SOCIAL ORGANIZATION AND GROUP FORMATION OF RACCOONS AT THE EDGE OF THEIR DISTRIBUTION

Transcription

1 Journal of Mammalogy, 89(3): , 2008 SOCIAL ORGANIZATION AND GROUP FORMATION OF RACCOONS AT THE EDGE OF THEIR DISTRIBUTION JUSTIN A. PITT,* SERGE LARIVIÈRE, AND FRANÇOIS MESSIER Department of Biological Sciences, University of Alberta, Z-907 Biological Sciences Buildings, Edmonton, Alberta T6G 2E9, Canada (JAP) Cree Hunters and Trappers Income Security Board, Edifice Champlain, Bureau 110, 2700 Boulevard Laurier, Sainte-Foy, Quebec G1V 4K5, Canada (SL) Department of Biology, University of Saskatchewan, 112 Science Place, Saskatoon, Saskatchewan S7N 5E2, Canada (FM) We examined spatial and temporal relationships of 104 raccoons (Procyon lotor) at the northern edge of their distribution to report social organization and to test hypotheses regarding the formation of coalition groups among adult male carnivores. Social tolerance among females varied among years, but adults maintained relatively exclusive home ranges (overlap ranged from 2.2% to 12.5%) that were dispersed in a uniform pattern throughout the duration of the study (Clark Evans ratios ranging from 1.32 to 1.87, P, ). Social interactions among males were more complex than previously described for low-density populations of raccoons, with most adults (approximately 80%) forming coalition groups. We identified 9 unique coalitions that had extensively overlapping home ranges and positively associated nightly movements within group members (overlap indices ranging from 85.3% to 97.3%). Coalitions maintained exclusive territories between groups (overlap indices ranging from 0% to 3.2%) and formed distinct spatial boundaries that were highly correlated with home-range boundaries of females. Male coalition groups in the order Carnivora are hypothesized to form in response to aggregations of females, but examination of our data suggests that this is not a prerequisite for their formation. We propose a dominance hierarchy where subordinate males benefit through increased likelihood of inheritance of territories, whereas dominant males benefit through increased efficiency of territorial defense. Key words: coalition groups, dynamic interaction, group structure, Procyon lotor, raccoons, social organization, spatial dispersion, territoriality Understanding the factors that dictate spatial organization and social interactions is a fundamental pursuit of animal ecology (Adams 2001). Space use and spatial patterns are influenced by the distribution of resources as well as the presence and characteristics of other individuals (Shier and Randall 2004). Males and females commonly exhibit varying spatial patterns because of differences in fitness pressures between sexes (Lott 1991). Typically, space use by females is influenced by the dispersion of resources and their associated requirements. However, space use by males is coupled with the ability not only to exploit these same resources, but also to mate with females. As such, spatial patterns among males also become a function of the spatial distribution of females (Macdonald 1983). * Correspondent: pitt@ualberta.ca Ó 2008 American Society of Mammalogists Territory ownership is a major determinant of fitness in territorial animals (Morrell and Kokko 2004). Social tolerances within both sexes should be linked to resource availability, because defense of territories costs energy (Davies and Houston 1984; López-Sepulcre and Kokko 2005). Territorial defense would not be beneficial if the resource of interest is spread thinly over a large area that cannot compensate for the cost of defense, and the resource is overly abundant and not worth defending (Krebs and Davies 1993; Samson and Huot 2001). Conversely, it has been hypothesized that grouping behavior and aggregations are favored when resources are highly concentrated (Gehrt and Fritzell 1998). If the resource is not easily depleted, the benefits accrued to defend this resource would not be worth the cost of defending, and we would expect nonassociated aggregations of individuals (i.e., static interactions see Doncaster 1990). Alternatively, if the resource is highly concentrated but diminishable, an individual may not be able to maintain exclusive access to that resource. In such instances, individuals may benefit by forming coalition groups that cooperatively defend that resource (Durant et al. 2004; 646

2 June 2008 PITT ET AL. SOCIAL ORGANIZATION OF RACCOONS 647 McDonald and Potts 1994). Because spatial organization of males is a function of dispersion of females in many mammalian systems, it has been hypothesized that spatial organization of females may be the determinant of sociality and cooperation in adult male carnivores (Caro 1994). As females attain localized high densities, the ability of a male to successfully monopolize access to a home range of a female decreases (Gehrt and Fritzell 1998). As competition among males increases, the benefit of forming a coalition group becomes greater (Caro 1989, 1994; McDonald and Potts 1994). In such cases, we would expect associated aggregations (dynamic interactions) among males that cooperatively defend access to females while maintaining mutually exclusive territories between groups. The raccoon (Procyon lotor) is an omnivorous carnivore that has a distribution that spans most of North America. Space use and social organization of the raccoon have been studied extensively in