Lethal and sublethal effects of black-backed jackals on cape ... - BioOne

Journal of Mammalogy, 94(2):295–306, 2013

Lethal and sublethal effects of black-backed jackals on cape foxes and bat-eared foxes JAN F. KAMLER*, UTE STENKEWITZ,

AND

DAVID W. MACDONALD

Wildlife Conservation Research Unit, University of Oxford, Department of Zoology, The Recanati-Kaplan Centre, Tubney House, Abingdon Road, Tubney, Abingdon OX13 5QL, United Kingdom * Correspondent: [email protected] Little is known about the sublethal effects of mesocarnivores on small carnivores, which can have important implications regarding the ecology and behavior of the latter. We investigated the ecology of cape foxes (Vulpes chama) and bat-eared foxes (Otocyon megalotis) in the absence of black-backed jackals (Canis mesomelas), a dominant mesocarnivore and predator of both fox species. Results were compared with a concurrent study that investigated the ecology of both fox species in the presence of jackals, at a site ,5 km away. In the absence of jackals, densities of cape foxes increased 64% despite similar food and habitat resources between sites, indicating that jackals suppressed cape fox populations. In contrast, jackals did not suppress populations of bateared foxes. For both fox species, the absence of jackals resulted in smaller home-range sizes and nonselective use of habitats for den sites, indicating that jackals had sublethal effects on the ecology and behavior of both fox species. Additionally, in the absence of jackals, cape foxes were marginally more active during daytime, whereas bat-eared foxes exhibited smaller group sizes. The 2 fox species became more segregated in the absence of jackals, indicating that jackals also had sublethal effects on the interspecific relationships of fox species. Our results showed that the effects of a mesocarnivore can extend well beyond population suppression of small carnivores, although sublethal effects varied in intensity and often were species specific. Sublethal effects on small carnivores can occur even if population suppression by a mesocarnivore is not occurring. Key words: anti-predator behavior, bottom-up factors, Canis mesomelas, carnivore interactions, Otocyon megalotis, predator avoidance, South Africa, top-down factors, Vulpes chama Ó 2013 American Society of Mammalogists

DOI: 10.1644/12-MAMM-A-122.1

Brown et al. 1988; Creel et al. 2005; Owen-Smith and Mills 2006; Schmitz et al. 2004; Sheriff et al. 2009). Similar to traditional predator–prey relationships, the effects of mesocarnivores on smaller carnivores may extend well beyond numerical suppression, and could include sublethal effects on important aspects of their ecology and behavior. Several sublethal effects associated with interspecific killing are assumed to occur among mammalian carnivores, such as spatial and temporal segregation and group size, although evidence is mostly circumstantial (Palomares and Caro 1999; Ritchie and Johnson 2009). Although interspsecific killing and population suppression appear to be relatively common among carnivores in ecosystems worldwide (Palomares and Caro 1999), the behavioral responses of subordinate carnivores are rarely directly addressed (Thompson and Gese 2007). Among

The important role of mesocarnivores in ecosystems has recently received increased attention (Prugh et al. 2009; Ritchie and Johnson 2009; Roemer et al. 2009). Mesocarnivores can decrease prey populations (Prugh et al. 2009) and otherwise can function in ecological roles similar to apex predators (Roemer et al. 2009). Mesocarnivores also can influence community structure by suppressing populations of smaller carnivores (Ritchie and Johnson 2009), primarily through interspecific killing or spatial displacement (Kamler et al. 2003b, 2003c). However, little is known about the sublethal, or nonlethal (Lima 1998), effects that mesocarnivores have on smaller carnivores. In contrast, several studies have shown that sublethal effects of predators on prey can be as great as, or greater than, direct effects (Brown et al. 1999; Creel and Christiansen 2008; Cresswell et al. 2010; Lima 1998). For example, predators can have significant sublethal effects on the behavior, physiology, habitat use, activity, reproductive success, and community structure of prey (Berger 2007;

www.mammalogy.org 295

296

Vol. 94, No. 2

JOURNAL OF MAMMALOGY

carnivores, only a few subordinate species changed their movements or habitat use in the presence of dominant species, such as African wild dogs (Lycaon pictus) and cheetahs (Acinonyx jubatus) actively avoiding areas with lions (Panthera leo) and spotted hyenas (Crocuta crocuta—Creel and Creel 2002; Durant 1998), and coyotes (Canis latrans) changing their habitat use in the presence of wolves (C. lupus—Miller et al. 2012). Although swift foxes (Vulpes velox) changed their mating system and increased group sizes in response to lower coyote numbers, it was not clear if swift foxes were simply responding to increases in their own density (due to relaxation of predation) or actively changing their behavior due to fewer encounters with coyotes (Kamler et al. 2004). More research on carnivore communities is needed to determine if mesocarnivores have significant sublethal effects on smaller carnivores, as such information is important for a more comprehensive understanding of carnivore interactions and community dynamics. Throughout most of southern Africa, 3 species of canids are widespread and sympatric, the cape fox (V. chama; 2–4 kg), bat-eared fox (Otocyon megalotis; 3–5 kg), and black-backed jackal (C. mesomelas, 6–12 kg—Skinner and Chimimba 2005). Classification of mesocarnivores is dependent on context (Prugh et al. 2009; Ritchie and Johnson 2009), and for this study we classified the black-backed jackal as a mesocarnivore on the basis of its body size and ecological niche, whereas we classified both fox species as small carnivores. Kamler et al. (2012b) studied the ecology and resource partitioning of these 3 canid species in central South Africa. Results showed that jackals killed both fox species, probably for competition reasons because jackals often did not consume fox carcasses (Kamler et al. 2012b). Cape foxes appeared to coexist with jackals by partitioning space, habitat, time, and diets to various degrees. For example, cape foxes spatially avoided jackal core areas when foraging and establishing den sites, and they selected habitats for den sites that were used least by jackals. Cape foxes also were more nocturnal and less diurnal than jackals, and they consumed primarily small rodents, whereas jackals consumed primarily large ungulates (Kamler et al. 2012b). Bat-eared foxes appeared to coexist with jackals by spatially avoiding jackal core areas when establishing den sites, consuming primarily insects and fruits, and by increasing group sizes to deter predation (Kamler et al. 2012b). Additionally, both fox species had positive associations with each other while foraging and establishing den sites, likely because both were using the same areas outside of jackal core areas as refuges (Kamler et al. 2012b). However, it was not known if jackals suppressed fox populations or had sublethal effects on the ecology and behavior of the foxes. The purpose of this paper was to determine the density, ecology, and interspecific space use of cape foxes and bateared foxes in the absence of jackals. Our study period was concurrent with that of Kamler et al. (2012b), and the 2 sites were ,5 km apart in similar habitat. We hypothesized that in the absence of jackals: densities would increase for both fox species due to lack of predation and competition from jackals;

home range sizes would decrease for both fox species because there were no jackals for foxes to avoid; habitat selection of den sites would change for both fox species because it was no longer necessary for foxes to avoid or detect jackals; group sizes of bat-eared foxes would decrease due to lack of predation from jackals; cape foxes would become more diurnal and less nocturnal due to lack of encounters with jackals; and positive associations between fox species would decrease because they would no longer have to share the same refuge areas from jackals. Although this was an observational study, the comparison of adjacent populations of foxes exposed to the presence or absence of jackals offered a good approximation to experimental testing of a carnivore community. This is because experimental research on carnivores is severely hampered by logistical and ethical difficulties associated with large-scale manipulation of carnivore numbers, necessitating the use of observational studies when ‘‘natural’’ experiments exist (Berger et al. 2008; Thompson and Gese 2007). Differences in bottom-up factors, such as the abundance and dispersion of food resources and habitats, can explain intraspecific differences in the density and ecology of carnivore species (Macdonald 1983). Additionally, grazing pressure from ungulates and associated effects on vegetation influenced numbers of cape foxes, bat-eared foxes, and their prey in southern Africa (Blaum et al. 2009). Therefore, we measured basal (i.e., main) prey of each fox species and grazing pressure on both sites to test alternative hypotheses that food resources or habitats could explain any observed differences in the density and ecology of foxes between sites.

