We studied sexual segregation in an endangered alpine ungulate, Sierra Nevada bighorn sheep (Ovis canadensis sierrae) in the Sierra Nevada, California, U.S.A., during winter 2005–2006. We tested hypotheses for sexual segregation to better understand that phenomenon and to obtain information critical for the conservation of these rare mammals. Females foraged in larger groups that were closer to escape terrain than did males. Areas used by males had higher biomass of vegetation and were less open than areas used by females. Males foraged more efficiently in larger groups, whereas females foraged more efficiently when close to escape terrain. Females exhibited a higher bite rate than did males. Males traveled farther per day and in more open terrain than did females. Sexes of bighorn sheep also differed in their dietary niches. Those niches differed most where sexes of bighorn sheep overlapped more in spatial distribution, and differed less where spatial separation was more pronounced. These outcomes are most parsimoniously explained by the gastrocentric and predation hypotheses. In addition, sexes of bighorn sheep behaved as if they were separate species by exhibiting avoidance on one niche axis (space) when there was overlap on another axis (diet). Management and conservation plans must consider the disparate requirements of males and females to help assure the viability of these endangered mountain ungulates.
Introduction
Populations of Sierra Nevada bighorn sheep (Ovis canadensis sierrae) at one time occurred in ≥16 locations throughout that rugged mountain range. By the late 1940s, the number of extant populations had been reduced to 5 (Jones, 1950), and additional extirpations occurred in the following decades; by 1976, only two populations remained (Wehausen, 1980). A program to reestablish those alpine ungulates in areas from which they had been extirpated began in 1979 (Bleich et al., 1991), and three additional populations, totaling ∼120 individuals, had been established by 1986 (U.S. Fish and Wildlife Service, 2008). During the late 1980s, the total number of Sierra Nevada bighorn sheep declined substantially, reaching a low of ∼100 individuals by 1995; they were listed as endangered in 2000 (U.S. Fish and Wildlife Service, 2008).
Several factors may have contributed to population declines, including drought and predation (Wehausen, 1996), pneumonia epizootics (Onderka et al., 1988; Coggins and Matthews, 1992), and small group sizes, which may have resulted in decreased foraging efficiency (Berger, 1978; Molvar and Bowyer, 1994; Ruckstuhl and Festa-Bianchet, 2001; Bowyer and Kie, 2004). Nonetheless, little is known about how predation risk or habitat use and availability influence behaviors of these mountain ungulates, which are among the rarest large mammals in the world (U.S. Fish and Wildlife Service, 2008). Failure to account for life-history characteristics in the conservation of bighorn sheep could result in inappropriately designed reintroduction efforts and, thereby, hamper recovery of this endangered taxon (Festa-Bianchet and Apollonio, 2003; Whiting et al., 2008).
Foraging behavior and risk of predation play substantial roles in shaping life-history strategies for many ungulates (Bowyer, 1984, 2004; Bowyer et al., 1998, 1999; Bleich et al., 1997; Rachlow and Bowyer, 1998; Gaillard et al., 2000; Barten et al., 2001). Diet and predation risk have been studied for populations of bighorn sheep inhabiting the Sierra Nevada, California, U.S.A. (Wehausen, 1980, 1992, 1996; Chow, 1984), but those studies did not distinguish between habitats and forages used by adult males and adult females or the responses of the sexes to predators. Mountain lions (Puma concolor) and coyotes (Canis latrans) are relatively abundant in the Sierra Nevada (Pierce et al., 2000a, 2000b). Studies of foraging behavior by bighorn sheep have addressed ecological differences between sexes for desert-dwelling populations (Bleich et al., 1997; Mooring et al., 2003), but those bighorn sheep likely differ in their diets and habitat requirements from endangered bighorn sheep inhabiting alpine zones of the Sierra Nevada.
Several investigators (Main et al., 1996; Ruckstuhl and Neuhaus, 2002) previously summarized hypotheses regarding sexual segregation in ruminants, and Bowyer (2004) identified problems with some of those hypotheses and emphasized the importance of understanding evolutionary underpinnings in obtaining critical tests. The gastrocentric hypothesis predicts that adult males will forage in habitats where food is more abundant but may be of lower nutritional quality than habitats used by adult females during periods of segregation (Barboza and Bowyer, 2000, 2001). Indeed, males require larger amounts of food per day than do females (Demment and Van Soest, 1985; Illius and Gordon, 1987; Gross et al., 1996; Ruckstuhl and Neuhaus, 2002), which may indirectly influence spatial distributions, movements, and proximity to escape terrain (Ruckstuhl and Neuhaus, 2001; Berger, 1978, 1991). Females also undergo changes to their digestive tract to assimilate high-quality forage required to meet physiological needs during late gestation and lactation (Barboza and Bowyer, 2000, 2001; Zimmerman et al., 2006). Even where no differences occur in habitat selection between sexes (Bowyer, 1986), spatial separation of sexes within habitats may be present (Bowyer, 1984). This outcome is likely driven by differing dietary requirements of sexes (Barboza and Bowyer, 2000, 2001).
Risk of predation also may contribute to sexual segregation and can play an important role in differences in habitat selection by the sexes (Berger, 1991; Bleich et al., 1997; Kie and Bowyer, 1999; Hay et al., 2008); females seek areas that provide security for themselves and their offspring while maintaining nutritional reserves necessary to support the high costs of reproduction (Bowyer, 1984; Bleich et al., 1997). Thus, females incur fitness benefits directly through survival of offspring as well as indirectly via future reproductive success. In contrast, males invest more resources in body growth and secondary sexual traits, allowing them to compete effectively for access to estrous females (Ralls, 1977; Loison et al., 1999; Weckerly, 1998; Perez-Barberia et al., 2002; Spaeth et al., 2004).
