Dietary Patterns Suggest West Virginia Bobcats Are Generalist Carnivores

Understanding a species' dietary patterns can guide management of that species and its biological community by improving our understanding of its ecology, trophic interactions, and functional role on the landscape. Bobcats Lynx rufus are opportunistic hunters that take advantage of readily available prey (Baker et al. 2001). Common prey items throughout their range in North America include ungulates (family Cervidae), lagomorphs (family Leporidae), rodents (order Rodentia), and birds (class Aves), with the selection of each prey species varying by locality and bobcat sex, size, and age (Litvaitis et al. 1986; Labisky and Boulay 1998; Brockmeyer and Clark 2007; Newbury 2013). In the mid-Atlantic region of the United States, bobcats consume white-tailed deer (hereafter deer) Odocoileus virginianus, lagomorphs, and small rodents most frequently, with females selecting for smaller prey items more often than males (Fox and Fox 1982; McLean et al. 2005; Rose and Prange 2015).

Common bobcat prey species often prefer early successional stages of forests with regular disturbance, such as provided by clearcutting and burning, for food and cover rather than old-growth forests lacking groundcover and understory structure (Boyle and Fendley 1987; Brockmeyer and Clark 2007). Mast—a term used for the fruits (nuts and seeds) produced by trees—is an important and common food item for many wildlife species and is especially critical for overwinter survival of squirrels and other small rodents. However, forests throughout West Virginia have matured since the last bobcat food habits study nearly 40 y ago, which found that winter diets of bobcats were dominated by deer, small rodents, and rabbits (family Leporidae; Fox and Fox 1982). Forest maturation has led to increases in average diameter at breast height of trees, whereas the total number of trees has decreased because of overcrowding (Widmann et al. 2007). Thus, important mast-producing tree species, such as oak Quercus spp., American beech Fagus grandifolia, and hickory Carya spp. (Schuler et al. 2017), are declining disproportionately within overstories and understories of forests in West Virginia. These losses are due to many factors, including high-grade timbering and intense deer browsing (Grushecky et al. 2006; Miller et al. 2009; Rentch et al. 2010), competition with shade-tolerant species (Gundy et al. 2014; Schuler et al. 2017), and lack of fire regimes (Saladyga 2017). Because of these changes in forest composition in West Virginia, we expected to see a shift in bobcat dietary composition from the previous findings of the 1977–1978 harvest season due to decreases in mast-dependent prey species such as deer and small rodents.

Density-dependent and -independent factors may also play a role in bobcat prey consumption. Density-dependent factors may limit the abundance of prey available to specific individuals. Bobcats are territorial and maintain home ranges of approximately 12.1 km² as adult females and 33.9 km² as adult males according to a study in nearby western Virginia (McNitt et al. 2020). Of interest, a study in eastern West Virginia estimated average home range sizes of 20.0 km² for adult females, 41.1 km² for adult males, 11.7 km² for juvenile females, and 24.1 km² for juvenile males (Edwards et al. 2021). Although individuals of the same sex rarely overlap their home ranges, males will often overlap with one or more females' home ranges (McCord and Cardoza 1982; Rolley 1985; Rucker et al. 1989). Therefore, there may be increased competition for food resources between adults whose territories overlap. Similarly, dispersing juveniles will often cross multiple adult bobcats' home ranges in search of suitable territory (Kitchings and Story 1984; Rucker et al. 1989), creating additional competition among individuals for food resources.

Juvenile bobcats eat medium-sized or smaller prey twice as often as adults, likely because younger bobcats have not developed the skills necessary to capture larger prey items (Fritts and Sealander 1978; Litvaitis et al. 1986; McLean et al. 2005). However, diets of juvenile and adult bobcats, regardless of sex, may not differ over winter (Godbois et al. 2003; McLean et al. 2005; Rose and Prange 2015) as individuals will ingest carrion when available (DeVault and Rhodes 2002; Brockmeyer and Clark 2007; Hansen 2007; Platt et al. 2010). Bobcats also exhibit sexual dimorphism throughout most of their geographic range such that adult males tend to be larger than adult females. Thus, sex may also influence the bobcat's consumption of certain prey items, as adult males often hunt and eat larger prey items such as deer (Fritts and Sealander 1978; Litvaitis et al. 1984; Matlack and Evans 1992; Anderson and Lovallo 2003).