MATERIALS

AND

METHODS

Study area.—Several private ranches (hereafter, PR) occurred within our study site (28859 0 S, 24848 0 E), near Kimberley, South Africa (Fig. 1). The PR (81 km 2) represented the site where jackals were absent for the purposes of this paper (see below). The site was managed primarily for livestock production, mainly domestic sheep and small numbers of domestic goat and cattle. The PR also was managed for small-scale safari hunting of wild ungulates that included steenbok (Raphicerus campestris), common duiker (Sylvicapra grimmia), springbok (Antidorcas marsupialis), blesbok (Damaliscus dorcas), impala (Aepyceros melampus), and oryx (Oryx gazella), as well as common ostrich (Struthio camelus). Due to the threat of predation on livestock by jackals, the PR was surrounded by ‘‘predator-proof’’ fencing, where mesh size prevented jackal-sized carnivores from passing through, although smaller carnivores such as foxes could move freely. Thus, jackals were not normally present on PR, although transients occasionally entered through temporary holes under the fences. When jackals were detected on PR, landowners made considerable efforts to hunt them until the jackals left or were killed. Foxes and other small carnivores were not intentionally persecuted on PR. Benfontein Game Farm (BGF, 110 km2), the comparison site that occurred 4.7 km north of PR (Fig. 1), had relatively

April 2013

KAMLER ET AL.—EFFECTS OF JACKALS ON FOXES

FIG. 1.—Locations of the private ranches (PR) and the comparison site, Benfontein Game Farm (BGF), in central South Africa.

high jackal numbers because this species was not persecuted and the site was not surrounded entirely by predator-proof fencing. Small carnivore diversity, vegetation, and climate were similar between sites (see Kamler et al. 2012b for more details). Ungulate diversity was similar between sites, except that black wildebeest (Connochaetes gnou) occurred only on BGF, whereas impala, oryx, domestic goats, and domestic sheep occurred only on PR. Otherwise, the main difference between PR and BGF was that the former was managed primarily for sheep production with some hunting of wild ungulates and cattle production, whereas BGF was managed primarily for hunting of wild ungulates with some cattle production. All large (.20 kg) carnivore species, including lions, leopards (P. pardus), cheetahs, spotted hyenas, brown hyenas (Hyaena brunnea), and African wild dogs, were extirpated from this region before 1900 (Skinner and Chimimba 2005), leaving black-backed jackals and caracals (Caracal caracal) as the largest potential carnivores on both sites. Capture and radiotelemetry.—We captured cape foxes and bat-eared foxes using wire box traps (50 3 50 3 120 cm) baited with meat scraps, which were placed along dirt roads and intersections throughout the study site, with .0.5 km separating each trap, similar to that reported for BGF (Kamler et al. 2012b). We fitted foxes with radiocollars weighing 1–2% of their body mass. All study animals were sexed, weighed to the nearest 0.1 kg, and classified as adult (12 months old) or yearling (9–11 months) on the basis of tooth wear, body size, and reproductive condition, then released immediately at the capture site. Our research and handling protocol followed the animal care and use guidelines of the American Society of Mammalogists (Sikes et al. 2011).

297

We radiotracked study animals 2–3 times per week from a vehicle using a null-peak system consisting of dual 4-element yagi antennas. Radiotracking also occurred on foot using 3element hand-held antennas to locate den sites. When locating study animals, observers took 2 readings from known telemetry stations ,5 min apart, similar to that reported for BGF (Kamler et al. 2012b). Active locations were assumed to be independent for individual foxes because sequential locations were obtained .12 h apart. Density estimation.—We estimated the prewhelping population density of both fox species on the basis of the number of family groups collared, multiplied by the mean number of adults per group, divided by the total area occupied by collared foxes. Cape foxes were monogamous and territorial, and home ranges of the fox pairs were adjacent to each other; thus we were confident that at least 1 adult from all mated pairs was collared within our trapping area. Bat-eared foxes were numerous and easily observed; thus we attempted only to collar 1 member of each group, and we focused on monitoring adjacent groups at the center of PR. We determined the mean number of adult bat-eared foxes per group by using radiotelemetry homing to walk in on collared foxes on a monthly or bimonthly basis and counting other adults within the group. Uncollared groups within the trapping area were noted and included in the density estimate. Mean adult group size of bat-eared foxes was compared between BGF and PR using t-tests. Density of black-backed jackals on BGF was estimated from movements of radiocollared jackals and group sizes (Kamler et al. 2012b). Black-backed jackals did not consistently occur on PR; thus radiotracking could not be used to estimate their density there. Instead, on both sites we used 2 indices of relative abundance previously shown to be effective for Canis species: scat-deposition rates (Gese 2004; Knowlton 1984) and scent-station visitations (Gese 2004; Roughton and Sweeny 1982). For scat-deposition rates, we randomly established 2-km transects along dirt roads throughout BGF (n ¼ 4) and PR (n ¼ 3). After clearing the transects 4 weeks prior, transects were walked once per season from May 2006 to February 2007, and the number of jackal scats deposited per transect was counted. Scats from jackals were distinguished from those of foxes and other smaller carnivores on the basis of size and shape. For scent-station visitations, we established three 5.5-km transects per study site, with .2 km separating each transect. Along each transect, scent stations were placed every 0.5 km, resulting in 12 scent stations per transect. Scent stations consisted of a 1-m circle of sifted soil, baited with a cotton pad soaked in synthetic fermented egg solution (Schmitt Enterprises, Inc., New Ulm, Minnesota) and commercially available fish oil. On both sites, we baited stations and checked them for 3 consecutive mornings in July 2007, and replicated this process in August 2007. Each morning, tracks were identified to species on the basis of shape and size (Liebenberg 1990), and then the soil was resifted to erase all tracks. For each transect, we calculated a scent-station index (SSI) on the basis of the total number of stations with jackal visits divided by the total

298


number of operable stations, multiplied by 10 (Roughton and Sweeny 1982). Number of scats per transect and SSIs were compared between sites using Mann–Whitney U-tests because data were not normally distributed. Although the SSIs were only calculated over a 2-month period, the primary purpose was to compare this method with scat-deposition rates to determine if they showed similar trends. We estimated the density of jackals on PR by taking differences between sites in scat-deposition rates and SSIs, and extrapolating from the density reported on BGF. Home range and spatial analyses.—We determined annual home ranges using the 96% minimum convex polygon (MCP) method as calculated by using the animal movement extension (Hooge and Eichenlaub 1997) in ArcView geographic information system (GIS) software (version 3.2, Environmental Systems Research Institute, Inc., Redlands, California). We used 96% MCP because when compared with other methods (i.e., kernel and harmonic mean) and other percentages, it was shown to be the most accurate method for fitting estimated home ranges to actual territories for coyotes (Shivik and Gese 2000); thus it would likely work best for other territorial canids as well (see Kamler et al. 2012b for more details). We determined core areas, which were areas of concentrated use within home ranges, using the 50% MCP method. Core areas were biologically relevant because they encompassed all known natal dens of each fox, and were generally centered within the home ranges. Area-observation curves showed that home ranges for individuals were effectively determined by the first 30 locations; thus we included in analyses only foxes with .30 locations and 9 months of tracking. If foxes were tracked across multiple years, only the home range calculated during the 1st year was used in analyses, because data from the same fox across years were not independent. For both species, mean annual sizes of home ranges were compared with that reported on BGF using t-tests. We compared spatial overlap of annual home ranges and core areas between species. After initial analysis showed complete interspecific overlap of annual home ranges, we calculated the mean percent overlap of individual core areas between species. Spatial comparisons were analyzed only if: different individuals were monitored simultaneously for .1 month, and we were confident that all individuals or groups from other species that occurred within any individual core area were radiocollared. Following Kamler et al. (2012b), percent point overlap was determined for each fox by summing the number of its locations that fell within the overlap area, divided by the total number of locations within its core area. Interspecific avoidance of core areas was examined on the basis of 2 types of data: active locations and den site locations. To determine if foxes avoided other core areas within their home ranges while active, the number of active locations within other core areas was compared with that expected (on the basis of percentage of other core areas within individual home ranges) using compositional analysis (Aebischer et al. 1993; see ‘‘Habitat selection’’ below for more details). Individual foxes were the sampling unit, and their movements