the southern portion of its distribution where population density is high (Zeveloff 2002). Raccoons have the ability to display a wide array of social tolerances, ranging from completely tolerant, to strict territoriality, to complex systems of associated and nonassociated aggregations (Chamberlain and Leopold 2002; Gehrt and Fritzell 1998; Kamler and Gipson 2003; Walker and Sunquist 1997). Ecological factors, such as dispersion of resources, may in part explain this variability observed across the species distribution (Zeveloff 2002). Raccoons are known to have a high affinity for water and its associated resources (Gehrt and Fritzell 1998; Lotze and Anderson 1979). As such, it has been hypothesized that spatial dispersion of females in portions of their distribution is dictated by the dispersion of water resources (Gehrt and Fritzell 1998). When this resource becomes limiting and aggregated in the landscape, females respond by becoming aggregated, causing nonassociated localized high densities (Gehrt and Fritzell 1998). This aggregation of females has been hypothesized as the underlying mechanism dictating the recently discovered formation of coalition groups of adult males in southern populations of raccoons (Chamberlain and Leopold 2002; Gehrt and Fritzell 1998). Male coalition groups have positive dynamic interactions (i.e., temporal association of movements) and maintain exclusive territories to other groups of males (Gehrt and Fritzell 1998). Such cooperative behavior among adult males is relatively rare among mammalian populations and few studies have critically evaluated possible underlying mechanisms. We report on the spatial organization and social interactions of raccoons to test hypotheses regarding the social behavior of females and males and discuss this in the context of the formation of coalition groups of males. Little is known about the social ecology of raccoons in the prairies and parklands of Canada, with virtually no spatial data existing. Populations of raccoons have only recently become established in this portion of their distribution (Larivière 2004). However, because resources are sparse in this part of their range (Fritzell 1978a), we hypothesized females would maintain relatively exclusive home ranges dispersed in a regular pattern. Because grouping behavior of males has been hypothesized to be a result of aggregations of females, we expected males to maintain exclusive territories and remain asocial as was previously described in their northern distribution (Fritzell 1978a). Conversely, if we find that females are aggregated in the landscape, we would expect males to form coalition groups to cooperatively defend territories and access to females as found in their southern distribution (Chamberlain and Leopold 2002; Gehrt and Fritzell 1998). MATERIALS AND METHODS Study area. We conducted our study in the prairie parkland region of southwestern Manitoba, Canada ( N, W), at the northern edge of the distribution of raccoons. The area was gently rolling and characterized by an intensively cultivated landscape (63%). Small grain (wheat, barley, and oats) and oil (mainly canola and flax) crops occupied much of the study site. Grassland sites and areas managed for upland nesting waterfowl comprised 15.4% of the landscape. Numerous wetlands with narrow fringes of vegetation occurred throughout the study site and represented 14% of the landscape. Intermixed were small wooded bluffs (3.4%) composed largely of quaking aspen (Populus tremuloides), bur oak (Quercus macrocarpa), and chokecherry trees (Prunus virginiana). These bluffs typically ranged in size from 10 to 30 ha. A network of gravel roads (3%) divided the interior of the study site. Farmsteads, both active and abandoned, comprised 1.2% of the landscape. Snow and rainfall averaged 107 mm and 387 mm per year, respectively. Mean daily temperature averaged 18C and ranged monthly from 18.48C in January to 17.38C in July. Capture and radiotracking. We captured 104 raccoons in commercial mesh-wire live traps (Billman Supplies, Columbus, Ohio) during 4 annual 10-day trapping sessions that occurred from 2002 through 2004 in April, May, June, and October. Minimum density estimates in our study site ranged from 0.32 to 0.71 raccoons/km 2 from 2002 to Traps were deployed in a grid fashion and were placed in every quarter section over a 78-km 2 study site and baited with canned cat food. We checked traps each morning because raccoons are primarily nocturnal (Greenwood 1982). All raccoons were anesthetized using a standard dose of 20 mg of Zoletil (Vibrac, Carros, France) that was administered to the hind leg (Pitt et al. 2006a). Upon inoculation, raccoons were weighed and ear-tagged (National Band and Tag Co., Newport, Kentucky), sex was determined, and body condition was assessed using bioelectrical impedance analysis (Pitt et al. 2006b). We determined age by examining tooth wear (Grau et al. 1970) and reproductive status was verified by observing size and color of female teats and size and distention of male testicles. Raccoons were equipped with radiocollars ( MHz; Advanced Telemetry Systems, Isanti, Minnesota) and released at the site of capture. We radiotracked raccoons each night ( h) from mid-april to the beginning of August using a truckmounted null peak system (Spencer et al. 1987) and found raccoons daily from September to the end of November. Because raccoons become inactive during the winter, they were found bimonthly for the remainder of the year. All procedures

