elk numbers were at their lowest and the bison.elk ratios were near their highest. This was further corroborated by selection indices indicating continued high preference for elk calves with declining elk availability both within and among winters. Thus, it appears that wolf prey selection in this system is driven by a strong preference for the most vulnerable prey items (i.e., elk and elk calves in particular) and changes in prey selection are driven largely by circumstance (i.e., high bison:elk ratios; high wolf abundances; severe winters) rather than by a density-dependent change in wolf preference. Consequently the ecological relevance of prey-switching, namely its density-dependent stabilizing effects, do not appear to be present in this system at this time, perhaps best evidenced by a continued �330 Becker et al. decline of elk due to wolf predation (Chapters 23 and 24 by Garrott et al., this volume). Most natural systems with wolves, whose abundance has a strong positive relationship to prey biomass (Fuller 1989), are unlikely to have constant predator or prey abundance and stochastic processes can contribute to variable handling time. Thus we suggest continued refinement of prey-switching evaluations to account for this variability, and that evaluations with the definition provided by Murdoch (1969) should also consider whether there is a concurrent decrease in predation on the formerly more abundant and preferred prey, as herein lies the density-dependent stabilizing effect that is of primary ecological interest. Prior to reintroduction investigators predicted wolves would reduce the Yellowstone bison population by <15% (Boyce and Gaillard 1992, Boyce 1993). Bison predation park-wide was actually considerably less (<1%) during 1995–2000 (Smith et al. 2004). However Boyce (1995) did predict that prey switching from elk to bison could possibly occur in the Madison headwaters area in late winter. While we observed increased wolf predation on bison in late winter, it was unclear prior to our analyses whether this trend was driven by circumstance or by prey-switching to the increasing bison population and concurrently switching away from elk. Wolves are capable of subsisting almost exclusively on bison as evidenced in Wood Buffalo National Park (Carbyn et al. 1993). However, at this time wolves appear to primarily kill bison at high relative abundance ratios, particularly in severe winters and in times of high wolf abundances, with no indication that preference for elk has changed. Given this strong preference for elk it seems likely that elk numbers in the Madison headwaters area will continue to decrease to a low equilibrium, depending on their ability to escape predation via behavior or use of refuges (Creel et al. 2005, Hebblewhite et al. 2005, Chapters 18, 19, and 20 by Gower et al., Chapter 21 by White et al., and Chapter 24 by Garrott et al., this volume) that could produce pronounced switching away from elk and elk calves in particular. It is even possible that local extirpation of the Madison headwaters elk population could occur (Chapter 24 by Garrott et al., this volume). However, the dynamics of wolves, elk, and bison in the Madison headwaters area are still those of a developing system, with wolves present for little over a decade. Understanding patterns of prey selection, preference, and the presence or absence of prey-switching and their effects on community stability and persistence will require subsequent years of study to distinguish between transitory phenomena and the myriad influences of predator, prey, and environment in a newly-established large mammal system. V. SUMMARY 1. 2. 3. 