Vol. 94, No. 2

were assumed to be independent of each other, especially because we did not include group members that foraged together. Den sites included burrows that were dug by other animals (e.g., springhares [Pedetes capensis] and aardvarks [Orycteropus afer]) and small depressions above ground, primarily underneath bushes and shrubs and sometimes in hollowed-out termite (Trinervitermes trinervoides) mounds. The locations of den sites were not independent across individuals within each species, because dens were used alternatively, and sometimes simultaneously, by different individuals of the same species, and the same individual may have used a particular den on several different occasions for various periods of time. Therefore, to determine if foxes avoided other core areas at the study-site level, we grouped all den locations for each fox species, and considered the den locations as the independent sampling units. Dens found within core areas of other species were compared with that expected using chi-square goodnessof-fit tests. To calculate expected values, we determined the percentage of other core areas within the entire area occupied by each fox species, using 100% MCP on the basis of all locations pooled across individuals for each species (Kamler et al. 2003a), then multiplied the percentage by the total number of dens. Habitat selection.—Analysis of habitat selection was based on 2 types of data: active locations and den-site locations. For both types of data, habitat types used and available were generated using ArcView 3.2. Habitat types were delineated using GIS data created from satellite images (3-m resolution) verified and corrected by ground truthing. Habitat types were classified as pan basin, pan slope, bushveld, savanna, and other. Pan basin (2.4% of PR) occurred on an ephemeral lake (i.e., pan) and consisted of Nama-Karoo vegetation, dominated by small shrubs and short (,30 cm) grasses. Pan slope (76.1%) consisted of Nama-Karoo vegetation, dominated by small shrubs and short grasses. Bushveld (13.2%) consisted of savanna vegetation, dominated by scattered trees and tall grasses. Pure savanna (4.5%) was dominated by tall grasses mixed with some shrubs. Others included various habitats with low (,3%) occurrence such as wetland and agricultural land near a human residence. To determine if habitat types differed between sites, we compared frequencies (in km2) of different habitat types between PR and those reported on BGF (Kamler et al. 2012b) using a chi-square contingency table. To determine if the canid species partitioned habitats while active, we used compositional analysis (Aebischer et al. 1993) to compare habitat use versus available for each species at 2 spatial levels: within the study site and within individual home ranges. Compositional analysis uses the animal as the sampling unit; thus it avoids problems related to sampling level, nonindependence of proportions, differential use by groups of animals, arbitrary definition of habitat availability, and pseudoreplication (Aebischer et al. 1993). We determined habitat selection of active locations for all foxes that met the requirements for home-range calculations (see above). We assumed there were no between-year differences in habitat use,

April 2013


especially given that canid densities remained the same during the study; thus, we pooled individuals across years for each species. We pooled across sexes for each species because bateared foxes foraged in groups; thus the location of 1 collared bat-eared fox represented the location of several individuals of different sexes. All species had access to all habitat types at the study-site level. The locations of den sites were not independent across individuals within each species (see ‘‘Home ranges and spatial relationships’’ above for more details). Therefore, to determine if foxes exhibited habitat partitioning for den sites, we grouped all den locations for each species, and considered the den locations as the independent sampling units. For each species, habitat use of den sites was compared with that expected using chi-square goodness-of-fit tests. To calculate expected values, we determined the percentage of different habitat types within an availability polygon, using 100% MCP based on all locations pooled across individuals for each species (Kamler et al. 2003a), then multiplied the habitat percentages by the total number of den sites. Activity patterns.—We determined activity patterns of foxes on PR and BGF by calculating the proportion of radiolocations, or fixes, during which each individual was moving. When obtaining the last azimuth for triangulation of study animals, observers listened for 1 min and recorded the signal as active if it had distinct attenuation, or inactive if it had indistinct or absent attenuation. Three periods were used for comparative analyses: day, sunset, and night. Sunset was classified as a 2-h period, 1 h before and 1 h after sunset. Day was classified as a 6-h period preceding the sunset period, whereas night was classified as a 6-h period succeeding the sunset period. Sample sizes were too low to include a sunrise period. Data were pooled across years, and the percent active fixes was calculated for each study animal during each period. We used general linear models (GLM) on variables (i.e., proportion of active fixes per period) that were arcsine-transformed (Grafen and Hails 2002), weighted relative to the number of fixes per individual (=n, n ¼ number of fixes), to evaluate relationships of activity with site (PR and BGF), time period, and individual. Thus, individuals were the sampling unit, and these data were nested within site to account for repeated use of the same individuals over time within a site. Activity data were weighted to avoid biasing results toward individuals with more data points (Thomas and Taylor 2006). We analyzed activity data in Minitab 15 (Minitab Inc. 2007). Initial results of activity analysis on BGF showed that there was a significant species 3 season interaction (Kamler et al. 2012b). Therefore, we analyzed data for both fox species into 2 broad seasons: wet season (September–February) and dry season (March–August). Basal prey estimation.—Bat-eared foxes are primarily insectivorous, with specialized adaptations for feeding on termites (Maas and Macdonald 2004). Consequently, termites are usually the main prey of bat-eared foxes (Maas and Macdonald 2004; Skinner and Chimimba 2005), similar to that reported on BGF (Klare et al. 2011). Therefore, we estimated the relative abundance of northern harvester termites

299

(Hodotermes mossambicus) on both sites using pitfall traps. Northern harvester termites (hereafter termites) forage colonially, yet in unpredictable locations above ground (Wilson and Clark 1977); therefore presence or absence in pitfall traps (in contrast to numbers of individual termites) was used as an index to relative abundance. In areas occupied by bat-eared foxes on each study site, we randomly established 6 transects with .2 km separating transects. Along each transect, we placed 5 pitfall traps every 100 m. We set pitfall traps and checked them after 4 days, and we summed for each transect the number of pitfall traps that contained termites. We conducted surveys monthly from June to September 2007 (i.e., 4 replicates), because this coincided with the winter season when northern harvester termites are most active (Coaton 1958). Total number of pitfall traps per transect that contained termites was combined across months and compared between sites using a Mann–Whitney U-test because data were not normally distributed. Cape foxes frequently consume small (,1 kg) to mediumsized (1–3 kg) prey (Skinner and Chimimba 2005), and on BGF they preyed primarily on small rodents and medium-sized mammals such as hares (Lepus) and springhares (Kamler et al. 2012b). Because quantifying the biomass of a diverse array of small rodents, each heterogeneously distributed, was not logistically possible during this study, we used biomass of hares and springhares as proxies for total biomass of basal prey. These species were chosen because they are conspicuous and relatively easy to survey, and both species are generalist feeders that consume a wide variety of vegetation types (Anderson 1996; Skinner and Chimimba 2005). Consequently, their numbers likely reflect primary productivity and vegetation biomass, similar to that reported for leporid species in other areas (Cypher et al. 2000; Desmond 2004; MacCracken and Hansen 1982). To estimate densities of hares (primarily cape hares [Lepus capensis] and some scrub hares [L. saxatilis]) and springhares at both sites, we used line-transect sampling during 2007 and analyzed data using program Distance (Buckland et al. 2001). On PR, 11 line transects (1.2–2.8 km in length) were established along dirt tracks, totaling 21 km. On BGF, 12 line transects (1.4–4.1 km in length) were established along dirt tracks, totaling 28 km. Although roads can sometimes bias results in distance sampling (Buckland et al. 2001), the dirt tracks within our study sites were seldom used, and they occurred in homogeneous and open habitats. Thus, we were confident that our density estimates were representative of the entire study area (see Stenkewitz et al. 2010 for more details). For each species, the biomass equaled density multiplied by mean body mass. We used body mass estimates of 2.0 kg for cape hares (Stuart and Stuart 2001) and 2.6 kg for springhares (Anderson 1996), and summed both species for a single biomass estimate of basal prey. We calculated for both sites the stocking rate of all ungulates on the basis of large stock units (LSU), defined as livestock and wild ungulates .35 kg (Blaum et al. 2009). Additionally, we calculated total ungulate biomass/km2 for large (.35 kg) ungulates, because this might influence grazing pressure more