Although a number of competing hypotheses potentially explain sexual segregation (Main, 2008), many are neither independent nor, in some instances, even testable (Bowyer, 2004). We focus herein mostly on ecological hypotheses of value in understanding the conservation of bighorn sheep. Consequently, we limited our efforts to the gastrocentric and predation hypotheses, because they make clear predictions concerning sexual segregation, and have not been eliminated as an explanation by empirical research (for review, see Bowyer, 2004). Resurrecting old rejected hypotheses is unlikely to advance our knowledge of sexual segregation (Bleich et al., 1997; Bowyer, 2004). Other investigators recently have adopted this same approach successfully for framing viable hypotheses concerning sexual segregation (Long et al., 2009).
Bighorn sheep exhibit extreme sexual dimorphism in body size (Bleich et al., 1997; Weckerly, 1998; Loison et al., 1999). Adult males and females separate spatially for much of the year (Bleich et al., 1997), and sexes of these ungulates are well known for differential habitat use outside the mating season (Bleich et al., 1997; Krausman and Bowyer, 2003; Mooring et al., 2003). Moreover, bighorn sheep are morphologically and behaviorally well adapted to minimize risk of predation through vigilance and for their ability to evade predators in precipitous terrain (Berger, 1978, 1991; Festa-Bianchet, 1988; Bleich et al., 1997; Bleich, 1999). Finally, the extreme sexual size dimorphism and accompanying allometric differences in many ruminants may lead to sexes behaving as if they were separate but coexisting species (Kie and Bowyer, 1999; Bowyer, 2004; Bowyer and Kie, 2004). This potential outcome necessitates a niche-based approach for understanding the ecology of the sexes, wherein overlap by sexes on one niche axis may result in avoidance on another (Kie and Bowyer, 1999; Bowyer, 2004; Bowyer and Kie, 2004).
Our objectives were to provide a quantitative description of foraging behavior of adult male and adult female bighorn sheep, determine the specific habitat characteristics important to each sex, and test predictions of the gastrocentric and predation-risk hypotheses that have been posited to explain sexual segregation in dimorphic ruminants. We then discuss the relevance of those results to the conservation of Sierra Nevada bighorn sheep.
Hypotheses related to the gastrocentric hypothesis include the prediction that females should obtain a higher-quality diet than would males (as indexed by fecal nitrogen). We also postulated that males would feed in areas with a greater biomass of forage that areas used by females. Moreover, we predicted that effects of predation on sexual segregation should result in females occurring in larger groups than males, because of their greater vulnerability. Likewise, females should occur closer to escape terrain, and occupy more rugged areas with greater visibility than those used by males. Variables related to risk of predation also should modify foraging efficiency (percent of active time spent foraging), with group size, distance to escape terrain, ruggedness of terrain, and visibility all playing a greater role in females. Because of their vulnerability to predation, females would be expected to travel less than males to reduce risk; males would be predicted to travel greater distances than females to meet their absolute needs for more forage—predictions that could support both gastrocentric and predation hypotheses. Similarly, bite rate would be expected to be greater for females, which forage more selectively than males. Nonetheless, females also may expend less time foraging (resulting in a higher bite rate) because of increased vigilance for predators.
As a result of morphological and behavioral disparities, sexes of bighorn sheep should differ in their dietary niches. We predicted that differences in dietary niche would be most marked where sexes of bighorn sheep overlapped more in spatial distribution, and would differ less where spatial separation was more pronounced. We further postulated that sexes of bighorn sheep should behave as if they were separate species by exhibiting avoidance on one niche axis (diet) when there was overlap on another axis (space).
Methods
STUDY AREA
The Sierra Nevada (37°24′N, 118°41′W) is a young (<5,000,000 yr BP), rugged mountain range of far western North America, approximately 650 km long and from 75 to 125 km wide (Hill, 1975). Snow is a significant source of winter precipitation in the Sierra Nevada (National Oceanic and Atmospheric Administration, 2006), which affects availability of vegetation and timing of green-up for mountain ungulates (Festa-Bianchet, 1988; Rachlow and Bowyer, 1991, 1994; Bowyer et al., 1999); rain during spring and summer also is important for growth of forage plants used by bighorn sheep (Wehausen, 1992; Oehler et al., 2003).
Most storms form in the Pacific Ocean from November to March and move eastward over the Sierra Nevada, which casts a rain shadow and is responsible for the desert and steppe ecosystems typical of the eastern portion of that range (National Oceanic Atmospheric and Administration, 2006). Annual precipitation varies markedly with a range during autumn through early spring (1 October to 15 April) of 2.97–19.65 cm from 1993 to 2005 (Inyo County Water Department, 2006). Snowfall also varies markedly with an annual range of 0.25–21.4 cm during winter (average minimum to average maximum from 1 December to 31 May, 1948–2006; National Oceanic Atmospheric and Administration, 2006). Temperature is variable, ranging from 5.7 to 27.1°C (average minimum to average maximum during winter; National Oceanic Atmospheric and Administration, 2006).
Bighorn sheep currently are distributed in 4 subpopulations occupying 4 distinct areas (Mono Basin, Wheeler Ridge, Mt. Baxter, and Mt. Langley; Fig. 1) along the eastern slope of the Sierra Nevada (Wehausen, 1996). The Mt. Warren and the Mt. Gibbs subpopulations occupy the Mono Basin in the north. The Southern Region is occupied by 3 subpopulations, which we designated for separate study: Wheeler Ridge; Mt. Langley; and Mt. Baxter–Sawmill Canyon (Wehausen, 1996; U.S. Fish and Wildlife Service, 2008).
Vegetation in the eastern Sierra Nevada on lower-elevation (1500–2000 m) winter ranges (Wheeler Ridge, Mt. Langley, and Mt. Baxter–Sawmill Canyon) is typical of the Great Basin and is characterized by sagebrush steppe and pinyon-juniper forest (Chow, 1984). The overstory is dominated by single leaf pinyon (Pinus monophylla), big sagebrush (Artemisia tridentata), bitterbrush (Purshia tridentata), rabbitbrush (Chrysothamnus nauseosus), mountain whitethorn (Ceanothus cordulatus), and horsebrush (Tetradymia canescens), whereas the understory is characterized by needle-grass (Achnatherum speciosa), buckwheat (Eriogonum spp.), beardtongue (Penstemon spp.), and sedges (Carex spp.; Chow, 1984). At higher elevations (3300–4300 m), vegetation is typified by alpine communities such as prickly phlox (Leptodactylon spp.), crested wheatgrass (Agropyron spp.), sedges (Carex spp.), and vetches (Astragalus spp.). Plant nomenclature follows Munz (1974).
