SEXUAL DIMORPHISM, SOCIAL BEHAVIOR, AND INTRASEXUAL COMPETITION IN LARGE PLEISTOCENE CARNIVORANS

Journal of Vertebrate Paleontology 22(1):164–169, March 2002 q 2002 by the Society of Vertebrate Paleontology

SEXUAL DIMORPHISM, SOCIAL BEHAVIOR, AND INTRASEXUAL COMPETITION IN LARGE PLEISTOCENE CARNIVORANS BLAIRE VAN VALKENBURGH and TYSON SACCO Department of Organismic Biology, Ecology and Evolution, University of California, 621 Charles Young Drive South, Los Angeles, California 90095-1606, [email protected]

ABSTRACT—Among living carnivorans, the degree of sexual dimorphism in canine tooth size is correlated with breeding system. Monogamous pair-bonding species and those that form multi-male, multi-female groups tend to be less dimorphic than uni-male, group living species. This correspondence between social behavior and dental dimensions suggests that the inference of sexual dimorphism in extinct carnivorans could shed light on their breeding structure. In this paper, we estimate level of sexual dimorphism in skull length, canine tooth size, and lower molar length for two extinct species, the dire wolf, Canis dirus, and the sabertooth cat, Smilodon fatalis. Three methods are employed to estimate sexual dimorphism: extrapolation from coefficients of variation, division of the sample about the mean, and finite mixture analysis. Results indicate that dire wolves were similar to most canids in their low level of sexual dimorphism, suggesting a pair-bonded breeding structure. Smilodon fatalis appears to have been significantly less dimorphic than living or fossil lions and more comparable to solitary living felids in canine and skull size dimorphism. Thus it seems unlikely that S. fatalis had a polygynous breeding structure like lions in which males compete intensely for access to females. Instead, if S. fatalis lived in groups, these would have been composed of a monogamous pair and their offspring from current and perhaps previous years.

INTRODUCTION Marked sexual dimorphism in mammals is often associated with relatively intense competition among males for access to females (Short and Balaban, 1994; Weckerly, 1998). In carnivorans and primates, males tend to be larger overall than females and may have enhanced jaw musculature and enlarged upper canine teeth (Ewer, 1973; Martin et al., 1994). In these groups, strong dimorphism is associated with polygyny and an increased incidence of infanticide relative to monogamous species (Harvey et al., 1978; Kay et al., 1988, Plavcan and van Schaik, 1992; Gittleman and Van Valkenburgh, 1997). A survey of sexual dimorphism in skull and tooth size in 45 species of extant carnivorans (Gittleman and Van Valkenburgh, 1997) found that dimorphism was more pronounced in canine tooth size than either cheek tooth size or skull length. Moreover, felids, mustelids, and procyonids tended to be more dimorphic than canids, viverrids, and hyaenids. After accounting for phylogenetic effects, Gittleman and Van Valkenburgh (1997) found that the extent of canine dimorphism was not related to diet, body size, activity pattern, or habitat. The only factor of relevance appeared to be breeding system. Uni-male, group living species tend to be more dimorphic than either multi-male, multi-female groups or monogamous pair-bonding species. This correspondence between breeding system and canine dimorphism in extant carnivorans suggests that the reconstruction of sexual dimorphism of extinct species could provide a window into their social behavior. This has been done for Miocene chalicotheres (Coombs, 1975), Paleocene mesonychids (O’Leary et al., 2000), and a variety of extinct primate taxa including Eocene omomyids (Krishtalka et al., 1990), Oligocene anthropoids (Fleagle et al., 1980), and Pliocene hominids (McHenry, 1991). Here we estimate the extent of sexual dimorphism in the two most abundant species preserved in the late Pleistocene Rancho La Brea asphalt deposits, the dire wolf, Canis dirus, and the sabertooth cat, Smilodon fatalis. The Rancho La Brea asphalt deposits (Los Angeles, California) represent a predator-trap deposit in which numerous carnivorans and scavenging birds apparently were attracted by the