300


Vol. 94, No. 2

than numbers of individual grazers, especially if grazing species exhibit a wide range of different body masses. For calculating LSUs, we used data provided by the landowners on PR on the basis of complete annual counts of livestock and wild ungulates, whereas on BGF we used data from complete annual counts given by Klare et al. (2010). To calculate total ungulate biomass per site, we multiplied ungulate numbers by body mass of ungulate species. Body mass estimates of wild ungulate species were provided by Stuart and Stuart (2001), whereas body mass estimates of domestic ungulates were provided by Kamler et al. (2012a).

RESULTS Results are given for data collected on PR, and, where appropriate, data are statistically compared with those from BGF. For consistency and ease of comparison between sites, results from BGF given by Kamler et al. (2012b) are shown after similar data are given for PR, even if data were not statistically compared. Densities.—We captured, radiocollared, and monitored 9 cape foxes and 18 bat-eared foxes on PR from August 2005 to March 2008. On BGF, 11 cape foxes and 22 bat-eared foxes were radiocollared from June 2005 to February 2008 (Kamler et al. 2012b). Adult cape foxes occurred only in monogamous mated pairs on both PR and BGF. Mean (6 SE) adult group size of bat-eared foxes was significantly lower (t ¼ 4.410, P , 0.001) on PR (2.28 6 0.24 fox/group, range ¼ 1–5, n ¼ 25) than on BGF (4.37 6 0.30 fox/group, range ¼ 1–12, n ¼ 54). Estimated densities on PR were 1.36 fox/10 km2 for cape foxes, and 6.77 fox/10 km2 for bat-eared foxes. On BGF, estimated densities were 0.49 fox/10 km2 for cape foxes, and 10.70 fox/10 km2 for bat-eared foxes (Kamler et al. 2012b). The estimated density of black-backed jackals was 3.25 jackal/10 km2 on BGF, on the basis of movements and group sizes of radiocollared jackals (Kamler et al. 2012b). Scat deposition rates were significantly lower during every season (Mann–Whitney: all Z ¼ 2.223 to 2.141, all P ¼ 0.003 to 0.032) on PR (mean seasonal range ¼ 0–2.0 scat/transect) compared with BGF (mean seasonal range ¼ 5.3–17.5 scat/ transect). For scent-station surveys, the mean (6 SE) SSI during the 1st survey was significantly lower (Mann–Whitney: Z ¼ 2.087, P ¼ 0.037) on PR (0.00 6 0.00) than BGF (9.00 6 3.51). Similarly, the SSI during the 2nd survey was significantly lower (Z ¼ 1.993, P ¼ 0.046) on PR (0.67 6 0.67) than BGF (7.67 6 2.40). Overall, the mean seasonal scat deposition rate on PR was 6.3% (range ¼ 0–11.6%) of that on BGF, whereas the mean SSI on PR was 4.4% (range ¼ 0–8.7%) of that on BGF. When results of both survey methods were weighted equally, jackal abundance on PR was 5.4% of that on BGF, resulting in an estimated density of 0.18 jackal/10 km2 on PR (extrapolated from density of 3.25 jackal/10 km2 on BGF). Because both indices were independent yet showed similar differences in jackal abundance between sites, these indices likely reflected actual differences in jackal densities. Home range and spatial analyses.—Mean (6 SE) annual home-range size for cape foxes on PR was 9.20 6 0.82 km2 (n

FIG. 2. The a) annual home ranges and b) core areas of cape foxes (Vulpes chama, n ¼ 4) and bat-eared foxes (Otocyon megalotis, n ¼ 10) monitored simultaneously on private ranches (PR) in South Africa, 2006–2007. To better illustrate spacing patterns of all collared foxes, data from 2 bat-eared foxes are shown on the basis of 17 locations each (marked with *), although these data were not used in statistical analyses.

¼ 8), which was significantly smaller (t ¼ 12.029, P , 0.001) than that reported on BGF (27.68 6 1.45 km2, n ¼ 5). Mean annual home-range size for bat-eared foxes on PR was 2.79 6 0.30 km2 (n ¼ 11), which was significantly smaller (t ¼ 4.749, P , 0.001) than that reported on BGF (4.96 6 0.32 km2, n ¼ 16). Mean (6 SE) size of annual core areas on PR was 1.58 6 0.35 km2 for cape foxes and 0.73 6 0.16 km2 for bat-eared foxes. On BGF, mean annual core areas were 5.29 6 0.32 km2 for cape foxes and 0.72 6 0.10 km2 for bat-eared foxes (Kamler et al. 2012b). Because 1 bat-eared fox dispersed, and another died, before 30 locations were obtained, we estimated their minimum home ranges for illustrative purposes on the basis of 17 locations each, although the data were not included in statistical analyses (Fig. 2). Annual home ranges on PR overlapped completely between fox species (Fig. 2), which was

April 2013

301


Table 1.—Results of compositional analysis showing the matrix of means and standard errors of habitat selection at the study-site and homerange levels for cape foxes (Vulpes chama) and bat-eared foxes (Otocyon megalotis) monitored on private ranches in South Africa, 2005–2008. Under each species, the selection rank of different habitats is given in parentheses (higher ranking is denoted by higher number). Note that cape foxes used habitats in proportion to availability (P ¼ 0.306) at the home-range level. Pan basin

Pan slope

Bushveld

3.244 6 1.711

2.405 6 2.061 0.839 6 0.463

Savanna

Other

Study-site level Cape foxa Pan basin (2) Pan slope (4) Bushveld (3) Savanna (1) Other (0)

3.244 2.405 –1.323 –3.223

6 6 6 6

1.711 2.061 1.906 2.586

1.918 –0.359 –4.140 –5.151

6 6 6 6

1.368 1.264 1.855 2.039

0.839 6 0.463 4.568 6 0.792 6.468 6 1.341

3.728 6 1.046 5.628 6 0.947

1.323 6 1.906 4.568 6 0.792 3.728 6 1.046

3.223 6.468 5.628 1.900

6 6 6 6

2.586 1.341 0.947 1.919

5.151 7.070 4.793 1.011

6 6 6 6

2.039 1.034 1.122 1.384

1.900 6 1.919

Bat-eared foxa Pan basin (3) Pan slope (4) Bushveld (2) Savanna (1) Other (0) Home-range level