Density-independent drivers of bobcat prey consumption include factors that may affect annual mast production and a bobcat's ability to hunt during winter, such as heavy snowfall impeding movement. Many common prey species are dependent on hard (i.e., nuts) and soft (i.e., fruits) mast (Wentworth et al. 1990; Feldhamer 2002; Steffen et al. 2002), with winter survival of these species dependent upon the previous year's mast production (Litvaitis et al. 1986; McShea and Healy 2002; Ryan et al. 2004). Mast drives reproductive success for many mast-dependent species such that an increase in mast production enhances mast-dependent species reproduction (Holling 1959; McShea and Healy 2002), providing increased prey for bobcats. Thus, years of low mast production may result in fewer food resources for bobcats the following year. Additionally, bobcats are not equipped with wide, heavily furred paws to support hunting on heavy snowfall like their nearby counterparts—Canada lynx Lynx canadensis (Fritts and Sealander 1978; Mautz and Pekins 1989; Peers et al. 2013)—which may increase their consumption of readily available items, such as carrion from hunter-harvested game or bait piles placed by trappers.

Using stomach contents of bobcats harvested in West Virginia, our research objectives were to 1) identify important winter prey items for bobcats; 2) compare the relative occurrences of prey items among bobcats of differing ages, sexes, and body conditions, and across harvest seasons and ecoregions; 3) estimate overlap and diversity indices of prey items; 4) determine effects of hard mast and snowfall on bobcat prey consumption; and 5) compare results with a previous study from the late 1970s. Identifying dietary habits of differing bobcat sexes and stage classes over winter provides information that can be later used to evaluate differing nutritional requirements of individuals and to identify factors that may influence resource selection and winter survival across bobcat sexes and stage classes. Additionally, understanding changes in observed bobcat dietary patterns over time can help guide management actions to increase prey availability for bobcats by managing for the dietary requirements of their common prey species.

West Virginia (62,755 km²) consists of 55 counties that are divided into six ecoregions (Uhlig and Wilson 1952): Appalachian Ridges and Valleys (ecoregion 1), Allegheny Plateau (ecoregion 2), Cumberland Mountains (ecoregion 3), Monongahela and Upper Ohio (ecoregion 4), Northern Ohio and West Virginia Hills (ecoregion 5), and Southern Ohio and West Virginia Hills (ecoregion 6). The state is 78% forested, with 77% of those forested areas dominated by oak–hickory forests (Widmann et al. 2007). Trees with the highest production of hard mast in West Virginia include beech (Fagus spp.), black oak Quercus velutina, red oak Quercus rubra, chestnut oak Quercus montana, hickory, scarlet oak Quercus coccinea, scrub oak Quercus ilicifolia, walnut (Juglans spp.), and white oak Quercus alba (Richmond et al. 2013, 2014).

Elevations in West Virginia range from 73 m above sea level at Harper's Ferry near the Potomac River in Jefferson County (ecoregion 1) to 1,482 m above sea level at Spruce Knob in the Allegheny Mountains of Pendleton County (ecoregion 1; Mullennex 2010). West Virginia has four distinct seasons, with mean annual snowfall ranging from 51 to 368 cm among ecoregions (NCDC 2016). Annual precipitation ranges from 137 cm in the central Allegheny Plateau, where annual temperatures average 9.4°C, to 93 cm in the Appalachian Ridges and Valleys, where annual temperatures average 11.7°C (NCDC 2016).

We obtained bobcat carcasses from hunters and trappers during the 2014–2015 and 2015–2016 harvest seasons (November–February), then randomly selected an equal subset of individuals from each season to determine winter dietary habits. We necropsied carcasses to obtain sex, age, stomach content, and morphometric data, which included body mass (without pelt) and mean kidney and kidney fat masses for each carcass. We extracted and sent lower canines to Matson's Laboratory (Manhattan, Montana) for age analysis. We sorted individuals into three stage classes according to growth patterns, maturity, and expected foraging behaviors: juveniles (< 1 y), yearlings (1–2 y), and adults (> 2 y; Landry 2017).

To evaluate dietary composition, we flushed stomach contents through a 0.5-mm mesh sieve, then dried the remaining content in an oven at 65–70°C for 12–24 h following methods of Korschgen (1980) and Litvaitis et al. (1984). We then weighed and recorded stomach contents following the point-frame method of Chamrad and Box (1964). We recorded contents as meat, bone, feather, hair, vegetation, or trap debris using 25 sampling points per frame. Trap debris consisted of wire, mesh, cotton fibers, netting, rocks, foam, rope, rubber, plastic, cloth, tape, and paper that were likely ingested by the individual while in containment (Fritts and Sealander 1978). West Virginia Department of Natural Resources (WVDNR) biologists submitted an anonymous survey to 133 trappers who donated bobcat carcasses to the project over the 2015–2016 harvest season to determine trapper bait use. We used results from this survey to determine potential biases in stomach content of bobcats that may have ingested trap bait and vegetation or other materials at the trap site.