4. We contributed new insights into wolf prey selection by investigating a temporally and spatially dynamic system in the Madison headwaters area of Yellowstone National Park that was unaffected by human harvests and contained two prey species (bison, elk) with age classes differing in their sizes and defenses. Prey selection by wolves was influenced by the absolute and relative abundance of prey types, the abundance of predators, and the duration of snow pack. Wolves strongly preferred elk calves relative to all other prey types, and elk calf abundance was inversely related to the occurrence of bison calves and adults in wolf diets. An increase in predator numbers, reflected in the wolf:ungulate ratio, resulted in a broadening of wolf prey selection from elk calves, with increasing probabilities of different prey types in the diet with increasing ratios. The probability of predation on both bison age classes and adult elk increased with increasing snow pack accumulation and duration, likely due to its long-term debilitating influence on ungulates that increased their vulnerability to wolves. �Chapter 16 5. 6. . Wolf Prey Selection: Choice or Circumstance? 331 We evaluated whether wolves switched prey from elk to bison using Murdoch’s (1969) equation and further evaluated potential changes in wolf preference using selection indices. While a curvilinear relationship existed between the ratio of bison to elk in wolf diets versus the ratio of bison to elk available in the population that suggested prey switching, confounding variability in wolf and prey numbers concurrent with no detected decrease in wolf preference away from elk did not support this stabilizing behavior. Comparative investigations of prey selection by wolves are complicated by the frequent use of terminology without consistent and explicit definitions and distinctions. Further investigations into evaluating prey-switching and wolf preference are recommended and utilizing Murdoch’s (1969) switching equation provides a rigorous evaluation of wolf preference and prey-switching, while eliminating inconsistencies in terminology. However, biologists should also consider whether there is a concurrent decrease in predation on the formerly more abundant and preferred prey to address confounding variability in predator and prey numbers common in natural systems. VI. REFERENCES Andersson, M., and S. Erlinge. 1977. Influence of predation on rodent populations. Oikos 29:591–597. Begon, M., J. L. Harper, and C. R. Townsend. 1996. Ecology. Blackwell Science, Malden, MA. Bergerud, A. T., H. E. Butler, and D. R. Miller. 1984. Antipredator tactics of calving caribou: Dispersion in the mountains. Canadian Journal of Zoology 62:1566–1575. Bergman, E. J., R. A. Garrott, S. Creel, J. J. Borkowski, and R. M. Jaffe. 2006. Assessment of prey vulnerability through analysis of wolf movements and kill sites. Ecological Applications 16:273–284. Bonsall, M. B., and M. P. Hassell. 1997. Apparent competition structures ecological assemblages. Nature 338:371–373. Boyce, M. S. 1993. Predicting the consequences of wolf recovery to ungulates in Yellowstone National Park. Pages 234–269 in R. S. Cook (Ed.) Ecological Issues on Reintroducing Wolves into Yellowstone National Park. Scientific Monograph NPS/ NRYELL/NRSM-93/22, USDI National Park Service, Denver, CO. Boyce, M. S. 1995. Anticipating consequences of wolves in Yellowstone: Model validation. Pages 199–209 in L. N. Carbyn, S. H. Fritts, and D. R. Seip (Eds.) Ecology and Conservation of Wolves in a Changing World. Canadian Circumpolar Institute, Edmonton, Alberta, Canada. Boyce, M. S., and J. -M. Gaillard. 1992. Wolves in Yellowstone, Jackson Hole, and the north fork of the Shoshone river: Simulating ungulate consequences of wolf recovery. Pages 4–71 to 4–115 in J. D. Varley and W. G. Brewster (Eds.) Wolves for Yellowstone? A Report to the United States Congress, Research and Analysis, Vol. 4. USDI National Park Service, Yellowstone National Park, WY. Burnham, K. P., and D. R. Anderson. 2002. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. Springer, New York, NY. Carbyn, L. N. 1974. Wolf Predation and Behavioral Interactions with Elk and other Ungulates in an Area of High Prey Diversity. Canadian Wildlife Service, Edmonton, Alberta, Canada. Carbyn, L. N. 1983. Wolf predation on elk in Riding Mountain National Park, Manitoba. Journal of Wildlife Management 47:963–976. Carbyn, L. N., and T. Trottier. 1987. Responses of bison on their calving grounds to predation by wolves in Wood Buffalo National Park. Canadian Journal of Zoology 65:2072–2078. Carbyn, L. N., S. M. Oosenbrug, and D. W. Anions. 