3 648 JOURNAL OF MAMMALOGY Vol. 89, No. 3 were approved by Manitoba Wildlife Animal Care Committee ( ) and followed guidelines approved by the American Society of Mammalogists (Gannon et al. 2007). An important assumption of our analysis is that all or nearly all raccoons in the study area were radiocollared. Examination of our trapping data suggested that most if not all yearling and adult raccoons were radiomarked in our study site. After the 3rd trapping session (June 2002), we captured only 1 adult female in the subsequent 9 trapping sessions that was not previously marked as a juvenile or yearling. That particular female was captured at the southern edge of our trapping grid. The 3rd trapping session coincided with the asymptote of new captures for both adult males and females. Yearling migrant males were captured continually throughout the study but rarely established themselves as residents. General spatial analyses. Locations were obtained by triangulation with a minimum 2-h interval between successive fixes (Salvatori 1999). We used at least 3 directional bearings to estimate an individual location, with bearings generally taken within 5 8 min of each other (White and Garrott 1990). We calculated the standard deviation of the bearing error (Lee et al. 1985), which was used to calculate 95% confidence ellipses in program LOCATE II (Nams 1990). Locations were estimated in Universal Transverse Mercators using the maximum-likelihood estimator available in LOCATE II software. These data were entered into ArcView 3.2 geographic information system software (Environmental Systems Research Institute, Redlands, California). Home ranges were estimated using the Animal Movement extension in ArcView (Hooge and Eichenlaub 1997). We calculated 100% minimum convex polygons to estimate home ranges, which were used for all spatial analyses. We tested for correlation between number of locations and range size using Pearson s product moment to determine if there was a relationship between sampling effort and estimated home-range size (Larivière and Messier 1998). We created area observation curves to determine the number of locations required to accurately depict the home range without being biased by sample size (Odum and Kuenzler 1955). All values are reported as mean 6 SE unless stated otherwise. Nearest-neighbor analysis. To assess the spatial dispersion of home ranges within the study site we used the Clark Evans ratio (R) of arithmetic centers to determine the distance of each raccoon from other conspecifics using the following equation (Clark and Evans 1954): X E X =SE X : ð1þ We calculated the expected distance (E(X)) and the standard error (SE) using corrections that accounted for both edge effects and correlations among nearest-neighbor distances using the following equation (Donnley 1978; Sinclair 1985): E X pffiffiffiffiffiffiffiffi pffiffi ¼ 0:5 A=n þ 0:051 þ 0:041= n L=n; ð2þ where n is the number of observed points (arithmetic centers in this case), A is the area for which n points fall, and L is the length of the boundary of the region (Donnley 1978). Standard error was calculated using equation 3: SE X ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi p 0:07A þ 0:037L ffiffiffiffiffiffiffiffi A=n=n: Clark Evans ratios (Rs) range in value from 0 to 2.15, with 0 indicating maximum aggregation, 1 indicating a random distribution, and 2.15 representing perfect uniformity or a regular pattern (Clark and Evans 1954). If raccoons were territorial we would expect home ranges to be uniformly distributed within the study site; conversely, if we found high social tolerance we would expect to observe aggregations of individuals. This analysis was used as one line of evidence for describing sociospatial patterns of raccoons, while also using it to test hypotheses regarding group formation by males and for comparison to studies of raccoons in high-density portions of their range. Static interaction. Static interaction in animal movements is defined as the spatial overlap of 2 home ranges and congruence in at least some part of their utilization distribution of location data (Doncaster 1990). For 1st approximation of static interaction, we report the intrasexual percent overlap of home ranges. Overlap area was calculated in ArcView using the merge and intersect functions in the Xtools extension (DeLuane 2003). Because this measure of sociospatial interactions does not take into account the utilization of the shared areas, we also calculated home-range overlap indices to determine intensity of use for raccoons with adjacent overlapping home ranges. Association was quantified by using the modified ratio presented by Ginsberg and Young (1992) in equation 4: ½ðn 1 þ n 2 Þ=ðN 1 þ N 2 ÞŠ 100; ð4þ where n 1 and n 2 relate to the number of locations that occur in the area of overlap and N 1 and N 2 refer to the number of total locations for that individual. This analysis is a coarse-scale approach to examining use of shared areas. To determine if raccoons were able to partition shared areas on a finer scale, we examined the correlation in home-range use of each range using a grid cell method of analysis (see Doncaster 1990). We used Spearman s rank coefficient, which was calculated on the pairs of fix scores obtained from all of the grid-cells (64 ha) frequented by 1 or both animals, and then tested for