We used meat, bone, and feather samples to assist with macroscopic identification of prey items, and we created hair slides and cuticle impressions to identify hair samples microscopically following the methods of Litvaitis et al. (1984) and Quadros (2006). We identified hair samples using guard-hair medulla and scale-pattern reference keys (Spence 1963; Teerink 1991; Lanszki et al. 2015) and medulla reference slides. We identified hair samples to the lowest possible taxonomic level, then separated them into their taxonomic family for ease of identification. We could only identify meat samples that contained hair or bone on the sample. We identified feathers to taxonomic class. After identification, we categorized prey items into 10 core groups for statistical analyses: birds, deer, rabbits (including hares), mice and rats (families Cricetidae, Dipodidae, and Muridae), squirrels (tree squirrels [Sciurus spp.], Eastern chipmunks Tamias striatus, and groundhogs Marmota monax), moles and shrews (families Soricidae and Talpidae [or order Insectivora]), Virginia opossum Didelphis virginiana, skunks Mephitis mephitis and Spilogale putorius, raccoon Procyon lotor, and “others” (Table S1). Others consisted of prey items that occurred in < 5% of stomachs (i.e., porcupine Erethizon dorsatum, North American beaver Castor canadensis, canids (family Canidae), felids (family Felidae), and mustelids (family Mustelidae).

We determined frequencies of bobcats collected across West Virginia by sex and stage class, and by harvest method (i.e., trapping vs. hunting), ecoregion, year, and month. We created an inverse kidney fat index (KFI_i) for body condition using mean kidney fat mass, measured following the methods of Finger et al. (1981), divided by the sum of mean kidney mass and mean kidney fat mass, then multiplied by 100 (Landry 2017). We assessed correlation between morphometric variables using Pearson's r coefficient. We then evaluated bobcat body condition (dependent variable) and mass (dependent variable) by stage class and sex (independent variables) using factorial analyses of variance (ANOVAs).

We calculated relative frequency of occurrence of prey items as the number of individual bobcats that ingested a certain prey item divided by the total number of bobcats sampled (Korschgen 1980; Corbett 1989; Van Dijk et al. 2007) to determine bobcat dietary consumption rates (Rose and Prange 2015). We evaluated stomach contents as presence or absence of specific prey group items, rather than measuring volume of differing prey items per stomach because of differences in digestibility among individuals, prey type, and amounts of prey items ingested at time of harvest (Baker et al. 1993; Brockmeyer and Clark 2007; Cherry et al. 2016). We evaluated the presence of vegetation, debris, and each taxonomic prey group using binomial generalized linear models (hereafter, logistic regression) to evaluate occurrence rates of prey items eaten by bobcats as a function of sex, stage class, and morphometric variables. We also evaluated prey item occurrence rates by harvest month, harvest season, and ecoregion to identify temporal or spatial effects on observed bobcat dietary content. We then used the “predict.glm” function to obtain fitted estimates and standard errors from the glm object, which we then used to evaluate and interpret 95% confidence intervals for the estimates of prey items that occurred at significantly different rates (α < 0.05).

We used function “prop.test” to model tests of independent proportions (χ²) to determine whether there were significant differences in overall stomach content by sex, stage class, and ecoregion of harvest. Pianka's index, which evaluates overlap in niche partitioning between categories (i.e., species and sex; Pianka 1973), was used to determine the amount of dietary overlap between males and females and among the three stage classes (Hill 1973; McNeil et al. 2017). We calculated Simpson's index of diversity and evenness to evaluate the diversity of ingested prey items. Simpson's index of diversity (1 − D) determines the probability that two randomly selected prey items eaten by bobcats belong to different taxonomic prey species (Simpson 1949). Evenness (E) measures how similar the occurrences of prey items are for each taxonomic prey group, such that similar proportions of all prey groups result in an evenness value of one (Hill 1973). Simpson's index is influenced by species' evenness and richness, such that the index value will decrease as percent evenness increases for a given species richness (Whittaker 1965; Hill 1973).