CAPTURE AND RADIOTELEMETRY
Beginning in 2003, the California Department of Fish and Game captured male and female bighorn sheep ≥1-yr-old via net-gunning from a helicopter (Krausman et al., 1985). The department held the necessary permits, and all research was consistent with guidelines published by the American Society of Mammalogists for research on wild mammals (Gannon et al., 2007). Our research also was approved by an independent Animal Care and Use Committee at Idaho State University (protocol no. 616). We used GPS collars to monitor winter movements and survivorship of this exceptionally rare mammal. Data were gathered via GPS collars from 6 adult female and 3 adult male bighorn sheep beginning on 1 January 2003 and ending 1 May 2005. These radio collars were programmed to record positions every 6–10 min for an 8-h period, 1 day per week for 2 years. These 8-h periods varied over time. Hourly locations were used to analyze movements.
We combined telemetry coordinates from GPS data and Digital Elevation Models (DEM) with ArcGIS 9.1 (Environmental Systems Research Institute Inc., Redlands, California, U.S.A.); Spatial Analyst Extension was employed to estimate daily movements of bighorn sheep. We determined mean distance traveled per day for each bighorn sheep with the Home Range Extension for ArcGIS™ (Rodgers and Carr, 1998). The elevation of each GPS location was calculated by overlaying a shapefile for each individual with a 10-m DEM, and the Surface Spot function in ArcMap then was used to derive the elevation of each location from the underlying DEM.
A terrain-ruggedness index (Nicholson et al., 1997) was developed for locations of male and female bighorn sheep. We used the program ArcGIS to construct 3 layers derived from the DEM to calculate slope, aspect, and elevation for each 10 × 10 pixel in the grid. Locations from all bighorn sheep then were overlaid on the terrain layers and a 300-m radius buffer was delineated around each point to obtain an adequate representation of terrain and habitat features available to bighorn sheep. The Grid Statistics tool was used to calculate the mean and SD for physical characteristics (slope, aspect, and elevation) of each circle.
COLLECTION OF FECES AND HABITAT VARIABLES
We collected fecal pellets during winter (1 January to 5 May) 2006 from the 4 subpopulations (Mt. Baxter, n = 28; Mt. Langley, n = 31; Wheeler Ridge, n = 46; Mono Basin, n = 38), and data were pooled by area, sex, and month. We collected fresh (≤3 days old) fecal pellets, confirmed organoleptically, from observed individuals or from live animals that had been captured. Each pellet group (from a single animal) was considered an independent sample.
Diet quality was indexed for adult male and adult female bighorn sheep from levels of nitrogen in feces (Blanchard et al., 2003; Leslie et al., 2008). We analyzed fecal pellets for percent nitrogen on an ash-free, dry-matter basis. Diets were estimated microhistologically (Sparks and Malechek, 1968; Hodgman et al., 1996), and plants were identified to genus. Fecal pellets were analyzed under contract by the Wildlife Habitat Laboratory at Washington State University, Pullman, Washington, U.S.A.
We measured characteristics of habitat for bighorn sheep at random locations and at locations where bighorn sheep were observed. We determined biomass (wet weight) of forage with double sampling (Reese et al., 1980; Reich et al., 1993; Barten et al., 2001) within 1-m2 plots. We sampled wet-weight biomass of forbs, grasses, and shrubs from male (n = 11), female (n = 20), and random (n = 18) locations during winters 2005 and 2006, and pooled years to obtain sample sizes sufficient for statistical analyses.
We calculated the greatest distance bighorn sheep were observed foraging from escape terrain by measuring the distance from the central point of the group to the nearest escape terrain with either a tape measure or from GPS coordinates. Escape terrain was defined as any geomorphic feature >5 m in height and diameter, which was part of a larger, contiguous geological formation (Risenhoover and Bailey, 1985; DeCesare and Pletscher, 2006). Measurements of distance to escape terrain were determined immediately following behavioral observations of bighorn sheep, or within 7 days of the initial observation where it seemed likely our immediate presence would disturb bighorn sheep (Rachlow and Bowyer, 1998). We estimated visibility at locations of male and female groups with a 2-m cover pole to make our data comparable with other research on bighorn sheep. Measurements were collected from a distance of 15 m in 4 cardinal directions, and percent of the pole (in 25-cm increments) obscured by vegetation or geomorphic features was recorded (Griffith and Youtie, 1988; Bleich et al., 1997; Bowyer et al., 1999). The mean percentage of the cover pole visible from 4 directions was used as the value for visibility.
FIELD OBSERVATIONS
We observed the behavior of bighorn sheep with 10 × 40 binoculars or a 20 × 60 spotting scope. A radio-collared bighorn sheep was selected randomly from each of the 4 general study areas (Mt. Langley, Mt. Baxter, Wheeler Ridge, and Mono Basin) and an attempt was made to locate that bighorn sheep via radio telemetry. Once that marked bighorn sheep was observed (≤300 m) and group size recorded, we selected an individual from the group containing that animal with a random numbers table and behavioral observations began. Bighorn sheep were categorized into sex and age classes according to Geist (1968): (1) adult females, (2) young, (3) class I males (yearlings), and (4) class II–IV males (adults).
We recorded activities of bighorn sheep on a hand-held personal computer, with an internal clock, using a combination of scan- and focal-animal sampling (Altmann, 1974). We used scans at 10-min intervals (Altmann, 1974) to categorize bighorn sheep activities, in a manner similar to that of Risenhoover and Bailey (1985): foraging (head down in feeding posture); bedding; aggressive; vigilant (i.e., alert; head in upward position with ears erect and directed forward); or alarmed (running to escape terrain). We used focal sampling (Altmann, 1974) of randomly selected individuals to collect data on bites taken by males or females stratified by forage class. The individual was then viewed continuously for at least 3 min and the number of bites of each forage class (forb, graminoid, shrub, or unknown plant) was recorded. We subsequently calculated foraging efficiency (percent of active time spent foraging; Berger, 1978) for each individual.