1.918 6 1.368 2.277 6 0.998 6.059 6 0.705 7.070 6 1.034

0.359 6 1.264 2.277 6 0.998 3.781 6 1.543 4.793 6 1.122

4.140 6 1.855 6.059 6 0.705 3.781 6 1.543 1.011 6 1.384

Bat-eared foxb Pan basin (0) Pan slope (1) Bushveld (2) a b

0.083 6 0.129 0.083 6 0.129 0.517 6 0.206

0.434 6 0.204

0.517 6 0.206 0.434 6 0.204

— — —

— — —

P , 0.001. P , 0.100.

similar to that reported on BGF (Kamler et al. 2012b). For cape foxes, mean point overlap of core areas with bat-eared foxes was 20.3% (range ¼ 0–61%) on PR (Fig. 2), compared with 24.3% (range ¼ 20–28%) on BGF (Kamler et al. 2012b). For bat-eared foxes, mean overlap of core areas with cape foxes was 18.0% (range ¼ 0–47%) on PR (Fig. 2), compared with 44.1% (range ¼ 0–100%) on BGF (Kamler et al. 2012b). When active, cape foxes used core areas of bat-eared foxes within their home ranges in proportion to availability (v21 ¼ 1.69, P ¼ 0.194) on PR, similar to that reported on BGF (Kamler et al. 2012b). When bat-eared foxes were active, they used core areas of cape foxes within their home ranges in proportion to availability (v21 ¼ 0.83, P ¼ 0.364) on PR, whereas on BGF bat-eared foxes used core areas of cape foxes more than expected (Kamler et al. 2012b). At the study-site level, den sites (n ¼ 40) of cape foxes occurred in the core areas of bat-eared foxes in proportion to availability (v21 ¼ 0.092, P ¼ 0.762) on PR, whereas on BGR they occurred in core areas of bat-eared foxes more than expected (Kamler et al. 2012b). Similarly, den sites (n ¼ 69) of bat-eared foxes occurred in the core areas of cape foxes in proportion to availability (v21 ¼ 0.427, P ¼ 0.514) on PR, whereas on BGR they occurred in core areas of cape foxes more than expected (Kamler et al. 2012b). Habitat selection.—At the study-site level, cape foxes (v24 ¼ 27.94, P , 0.001) and bat-eared foxes (v24 ¼ 41.32, P , 0.001) showed selection for habitats on PR, with pan slope ranking 1st for both species (Table 1), similar to that reported on BGF (Kamler et al. 2012b). At the home-range level, habitats of active locations were used in proportion to availability for cape foxes on PR (v22 ¼ 2.37, P ¼ 0.306),

similar to that reported on BGF (Kamler et al. 2012b). Bateared foxes showed marginal selection for habitats of active locations within their home ranges on PR (v22 ¼ 5.33, P ¼ 0.069), with bushveld ranking 1st (Table 1), whereas on BGF bat-eared foxes showed marginal selection with pan slope ranking 1st (Kamler et al. 2012b). The habitat of cape fox dens was in proportion to availability (v22 ¼ 4.08, P ¼ 0.130) on PR, whereas on BGF the dens of cape foxes were located in savanna habitat more than expected and bushveld less than expected (Kamler et al. 2012b). The habitat of bat-eared fox dens was used in proportion to availability (v22 ¼ 2.25, P ¼ 0.325) on PR, whereas on BGF the dens of bat-eared foxes were located in pan slope more than expected and bushveld less than expected (Kamler et al. 2012b). Overall, frequencies of habitat types available on the study sites were not significantly different (v24 ¼ 5.85, P ¼ 0.211) between PR and BGF. Activity patterns.—Activity patterns of cape foxes did not signficantly differ between sites in the wet season (GLM: F2,16 ¼ 2.16, P ¼ 0.147), whereas there was a marginally significant difference between sites in the dry season (F2,26 ¼ 2.55, P ¼ 0.097). During both seasons, cape foxes were more active during the day and less active at night on PR compared with BGF (Fig. 3). Activity patterns of bat-eared foxes significantly differed between sites in the wet season (GLM: F2,52 ¼ 5.50, P ¼ 0.007), but not in the dry season (F2,53 ¼ 0.46, P ¼ 0.636). The significant difference in the wet season was due to bateared foxes being less active at sunset on PR compared with BGF; however, there were no consistent patterns in differences across seasons for any time period (Fig. 3).

302


Vol. 94, No. 2

FIG. 3.Activity patterns of a) cape foxes (Vulpes chama) and b) bat-eared foxes (Otocyon megalotis) in the wet (September–February) and dry (March–August) seasons on private ranches (PR) and Benfontein Game Farm (BGF), South Africa, 2005–2008. Data are mean proportion (6 SE) of fixes classified as active (n ¼ number of foxes included in each season).

Basal prey estimation.—Overall, the relative abundance of termites was lower (Mann–Whitney: Z ¼2.906, P ¼ 0.004) on PR (X¯ ¼ 1.17 6 0.26 pitfalls/transect) compared with BGF (X¯ ¼ 2.63 6 0.37). This trend was consistent across months, with monthly relative abundance of termites 49.9–60.0% lower on PR (range of monthly means ¼ 1.00–1.33) than on BGF (2.33– 3.00). For hares and springhares, the estimated densities on PR were 10.2 hare/km2 (95% confidence interval [CI] ¼ 4.3–23.9, detection probability [DP] ¼ 0.53) and 12.9 springhare/km2 (95% CI ¼ 8.0–20.8, DP ¼ 0.65), compared with the reported estimates of 16.5 hare/km2 and 5.8 springhare/km2 on BGF (Stenkewitz et al. 2010). The detection functions for both species on PR were acceptable; thus we were confident that our density estimates were relatively accurate, especially because this method was shown to be acceptable for hares and springhares on BGF (Stenkewitz et al. 2010). On the basis of estimated densities 3 body mass, the total biomass of basal prey was 53.9 kg/km2 on PR and 48.1 kg/km2 on BGF. The stocking rates of ungulates was slightly higher on PR (LSU ¼ 29.5/100 ha) compared with BGF (LSU ¼ 27.2/100 ha). In contrast, total ungulate biomass was slightly lower on PR (1,789 kg/km2) compared with BGF (1,858 kg/km2), which indicated similar grazing pressure between sites.

DISCUSSION Black-backed jackals appeared to suppress populations of cape foxes, but not bat-eared foxes. In the near absence of

black-backed jackals on PR, the cape fox density increased nearly 3-fold compared with BGF, most likely due to the lack of predation and competition from jackals. The increased density was not likely due to differences in bottom-up factors such as habitat types, basal prey biomass, and grazing pressure (which may influence habitat and prey), because these factors were similar between sites. Our result was consistent with previous research that documented negative relationships between numbers of cape foxes and black-backed jackals across 22 sites in South Africa, associated with intensity in jackal control (Blaum et al. 2009). In contrast to our prediction, the density of bat-eared foxes decreased 37% on PR compared with BGF, despite the lower density of jackals on PR. The difference in bat-eared fox density was likely due to differences in termite abundance, the basal prey of bat-eared foxes, which was 56% lower on PR compared with BGF. Bat-eared fox abundance and reproductive success was related to termite abundance in other areas (Maas and Macdonald 2004; Nel et al. 1984; Waser 1980). Therefore, we conclude that the densities of bat-eared foxes on our study sites were related most to termite abundance, rather than top-down factors such a predation or competition from jackals. Adult group sizes of bat-eared foxes were significantly lower by 48% on PR compared with BGF, which supported our prediction. Bat-eared foxes mob predators when threatened (Kamler et al. 2012b; Malcolm 1986), and form larger groups to mob larger predators (Maas and Macdonald 2004); thus in