We obtained annual values of mast production (i.e., indices of mast abundance) collected by WVDNR biologists from 2013 and 2014 to represent potential changes in prey occurrence rates of bobcat dietary contents the following year (Richmond et al. 2013, 2014). Mast indices, ranging from 0 to 100, provide an annual measure of relative abundance for 18 different soft and hard mast-producing species throughout West Virginia. They are calculated using subjective measurements of mast abundance (i.e., abundant, common, or scarce) for each species per ecoregion (Richmond et al. 2013, 2014). Our analyses did not indicate that soft mast was a significant driver of dietary content in bobcats, which is to be expected since hard mast (i.e., nuts) is the dominant foraging item of the most common bobcat prey items in the literature, including small rodents, rabbits, and deer (Table S1; McLean et al. 2005; Rose and Prange 2015). Thus, we removed soft mast from further analyses and averaged index values of hard mast-producing species by ecoregion and sample year for evaluation (Table S2). We also obtained monthly mean snowfall data by ecoregion from the National Oceanic and Atmospheric Administration (NOAA 2016) National Centers for Environmental Information to analyze environmental influences on prey item occurrence in bobcat stomachs (Table S3). We evaluated relationships between 2 y of hard mast indices (dependent variable) and mean monthly snowfall (dependent variable) among ecoregions (independent variable) using one-way ANOVAs. We then compared prey frequencies of occurrence with mast indices recorded from the year before harvest using logistic regression to determine relationships between the rate of prey species occurrence and the amount of hard mast produced per ecoregion across the two seasons. Similarly, we evaluated prey frequencies of occurrence by mean monthly snowfall for each bobcat harvest season using logistic regression.

Fox and Fox (1982) used stomach contents from 172 bobcats to determine percent occurrence of prey items consumed over the 1977–1978 harvest season in West Virginia, but they only reported contents from 170 individuals when documenting prey item occurrence across bobcat sex and stage classes. The authors did not include information about potential bait trap materials, so we must assume that their estimates represent prey consumption via hunting or scavenging rather than trap bait. We combined prey items into similar categories as Fox and Fox (1982) for comparison: deer, rabbit, squirrel, rodent, opossum, raccoon, shrew, bird, and other. We compared their published prey item proportions with our proportions using paired t-tests to determine if there have been any substantial shifts in the dietary contents of bobcats over the last 38 y in West Virginia; then we further compared each dietary item across sex and stage classes.

We conducted all analyses using the statistical software R (ver. 4.1.3; R Core Team 2022) with a statistical significance threshold of α = 0.05. We evaluated logistic regression analyses using function glm with family = binomial, and we evaluated confidence intervals using function predict.glm with type = response. We evaluated tests of independent proportions using function prop.test. All functions listed are found in package stats. We evaluated all significant ANOVA results using Tukey's honestly significant difference post hoc analyses to confirm where differences occurred among groups.

We analyzed 300 bobcat stomachs from the 2014–2015 (n = 150) and 2015–2016 (n = 150) hunting and trapping seasons that were collected from 43 of the 55 counties in West Virginia (Figure 1). Most bobcats were trapped (n = 277), whereas 17 were taken by gun, four by bow, and two unknown (Data S1). The sex ratio of the samples was 1:1.1 males to females, with juveniles representing 21%, yearlings 29%, and adults 50% (Table 1). Bobcats in West Virginia exhibited sexual dimorphism and maturational differences in body mass and condition: males had a larger mean mass (F_1,298 = 150.5, P < 0.001) and slightly higher mean KFI_i (F_1,298 = 4.21, P = 0.041) than females; juveniles had a smaller mean mass than yearlings and adults (F_2,297 = 70.0, P < 0.001), whereas both juveniles and adults had higher mean KFI_i values than yearlings (F_2,297 = 3.92, P = 0.021; Table 2). Bobcat body mass and KFI_i were not highly correlated (r = 0.21; Landry 2017). Body mass differed significantly between the sex : stage interaction term, indicating that bobcat sex and stage class are strong categorical predictors of body mass (F_2,294 = 17.82, P < 0.001); meanwhile, KFI_i did not differ significantly by the interaction term (F_2,294 = 2.09, P = 0.126). Thus, we removed body mass from further dietary content analyses.

Deer, squirrels, mice and rats, rabbits, and opossum were the major prey groups by frequency of occurrence in sampled bobcats (Table 3). Birds, moles and shrews, raccoons, skunks, and others occurred less frequently overall (≤10%). Old world mice and rats (family: Muridae), beaver, canids, and porcupine were only found in stomachs from the 2014–2015 harvest season. Moles (Talpidae) and jumping mice (Dipodidae) were the least commonly occurring taxonomic family groups over the 2015–2016 harvest season (Data S1). The WVDNR survey received 68 anonymous responses (51% response rate) from trappers that donated bobcat carcasses to the project over the 2015–2016 harvest season. Fifty-five of the trappers used bait (81%), with 16.5% of that as deer parts. Rabbits, beaver, feathers, mice and rats, squirrels, and other items were also used as bait (Table S4).