correlation in use between the 2 utilization distributions (Doncaster 1990). Dynamic interaction. To assess temporal interactions and association in movements among raccoons that shared portions of overlapping home ranges, we used a nonparametric test using simultaneous telemetry locations with the Animal Movement extension in ArcView and program DYNAMIC (Doncaster 1990). The dynamic interaction test was used to quantify and determine if movements of 2 raccoons that were monitored simultaneously were moving independently of each other by determining if the 2 animals were located within a critical distance more or less often than expected by chance (Doncaster 1990). We constructed cumulative frequency histograms of distance intervals separating paired locations that would be expected if animals were moving at random as ð3þ

4 June 2008 PITT ET AL. SOCIAL ORGANIZATION OF RACCOONS 649 TABLE 1. Summary of 105 estimated minimum convex polygon home ranges (km 2 ) for raccoons during the spring and summer in Manitoba, Sex and year n X SE Females All years Pregnant Nonpregnant Total Males All years Dyad Nondyad Total well as the actual observed distance between locations. We then determined if the observed distance of simultaneous paired locations differed from the expected frequency. A distance of 100 m was used to classify temporally paired locations following Gehrt and Fritzell (1998). If the observed frequency distribution of locations that fell within the critical distance was greater than what we would have expected by chance, it was considered a positive interaction, indicating that the movements of the 2 animals were positively associated with each other. Conversely, a significant negative interaction occurred if the expected frequencies (unpaired locations) were greater than the observed (paired) frequencies for each distance interval. The use of the dynamic interaction test was employed only if the observed area of overlap was at least 20% of the individual s home range. A chi-square test was used to determine whether a positive or negative dynamic interaction had occurred, with all tests considered significant at a ¼ We coupled this test with denning data obtained concurrently throughout the study to determine if individuals that were associated with each other during nightly activity also were selecting daytime resting sites with the same conspecific as supportive evidence for positive dynamic interactions. RESULTS General space use. We calculated 105 home ranges (58 of males and 47 of females) during the spring and summer of based on 7,705 locations. Mean number of locations used to estimate a home range was SD. Area observation curves consistently yielded asymptotes at approximately 60 locations per home range, which was used as our benchmark for data analyses. Additionally, we found no relationship between home-range size and the number of estimated locations (r 2 ¼ 0.08, P ¼ 0.51, n ¼ 105), indicating that home-range estimation was not influenced by sampling effort. Minimum convex polygon estimates of home ranges of males averaged 15.6 km SE, which were larger than home ranges of females ( km 2 ; F ¼ 22.7, d.f. ¼ 1, 103, P, 0.001; Table 1). For both sexes, minimum convex polygon estimates of home-range size were similar among years (F ¼ 0.91, d.f. ¼ 1, 103, P ¼ 0.41). Nearest-neighbor analysis. Home ranges of female raccoons were spaced in a regular pattern throughout most of the study (Table 2). Clark Evans ratios for females ranged in value from 1.32 to 1.87, with home-range dispersion in 1 year being only marginally different from a random distribution, but still tended toward a regular pattern (Table 2). Arithmetic center of home ranges of males ranged from a clumped to random distribution when not accounting for male coalition groups, with Clark Evans ratios ranging from 0.72 to However, when we merged male coalition groups and treated them as a single unit, dispersion of males tended toward a regular pattern, significantly so in all but 1 year (2003) that was only marginally significant toward a uniform distribution (Table 2). Static interaction. Among females inhabiting the study site, percent overlap of home ranges averaged 12.3% 6 2.3% SE for all years, and ranged annually from 6.7% to 21.1%. When considering only adult females, percent home-range overlap was less (U ¼ 727, n 1 ¼ 58, n 2 ¼ 34, P ¼ 0.04), ranging from 2.2% to 12.5% and averaging 5.6% 6 1.3% (Table 3). Indices of home-range overlap followed the same trends as percent overlap for adults and yearlings, with overlap indices being higher for all females (15.1% 6 3.4%) compared to only adults (7.5% 6 2.1%). Overlap indices followed no trend of either being consistently lower or higher compared to percent overlap (U ¼ 1,903, n 1 ¼ 64, n 2 ¼ 64, P ¼ 0.59), suggesting that percent overlap closely approximated intensity of use in these overlap zones as well. Correlations in utilization TABLE 2. Spatial distribution of male and female raccoon home ranges in Manitoba, R is the Clark Evans ratio (Clark and Evans 1954), which is an index of dispersion. A significant value. 1 indicates a uniform or regular pattern, and a value, 1 indicates a clumped distribution. A value not different from 1 indicates a random distribution. Year Season Female Male a Male b N R P N R P N R P 2002 Spring summer , ¼ , Spring summer , ¼ ¼ Spring summer ¼ ¼ ,0.001 a Male dyads treated independently. b Male dyads merged.