1993. Wolves, Bison and the Dynamics Related to the Peace-Athabasca Delta in Canada’s Wood Buffalo National Park. Canadian Circumpolar Institute, Edmonton, Alberta, Canada. Caswell, H. 1978. Predator-mediated coexistence a nonequilibrium model. American Naturalist 110:141–151. Charnov, E. L., G. H. Orians, and K. Hyatt. 1976. Ecological implications of resource depression. American Naturalist 110:247–259. Cheatum, E. L. 1949. Bone marrow as an index of malnutrition in deer. New York State Conservation 3:19–22. Chesson, J. 1978. Measuring preference in selective predation. Ecology 59:211–215. Craighead, J. J., F. C. Craighead, Jr., R. L. Ruff, and B. W. O’Gara. 1973. Home ranges and activity patterns of nonmigratory elk of the Madison drainage herd as determined by biotelemetry. Wildlife Monographs 33:1–50. Creel, S., and N. M. Creel. 2002. The African Wild Dog: Behavior, Ecology, and Conservation. Princeton University Press, Princeton, NJ. �332 Becker et al. Creel, S., J. A. Winnie, Jr., B. Maxwell, K. Hamlin, and M. Creel. 2005. Elk alter habitat selection as an antipredator response to wolves. Ecology 86:3387–3397. Creel, S. R., and N. M. Creel. 1995. Communal hunting and pack size in African wild dogs, Lycaon pictus. Animal Behaviour 50:1325–1339. Dale, B. M., L. G. Adams, and R. T. Bowyer. 1994. Functional response of wolves preying on barren-ground caribou in a multipleprey ecosystem. Journal of Animal Ecology 63:644–652. Dale, B. M., L. G. Adams, and R. T. Bowyer. 1995. Winter wolf predation in a multiple ungulate prey system: Gates of the Arctic National Park, Alaska. Pages 223–230 in L. N. Carbyn, S. H. Fritts, and D. R. Seip (Eds.) Ecology and Conservation of Wolves in a Changing World. Canadian Circumpolar Institute, Edmonton, Alberta, Canada. Denno, R. F., C. Gratton, M. A. Peterson, G. A. Langellotto, D. L. Finke, and A. F. Huberty. 2002. Bottom-up forces mediate natural-enemy impact in a phytophagous insect community. Ecology 83:1443–1458. Denno, R. F., and M. A. Peterson. 2000. Caught between the devil and the deep blue sea: Mobile planthoppers elude natural enemies and deteriorating host plants. American Entomology 46:95–109. Elliott, J. M. 2004. Prey switching in four species of carnivorous stoneflies. Freshwater Biology 49:709–720. Endler, J. A. 1991. Interactions between predators and prey. Pages 169–196 in J. R. Krebs and N. B. Davies (Eds.) Behavioural Ecology: An Evolutionary Approach. Blackwell, Oxford, United Kingdom. Errington, P. L. 1946. Predation and vertebrate populations. Quarterly Review of Biology 21:145–177, 221–245. Eubanks, M. D., and R. F. Denno. 2000. Health food versus fast food: The effects of prey quality and mobility on prey selection by a generalist predator and indirect interactions among prey species. Ecological Entomology 25:140–146. Fitzgibbon, C. D., and J. Lazarus. 1995. Antipredator behavior of Serengeti ungulates: Individual differences and population consequences. Pages 274–296 in A. R. E. Sinclair and P. Arcese (Eds.) Serengeti II: Dynamics, Management, and Conservation of an Ecosystem. University of Chicago Press, Chicago, IL. Frank, D. A., and S. J. McNaughton. 1992. The ecology of plants, large mammalian herbivores, and drought in Yellowstone National Park. Ecology 73:2043–2058. Fryxell, J. M., and P. Lundberg. 1994. Diet choice and predator–prey dynamics. Evolutionary Ecology 7:379–393. Fuller, T. K. 1989. Population dynamics of wolves in north-central Minnesota. Wildlife Monographs 105:1–41. Fuller, T. K., and L. B. Keith. 1980. Wolf population dynamics and prey relationships in northeastern Alberta. Journal of Wildlife Management 44:583–602. Fuller, W. A. 1959. The horns and teeth as indicators of age in bison. Journal of Wildlife Management 23:342–344. Garrott, R. A., J. E. Bruggeman, M. S. Becker, S. T. Kalinowski, and P. J. White. 2007. Evaluating prey switching in wolf-ungulate systems. Ecological Applications 17:1588–1597. Garrott, R. A., L. L. Eberhardt, J. K. Otton, P. J. White, and M. A. Chaffee. 2002. A geochemical trophic cascade in Yellowstone’s geothermal environments. Ecosystems 5:659–666. Garrott, R. A., L. L. Eberhardt, P. J. White, and J. J. Rotella. 