Deer (|z| = 2.73, P = 0.006) and opossum (|z| = 2.04, P = 0.042) occurred more often in the diets of male bobcats than female bobcats, whereas rabbits (|z| = 2.34, P = 0.019) occurred more often in the diets of female bobcats than male bobcats. Deer (|z| = 2.80, P = 0.005) also occurred more often in the diets of adults than juveniles, whereas raccoon (|z| = 2.12, P = 0.034) occurred more often in the diets of yearlings than adults. Mice and rats (|z| = 2.31, P = 0.021) occurred more often in the diets of juveniles than yearlings, and both mice and rats (|z| = 3.14, P = 0.002) and squirrels (|z| = 3.51, P < 0.001) occurred more often in the diets of juveniles than adults (Table 3). Bobcats with better body conditions (i.e., higher KFI_i values) had a higher frequency of occurrence of Virginia opossum (|z| = 3.42, P < 0.001) in their diets, whereas those with less optimal body conditions had a higher occurrence of rabbits (|z| = 2.34, P = 0.019; Figure 2).

Most bobcat stomachs contained hair samples (97% occurrence). Trap debris (9%) and vegetation (81%) did not occur at significantly different rates in bobcat stomachs by stage class, sex, body condition, and body size, nor did they differ across harvest seasons or ecoregions (|z| ≤ 1.89, P ≥ 0.058; Data S1). Vegetation occurred 40% more often in the diets of bobcats harvested via trap than via gun or bow (|z| = 4.00, P ≤ 0.001), whereas debris occurred only in trapped individuals (|z| = 0.02, P = 0.986). Birds, skunks, and others did not occur at significantly different rates among bobcat stage classes, sexes, body conditions, or body sizes, nor did they differ across harvest seasons or ecoregions (|z| ≤ 1.50, P ≥ 0.134).

Tests of independent proportions resulted in no significant differences for overall prey frequencies of occurrence in stomachs among bobcat sexes (χ²₁ = 0.89, P = 0.345), stage classes (χ²₂ = 1.51, P = 0.470), or across ecoregions (χ²₅ = 4.16, P = 0.527). Pianka's index showed a 92.3% dietary overlap between bobcat sexes and a 34.7% overlap among bobcat stage classes. Simpson's index of diversity indicated an 87% probability that two randomly selected prey items belong to different species. Finally, evenness exhibited 80% similarity among all core prey groups.

Hard mast indices for West Virginia increased from 2013 to 2014 when analyzed across all ecoregions (F_1,298 = 120.10, P < 0.001), where mean index values for hard mast ranged from 26.92 to 55.33 across ecoregions and years (Table S2). Bobcats had a higher frequency of occurrence of squirrels in their diets over the 2015–2016 harvest season when accounting for increased hard mast production in 2014 (+56.9% predicted slope; |z| = 2.26, P = 0.024; Figure 3). Snowfall ranged from 0 cm in December to 58.5 cm in February of the 2014–2015 harvest season and from 0 cm in November to 49.9 cm in January of the 2015–2016 harvest season (Table S3). Total snowfall varied between harvest seasons (F_1,295 = 11.23, P ≤ 0.001) and among harvest months (F_3,293 = 65.80, P ≤ 0.001), yet there were no significant dietary differences found with varying snowfall.

When comparing the presence of prey groups found in the historical study from the 1977 harvest season (Fox and Fox 1982) with our current study, we found multiple differences between overall prey presence in the diet composition of bobcats and differences of prey presence by sex and stage classes (Table 4). Presence of deer (t = 3.67, P = 0.003) and rodents (t = 2.10, P = 0.036) decreased over time, whereas the presence of opossum (t = 4.09, P < 0.001) and raccoon (t = 2.60, P = 0.009) increased. When observing differences in diet composition between sex and stage class from each study, we found that opossum presence increased in both males (t = 3.37, P < 0.001) and females (t = 2.40, P = 0.017) over time. Similarly, presence of raccoon increased significantly in males (t = 2.31, P = 0.022) over time, whereas the presence of deer decreased by nearly half in females (t = 4.12, P < 0.001). Deer presence also decreased substantially between juveniles (t = 3.90, P < 0.001) and yearlings (t = 2.42, P = 0.017), whereas opossum presence increased for juveniles (t = 2.44, P = 0.016) and yearlings (t = 2.73, P = 0.007). Similarly, raccoon presence increased for yearlings (t = 2.09, P = 0.039) and shrew presence decreased for juveniles (t = 2.04, P = 0.043) between the two studies.