5 650 JOURNAL OF MAMMALOGY Vol. 89, No. 3 TABLE 3. Intrasexual percent overlap, overlap indices, and correlation in utilization distributions (UDs) for 105 estimated home ranges of raccoons in Manitoba, Correlations in UD calculations are based on Spearman s rank correlation and all values for males are calculated excluding values for dyads. Males Females All males Within dyads Among dyads All females Adult females Year X SE X SE X SE X SE X SE 2002 % overlap Overlap indices Correlation in UD þ % overlap Overlap indices Correlation in UD þ % overlap Overlap indices Correlation in UD þ distributions were typically negative among females, with mean correlation values ranging from 0.44 to 0.53 for all females (Table 3). Similarly, when considering only adult females, values were lower, ranging from 0.51 to 0.67 (Table 3). These negative values indicate repulsion or dissimilar use in shared utilization distributions. Percent overlap for home ranges of males was 19.1% 6 3.7%, and ranged annually from 18.1% to 22.1% when not accounting for coalition groups. Males formed distinct spatial groups (Fig. 1) with overlap within these spatial groups being high, ranging from 81.6% to 95.6% annually. Home-range overlap among male coalition groups was low, with percent overlap ranging from 0% to 3.2% (Table 3). Overlap indices were similar to values for percent overlap when calculated both among and within male coalition groups (U ¼ 3,437, n 1 ¼ 83, n 2 ¼ 83, P ¼ 0.68), suggesting that percent overlap was a close approximation of how intensively these shared areas were used. Spearman s rank correlation values were significantly positive for all-male coalition groups, with annual values ranging from 0.77 to 0.88 (Table 3), which indicated that utilization distributions were used similarly. Correspondingly, significantly negative values were recorded for among coalition-group tests (range 0.79 to 0.84). All values were significant at P, Dynamic interaction. All tests for dynamic interactions among males were positive (Fig. 2) and we observed no negative interactions among males. We identified 9 unique pairs of males (3 in 2002, 7 in 2003, and 5 in 2004) that were positively associated with each other and formed coalition groups, with group size never exceeding 2 individuals. Once a coalition was formed, males remained grouped during all seasons and were found together at daytime rest sites 94% of the time during spring and summer months and also were found together at winter denning locations. On several instances we also observed paired males at daytime rest locations during the breeding season in the presence of females. Males that were associated with coalition groups had larger home ranges ( km 2 ) than those that were not ( km 2 ; F ¼ 12.3, d.f. ¼ 1, 47, P, 0.001). Home ranges of male coalition groups overlapped home ranges of multiple females, whereas home ranges of females typically showed extensive overlap (X ¼ 80.3% 6 5.0%, range %) and had nearly congruent external boundaries with the home range of a primary male coalition group (e.g., Fig. 1). Overlap of home ranges of females with nonprimary coalition groups was negligible (X ¼ 6.4% 6 1.8%). Lone adult males had few known females FIG. 1. Home ranges of adult male (n ¼ 12) and female (n ¼ 17) raccoons in Manitoba, 2004, with georeferenced aerial photo in background. Polygons with dark lines (no fill) represent home ranges of males and shaded polygons represent home ranges of females. Dark rectangular outline represents trapping grid.