2003. Climate-induced variation in vital rates of an unharvested largeherbivore population. Canadian Journal of Zoology 81:33–45. Gittleman, J. L. 1989. Carnivore group living: Comparative trends. Pages 183–207 in J. L. Gittleman (Ed.) Carnivore Behavior, Ecology and Evolution. Cornell University Press, Ithaca, NY. Greene, C. H. 1986. Patterns of prey selection: Implications of predator foraging tactics. American Naturalist 128:824–839. Greenwood, J. J. D., and R. A. Elton. 1979. Analysing experiments on frequency-dependent selection in predators. Journal of Animal Ecology 48:721–737. Hamlin, K. L., D. F. Pac, C. A. Sime, R. M. DeSimone, and G. L. Dusek. 2000. Evaluating the accuracy of ages obtained by two methods for Montana ungulates. Journal of Wildlife Management 64:441–449. Hanski, I., L. Hansson, and H. Henttonen. 1991. Specialist predators, generalist predators, and the microtine rodent cycle. Journal of Animal Ecology 60:353–367. Hayward, M. W., P. Henschel, J. O’Brien, G. Hofmeyr, G. Balme, and G. I. H. Kerley. 2006. Prey preferences of the leopard (Panthera pardus). Journal of Zoology 270:298–313. Hebblewhite, M., P. C. Paquet, D. H. Pletscher, R. B. Lessard, and C. J. Callaghan. 2003. Development and application of a ratio estimator to estimate wolf kill rates and variance in a multiple-prey system. Wildlife Society Bulletin 31:933–946. Hebblewhite, M., C. A. White, C. G. Nietvelt, J. A. McKenzie, T. E. Hurd, J. M. Fryxell, S. E. Bayley, and P. C. Paquet. 2005. Human activity mediates a trophic cascade caused by wolves. Ecology 86:2135–2144. Hobbs, N. T. 1996. Modification of ecosystems by ungulates. Journal of Wildlife Management 60:695–713. Holling, C. S. 1959. The components of predation as revealed by a study of small mammal predation of European pine sawfly. Canadian Entomology 91:293–320. Holt, R. D. 1977. Predation, apparent competition, and the structure of prey communities. Theoretical Population Biology 12:197–229. Holt, R. D., and J. H. Lawton. 1994. The ecological consequences of shared natural enemies. Annual Review of Ecology and Systematics 25:495–520. Hopcraft, J. G. C., A. R. E. Sinclair, and C. Packer. 2005. Planning for success: Serengeti lions seek prey accessibility rather than abundance. Journal of Animal Ecology 74:559–566. �Chapter 16 . Wolf Prey Selection: Choice or Circumstance? 333 Hudson, R. J., J. C. Haigh, and A. B. Bubenik. 2002. Physical and physiological adaptations. Pages 255–257 in D. E. Toweill and J. Ward Thomas (Eds.) North American Elk: Ecology and Management. Smithsonian Institution Press, Washington, DC. Huggard, D. J. 1993a. Prey selectivity of wolves in Banff National Park I. Prey species. Canadian Journal of Zoology 71:130–139. Huggard, D. J. 1993b. Effect of snow depth on predation and scavenging by gray wolves. Journal of Wildlife Management 57:382–388. Hunter, M. D., and P. W. Price. 1992. Playing chutes and ladders: Heterogeneity and the relative roles of bottom-up and top-down forces in natural communities. Ecology 73:724–732. Husseman, J. S., D. L. Murray, G. Power, C. Mack, C. R. Wenger, and H. Quigley. 2003. Assessing differential prey selection patterns between two sympatric large carnivores. Oikos 101:591–601. Ivlev, V. S. 1961. Experimental ecology of the feeding of fishes. Yale University Press, New Haven, CT. Jaffe, R. 2001. Winter wolf predation in an elk-bison system in Yellowstone National Park, Wyoming. Thesis. Montana State University, Bozeman, MT. Je˛drzejewski, W., B. Je˛drzejewska, H. Okarma, and A. L. Ruprecht. 1992. Wolf predation and snow cover as mortality factors in the ungulate community of the Bialoweiza National Park, Poland. Oecologia 90:27–36. Je˛drzejewski, W., B. Je˛drzejewska, H. Okarma, K. Schmidt, K. Zub, and M. Musiani. 2000. Prey selection and predation by wolves in Bialowieza Primeval Forest, Poland. Journal of Mammalogy 81:197–212. Karanth, K. U., and M. E. Sunquist. 1995. Prey selection by tiger, leopard, and dhole in tropical forests. Journal of Animal Ecology 64:439–450. Krivan, V., and J. Eisner. 2003. Optimal foraging and predator-prey dynamics III. Theoretical Population Biology 63:269–279. Krebs, C. J., S. Boutin, and R. Boonstra (Eds). 