Studies on bobcat dietary consumption rates of specific prey items have been performed in Ohio on nonharvested bobcats (Rose and Prange 2015), in Pennsylvania on hunter-harvested bobcats (McLean et al. 2005), and in Virginia on scat contents and hunter-harvested bobcats (Progulske 1955). We found that deer had the highest prevalence of any prey item consumed, such that it occurred in 32% of bobcat stomachs overall. Our results are similar to the study performed previously in West Virginia (Fox and Fox 1982) where deer occurred in 49% of sampled bobcat stomachs, and to the Pennsylvania study (McLean et al. 2005) where deer occurred in 41% of sampled bobcat stomachs. Similarly, we identified the presence of rabbits in nearly 22% of sampled bobcat stomachs, where Fox and Fox (1982) identified rabbits in 23%, McLean et al. (2005) in 22%, and Progulske (1955) in 39.5% of sampled bobcats. However, the West Virginia studies were the only two in the region to illustrate the importance of small rodents to bobcats, where we found nearly 31% occurrence and Fox and Fox (1982) found 40% occurrence.

Our results support previous findings that sexual dimorphism in bobcats explains the dietary differences observed between sexes, as males have a larger mean body size than females (Fox and Fox 1982; Sikes and Kennedy 1993; McLean et al. 2005; Rose and Prange 2015; Landry 2017). We showed that males consumed larger prey items (i.e., deer) and females consumed smaller prey items (i.e., rabbits). Similarly, young bobcats consumed small rodents, including squirrels, chipmunks, mice, and rats, over 50% more often than did adult bobcats. This is likely due to adults having better developed their skills as hunters over time. However, morphometric differences between stage classes may also explain variations in prey consumption for similar reasons, as we should expect larger individuals to be stronger and better suited at hunting and killing larger prey items (Fritts and Sealander 1978; Litvaitis et al. 1984; Matlack and Evans 1992; Anderson and Lovallo 2003; McLean et al. 2005).

Of interest, we found that juveniles and adults had better body conditions (i.e., higher kidney fat values) than yearlings, whereas weight increased linearly with age. Since yearlings are known to disperse large distances when searching for their own territories (Kamler et al. 2000), one could expect fat loss with little to no loss of weight due to the conversion of fat into muscle with increased exercise. Alternatively, yearlings are near or at reproductive maturity and may have lost their juvenile fat deposits during their growth into maturation (Landry 2017). Although we are only able to speculate about causes of variation in body conditions that we observed among differing stage classes of bobcats, we did identify patterns in the dietary consumption of prey items by individuals with higher vs. lower fat content. Specifically, we found that bobcats with higher fat content, which may allude to better body conditions during winter, consumed more opossum, whereas those with less fat, which may allude to either leaner or poorer body conditions, consumed more rabbits. Since rabbits are faster than opossums, especially regarding their use of speed rather than playing dead for predator evasion, perhaps we can attribute this pattern to thinner or leaner individuals being faster than individuals with heavier fat content; however, because of the nature of this study, we can only speculate as to why this pattern was observed.

Deer occurred in stomachs of bobcats of all stage classes and sexes, likely due to increases in carrion availability over winter (DeVault and Rhodes 2002; Brockmeyer and Clark 2007; Hansen 2007; Platt et al. 2010). Bobcat harvest season overlaps deer harvest seasons in West Virginia. In 2015, deer hunters harvested a total 138,493 deer (Crum 2016), adding to the amount of carrion readily available to bobcats due to the presence of gut piles, carrion, and unrecovered deer. Carrion consumption is a less common yet observed behavior of bobcats throughout their range in North America (Platt et al. 2010; King et al. 2015). Therefore, bobcat consumption of deer in eastern North America may not necessarily be linked to predatory skills or behavior when carrion is readily available. Moreover, large prey items such as deer are likely eaten by multiple bobcats, contributing to the high occurrence in their diet.

In addition to increased carrion availability over winter due to hunter harvest, bobcat stomach contents may be biased by bait contamination from the trapping methods used by trappers. Despite this, bait contamination is rarely discussed in the literature. Although we could not determine which stomach items were due to bait consumption, we were able to estimate the amount of bait used by a proportion of the trappers who submitted carcasses to the project over the 2015–2016 harvest season. We found that 81% of surveyed trappers reported using bait in their sets, with 16.5% of that reported bait having contained deer parts. Of interest, deer occurred 6.3% more often in the diets of individuals harvested via gun or bow than by trap, which indicates that bobcats consume deer regardless of bait use. Although we cannot evaluate an error rate caused by bait-induced bias in our findings, the data from this survey allow us to consider the impact of bait use on our results, which was minimal (Landry 2017). Thus, we assumed that all prey occurrences were the result of typical dietary food habits of the bobcat population in West Virginia over winter.

Although generally solitary aside from the short period that young remain with their mothers, female offspring often exhibit philopatry where their established home ranges overlap or are congruent with their mother's home range (Kapfer 2014). Males have larger home ranges than females, yet they tend to overlap one or more female home ranges throughout their life cycle (Ferguson et al. 2009). Despite these behaviors, which may result in sexes and stage classes being exposed to different prey types and abundances, tests of independent proportions resulted in no significant differences of diet diversity among sex, stage class, and ecoregion of harvest.