April 2013


the absence of their main predator, in this case jackals, the group size of bat-eared foxes decreased. Our result was consistent with data from BGF that showed that bat-eared foxes in larger groups had lower predation from jackals than those in smaller groups (Kamler et al. 2012b), suggesting that group size was an adaptive response to predation. Previous researchers speculated that group size in bat-eared foxes was antipredator behavior (Lamprecht 1979; Malcolm 1986), whereas others have shown that group sizes of prey species can be an adaptive response to predation (Jedrzejeski et al. 2006). Although multiple factors likely influence group sizes of bat-eared foxes, including food resources (Maas and Macdonald 2004; Nel et al. 1984), we conclude that group sizes of bateared foxes in our study were influenced most by differences in jackal numbers. Home-range sizes for both fox species were significantly smaller on PR than BGF, which supported our prediction. Because density of cape foxes was lower on PR, it was not surprising that their home ranges were smaller on PR, because home-range size is negatively related to density in red foxes (V. vulpes—Sargeant 1972; Trewhella et al. 1988), a closely related species. On BGF, cape foxes avoided jackal core areas when foraging and establishing dens, and consequently had unusually large home ranges for a carnivore of their body size (Gittleman and Harvey 1982; Kamler et al. 2012b). Therefore, the smaller home ranges on PR also may have been due to the lack of jackals, which resulted in more potential foraging areas and den sites for cape foxes, reducing their overall movements. In fact, Kamler et al. (2012b) classified the cape fox as a fugitive species (i.e., species that forages widely and selectively uses areas with low occurrence of intraguild predators—Creel and Creel 2002; Durant 1998), due to their relatively large home ranges and avoidance of jackal core areas. Similarly, home ranges of other fugitive carnivores, including African wild dogs and long-tailed weasels (Mustela frenata), increased across sites when numbers of their intraguild predators increased, despite similar habitat and food resources (Creel and Creel 2002; St-Pierre et al. 2006). We conclude that differences in home-range sizes of cape foxes in our study were influenced most by the presence of jackals, which affected both the density and space use of cape foxes. Interestingly, bat-eared foxes did not exhibit the typical negative relationship between density and home-range size observed in other fox species (Sargeant 1972; Trewhella et al. 1988); thus our results appeared to be counterintuitive. Similarly, previous researchers found negative relationships between termite abundance and home-range size of bat-eared foxes (Maas and Macdonald 2004; Mackie and Nel 1989), the opposite of that found in our study. However, because bateared foxes likely formed larger groups on BGF due to jackal predation, home ranges on BGF were probably larger to incorporate more group members, regardless of available food resources or bat-eared fox density. Additionally, bat-eared foxes on BGF avoided jackal core areas when establishing den sites; thus home ranges may have been smaller on PR because lack of jackals resulted in more potential areas for den sites,

303

reducing overall movements. Thus, we conclude that homerange sizes of bat-eared foxes in our study were influenced most by differences in jackal numbers. Both fox species had similar habitat selection at the studysite level across sites, indicating that jackals do not influence habitat selection of foxes at broad scales. At the home-range level, use of habitats by cape foxes was similar between sites, indicating that jackals did not influence habitat use of cape foxes while they foraged within their home ranges. In contrast, bat-eared foxes selected most for bushveld habitat while foraging within their home ranges on PR, whereas on BGF they selected most for pan slope. Reasons for the difference in habitat selection by bat-eared foxes were not clear, but we suspect it was related to differences in food resources between sites. On BGF, bat-eared foxes supplemented their diet of termites with berries, primarily bluebush (Diospyros lycioides—Klare et al. 2011), the dominant shrub in bushveld habitat. Thus, due to the lower abundance of termites on PR compared with BGF, bat-eared foxes likely foraged more in bushveld habitat to take advantage of bluebush berries, an important alternative food resource. Concerning den sites, both fox species on PR used habitats in proportion to availability, which supported our prediction. On BGF, both fox species used bushveld less than expected when establishing den sites, likely because the taller vegetation of this habitat made it more difficult for foxes to observe approaching jackals while resting or guarding dens (Kamler et al. 2012b). Additionally, cape foxes selected savanna habitat more than expected for den sites on BGF, which was the habitat most avoided by jackals for their respective den sites. Previous research showed that habitat partitioning was a common mechanism of coexistence between different-sized canids (Atwood and Gese 2010; Fedriani et al. 2000; Nelson et al. 2007; Thompson and Gese 2007), although it was not clear if smaller canids actively changed habitats in the presence of larger canids (but see Miller et al. 2012). On the basis of the differences between sites, we conclude that habitat selection of den sites by both fox species was influenced by jackal presence, as both fox species appeared to alter habitats of den sites to better avoid or detect jackals on BGF. Cape foxes were consistently more active during the day and less active at night on PR compared with BGF, which supported our prediction. In fact, cape foxes were active 10– 13% of the time during the day on PR across seasons, compared with only 0–2% active during the same time period on BGF. On BGF, cape foxes were more nocturnal and less diurnal than jackals, which likely helped reduce encounter rates between the species (Kamler et al. 2012b). Previous research showed that temporal partitioning facilitated coexistence between sympatric carnivores (Di Bitetti et al. 2009; Harrington et al. 2009), and that activity patterns of canids could be an adaptive response to mortality (Kitchen et al. 2000). Because cape foxes exhibited different activity patterns in the absence of jackals, and differences were consistent across seasons, we conclude that the presence of jackals influenced activity patterns of cape foxes on BGF.

304

Vol. 94, No. 2


The activity of bat-eared foxes in southern Africa was shown to follow the activity of their basal prey, northern harvester termites, which are nocturnal during most of the year, except in winter (i.e., dry season) when they become more diurnal (Nel and Maas 2004; Skinner and Chimimba 2005). Bat-eared foxes followed this same pattern on both PR and BGF; thus we conclude that their activity pattern was influenced most by termite activity, rather than by jackals. The core areas of cape foxes overlapped those of bat-eared foxes by a similar percentage on both sites. In contrast, the core areas of bat-eared foxes on PR overlapped those of cape foxes less than half as much as on BGF. Compared with results from BGF, bat-eared foxes on PR no longer selected for cape fox core areas when active, and both fox species on PR no longer selected for the other species’ core areas when establishing their respective den sites, which supported our prediction. On BGF, the positive associations between the fox species were unusual for canids, and Kamler et al. (2012b) speculated that this may have been due to their mutual avoidance of jackal core areas, especially when establishing den sites. Our results on PR showed that in the absence of jackals, the fox species became more segregated when foraging and when establishing den sites. Therefore, we conclude that the presence of jackals influenced the interspecific spatial associations between fox species. Because this was an observational study, some fox behaviors might have been confounded to some degree by site-specific characteristics that we did not measure. Nevertheless, the absence of jackals, a dominant mesocarnivore and predator of both fox species, appeared to affect a broad range of fox behaviors. Several sublethal effects were species specific (e.g., group size differences), probably due to different evolutionary histories of the fox species and related constraints on behavioral plasticity. The sublethal effects of jackals occurred in concert with population suppression for cape foxes, but not for bat-eared foxes. This indicates that sublethal effects of mesocarnivores on small carnivores can be present whether or not population suppression occurs. Our results were consistent with the suspected and confirmed sublethal effects associated with interspecific killing within invertebrate, avian, and mammalian predator communities (Palomares and Caro 1999; Polis et al. 1989; Sergio and Hiraldo 2008), as well as the various sublethal effects that predators have on their prey (Berger 2007; Brown et al. 1988; Creel et al. 2005; OwenSmith and Mills 2006; Schmitz et al. 2004; Sheriff et al. 2009). Sublethal effects of mesocarnivores likely vary across ecosystems, being dependent on numbers and diversity of different-sized carnivores, as well as food and habitat resources. We compared the ecology of foxes in ecosystems with large carnivores absent and jackal numbers manipulated to the extremes, while bottom-up factors were relatively constant, at least for cape foxes. Future research should investigate the ecology of foxes on additional sites that vary in jackal numbers, have large carnivores present, or where food and habitat resources are manipulated to the extremes. Such research will lead to a more comprehensive understanding of

the lethal and sublethal effects of mesocarnivores, as well as the other factors that regulate the density and ecology of small carnivores.