2001. Ecosystem Dynamics of the Boreal Forest: The Kluane Project. Oxford University Press, New York, NY. Kruuk, H. 1972. The Spotted Hyena. University of Chicago Press, Chicago, IL. 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. Langellotto, G. A., and R. F. Denno. 2004. Responses of invertebrate natural enemies to complex-structured habitats: A metaanalytical synthesis. Oecologia 139:1–10. Lechowicz, M. J. 1982. The sampling characteristics of electivity indices. Oecologia 52:22–30. Lind, J., and W. Cresswell. 2005. Determining the fitness consequences of anti-predation behavior. Behavioral Ecology 16:945–956. MacNulty, D. R. 2002. The predatory sequence and the influence of injury risk on hunting behavior in the wolf. Thesis. University of Minnesota, Saint Paul, MN. MacNulty, D. R., L. D. Mech, and D. W. Smith. 2007. A proposed ethogram of large-carnivore predatory behavior, exemplified by the wolf. Journal of Mammalogy 88:595–605. Makacha, S., and G. B. Schaller. 1969. Observations on lions in Lake Manyara National Park, Tanzania. East African Wildlife Journal 7:99–103. Matter, W. J., and R. W. Mannan. 2005. How do prey persist? Journal of Wildlife Management 69:1315–1320. McNaughton, S. J. 1985. Ecology of a grazing system: The Serengeti. Ecological Monographs 55:259–295. Mech, D. L. 1970. The Wolf: The Ecology and Behavior of an Endangered Species. University of Minnesota Press, Minneapolis, MN. Mech, L. D., and L. Boitani. 2003. Wolf social ecology. Pages 1–34 in L. D. Mech and L. Boitani (Eds.) Wolves: Behavior, Ecology, and Conservation. University of Chicago Press, Chicago, IL. Mech, L. D., T. J. Meir, J. W. Burch, and L. G. Adams. 1995. Patterns of prey selection by wolves in Denali National Park, Alaska. Pages 231–243 in L. N. Carbyn, S. H. Fritts, and D. R. Seip (Eds.) Ecology and Conservation of Wolves in a Changing World. Canadian Circumpolar Institute, Edmonton, Alberta, Canada. Mech, L. D., and R. O. Peterson. 2003. Wolf-prey relations. Pages 131–160 in L. D. Mech and L. Boitani (Eds.) Wolves: Behavior, Ecology, and Conservation. University of Chicago Press, Chicago, IL. Menard, S. 2002. Applied Logistic Regression Analysis. Sage Publications, Thousand Oaks, CA. Menge, B. A. 1972. Foraging strategy of a starfish in relation to actual prey availability and environmental predictability. Ecological Monographs 42:25–50. Messier, F. 1995. On the functional and numerical responses of wolves to changing prey density. Pages 187–197 in L. N. Carbyn, S. H. Fritts, and D. R. Seip (Eds.) Ecology and Conservation of Wolves in a Changing World. Canadian Circumpolar Institute, University of Alberta, Edmonton, Alberta, Canada. Moffitt, S. A. 1998. Aging bison by the incremental cementum growth layers in teeth. Journal of Wildlife Management 62:1276–1280. Moran, M. D., T. P. Rooney, and L. E. Hurd. 1996. Top-down cascade from a bitrophic predator in an old-field community. Ecology 77:2219–2227. Morse, D. H. 1980. Behavioral Mechanisms in Ecology. Harvard University Press, Cambridge, MA. Murdoch, W. W. 1969. Switching in general predators: Experiments on predator specificity and the stability of prey populations. Ecological Monographs 39:335–354. Murdoch, W. W., and J. R. Bence. 1987. General predators and unstable prey populations. Pages 17–30 in W. C. Kerfoot and A. Sih (Eds.) Predation: Direct and Indirect Impacts on Aquatic Communities. University Press of England, Hanover, NH. �334 Becker et al. Murdoch, W. W., and J. R. Marks. 1973. Predation by coccinellid beetles: Experiments in switching. Ecology 54:160–167. Murdoch, W. W., and A. Oaten. 1975. Predation and population stability. Advances in Ecological Research 9:1–131. Murie, A. 1944. The wolves of Mount McKinley. Fauna of the national parks of the United States. Fauna Series 5:1–238. Nelson, M. E., and D. L. Mech. 1981. Deer social organization and wolf predation in northeastern Minnesota. Wildlife Monographs Number 77. Nelson, M. E., and D. L. Mech. 1986. Relationship between snow depth and gray wolf predation on white-tailed deer. Journal of Wildlife Management 50:471–474. Neter, J., M. H. Kutner, C. J. Nachtsheim, and W. Wasserman. 1996. Applied Linear Statistical Models. McGraw-Hill, Boston, MA. Oaten, A., and W. W. Murdoch. 