6 June 2008 PITT ET AL. SOCIAL ORGANIZATION OF RACCOONS 651 FIG. 2. Example of a test for dynamic interaction between 2 adult male raccoons in Manitoba during the summer of In this case, a positive interaction is indicated by the observed cumulative probability being greater than the expected cumulative probability (P, 0.001). present within their home ranges and coalition groups secured access to more females (X ¼ ) than did lone males (X ¼ ; U ¼ 16, n 1 ¼ 30, n 2 ¼ 9, P ¼ 0.001). We only observed 1 instance of a positive interaction among female female dynamic interaction tests. This was a case in which a yearling female was still traveling with her presumed mother during the year after parturition. DISCUSSION Home ranges of both males and females were among the largest reported in the literature, which is consistent with the notion that resources are sparsely distributed at the northern edge of the distribution of raccoons (Fritzell 1978a). Home ranges of males were approximately 4 times larger than home ranges of females, which is approximately the same order of magnitude as was found for southern populations of raccoons (Gehrt and Fritzell 1997). The observed discrepancy between the sexes in home-range size is likely a function of attempts by male raccoons to encompass multiple females and represents different fitness pressures between sexes (Gehrt and Fritzell 1997). Spacing patterns of females. Home ranges of females tended toward a regular distribution that differed considerably from that of females in southern populations, which followed a clumped distribution with localized high densities centered around water resources (Gehrt and Fritzell 1998). For northern populations, it has been hypothesized that females in particular may have a strong reliance on farmsteads for both food and maternity dens (Fritzell 1978b; Larivière 2004). However, farmsteads in our study site were randomly distributed (R ¼ 0.99, P ¼ 0.91), suggesting that this resource is not the sole dictator of the dispersion of females in the landscape. Other mechanisms, such as other resources or social tolerances likely play a role in spatial patterns of females. Most studies of spacing behavior and social interactions of raccoons have found that females are typically gregarious and socially tolerant of conspecifics (Gehrt and Fritzell 1998; Kamler and Gipson 2003; Walker and Sunquist 1997). However, no studies have examined in close detail the spatial organization of females in low-density portions of their distribution. In our study, adult females maintained home ranges that were relatively exclusive of other adult females. Additionally, we found that when areas were shared, the areas were not used similarly by the females involved, with Spearman s rank values indicating repulsion in most cases. However, when we tested for temporal segregation within these areas, we found no evidence that it had occurred. This may be because the shared areas were used minimally, so there would be no reason for females to temporally segregate. Spacing patterns of males. Sociality is rare among carnivores, with only 10 15% of the 271 species in the order Carnivora forming groups outside of the mating season (Gittleman 1989; Valenzuela and Macdonald 2002). Moreover, few of these spatial groups are composed of adult males. For instance, in the family Procyonidae, coatis (genus Nasua) are known to form bands that may exceed 30 individuals (Gompper et al. 1997); however, these bands generally consist of many females and yearling males with typically only 1 adult breeding male. When male coalition groups do form, they generally form relative to spatial aggregations of females (Caro 1994; Gehrt and Fritzell 1998). In studies of male coalition groups in felid social systems, females were either temporarily or permanently aggregated and shared extensively overlapping home ranges (Caro 1989). Previous studies of northern populations of raccoons in North Dakota reported that males were territorial and remained asocial (Fritzell 1978a). In our study, we found that most adult males associated positively with conspecifics to form coalition groups. Members of male coalitions were positively associated with each other during both nighttime activity and at daytime resting sites and maintained territories exclusive of other adult males (both coalitions and singletons). Group association endured throughout the year, including during the breeding season. These findings are contradictory to the suggestion that male raccoons are asocial in low-density portions of their distribution (Fritzell 1978a; Zeveloff 2002). Coalition groups formed despite the fact that females were not aggregated in the landscape, suggesting that aggregation of females is not a prerequisite for the formation of male coalition groups in raccoons or carnivores in general, as previously hypothesized (Caro 1989; Gehrt and Fritzell 1998). Thus, the mechanisms and benefits accrued by the formation of coalition groups in raccoons remain unknown. Both lone males and coalition groups were able to gain access to females in our study, which was similar to results found in Mississippi (Chamberlain and Leopold 2002). However, this was not the case for raccoons in Texas, where lone males were unable to secure access to females (Gehrt and Fritzell 1998). In our study, males in coalition groups were able to secure access to more females per male than were lone males, even when accounting for the fact that multiple males