ACKNOWLEDGMENTS We thank I. Joubert and G. de Wett for allowing us access to their properties, and De Beers Consolidated Mines and the McGregor Museum for providing logistical support. We also thank L. Mayer and S. Visagie for making traps, Z. Davidson, A. Loveridge, and W. B. Ballard (deceased) for equipment loans, and E. Herrmann, B. Wilson, B. Dean, N. F. Jacobsen, U. Klare, and D. Nkosi for help with field research. Funding for JFK was provided by a Research Fellowship from the Wildlife Conservation Society, New York, and a Marie Curie Fellowship from the European Commission, Brussels, Belgium. Additional support was provided by British Airways, United Kingdom, and grants to DWM from the Peoples’ Trust for Endangered Species. Our research protocol (#0401/05) was approved by the Department of Tourism, Environment and Conservation, Kimberley, South Africa.

LITERATURE CITED AEBISCHER, N. J., P. A. ROBERTSON, AND R. E. KENWARD. 1993. Compositional analysis of habitat use from animal radio-tracking data. Ecology 74:1313–1325. ANDERSON, P. C. 1996. The population dynamics and ecological role of the springhare Pedetes capensis (Forster, 1778) in the Kimberley area, Northern Cape Province, South Africa. Ph.D. dissertation, University of the Orange Free State, Bloemfontein, South Africa. ATWOOD, T. C., AND E. M. GESE. 2010. Importance of resource selection and social behavior to partitioning of hostile space by sympatric canids. Journal of Mammalogy 91:490–499. BERGER, J. 2007. Carnivore repatriation and holarctic prey: narrowing the deficit in ecological effectiveness. Conservation Biology 21:1105–1116. BERGER, K. M., E. M. GESE, AND J. BERGER. 2008. Indirect effects and traditional trophic cascades: a test involving wolves, coyotes, and pronghorn. Ecology 89:818–828. BLAUM, N., B. TIETJEN, AND E. ROSSMANITH. 2009. Impact of livestock husbandry on small- and medium-sized carnivores in Kalahari Savannah Rangelands. Journal of Wildlife Management 73:60–67. BROWN, J. S., K. P. KOTLER, R. J. SMITH, AND W. O. WIRTZ II. 1988. The effects of owl predation on the foraging behavior of hetermyid rodents. Oecologia 76:408–415. BROWN, J. S., J. W. LAUNDRE´ , AND M. GURUNG. 1999. The ecology of fear: optimal foraging, game theory, and trophic interactions. Journal of Mammalogy 80:385–399. BUCKLAND, S. T., D. R. ANDERSON, K. P. BURNHAM, L. LAAKE, D. L. BORCHERS, AND L. THOMAS. 2001. Introduction to distance sampling. Oxford University Press, Oxford, United Kingdom. COATON, W. G. H. 1958. The hodotermitidae harvester termites of South Africa. Department of Agriculture Science Bulletin 375, Entomology Series 43:1–112. CREEL, S., AND D. CHRISTIANSON. 2008. Relationships between direct predation and risk effects. Trends in Ecology and Evolution 23:194–201. CREEL, S., AND N. M. CREEL. 2002. The African wild dog: behavior, ecology, and conservation. Princeton University Press, Princeton, New Jersey.

April 2013


CREEL, S., J. WINNIE, JR., B. MAXWELL, K. HAMLIN, AND M. CREEL. 2005. Elk alter habitat selection as an antipredator response to wolves. Ecology 86:3387–3397. CRESSWELL, W., J. LIND, AND J. L. QUINN. 2010. Predator-hunting success and prey vulnerability: quantifying the spatial scale over which lethal and non-lethal effects of predation occur. Journal of Animal Ecology 79:556–562. CYPHER, B. L., ET AL. 2000. Population dynamics of San Joaquin kit foxes at the Naval Petroleum Reserves in California. Wildlife Monographs 145:1–43. DESMOND, M. J. 2004. Habitat associations and co-occurrence of Chihuahuan Desert hares (Lepus californicus and L. callotis). American Midland Naturalist 151:414–419. DI BITETTI, M. S., Y. E. DI BLANCO, J. A. PEREIRA, A. PAVIOLO, AND I. J. PEREZ. 2009. Time partitioning favors the coexistence of sympatric crab-eating foxes (Cerdocyon thous) and pampas foxes (Lycalopex gymnocercus). Journal of Mammalogy 90:479–490. DURANT, S. M. 1998. Competition refuges and coexistence: an example from Serengeti carnivores. Journal of Animal Ecology 67:370–386. FEDRIANI, J. M., T. K. FULLER, R. M. SAUVAJOT, AND E. C. YORK. 2000. Competition and intraguild predation among three sympatric carnivores. Oecologia 125:258–270. GESE, E. M. 2004. Survey and census techniques for canids. Pp 273– 279 in Canids: foxes, wolves, jackals and dogs. Status survey and conservation action plan (C. Sillero-Zubiri, M. Hoffmann, and D. W. Macdonald, eds.). IUCN/SSC Canid Specialist Group, Gland, Switzerland. GITTLEMAN, J. L., AND P. H. HARVEY. 1982. Carnivore home-range size, metabolic needs and ecology. Behavioural Ecology and Sociobiology 10:57–63. GRAFEN, A., AND R. HAILS. 2002. Modern statistics for the life sciences. Oxford University Press, Oxford, United Kingdom. HARRINGTON, L. A., ET AL. 2009. The impact of native competitors on an alien invasive: temporal niche shifts to avoid interspecific aggression? Ecology 90:1207–1216. HOOGE, P. N., AND B. EICHENLAUB. 1997. Animal movement extension to arcview, ver. 1.1. Alaska Biological Science Center, U.S. Geological Survey, Anchorage. JEDRZEJEWSKI, W., H. SPAEDTKE, J. F. KAMLER, B. JEDRZEJEWSKA, AND U. STENKEWITZ. 2006. Group size dynamics of red deer in Bialowieza Primeval Forest, Poland. Journal of Wildlife Management 70:1054– 1059. KAMLER, J. F., W. B. BALLARD, E. B. FISH, P. R. LEMONS, K. MOTE, AND C. C. PERCHELLET. 2003a. Habitat use, home ranges, and survival of swift foxes in a fragmented landscape: conservation implications. Journal of Mammalogy 84:989–995. KAMLER, J. F., W. B. BALLARD, R. L. GILLILAND, P. R. LEMONS, AND K. MOTE. 2003b. Impacts of coyotes on swift foxes in northwestern Texas. Journal of Wildlife Management 67:317–323. KAMLER, J. F., W. B. BALLARD, R. L. GILLILAND, AND K. MOTE. 2003c. Spatial relationships between swift foxes and coyotes in northwestern Texas. Canadian Journal of Zoology 81:168–172. KAMLER, J. F., W. B. BALLARD, P. R. LEMONS, AND K. MOTE. 2004. Variation in the mating system and group structure in two populations of swift foxes, Vulpes velox. Animal Behaviour 68:83–88. KAMLER, J. F., U. KLARE, AND D. W. MACDONALD. 2012a. Seasonal diet and prey selection of black-backed jackals on a small-livestock farm in South Africa. African Journal of Ecology 50:299–307.