STATISTICAL ANALYSES
We used a mixed, generalized-linear model (PROC MIXED, Little et al., 1996; SAS Institute, 1999) to analyze data on percent fecal nitrogen (dependent variable), with sex as the independent variable, study area as the random effect, and Julian date and elevation as covariates. Percentage data were arcsine–square root transformed prior to analysis. We used Tukey's HSD (Zar, 1999) for multiple comparisons between groups when there were significant main effects (P ≤ 0.05).
We used multivariate analysis of variance (MANOVA) to compare differences in forage abundance between sites used by sexes of bighorn sheep (Zar, 1999). Independent variables were locations of males (≥2 yrs), adult females and young (including males <2 yrs old), and random sites. Dependent variables for the MANOVA were total wet weight of both live and dead biomass (g m–2), live and dead biomass of forbs and grasses, and live biomass of shrubs. Wet weight of biomass was used because that variable is more closely related to rumen fill than is dry weight (Belovsky, 1978). We employed Tukey's HSD for multiple comparisons between groups with significant main effects (male, female, and random locations). Berger (1991) and Bleich et al. (1997) reported that female bighorn sheep use more open areas than did males; thus, a 1-tailed t-test (which predicted males in less-open areas) was used to compare habitat openness at foraging locations of males and females.
We used principal components analysis (PCA) of diet composition for bighorn sheep, based on the variance-covariance matrix (McGarigal et al., 2000), to reduce dimensionality of those data. The number of species identified by microhistological analysis was reduced to 8 genera (sensu Kie and Bowyer, 1999) because they: (1) were consumed by >20% of all bighorn sheep; (2) comprised ≥30% of the diet in at least 1 bighorn sheep of each sex; and (3) composed ≥5% of diets averaged over all samples. The 8 genera selected were Agropyron (wheat grass), Artemisia (sagebrush), Astragalus (milk-vetch), Carex (sedge), Festuca (fescue grass), Leptodactylon (prickly phlox), Poa (blue grass), and Achnatherum (desert needlegrass).
We assigned subpopulations of bighorn sheep to 2 separate regions based primarily on elevation: (1) Mono Basin; and (2) Southern Region (consisting of Mt. Langley, Mt. Baxter–Sawmill Canyon, and Wheeler Ridge subpopulations). Bighorn sheep inhabiting the Mono Basin spent winters at high elevations, whereas bighorn sheep from the Southern Region occurred at low elevations during winter. We used MANOVA and Tukey's HSD for multiple comparisons to test for differences in diet composition of forage species among study areas and sexes. Percentage data were arcsine–square root transformed (Zar, 1999). We plotted means with 95% confidence intervals as bivariate ellipses for the first 2 principal components to examine differences between diets of sexes (i.e., the dietary niche). We used the Shannon-Weiner Index to infer differences in diet diversity between sexes and study areas, and H′ was transformed to eH′ to reflect the number of forage species (Ricklefs and Miller, 2001). We estimated diet diversity from genera of plants in the diets of bighorn sheep that met criteria used for PCA.
We examined mean daily distances traveled for individual male and female bighorn sheep with the Satherwaite 2-tailed t-test for unequal variances (Zar, 1999). Differences in elevations used by adult males and females during winter were tested with ANOVA with mean monthly elevation as the dependent variable, and sex, month, and year as main effects, as well as pairwise interactions between sex, month, and year. We used MANOVA to test for differences between sexes in use of slope, aspect, and a composite index based on the standard deviation of slope multiplied by the angular deviation in aspect (Nicholson et al., 1997; Bowyer et al., 1999; Pierce et al., 2004; Oehler et al., 2005). Dependent variables for this analysis were slope, aspect, and terrain ruggedness, and independent variables were sex, subpopulation, and a sex × subpopulation interaction.
Results
FECAL NITROGEN
Mean values of fecal nitrogen ranged from 1.4 to 2.8% for adult males, and from 1.2 to 2.8% for adult females across all subpopulations during winter (Table 1). Fecal nitrogen was not significantly different between sexes when controlled for elevation and Julian date, with sites used by bighorn sheep and subpopulation as random effects (MIXED GLM, F1,3 = 0.28, P = 0.635; Fig. 2). Nevertheless, fecal nitrogen varied for bighorn sheep in different subpopulations and by month after controlling for effects of elevation (ANCOVA overall F8,20 = 7.56, P < 0.001; herd F3,20 = 14.98, P < 0.0001, month F3,20 = 3.86, P = 0.0249; Fig. 2). Tukey's HSD revealed that fecal nitrogen in the Mt. Baxter subpopulation was significantly higher (P < 0.05) than for all other subpopulations, and fecal nitrogen in the Mono Basin subpopulation was lower (P < 0.05) than all subpopulations except Mt. Langley (Fig. 2).
TABLE 1
Fecal nitrogen (ash-free basis) of adult male and female bighorn sheep in the Sierra Nevada, Inyo and Mono Counties, California, U.S.A., and the elevation at which feces were collected during winter 2006.
FORAGE ABUNDANCE
Total biomass of vegetation ranged from 6 to 138 g m–2 at female locations, from 34 to 182 g m–2 at male locations, and from 13 to 127 g m–2 at random locations during winter. Biomass of forage classes (forbs, grasses, shrubs, and total) varied for locations of sexes and random locations (MANOVA; overall F8,68 = 2.12, P = 0.04; Fig. 3), and for subpopulations (MANOVA; F12,90 = 2.12, P = 0.002). Locations of males had higher biomass of shrubs (P < 0.05) and total vegetation (P < 0.05) than sites used by females or random locations (Fig. 3). Locations of females did not differ from random locations in shrub biomass or total biomass (Fig. 3). Moreover, biomass of forbs and grasses was not significantly different among male, female, or random locations (Tukey's HSD; P > 0.05; Fig. 3).