7 652 JOURNAL OF MAMMALOGY Vol. 89, No. 3 were associated with females residing within home ranges of coalition groups. Additionally, the number of females that males were able to secure access to within coalition groups ranged from 1 to 6 females. Home-range boundaries for male coalitions were highly associated with those of home ranges of females that they encompassed, and home ranges of females typically had minimal overlap with other, nonprimary coalitions. In avian systems, cooperative behavior among groups of males has been studied more extensively and 3 distinct systems have been described with dominant subordinate hierarchies (Krakauer 2005; Lank et al. 2002). The main differences among these systems are the age at group formation (through attrition or developed during adulthood), whether or not male coalitions actively defend territories, and the ability of group members to breed. The spacing system we described for raccoons resembles coalitions of Chiroxiphia manakins, where it is hypothesized that subordinate males benefit through increased likelihood of future inheritance of the occupied territory (McDonald and Potts 1994). In the manakins, group formation occurs after adulthood is reached, males are territorial, and subordinate males did not father offspring (McDonald and Potts 1994). For male raccoons, group formation also occurred after adulthood was reached, as opposed to through attrition. There were several instances when after the death of 1 male coalition member, the remaining male formed a new coalition with another adult conspecific in the study site. Additionally, we observed instances where the presumably dominant (older) male was eventually overthrown, left the territory, and became solitary. In all cases, the remaining male formed a new coalition in the same territory with another peripheral male that was marked as a yearling the previous year. In contrast, in systems where group formation occurs through attrition, once a member of a coalition is lost the remaining male remains solitary (Krakauer 2005). Also, in accordance with the increased likelihood of territory inheritance hypothesis, coalition groups were territorial among other coalitions, forming distinct spatial groups. However, we were unable to assess whether subordinates in male coalitions of raccoons were able to breed, or to assess paternity of offspring, which makes it difficult to assess direct or indirect fitness benefits of group formation. Nevertheless, we propose that in our study, group formation was likely due to the benefits that dominant males received through increased efficiency in territory maintenance and the increased likelihood of territory inheritance by subordinate males. ACKNOWLEDGMENTS We thank S. Cherry, F. Dillemuth, S. Pitt, T. Quirk, R. Tipton, L. Towns, N. Wiebe, K. Wlock, and R. Walrath for assistance in the field. Comments from S. Pitt and L. Pitt helped improve this manuscript. JAP was supported by an excellence scholarship through the Delta Waterfowl Foundation. This study was funded by the Delta Waterfowl Foundation, the University of Saskatchewan, and a discovery grant from the National Science and Engineering Research Council. LITERATURE CITED ADAMS, E. S Approaches to the study of territory size and shape. Annual Review of Ecology and Systematics 32: CARO, T. M Determinants of sociality asociality in felids. Pp in Comparative socioecology (V. Standen and R. A. Foley, eds.). Blackwell Scientific, Cambridge, Massachusetts. CARO, T. M Cheetahs of the Serengeti Plains: group living in an asocial species. University of Chicago Press, Chicago, Illinois. CHAMBERLAIN, M. J., AND B. LEOPOLD Spatio-temporal relationships among adult raccoons (Procyon lotor) in central Mississippi. American Midland Naturalist 147: CLARK, P. J., AND F. EVANS Distance to nearest neighbor as a measure of spatial relationships in populations. Ecology 35: DAVIES, N. B., AND A. I. HOUSTON Territory economics. Pp in Behavioural ecology: an evolutionary approach (J. R. Krebs and N. B. Davies, eds.). Blackwell Scientific, Oxford, United Kingdom. DELUANE, M Xtools extension to ArcView, version 9/15/2003. Department of Forestry, Portland, Oregon. DONCASTER, C. P Non-parametric estimates of interaction from radio-tracking data. Journal of Theoretical Biology 143: DONNLEY, K Simulations to determine the variance and edge effect of total nearest neighbor distance. Pp in Simulation methods in archaeology (I. Hodder, ed.). Cambridge University Press, London, United Kingdom. DURANT, S. M., M. KELLY, AND T. M. CARO Factors affecting life and death in Serengeti cheetahs: environment, age, and sociality. Behavioral Ecology 15: FRITZELL, E. K. 