305

KAMLER, J. F., U. STENKEWITZ, U. KLARE, N. F. JACOBSEN, AND D. W. MACDONALD. 2012b. Resource partitioning among cape foxes, bateared foxes, and black-backed jackals in South Africa. Journal of Wildlife Management 76:1241–1253. KITCHEN, A. M., E. M. GESE, AND E. R. SCHAUSTER. 2000. Changes in coyote activity patterns due to reduced exposure to human persecution. Canadian Journal of Zoology 78:853–857. KLARE, U., J. F. KAMLER, AND D. W. MACDONALD. 2011. The bat-eared fox: a dietary specialist? Mammalian Biology 76:646–650. KLARE, U., J. F. KAMLER, U. STENKEWITZ, AND D. W. MACDONALD. 2010. Diet, prey selection, and predation impact of black-backed jackals in South Africa. Journal of Wildlife Management 74:1030– 1042. KNOWLTON, F. F. 1984. Feasibility of assessing coyote abundance on small areas. Final Report, Work Unit 909:01. Denver Wildlife Research Center, Denver, Colorado. LAMPRECHT, J. 1979. Field observations on the behaviour and social system of the bat-eared fox Otocyon megalotis. Zeitschrift fur Tierpsychologie 49:260–284. LIEBENBERG, L. 1990. A field guide to the animal tracks of southern Africa. David Philip Publishers, Cape Town, South Africa. LIMA, S. L. 1998. Nonlethal effects in the ecology of predator–prey interactions. BioScience 48:25–34. MAAS, B., AND D. W. MACDONALD. 2004. Bat-eared foxes ‘insectivory’ and luck: lessons from an extreme canid. Pp 227–242 in The biology and conservation of wild canids (D. W. Macdonald and C. Sillero-Zubiri, eds.). Oxford University Press, Oxford, United Kingdom. MACCRACKEN, J. G., AND R. M. HANSEN. 1982. Herbaceous vegetation of habitat used by blacktail jackrabbits and nuttall cottontails in southeastern Idaho. American Midland Naturalist 107:180–184. MACDONALD, D. W. 1983. The ecology of carnivore social behaviour. Nature 301:379–384. MACKIE, A. J., AND J. A. J. NEL. 1989. Habitat selection, home range size, and group size of bat-eared foxes in the Orange Free State. South African Journal of Wildlife Research 19:135–139. MALCOLM, J. R. 1986. Socio-ecology of bat-eared foxes (Otocyon megalotis). Journal of Zoology (London) 208:457–467. MILLER, B. J., H. J. HARLOW, T. S. HARLOW, D. BIGGINS, AND W. J. RIPPLE. 2012. Trophic cascades linking wolves (Canis lupus), coyotes (Canis latrans), and small mammals. Canadian Journal of Zoology 90:70–78. MINITAB INC. 2007. Minitab statistical softward, version 15. Minitab, Inc., State College, Pennsylvania. NEL, J. A. J., AND B. MAAS. 2004. Bat-eared fox (Otocyon megalotis. Pp 183–189 in Canids: foxes, wolves, jackals and dogs. Status Survey and Conservation Action Plan (C. Sillero-Zubiri, M. Hoffmann, and D. W. Macdonald, eds.). IUCN/SSC Canid Specialist Group, Gland, Switzerland. NEL, J. A. J., M. G. L. MILLS, AND R. J. VAN AARDE. 1984. Fluctuating group size in bat-eared foxes (Otocyon m. megalotis) in the southwestern Kalahari. Journal of Zoology (London) 203:294–298. NELSON, J. L., B. L. CYPHER, C. D. BJURLIN, AND S. CREEL. 2007. Effects of habitat on competition between kit foxes and coyotes. Journal of Wildlife Management 71:1467–1475. OWEN-SMITH, N., AND M. G. L. MILLS. 2006. Manifold interactive influences of the population dynamics of a multispecies ungulate assemblage. Ecological Monographs 76:73–92. PALOMARES, F., AND T. M. CARO. 1999. Interspecific killing among mammalian carnivores. American Naturalist 153:492–508.

306


POLIS, G. A., C. A. MYERS, AND R. D. HOLT. 1989. The ecology and evolution of intraguild predation: potential competitors that eat each other. Annual Review of Ecology and Systematics 20:297–330. PRUGH, L. R., ET AL. 2009. The rise of the mesopredator. BioScience 59:779–791. RITCHIE, E. G., AND C. N. JOHNSON. 2009. Predator interactions, mesopredator release and biodiversity conservation. Ecology Letters 12:982–998. ROEMER, G. W., M. E. GOMPPER, AND B. VAN VALKENBURGH. 2009. The ecological role of the mammalian mesocarnivore. BioScience 59:165–173. ROUGHTON, R. D., AND M. W. SWEENY. 1982. Refinements in scentstation methodology for assessing trends in carnivore populations. Journal of Wildlife Management 46:217–229. SARGEANT, A. B. 1972. Red fox spatial characteristics in relation to waterfowl predation. Journal of Wildlife Management 36:225–236. SCHMITZ, O. J., V. KRIVAN, AND O. OVADIA. 2004. Trophic cascades: the primacy of trait-mediated indirect interactions. Ecology Letters 7:153–163. SERGIO, F., AND F. HIRALDO. 2008. Intraguild predation in raptor assemblages: a review. Ibis 150 (S1):132–145. SHERIFF, M. J., C. J. KREBS, AND R. BOONSTRA. 2009. The sensitive hare: sublethal effects of predator stress on reproduction in snowshoe hares. Journal of Animal Ecology 78:1249–1258. SHIVIK, J. A., AND E. M. GESE. 2000. Territorial significance of home range estimators for coyotes. Wildlife Society Bulletin 28:940–946. SIKES, R. S., W. L. GANNON, AND THE ANIMAL CARE AND USE COMMITTEE OF THE AMERICAN SOCIETY OF MAMMALOGISTS. 2011. Guidelines of the American Society of Mammalogists for the use of wild mammals in research. Journal of Mammalogy 92:235–253.

Vol. 94, No. 2

SKINNER, J. D., AND C. T. CHIMIMBA. 2005. The mammals of the southern African subregion. Cambridge University Press, Cambridge, United Kingdom. STENKEWITZ, U., E. HERRMANN, AND J. F. KAMLER. 2010. Distance sampling for estimating springhare, cape hare and steenbok densities in South Africa. South African Journal of Wildlife Research 40:87–92. ST-PIERRE, C., J. P. OUELLET, AND M. CRETE. 2006. Do competitive intraguild interactions affect space and habitat use by small carnivores in a forested landscape? Ecography 29:487–496. STUART, C., AND T. STUART. 2001. Field guide to mammals of southern Africa. Struik Publishers, Cape Town, South Africa. THOMAS, D. L., AND E. J. TAYLOR. 2006. Study designs and tests for comparing resource use and availability II. Journal of Wildlife Management 70:324–336. THOMPSON, C. M., AND E. M. GESE. 2007. Food webs and intraguild predation: community interactions of native mesocarnivore. Ecology 88:334–346. TREWHELLA, W. J., S. HARRIS, AND F. E. MCALLISTER. 1988. Dispersal distance, home range size and population density in the red fox (Vulpes vulpes): a quantitative analysis. Journal of Applied Ecology 25:423–434. WASER, P. M. 1980. Small nocturnal carnivores: ecological studies in the Serengeti. African Journal of Ecology 18:167–185. WILSON, D. S., AND A. B. CLARK. 1977. Above ground defense in the harvester termite Hodotermes mossambicus (Hagen). Journal of the Entomological Society of Southern Africa 40:271–282.

Submitted 7 May 2012. Accepted 24 August 2012. Associate Editor was Dirk H. Van Vuren.