BITE RATE
Females had higher bite rates (bites min–1) for all forage classes during winter than did males (Fig. 4). The overall bite rate (mean ± SE) for female bighorn sheep (31.31 ± 1.26) was higher than for males (24.46 ± 0.96) for all forage classes combined (ANOVA; F1,95 = 7.36, P = 0.007; Fig. 4). Bite rates for bighorn sheep were not significantly different between study areas (ANOVA; F3,95 = 0.85, P = 0.467) or forage classes (ANOVA; F3,95 = 1.86, P = 0.140; Fig. 5) for both sexes combined. Similarly, there was no significant interaction between sex and forage class (ANOVA; F3,95 = 0.22, P = 0.88).
DIETARY COMPOSITION AND NICHE
Bighorn sheep exhibited diverse diets as estimated from microhistological analyses of feces; 57 genera of plants were detected in fecal samples from males and females. Bighorn sheep in the Mono Basin consumed mostly forbs (Astragalus and Leptodactylon) during winter (Table 2). Bighorn sheep ate mostly shrubs (Artemisia) and graminoids (Achnatherum) in the Southern Region (Table 3, Fig. 5). Principal components analysis (PCA) indicated that bighorn sheep in the Mono Basin exhibited differential patterns in diet composition compared with bighorn sheep in the Southern Region (Fig. 5). Principal component 1 explained 43.9% of the variation in diets among bighorn sheep and likely represented a continuum in altitude from lower elevations (positive loadings) to higher elevations (negative loadings). Principal component 2 explained 24.0% of the variation in diets among bighorn sheep and probably represented a continuum from browsing (negative loadings) to grazing (positive loadings). Diets of male and female bighorn sheep overlapped in both study regions, but mean PCA scores and their 95% confidence intervals indicated that patterns of separation in diets existed (Fig. 6). There was an overall effect of sex on mean PCA scores reflecting diet composition (MANOVA; F2,77 = 4.09, P = 0.021). Similarly, mean PCA scores of diet composition differed between bighorn sheep from the Mono Basin and those from the Southern Region (F2,77 = 66.20, P < 0.0001). The sex by region interaction, however, was not significant (F2,78 = 2.01, P = 0.14).
TABLE 2
Descriptive statistics for percent of plant genera in diets (estimated from microhistological analyses of feces) of male and female bighorn sheep in the Mono subpopulation in the Sierra Nevada, Mono County, California, U.S.A., during winter 2006.
TABLE 3
Descriptive statistics for percent of plant genera in diets (estimated from microhistological analyses of feces) of male and female bighorn sheep in the Southern Region of the Sierra Nevada, Inyo County, California, U.S.A., during winter 2006.
The Shannon-Weiner index indicated that overall diet diversity in plant genera was similar between sexes (eH′males = 9.68, eH′females = 9.85). Nonetheless, this same index also revealed that diet diversity was higher for male bighorn sheep (eH′ = 34.42) than for females (eH′ = 26.95) inhabiting the Mono Basin, but was more similar between sexes in the Southern Region (male eH′ = 18.80, female eH′ = 21.97).
DAILY MOVEMENTS
Hourly movements ( ± SE) for males ranged from 27 to 998 m ( = 208 ± 20.6 m). Hourly movements for females ranged from 16 to 301 m ( = 69 ± 8.3 m). Daily movements for both sexes combined ranged from 1.12 to 14.50 km ( = 4.10 ± 3.69 km). Although sample sizes were small, mean daily movements of males (5.27 ± 1.85 km) also were significantly greater (t7 = –2.39, P = 0.048) than for females (2.36 ± 0.096 km).
TERRAIN RUGGEDNESS, VISIBILITY, AND ELEVATION
Differences in the use of rugged terrain occurred between sexes of bighorn sheep (MANOVA, F3,54 = 3.26, P = 0.02; Fig 7). Moreover, a significant difference in this ruggedness index occurred among areas occupied by subpopulations (ANOVA, F3,56 = 10.04, P < 0.0001). Mono Basin had the most rugged terrain (P < 0.05), and Wheeler Ridge had more rugged terrain (P < 0.05) than Mt. Langley and Mt. Baxter. Females used more rugged terrain (2924.1 ± 1628.05; mean ± SE) than did males (2096.4 ± 1104.09; ANOVA F 1,56 = 9.19, P = 0.003). Females also used areas with significantly more variation in slope (70.8 ± 27.11) than did males (66.7 ± 13.79, ANOVA F1,56 = 9.66, P = 0.003). Angular deviation of aspect in areas used by females (124.2 ± 26.02) did not differ from that used by males (117.4 ± 29.59; ANOVA, F 1,56 = 0.15, P = 0.703). With the exception of the Mt. Langley subpopulation, females used more rugged terrain than did males (Fig. 7). Visibility ranged from 62% to 91% at female locations, and 28% to 93% at male locations. The t-test indicated that mean (±SE) visibility at locations used by females (76 ± 9%) was significantly greater (P = 0.05) than locations used by male bighorn sheep (56 ± 25%).
We used locations of 29 male and 35 female groups of bighorn sheep during winter to compare differences in elevations used by sexes. Overall (ANOVA, F1,56 = 4.72, P = 0.034), females occurred at higher mean (±SE) elevations (2119 ± 60.7 m) than did males (2049 ± 73.6 m). Bighorn sheep inhabiting the Mono Basin used the highest (2826 ± 51.5 m), whereas bighorn sheep in the Mt. Baxter area used the lowest elevations (1662 ± 19.6 m). Indeed, a significant difference in elevations used by bighorn sheep occurred among subpopulations (ANOVA, F3,56 = 47.06, P < 0.0001). Tukey's HSD indicated that bighorn sheep in the Mono Basin used significantly (P < 0.05) higher elevations than all other areas, while bighorn sheep inhabiting Mt. Baxter used lower elevations than all other subpopulations (P < 0.05). Bighorn sheep on Wheeler Ridge (2110 ± 37.4 m) and Mt. Langley (2031 ± 51.5 m) did not differ significantly in use of elevation (P > 0.05).