1978a. Aspects of raccoon (Procyon lotor) social organization. Canadian Journal of Zoology 56: FRITZELL, E. K. 1978b. Habitat use of prairie raccoons during the waterfowl breeding season. Journal of Wildlife Management 42: GANNON, W. L., R. S. SIKES, AND THE ANIMAL CARE AND USE COMMITTEE OF THE AMERICAN SOCIETY OF MAMMALOGISTS Guidelines of the American Society of Mammalogists for the use of wild mammals in research. Journal of Mammalogy 88: GEHRT, S. D., AND E. K. FRITZELL Sexual differences in home ranges of raccoons. Journal of Mammalogy 78: GEHRT, S. D., AND E. K. FRITZELL Resource distribution, female home range dispersion and male spatial interactions: group structure in a solitary carnivore. Animal Behaviour 55: GINSBERG, J. R., AND T. P. YOUNG Measuring association between individuals or groups in behavioral studies. Animal Behaviour 44: GITTLEMAN, J. L Carnivore group living: comparative trends. Pp in Carnivore behavior, ecology, and evolution (J. L. Gittleman, ed.). Cornell University Press, Ithaca, New York. GOMPPER, M. E., J. L. GITTLEMAN, AND R. K. WAYNE Genetic relatedness, coalitions, and social behavior of white-nosed coatis, Nasua narica. Animal Behaviour 53: GRAU, G. A., G. C. SANDERSON, AND J. P. ROGERS Age determination of raccoons. Journal of Wildlife Management 34: GREENWOOD, R. J Nocturnal activity and foraging of prairie raccoons (Procyon lotor) in North Dakota. American Midland Naturalist 107: HOOGE, P. N., AND B. EICHENLAUB Animal Movement extension to ArcView, version 1.1. Alaska Biological Science Center, United States Geological Survey, Anchorage. KAMLER, J. F., AND P. S. GIPSON Space and habitat use by male and female raccoons, Procyon lotor, in Kansas. Canadian Field- Naturalist 117:

8 June 2008 PITT ET AL. SOCIAL ORGANIZATION OF RACCOONS 653 KRAKAUER, A Kin selection and cooperative courtship in wild turkeys. Nature 434: KREBS, J. R., AND N. B. DAVIES An introduction to behavioural ecology. 3rd ed. Blackwell Scientific, Oxford, United Kingdom. LANK, D. B., C. M. SMITH, O. HANOTTE, A. OHTONEN, S. BAILEY, AND T. BURKE High frequency of polyandry in a lek mating system. Behavioral Ecology 13: LARIVIÈRE, S Range expansion of raccoons in the Canadian prairies. Wildlife Society Bulletin 32: LARIVIÈRE, S., AND F. MESSIER Spatial organization of a prairie striped skunk population during the waterfowl nesting season. Journal of Wildlife Management 62: LEE, J. E., G. C. WHITE, R.A.GARROTT, R.M.BARMANN, AND A. W. ALLREDGE Assessing the accuracy of a radio telemetry system for estimating mule deer locations. Journal of Wildlife Management 49: LÓPEZ-SEPULCRE, A., AND H. KOKKO Territorial defense, territory size, and population regulation. American Naturalist 166: LOTT, D Intraspecific variation in the social systems of wild vertebrates. Cambridge University Press, New York. LOTZE, J. H., AND S. ANDERSON Procyon lotor. Mammalian Species 119:1 8. MACDONALD, D. W The ecology of carnivore social behavior. Nature 301: MCDONALD, D. B., AND W. K. POTTS Cooperative display and relatedness among males in a lek-mating bird. Science 266: MORRELL, L. J., AND H. KOKKO Can too strong female choice deteriorate male ornamentation? Proceedings of the Royal Society of London, B. Biological Sciences 271: NAMS, V. O Locate II user s guide. Pacer Computer Software, Truro, Nova Scotia, Canada. ODUM, E. P., AND E. J. KUENZLER Measurement of territory and home range size in birds. Auk 72: PITT, J. A., S. LARIVIÈRE, AND F. MESSIER. 2006a. Efficacy of ZoletilÒ for field immobilization of raccoons. Wildlife Society Bulletin 34: PITT, J. A., S. LARIVIÈRE, AND F. MESSIER. 2006b. Condition indices and bioelectrical impedance analysis to predict body condition of small carnivores. Journal of Mammalogy 87: SALVATORI, V., A. K. SKIDMORE, F. CORSI, AND F. VAN DER MEER Estimating temporal independence of radio-telemetry data on animal activity. Journal of Theoretical Biology 198: SAMSON, C., AND J. HUOT Spatial and temporal interaction between female American black bears in mixed forests in eastern Quebec. Canadian Journal of Zoology 79: SHIER, D. M., AND J. A. RANDALL Spacing as a predictor of social organization in kangaroo rats (Dipodomys heermanni arenae). Journal of Mammalogy 85: SINCLAIR, D. F On tests of spatial randomness using mean nearest neighbor distance. Ecology 66: SPENCER, H., L. GODFREY, AND P. O CONNOR A remotely switched passive null-peak network for animal tracking and radio direction finding. Australian Wildlife Research 14: VALENZUELA, D., AND D. W. MACDONALD Home-range use by white-nosed coatis (Nasua narica): limited water and a test of the resource dispersion hypothesis. Journal of Zoology (London) 258: WALKER, S., AND M. SUNQUIST Movement and spatial organization of raccoons in north central Florida. Florida Field- Naturalist 25: WHITE, G. C., AND R. A. GARROTT Analysis of wildlife radiotracking data. Academic Press, San Diego, California. ZEVELOFF, S. I Raccoons: a natural history. Smithsonian Institution, Washington, D.C. Submitted 28 July Accepted 8 October Associate Editor was Paul T. Stapp.