1975. Switching, functional response, and stability in predator-prey systems. American Naturalist 109:299–319. Paine, R. T. 1966. Food web complexity and species diversity. American Naturalist 100:65–75. Parker, K. L., C. T. Robbins, and T.A Hanley. 1984. Energy expenditures for locomotion by mule deer and elk. Journal of Wildlife Management 48:474–488. Pastorok, R. A. 1981. Prey vulnerability and size selection by Chaoborus larvae. Ecology 62:1311–1324. Peckarsky, B., and M. A. Penton. 1989. Mechanisms of prey selection by stream-dwelling stoneflies. Ecology 70:1203–1218. Peterson, R. O. 1977. Wolf Ecology and Prey Relationships on Isle Royale. National Park Service Scientific Monograph Series 11, Washington, DC. Post, D. M., M. E. Conners, and D. S. Goldberg. 2000. Prey preference by a top predator and the stability of linked food chains. Ecology 81:8–14. Potvin, F., H. Jolicoeur, and J. Huot. 1988. Wolf diet and prey selectivity during two periods for deer in Quebec: Decline versus expansion. Canadian Journal of Zoology 66:1274–1279. Power, M. E., J. C. Marks, and M. S. Parker. 1992. Variation in the vulnerability of prey to different predators: Community level consequences. Ecology 73:2218–2223. Prugh, L. 2005. Coyote prey selection and community stability during a decline in food supply. Oikos 110:253–264. Quinn, J. L., and W. Cresswell. 2004. Predator hunting behaviour and prey vulnerability. Journal of Animal Ecology 73:143–154. R Development Core Team. 2006. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria<http://www.R-project.org>. Rodgers, A. R. 1990. Evaluating preference in laboratory studies of diet selection. Canadian Journal of Zoology 68:188–190. Rosenheim, J. A., T. E. Glik, R. E. Goeriz, and B. Rämert. 2004. Linking a predator’s foraging behavior with its effects on herbivore population suppression. Ecology 85:3362–3372. Rosenzweig, M. L. 1966. Community structure in sympatric carnivores. Journal of Mammalogy 47:602–612. Schaller, G. B. 1972. The Serengeti Lion: A Study of Predator–Prey Relations. The University of Chicago Press, Chicago, IL. Scheel, D. 1993. Profitability, encounter rates, and prey choice of African lions. Behavioral Ecology 4:90–97. Severinghaus, C. W. 1947. Relationship of weather to winter mortality and population levels among deer in the Adirondack region of New York. Transactions North American Wildlife Conference 12:212–223. Shupe, J. L., A. E. Olsen, H. B. Peterson, and J. B. Low. 1984. Flouride toxicosis in wild ungulates. Journal of the American Veterinary Medical Association 185:1295–1300. Sinclair, A. R. E., and P. Arcese. 1995. Population consequences of predation-sensitive foraging: The Serengeti wildebeest. Ecology 76:882–891. Sinclair, A. R. E., S. Mduma, and J. S. Brashares. 2003. Patterns of predation in a diverse predator-prey system. Nature 425:288–290. Smith, D. W., T. D. Drummer, K. M. Murphy, D. S. Guernsey, and S. B. Evans. 2004. Winter prey selection and estimation of wolf kill rates in Yellowstone National Park, 1995–2000. Journal of Wildlife Management 68:153–166. Smith, D. W., L. D. Mech, M. Meagher, W. E. Clark, R. Jaffe, M. K. Phillips, and J. A. Mack. 2000. Wolf-bison interactions in Yellowstone National Park. Journal of Mammalogy 81:1128–1135. Smith, D. W., R. O. Peterson, and D. B. Houston. 2003. Yellowstone after wolves. BioScience 53:330–340. Smuts, G. L. 1978. Interrelations between predators, prey, and their environment. BioScience 28:316–320. Stahler, D., B. Heinrich, and D. Smith. 2002. Common ravens, Corvus corax, preferentially associate with grey wolves, Canis lupus, as a foraging strategy in winter. Animal Behaviour 64:283–290. Stephens, D. W., and J. R. Krebs. 1986. Foraging Theory. Princeton University Press, Princeton, NJ. Synder, W. E., and A. R. Ives. 2001. Generalist predators disrupt biological control by a specialist parasitoid. Ecology 82:705–716. Taylor, R. J. 1984. Predation. Chapman & Hall, New York, NY. Temple, S. A. 1987. Do predators always capture substandard individuals disproportionately from prey populations? Ecology 68:669–674. Terborgh, J., L. Lopez, P. V. Nunez, M. Rao, G. Shahabuddin, G. Orihuela, M. Riveros, R. Ascanio, G. H. Adler, T. D. Lambert, and L. Balbas. 