GROUP SIZE, DISTANCE TO ESCAPE TERRAIN, AND FORAGING EFFICIENCY
Female bighorn sheep occurred in larger groups (6.40 ± 0.99; mean ± SE) than did males (3.78 ± 1.01; ANOVA, F1,27 = 5.04, P = 0.03; Fig. 8); 72% of female groups contained young. Females foraged closer to escape terrain (30.7 ± 5.10 m) than males (118.1 ± 18.92 m; ANOVA, F2,26 = 9.24, P < 0.001; Fig. 8). Foraging efficiency was positively related to group size for males, but that relationship was not significant for females (Fig. 9). Conversely, foraging efficiency was negatively related to the distance from escape terrain for females, but that relationship was not significant for males (Fig. 10). When controlled for group size and distance to escape terrain, foraging efficiency did not differ between sexes (ANCOVA overall F3,25 = 1.72, P = 0.18; sex F1,25 = 0.78, P = 0.39), indicating the importance of those covariates in affecting vigilance behavior. Similarly, when group size and distance to escape terrain were included as covariates, individual vigilance (percent of time alert) did not differ between sexes (ANCOVA, F3,25 = 1.76, P = 0.18; sex F1,25 = 1.03, P = 0.319).
Discussion
PREDICTIONS FROM HYPOTHESES
The relevance of forage acquisition and risk of predation to the evolution of life-history strategies of male and female ungulates has received increasing attention (Bowyer, 2004). In addition, differing strategies by the sexes are at the forefront of conservation issues relative to the maintenance of viable populations of large herbivores across increasingly fragmented landscapes (e.g., Bleich et al., 1997; Bowyer, 2004; Rubin and Bleich, 2005). Indeed, management plans implemented to enhance habitat for ungulates may inadvertently benefit one sex at the expense of the other (Bowyer et al., 2001; Stewart et al., 2003). Our results, derived specifically from research on one of the most endangered ungulates worldwide, provide additional insights into the evolutionary significance of sexual segregation, while simultaneously providing directions for implementing conservation strategies.
Predictions from predation and gastrocentric hypotheses (Bowyer, 2004) were generally supported by results from this study of bighorn sheep (Table 4). These hypotheses need not be mutually exclusive but the predation and gastrocentric hypotheses are necessary to explain observed patterns of niche partitioning between sexes, especially where tradeoffs occur (Bowyer, 2004). Competing hypotheses exist (Bowyer, 2004; Main, 2008), but predation and the gastrocentric hypothesis are most fruitful of viable hypotheses for framing research questions because they make testable predictions (Bowyer, 2004; Long et al., 2009).
TABLE 4
Predictions from the predation and gastrocentric hypotheses as related to variables sampled for bighorn sheep in Sierra Nevada, Inyo and Mono counties, California, U.S.A., during winter. na = not applicable.
GASTROCENTRIC HYPOTHESIS
Greater biomass of forage occurred at locations of males than those of females. Dietary separation between sexes, and male bighorn sheep traveling farther than females, also supported the gastrocentric hypothesis (Table 4). Predicted differences in fecal nitrogen, however, did not (Barboza and Bowyer, 2000, 2001). In addition, differences between sexes in daily distance traveled and bite rate potentially support both predation and gastrocentric hypotheses.
Although sample sizes were small, male bighorn sheep moved farther per day than did females. These larger daily movements may be explained by the need for males to acquire a greater amount of food per day than needed by females (Gross et al., 1996; Barboza and Bowyer, 2000, 2001). Male bighorn sheep, owing to their larger rumens, can use lower-quality forage with higher fiber content than can females (Barboza and Bowyer, 2000, 2001; Zimmerman et al., 2006). Further, allometric differences in digestive tract morphology between males and females predict male ruminants require longer periods of non-activity for rumination (Demment and Van Soest, 1985; Jenks et al., 1994; Ruckstuhl, 1998; Barboza and Bowyer, 2000, 2001). Adult female bighorn sheep are approximately 60% the weight of adult males (Bleich et al., 1997; Weckerly, 1998; LeBlanc et al., 2001), and spend more time foraging and less time ruminating throughout the day (Ruckstuhl, 1998). The combination of these nutritional requirements (including males possessing a larger rumen than females) and sparsely vegetated winter ranges in alpine zones of the Sierra Nevada offer a reasonable explanation for why male bighorn sheep moved greater distances per day than did females.
As noted previously, diets of male and female bighorn sheep did not differ in diet quality during winter (as indexed by fecal nitrogen). No difference in diet quality between sexes may not be a reasonable conclusion because of previous contradictory results (Barboza and Bowyer, 2000, 2001; Bowyer, 2004, for reviews) and the large differences in the dietary niches of male and female bighorn sheep (Fig. 6). Results for fecal nitrogen may be confounded by the role that sexual dimorphism and resultant differences in allometry play in digesting forages of varying fibrosity and quality (Jenks et al., 1994; Barboza and Bowyer, 2000, 2001). Females, because of changes in gut morphology associated with late gestation, may be better at extracting nitrogen from forage than are males (Jenks et al., 1994; Zimmerman et al., 2006; J. A. Jenks, personal communication). Differences in fecal nitrogen, however, occurred between subpopulations of bighorn sheep (Table 1, Fig 2), indicating that fecal nitrogen was reflecting overall changes in diet quality. For instance, diet quality in the Mono Basin was significantly lower than for all other subpopulations, likely a result of bighorn sheep in the Mono Basin occupying high elevations with low-quality forages during winter. Although levels of fecal nitrogen (1.2 to 2.4%) were lower for bighorn sheep occupying the Mono Basin than for other subpopulations, those levels are within the range or higher than other published values reported for bighorn sheep inhabiting the Rocky Mountains, U.S.A., during winter (0.8–3.2%, Irwin et al., 1993; 1.2–1.5%, Blanchard et al., 2003). Further, most fecal nitrogen levels were >1.3%, reported by Irwin et al. (1993) as necessary for a maintenance diet during winter, indicating bighorn sheep in the Mono Basin were not nutritionally deficient during winter 2005.