Simpson's index of diversity indicated an overall high diversity of prey items consumed by bobcats during winter in West Virginia, while evenness supported generality in bobcat dietary selection. Pianka's index of overlap suggested a nearly complete dietary overlap of male and female bobcats, despite sexual dimorphism, and only a third of dietary overlap between juvenile, yearling, and adult bobcats. A lower overlap of dietary content between stage classes is expected for young vs. adult, as young are not large or skilled enough to take down larger prey items (Anderson and Lovallo 2003; McLean et al. 2005). Overall, bobcats in West Virginia have considerably high dietary diversity, supporting the theory of bobcats as generalist carnivores.

Snowfall over the two harvest seasons did not affect prey consumption for many of the observed prey species. However, we found a positive correlation between increased hard mast production the year before increased squirrel consumption by bobcats. Winter survival and the following year's reproductive success for squirrels is dependent upon the production of hard mast over the previous growing season (Wentworth et al. 1990; Feldhamer 2002; Steffen et al. 2002). Therefore, further research and longer-term trend analyses may enable the prediction of feeding behavior and retention rates for bobcats as prey availability over winter months drives juvenile survival for bobcats (McCord and Cardoza 1982; Litvaitis et al. 1986; McShea and Healy 2002).

We found that the occurrence of deer and rodents in bobcat stomachs decreased significantly from the 1977–1978 harvest season (Fox and Fox 1982) to our current study, especially the amount of deer consumed by female bobcats and by juvenile and yearling bobcats. As mentioned, 138,493 deer were harvested in 2015, yet only 40,518 deer were reported over the 1977 harvest season in West Virginia (Michael et al. 2015; Crum 2016). Similarly, 1,380 bobcats were reported as harvested in 2015 (Crum 2016), whereas only 548 bobcats were reported in 1977 (Michael et al. 2015). Thus, total harvests increased for both deer and bobcats between 1977 and 2015. Anecdotally, a simple paired t-test using the difference in means between each year was not significant (t = −1.0171, P = 0.4946). Thus, our research found that deer occurrence in the diet decreased despite the increase in both deer and bobcat harvest. This leads us to believe that carrion consumption is not the only driver of deer consumption by bobcats.

It would be interesting for a future study to compare these results with observed changes in vegetation on the landscape (i.e., forest density or composition of open areas), deer densities, and bobcat densities over the last 40 y. Although West Virginia's total population numbers have decreased over the last 40 y (U.S. Census Bureau; census.gov), perhaps there has been a change in human encroachment that has played a role in the dietary changes of bobcats over time. Opossum and raccoon are known to be associated with human-populated areas (Klinkowski-Clark et al. 2010), and we found that these prey items have increased in the diets of male, female, juvenile, and yearling bobcats over the last 40 y. Further comparisons of these data may lead to a better understanding of how bobcat densities, diet, and habitat requirements have changed over time.

Bobcat habitat use is influenced by population density and prey availability (Benson et al. 2004). With increasing maturity of West Virginia's forests over the last 40 y since the Fox and Fox (1982) study, there have been shifts in dietary frequencies of occurrence for many of the core prey groups by differing bobcat demographics (i.e., stage class, sex). Current frequencies and diversity of prey species found in bobcat stomachs in West Virginia as compared with 40 y ago can only begin to highlight prey densities and habitat quality for bobcats across the state. Therefore, we recommend further and longer-term research on bobcat habitat suitability by evaluating prey and bobcat densities, habitat quality, and effects of mast production on prey survival and reproduction. Since bobcats are harvested in West Virginia, we recommend management of possible limiting factors (i.e., prey availability) and consideration of appropriate harvest yields that attain management goals for the bobcat population across the state.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Data S1. Capture details, demographic and morphometric data, environmental data, and dietary content data of bobcats Lynx rufus harvested by hunters and trappers over the 2014–2015 and 2015–2016 harvest seasons in West Virginia. Capture details include bobcat individual tag numbers, month and season (1 = 2014–2015, 2 = 2015–2016) of capture, capture method, and county and ecoregion of capture. Demographic and morphometric data include stage class (J = juvenile [< 1 y], Y = yearling [1–2 y], A = adult [> 2 y]), sex, body mass (kg), and inverse kidney fat index (KFI_i) of each bobcat. Environmental data include mean monthly snowfall (cm) per ecoregion and annual mean mast indices for both soft (i.e., fruits) and hard (i.e., nuts) mast per ecoregion. Dietary content data include occurrence values (1 = present, 0 = absent) for each prey group found in the stomach contents of each bobcat: birds, vegetation, debris, deer, moles and shrews, rabbits and hares, mice and rats, squirrels, opossum, raccoon, skunk, mustelids, beaver, canids, porcupine, and felids.