2001. Ecological meltdown in predator-free forest fragments. Science 294:1923–1925. Turchin, P., and I. Hanski. 1997. An empirically based model for latitudinal gradient in vole population dynamics. American Naturalist 149:842–874. �Chapter 16 . Wolf Prey Selection: Choice or Circumstance? 335 van Baalen, M., V. Krivan, and P. C. J. van Rijn. 2001. Alternative food, switching predators, and the persistence of predator-prey systems. American Naturalist 157:512–524. Weaver, J. L. 1994. Ecology of wolf predation amidst high ungulate diversity in Jasper National Park, Alberta. Dissertation. University of Montana, Missoula, MT. White, G. C., and R. A. Garrott. 1990. Analysis of Wildlife Radio-Tracking Data. Academic Press, San Diego, CA. Wittmer, H. U., A. R. E. Sinclair, and B. N. McLellan. 2005. The role of predation in the decline and extirpation of woodland caribou. Oecologia 144:257–267. APPENDIX Multinomial Model List Model Suite 1: Prey Models P1 ¼ PREY�ELKcalf P2 ¼ PREY�ELKcalf þ ELKad P3 ¼ PREY�ELKcalf þ ELKad þ BISONcalf P4 ¼ PREY�ELK P5 ¼ PREY�ELK þ BISON P6 ¼ PREY�BISON.ELK P7 ¼ PREY�ELKcalf þ BISONcalf Model Suite 2: Prey þ Snow Pack Models PS1 ¼ PREY�ELKcalf þ SWEacc PS2 ¼ PREY�ELKcalf þ ELKad þ SWEacc PS3 ¼ PREY�ELKcalf þ ELKad þ BISONcalf þ SWEacc PS4 ¼ PREY�ELK þ SWEacc PS5 ¼ PREY�ELK þ BISON þ SWEacc PS6 ¼ PREY�BISON.ELK þ SWEacc PS7 ¼ PREY�ELKcalf þ SWEmean PS8 ¼ PREY�ELKcalf þ ELKad þ SWEmean PS9 ¼ PREY�ELKcalf þ ELKad þ BISONcalf þ SWEmean PS10 ¼ PREY�ELK þ SWEmean PS11 ¼ PREY�ELK þ BISON þ SWEmean PS12 ¼ PREY�BISON.ELK þ SWEmean PS13 ¼ PREY�ELKcalf þ BISONcalf þ SWEacc PS14 ¼ PREY�ELKcalf þ BISONcalf þ SWEmean Model Suite 3: Prey þ Wolf Competition Models PW1 ¼ PREY�ELKcalf þ MULTPK PW2 ¼ PREY�ELKcalf þ ELKad þ MULTPK PW3 ¼ PREY�ELKcalf þ ELKad þ BISONcalf þ MULTPK PW4 ¼ PREY�ELK þ MULTPK PW5 ¼ PREY�ELK þ BISON þ MULTPK PW6 ¼ PREY�BISON.ELK þ MULTPK �336 PW7 ¼ PREY�ELKcalf þ WOLF:ELK PW8 ¼ PREY�ELKcalf þ ELKad þ WOLF:ELK PW9 ¼ PREY�ELKcalf þ ELKad þ BISONcalf þ WOLF:ELK PW10 ¼ PREY�ELK þ WOLF:ELK PW11 ¼ PREY�ELK þ BISON þ WOLF:ELK PW12 ¼ PREY�BISON.ELK þ WOLF:ELK PW13 ¼ PREY�ELKcalf þ WOLF:UNG PW14 ¼ PREY�ELKcalf þ ELKad þ WOLF:UNG PW15 ¼ PREY�ELKcalf þ ELKad þ BISONcalf þ WOLF:UNG PW16 ¼ PREY�ELK þ WOLF:UNG PW17 ¼ PREY�ELK þ BISON þ WOLF:UNG PW18 ¼ PREY�BISON.ELK þ WOLF:UNG PW19 ¼ PREY�ELKcalf þ BISONcalf þ MULTPK PW20 ¼ PREY�ELKcalf þ BISONcalf þ WOLF:ELK PW21 ¼ PREY�ELKcalf þ BISONcalf þ WOLF:UNG Model Suite 4: Prey þ Snow Pack þ Wolf Competition Models PSW1 ¼ PREY�ELKcalf þ MULTPK þ SWEacc PSW2 ¼ PREY�ELKcalf þ ELKad þ MULTPK þ SWEacc PSW3 ¼ PREY�ELKcalf þ ELKad þ BISONcalf þ MULTPK þ SWEacc PSW4 ¼ PREY�ELK þ MULTPK þ SWEacc PSW5 ¼ PREY�ELK þ BISON þ MULTPK þ SWEacc PSW6 ¼ PREY�BISON.ELK þ MULTPK þ SWEacc PSW7 ¼ PREY�ELKcalf þ WOLF:ELK þ SWEacc PSW8 ¼ PREY�ELKcalf þ ELKad þ WOLF:ELK þ SWEacc PSW9 ¼ PREY�ELKcalf þ ELKad þ BISONcalf þ WOLF:ELK þ SWEacc PSW10 ¼ PREY�ELK þ WOLF:ELK þ SWEacc PSW11 ¼ PREY�ELK þ BISON þ WOLF:ELK þ SWEacc PSW12 ¼ PREY�BISON.ELK þ WOLF:ELK þ SWEacc PSW13 ¼ PREY�ELKcalf þ WOLF:UNG þ SWEacc PSW14 ¼ PREY�ELKcalf þ ELKad þ WOLF:UNG þ SWEacc PSW15 ¼ PREY�ELKcalf þ ELKad þ BISONcalf þ WOLF:UNG þ SWEacc PSW16 ¼ PREY�ELK þ WOLF:UNG þ SWEacc PSW17 ¼ PREY�ELK þ BISON þ WOLF:UNG þ SWEacc PSW18 ¼ PREY�BISON.ELK þ WOLF:UNG þ SWEacc PSW19 ¼ PREY�ELKcalf þ MULTPK þ SWEmean PSW20 ¼ PREY�ELKcalf þ ELKad þ MULTPK þ SWEmean PSW21 ¼ PREY�ELKcalf þ ELKad þ BISONcalf þ MULTPK þ SWEmean PSW22 ¼ PREY�ELK þ MULTPK þ SWEmean PSW23 ¼ PREY�ELK þ BISON þ MULTPK þ SWEmean PSW24 ¼ PREY�BISON.ELK þ MULTPK þ SWEmean PSW25 ¼ PREY�ELKcalf þ WOLF:ELK þ SWEmean PSW26 ¼ PREY�ELKcalf þ ELKad þ WOLF:ELK þ SWEmean PSW27 ¼ PREY�ELKcalf þ ELKad þ BISONcalf þ WOLF:ELK þ SWEmean PSW28 ¼ PREY�ELK þ WOLF:ELK þ SWEmean PSW29 ¼ PREY�ELK þ BISON þ WOLF:ELK þ SWEmean PSW30 ¼ PREY�BISON.ELK þ WOLF:ELK þ SWEmean PSW31 ¼ PREY�ELKcalf þ WOLF:UNG þ SWEmean PSW32 ¼ PREY�ELKcalf þ ELKad þ WOLF:UNG þ SWEmean Becker et al. �Chapter 16 . Wolf Prey Selection: Choice or Circumstance? PSW33 ¼ PREY�ELKcalf þ ELKad þ BISONcalf þ WOLF:UNG þ SWEmean PSW34 ¼ PREY�ELK þ WOLF:UNG þ SWEmean PSW35 ¼ PREY�ELK þ BISON þ WOLF:UNG þ SWEmean PSW36 ¼ PREY�BISON.ELK þ WOLF:UNG þ SWEmean PSW37 ¼ PREY�ELKcalf þ BISONcalf þ MULTPK þ SWEacc PSW38 ¼ PREY�ELKcalf þ BISONcalf þ MULTPK þ SWEmean PSW39 ¼ PREY�ELKcalf þ BISONcalf þ WOLF:ELK þ SWEacc PSW40 ¼ PREY�ELKcalf þ BISONcalf þ WOLF:ELK þ SWEmean PSW41 ¼ PREY�ELKcalf þ BISONcalf þ WOLF:UNG þ SWEacc PSW42 ¼ PREY�ELKcalf þ BISONcalf þ WOLF:UNG þ SWEmean 337 � 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 Journal Articles Description An account of the resource CPW peer-reviewed journal publications Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. 