Principal Component Analysis indicated that diets of bighorn sheep from the two regions (Mono Basin and Southern Region) diverged markedly in forage composition (Fig. 5), probably a result of bighorn sheep foraging in differing plant communities associated with varying elevations. Bighorn sheep in the Mono Basin foraged mostly on plants characteristic of the alpine fell-fields community (Munz 1974), which included Leptodactylon, Astragalus, and Carex. Bighorn sheep in the Southern Region, however, foraged mostly on plants in the sagebrush-scrub community and included more Artemisia, Achnatherum, Festuca, and Agropyron in their diets during winter. Further, diets of male and female bighorn sheep in the Southern Region exhibited significant separation (Fig. 6); males had a broader dietary niche (PC2) than those of females and tended to eat more shrubs such as Artemisia, whereas females tended to eat more graminoids including Achnatherum, Poa, and Festuca. In the Mono Basin, diets of males were more characteristic of lower-elevation communities (PC1) in the Southern Region and contained more Artemisia, Achnatherum, Poa, and Festuca, whereas diets of females in the Mono Basin were associated with higher elevations and contained mostly Leptodactylon, Astragalus, and Carex.
Differences in diet composition between sexes of bighorn sheep are consistent with a niche-based approach to understanding sexual segregation (Kie and Bowyer, 1999; Bowyer, 2004; Bowyer and Kie, 2004), where overlap on one niche axis is accompanied by avoidance on another. Spatial separation of sexes in both the Mono Basin and the Southern Region were documented. There was, however, more spatial (differential use of rugged terrain by sexes) and less dietary separation between sexes in the Mono Basin than in the Southern Region. Conversely, the sexes of bighorn sheep in the Southern Region exhibited less separation in use of rugged terrain, with considerably less overlap on the dietary niche axis than for bighorn sheep occupying the Mono Basin. These results are consistent with previous studies where dietary and spatial niches between sexes resulted in overlap on one niche axis being accompanied by avoidance on another (Kie and Bowyer, 1999; Stewart et al., 2003). These outcomes also provide additional support for the concept that the sexes of dimorphic ungulates should be managed as if they were coexisting species (Kie and Bowyer, 1999; Bowyer et al., 2001; Stewart et al., 2003; Bowyer, 2004; Rubin and Bleich, 2005).
PREDATION HYPOTHESIS
Observations concerning the behavior of sexes of bighorn sheep that supported the predation hypothesis included differences in the distance sexes occurred from escape terrain, variation in group size by sex, differential visibility and ruggedness of terrain for habitats occupied by males and females, and sexual differences in foraging efficiency (Table 4). Females occurring in areas with less forage but forming larger groups than males indicated that predation risk rather than resources were responsible for this pattern of group formation in females. Others have reported that males foraged in smaller groups than females (Mooring et al., 2003). In addition, males foraged farther from escape terrain than did females (Fig. 8), which also was reported by Berger (1991) for bighorn sheep inhabiting a desert environment.
Females occupied more rugged terrain than did males (Fig. 7). Bleich et al. (1997) similarly reported that females used more rugged terrain, and that males occupied gentler slopes (and even flat terrain) when foraging. Terrain ruggedness was significantly different across geographic locations but, with the exception of the Mt. Langley subpopulation, the pattern of females using more rugged terrain than males was consistent. We hypothesize that females used more rugged terrain to reduce the risk of predation and for protection of their vulnerable offspring, which has been reported by others (Berger, 1991; Bleich et al., 1997; Bleich, 1999; Mooring et al., 2004).
Males foraged in areas with significantly lower visibility (56%) than areas where females foraged (74%). Visibility and distance to escape terrain influence how bighorn sheep allocate foraging time and time spent vigilant for detecting predators. Females typically select more open habitats than do males (Berger, 1991; Bleich et al., 1997). Further, because there was more variation in visibility in habitats used by males, we hypothesize that those individuals were willing to occupy sites with a greater risk of predation, but with abundant forage, to obtain the food necessary for maintenance of body reserves during winter. Indeed, a larger proportion of male bighorn sheep than females have been killed by mountain lions in the Sierra Nevada (California Department of Fish and Game, unpublished data). In addition, Bleich et al. (1997) reported that males occupied areas with more predators than did females when the sexes were spatially segregated.
Foraging efficiency was positively correlated with group size for males, but not for females (Fig. 9). Foraging efficiency was negatively influenced by distance from escape terrain for females, but not for males (Fig. 10). Females, regardless of group size, remained much closer to escape terrain than did males. These observations are unique to this study and emphasize the need to fully understand habitat requirements of the sexes. Others also have reported that foraging efficiency was positively related to group size (Berger, 1978; Rachlow and Bowyer, 1998) and negatively related to distance from escape terrain (Risenhoover and Bailey, 1985), but those authors did not differentiate between sexes. Indeed, Bleich et al. (1997) noted that proximity to escape terrain was more important in defining habitat selection by female than male bighorn sheep.
Female mountain sheep use steep, rocky terrain with good visibility ostensibly to detect and evade predators (Festa-Bianchet, 1988; Berger, 1991; Bleich et al., 1997; Rachlow and Bowyer, 1998; Bleich, 1999; Mooring et al., 2004). Consequently, females remain closer to escape terrain than do males. Males venture farther from escape terrain, and such movements likely increase the rate at which they encounter new foraging areas (Mysterud et al., 2001), enabling them to acquire resources important for weight gain and horn growth, both of which are essential for future reproductive success (LeBlanc et al., 2001). Differences in daily movements between sexes of bighorn sheep also might result from intersexual differences in activity budgets (Ruckstuhl, 1998), because males have a larger gut capacity (absolute rumen size) and require longer periods of rumination between foraging bouts than do females (Barboza and Bowyer, 2000, 2001).