Available: https://doi.org/10.3996/JFWM-22-001.S1 (35 KB CSV)

Table S1. Potential bobcat Lynx rufus prey species by taxonomy and common names in West Virginia from 2014 to 2016.

Available: https://doi.org/10.3996/JFWM-22-001.S2 (21 KB DOCX)

Table S2. Hard mast (i.e., nuts) species' index values averaged by ecoregion from 2013 and 2014 mast surveys in West Virginia (Richmond et al. 2013, 2014).

Available: https://doi.org/10.3996/JFWM-22-001.S2 (21 KB DOCX)

Table S3. Mean monthly snowfall (cm) by ecoregion for the entire state of West Virginia over the 2014–2015 (season 1) and 2015–2016 (season 2) bobcat Lynx rufus harvest seasons (November–February).

Available: https://doi.org/10.3996/JFWM-22-001.S2 (21 KB DOCX)

Table S4. Bait types, use, and proportions of all bait types from a survey of trappers (n = 68) that donated bobcat Lynx rufus carcasses to the project over the 2015–2016 harvest season (November–February) in West Virginia. Fifty-five of the trappers used bait (81%).

Available: https://doi.org/10.3996/JFWM-22-001.S2 (21 KB DOCX)

Reference S1. Boyle KA, Fendley TT. 1987. Habitat suitability models: bobcat. Washington, D.C.: U.S. Fish and Wildlife Service. Biological Report 82.

Available: https://doi.org/10.3996/JFWM-22-001.S3 (1.452 MB PDF) and https://digitalmedia.fws.gov/digital/collection/document/id/1709/

Reference S2. Crum JM. 2016. White-tailed deer. Pages 21–35 in West Virginia Big Game Bulletin 2015. South Charleston: West Virginia Division of Natural Resources. Wildlife Resources Section Bulletin 16-1.

Available: https://doi.org/10.3996/JFWM-22-001.S4 (3.572 MB PDF)

Reference S3. Edwards J, Rota C, Belcher K. 2021. Bobcat home range, resource selection, and survival in eastern West Virginia. Unpublished report. Morgantown: Division of Forestry and Natural Resources, West Virginia University.

Available: https://doi.org/10.3996/JFWM-22-001.S5 (776 KB PDF)

Reference S4. Richmond E, Ryan CW, Tucker RL, Peters ML. 2013. 2013 West Virginia mast survey and hunting outlook. South Charleston: West Virginia Division of Natural Resources. Wildlife Resources Section Bulletin Number 13-3.

Available: https://doi.org/10.3996/JFWM-22-001.S6 (18.055 MB PDF)

Reference S5. Richmond E, Ryan CW, Tucker RL, Peters ML. 2014. 2014 West Virginia mast survey and hunting outlook. South Charleston: West Virginia Division of Natural Resources. Wildlife Resources Section Bulletin Number 14-3.

Available: https://doi.org/10.3996/JFWM-22-001.S7 (6.096 MB PDF)

Reference S6. Widmann RH, Dye CR, Cook GW. 2007. Forests of the mountain state. Newtown Square, Pennsylvania: USDA Forest Service. NRS-17.

Available: https://doi.org/10.3996/JFWM-22-001.S8 (3.413 MB PDF)

“Gone, but never forgotten”: this manuscript is dedicated to the memory of our beloved coauthor, Richard E. Rogers, who passed away suddenly in February 2022; may he rest in eternal peace. We thank the Journal reviewers and Associate Editor for their time and effort spent reviewing the manuscript. We appreciate all of their valuable comments and suggestions, which helped us to improve the quality of the manuscript. This project was funded by the WVDNR through a Federal Aid in Wildlife Restoration grant from the U.S. Fish and Wildlife Service (Federal Aid Grant W-48-R). Funding provided by the Wildlife Restoration Program was made possible through the purchase of hunting licenses and equipment. We thank the WVDNR, West Virginia University's Davis College and School of Natural Resources, and the West Virginia Trappers' Association for project support. We also thank T. Rounsville for help with stomach content identification and K. Goins, K. Matt, S. Wilson, M. Stevens, B. von Blon, P. Pritt, B. Watson, S. Hall, and L. Price for help with necropsies and sample collection. J.T.A. was supported by the National Science Foundation (01A-1458952) and the National Institute of Food and Agriculture (McStennis Project WVA00812) during manuscript preparation.

Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.