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 Wolf prey selection in an elk-bison system: choice or circumstance? Description An account of the resource <span>What a predator eats when given choices, and the subsequent effects of this behavior on ecosystem stability, has long been a topic of interest for ecologists. <a href="https://www.sciencedirect.com/topics/earth-and-planetary-sciences/prey-selection" title="Learn more about Prey Selection from ScienceDirect's AI-generated Topic Pages" class="topic-link" target="_blank" rel="noreferrer noopener">Prey selection</a> is influenced by the absolute and relative abundances of prey types, the life history characteristics of predators and prey, and the attributes of the environment in which these interactions occur. Strong preference by a predator for a particular prey type can lead to ecosystem instability, while prey switching can lessen predation effects on the less abundant prey and enhance system stability. Evaluating prey selection in large mammal systems is difficult due to the broad spatial and temporal scales at which these predatory interactions occur, and investigations, particularly with wolf-ungulate systems, typically involve only the primary prey. Multiple prey species characterize most large mammal predator-prey systems, therefore research into predator-multiple prey dynamics has the potential to yield important ecological insights. We studied winter prey selection during 1996–1997 through 2006–2007 in a newly established wolf-elk-bison system where prey differed substantially in their vulnerability to wolf (</span><em>Canis lupus</em><span>) predation and wolves preyed primarily on elk (</span><em>Cervus elaphus</em><span>) but also used bison (</span><em>Bison bison</em><span>) to varying degrees within and among winters and packs. We analyzed the relative influences of prey abundance, predator abundance, and environmental variables on the selection of prey species and age classes and evaluated whether wolves exhibited prey switching from elk to bison.</span> Bibliographic Citation A bibliographic reference for the resource. Recommended practice is to include sufficient bibliographic detail to identify the resource as unambiguously as possible. Becker, M. S., R. A. Garrott, P. J. White, C. N. Gower, E. J. Bergman and R. Jaffe. 2008. Wolf prey selection in an elk-bison system: choice or circumstance? Pages 305-337 <em>in</em> Garrott, R.A., P.J. White and F.G.R. Watson, editors. The ecology of large mammals in central Yellowstone: sixteen years of integrated field studies. Academic Press, New York, New York, USA. <a href="https://doi.org/10.1016/S1936-7961(08)00216-9" target="_blank" rel="noreferrer noopener">https://doi.org/10.1016/S1936-7961(08)00216-9</a> Creator An entity primarily responsible for making the resource Becker, Matthew S. Garrott, Robert A. White, P.J. Gower, Claire N. Bergman, Eric J. Jaffe, Rosemary Subject The topic of the resource Prey selection Wolf Elk Bison Extent The size or duration of the resource. 33 pages Date Created Date of creation of the resource. 2008-11-14 Rights Information about rights held in and over the resource <a href="http://rightsstatements.org/vocab/InC-NC/1.0/" target="_blank" rel="noreferrer noopener">In Copyright - Non-Commercial Use Permitted</a> Format The file format, physical medium, or dimensions of the resource application/pdf Language A language of the resource English Is Part Of A related resource in which the described resource is physically or logically included. Terrestrial Ecology Type The nature or genre of the resource Article