OTHER TRADEOFFS
Female bighorn sheep had higher bite rates than males for all classes of forage (Fig. 4). We hypothesize that the higher bite rates detected in females likely occurred because of smaller bite sizes taken by females than by males, which may result in reduced handling time (i.e., mastication and rumination; Ruckstuhl et al., 2003). Females also may have exhibited higher bite rates because of the tradeoff between foraging efficiency and predation risk. Increasing bite rates combined with a more rapid rate of digestion by females (Barboza and Bowyer, 2000, 2001) would allow them to spend less time with the head in a foraging position, and thereby increase the chance of detecting a predator (Berger, 1978, 1991; Ruckstuhl et al., 2003). Females, however, were not significantly more vigilant than males, perhaps because they occurred in areas with greater visibility than did males.
Female bighorn sheep also may increase bite rates in late spring when energy requirements are higher because of costs of late gestation (Gross et al., 1996). Indeed, female ungulates of sexually dimorphic species typically take smaller bites and forage more selectively than do males (Miquelle et al., 1992; Spaeth et al., 2004). Additionally, males may be less selective for particular parts of plants because they can digest low-quality forage more easily than can females (Barboza and Bowyer, 2000, 2001); hence, males take larger bites with a concomitant increase in the amount of the time needed to masticate and break down plant tissue (or spend more time ruminating).
IMPLICATIONS FOR CONSERVATION
Several important differences in the foraging behavior and habitat use of male and female bighorn sheep inhabiting alpine areas of the Sierra Nevada were documented. Those differences provide conservationists and land managers with detailed knowledge of sex-specific habitat requirements for these endangered herbivores. Further, behavioral differences exist in mountain ungulates that could be used in population-level decisions for management and monitoring. For instance, male bighorn sheep used ranges with more shrub and overall biomass of vegetation than ranges used by females in winter. Additionally, females used more open habitats, whereas males used areas with lower visibility.
Group size and escape terrain should be considered when making some management decisions because of their effects on foraging behavior of bighorn sheep (Risenhoover and Bailey, 1985). Our results indicated that group size has a greater effect on foraging efficiency of males, but that proximity to escape terrain may be a more important factor increasing foraging efficiency of females. These differences could be meaningful when determining a minimum number of animals to be translocated or when deciding if potential translocation sites provide adequate habitat (e.g., escape terrain for females, abundant forage for males). In addition, locating and developing water sources for bighorn sheep in more arid landscapes than the Sierra Nevada might inadvertently favor one sex over the other (Whiting et al., 2009, 2010). Only 41 of 100 populations of bighorn sheep that were translocated in the western United States from 1923 to 1997 were successful (Singer et al., 2000). Considering behavioral and ecological differences between sexes may improve this record.
Habitat loss, protection, and management are key issues for the conservation of many wildlife species (Rubin and Bleich, 2005; Wearmouth and Sims, 2008), especially for bighorn sheep inhabiting alpine environments. Nonetheless, land managers traditionally have considered habitat needs of males and females to be identical (Bowyer et al., 2001; Stewart et al., 2003; Bowyer, 2004). The current recovery plan for the bighorn sheep in the Sierra Nevada does not make a distinction between the habitat requirements of the sexes (U.S. Fish and Wildlife Service, 2008). Our research indicated that critical habitat requirements are specific to each sex. For instance, protection of alpine meadow systems containing an abundance of forbs and grasses that are located immediately adjacent to escape terrain could be beneficial for female bighorn sheep to enhance foraging efficiency and reduce risk of predation. Habitat treatments that are designed with sex-specific considerations, such as removal of overstory vegetation (i.e., pinyon juniper and other conifers) adjacent to escape terrain, may be useful for enhancing habitat on winter ranges used by female bighorn sheep. In the Sierra Nevada, however, the use of fire as a range-management tool could potentially benefit females by increasing regeneration of forbs and grasses while simultaneously increasing visibility, but to the short-term detriment of males. Fires prescribed in areas occupied by male bighorn sheep initially may reduce the overall biomass of vegetation and reduce shrub cover. Even short-term detriments could be important in the conservation these rare alpine ungulates. Nonetheless, management that favors females may be prudent because of the polygynous mating system and the need to enhance population growth for recovery of bighorn sheep in the Sierra Nevada.
Recovery could be enhanced by implementation of management strategies that incorporate differential niche requirements of sexes of bighorn sheep. For example, proposals for translocation should consider that females prefer areas with a diversity of terrain features that provide security from predation. Further, spatial use of habitats may vary greatly between sexes. Males in our study exhibited much larger daily rates of movement than did females. These longer distances traveled should be considered when making land-management decisions such as determining risks associated with grazing allotments for domestic sheep, which can serve as a source of diseases for bighorn sheep (DeCesare and Pletscher, 2006).
Male and female ruminants use space differently during segregation, and this behavior has implications for wildlife biologists seeking to improve methods for population estimation (Bleich et al., 1997; Bowyer, 2004; Rubin and Bleich, 2005). During winter, bighorn sheep are sexually segregated, which may affect estimates of male to female ratios if one sex or the other is undercounted during surveys because of their disjunct distributions (Bleich et al., 1997). We demonstrated that females generally occupy more rugged terrain, which could affect visibility and cause a bias associated with population estimates. Estimating the number of males in a population also could be problematic if bighorn sheep are more likely to move greater distances on a daily basis, and thereby exacerbate the potential for double counting or under-estimation if surveys are not properly designed. Bighorn sheep in the Sierra Nevada are endangered and may require extreme measures to ensure their survival. Management and conservation plans must consider the disparate requirements of the sexes of this alpine-dwelling ungulate if they are to be successful.
Acknowledgments
This project was funded by California Department of Fish and Game (CDFG), the California Chapter of the Foundation for North American Wild Sheep, and the Department of Biological Sciences at Idaho State University. The CDFG provided field equipment, logistical support, and access to maps and some data used in analyses. D. German, H. Johnson, D. Jensen, J. Wehausen, L. Bowermaster, J. Villepique, and B. Pierce provided support and helped develop ideas for this research. K. M. Stewart, P. Murphy, J. G. Kie, D. J. Delehanty, K. L. Monteith, J. C. Whiting, and P. Atwood assisted with data analysis and reviewed this manuscript. This is a contribution from the CDFG Mountain Sheep Conservation Program, and is Professional Paper 066 from the Eastern Sierra Center for Applied Population Ecology.