presence of a dead or dying herbivore that had become mired in the sticky asphalt seeps (Stock, 1992). It is generally agreed that entrapments of herbivores occurred relatively rarely, and that each one attracted a number of predatory birds and mammals, some of which were themselves trapped (Marcus and Berger, 1984; Stock, 1992). Based on the numbers of herbivores preserved in the deposits and the total time span represented, Marcus and Berger (1984) estimated one entrapment event every 50 years. For each herbivore that is preserved at Rancho La Brea, there are nine to ten carnivorans (Stock, 1992). Assuming that these individuals represent only a subset of the total number of carnivorans that fed on a carcass, it seems likely that fairly sizeable groups of C. dirus or S. fatalis fed together on occasion. Alternatively, the carnivorans represented in the tar seeps might represent the accumulation of numerous solitary individuals that visited the carcass in succession. Arguing against this is the fact that solitary large carnivores, such as tigers, are territorial with relatively large home ranges (16 km2– 72 km2; Sandell, 1989). Moreover, given the large numbers of scavenging birds at La Brea, it seems unlikely that carcasses persisted for very long and thus solitary individuals must have followed one another in close temporal succession to produce the numbers preserved in the tar seeps. Thus, the most likely scenario is a predator trap in which groups of carnivorans fed together on a carcass. Among living carnivorans, social feeding is not associated with a single breeding system; hyenas, many canids, and a few felids will tolerate conspecifics at a kill, while their breeding systems range from monogamy to polygyny (Bertram, 1979; Gittleman, 1989). To better understand the social behavior of C. dirus and S. fatalis, we examined the extent of sexual dimorphism in skull and tooth size within one site or ‘‘pit,’’ Pit 13. Pit 13 was chosen because it is relatively fossiliferous and spans a fairly narrow time span (15,360–14,310 ybp), as determined by radiometric dating of four specimens (Marcus and Berger, 1984). The analysis was confined to a single locality to minimize the confounding effects of mixing populations from different times, given that mean body size of both C. dirus and

164

VAN VALKENBURGH AND SACCO—SEXUAL DIMORPHISM IN PLEISTOCENE CARNIVORANS TABLE 1. Sample size (N), mean (in mm), standard deviation (SD), and coefficient of variation (CV) for each measure in fossil taxa: skull length (SKL), upper canine anteroposterior diameter (C1AP), upper canine mediolateral diameter (C1ML), lower first molar length (m1L). Species Canis dirus SKL C1AP C1ML m1L Smilodon fatalis SKL C1AP CIML m1L

N

Mean

SD

CV

32 85 84 133

252.1 16.14 11.07 33.66

8.34 1.14 0.94 1.91

3.31 7.04 8.50 5.69

11 43 43 29

299.18 39.02 20.17 28.64

11.86 2.48 1.60 1.61

3.96 6.35 7.91 5.62

S. fatalis appears to vary over time at Rancho La Brea (Menard, 1947; Nigra and Lance, 1947). Estimating the degree of sexual dimorphism represented in a sample of individuals in which sex is unknown is not a trivial problem. If the sexes differ greatly in size, then it is easier because the sample may show a bimodal distribution. However, significant sexual dimorphism can be hidden within a unimodal distribution if there is substantial size overlap between males and females (Godfrey et al., 1993). Thus, the absence of a bimodal distribution is not definitive evidence of monomorphism. Fortunately, there are alternative approaches that are better at revealing weak to moderate dimorphism and these are used here. These rely on estimates of sample variation, sample mean, and sample range (Plavcan, 1994). Although the accuracy of each is reduced by strongly biased sex ratios and high intrasexual variability, Plavcan (1994) demonstrated that they could be used to set an upper bound on sexual dimorphism with confidence. MATERIALS AND METHODS All available complete and partial skulls and dentaries of Canis dirus and Smilodon fatalis from Pit 13 were included in the sample. All specimens are housed at the George C. Page Museum (Natural History Museum of Los Angeles County, California, USA). Using digital calipers, measurements were taken (to the nearest 0.01 cm) of skull length, upper (permanent) canine width and length, and lower carnassial (first molar) length. A total of 338 C. dirus and 111 S. fatalis specimens were measured, although sample sizes varied for each measurement (Table 1). The dimorphism estimates obtained for the two extinct species were compared with those of seven extant canid species and eleven extant felid species. Most of the raw data for the extant species are taken from Gittleman and Van Valkenburgh (1997). However, four species that were not examined in that paper have been added here: African wild dog (Lycaon pictus), puma (Puma concolor), jaguar (Panthera onca) and cheetah (Acinonyx jubatus). The measured skulls are housed in the mammalogy collections of the United States National Museum, the Natural History Museum of Los Angeles County, the Field Museum of Chicago, and the National Museums of Kenya, Nairobi. Sample sizes for each of the four species range from 12 to 26, split evenly or nearly so, between the sexes. The morphological variables discussed in this paper are described below and follow those used in Gittleman and Van Valkenburgh (1997). As in that study, we also took additional measurements such as the size of the lower canine and upper carnassial. The extent of dimorphism in both the upper and lower carnassial teeth was similar and so we present data only for the latter. Lower canine teeth were too poorly represented (n 5 6)

165

for S. fatalis to calculate a reasonable estimate of dimorphism and so that measure was excluded from the analysis. In the interests of brevity, we do not present the data for lower canine tooth dimorphism in C. dirus; it was similar to that of the upper canine. 1. Upper canine size: mediolateral breadth and anteroposterior length of upper canines were measured at the dentineenamel junction. Although it would have been desirable to also estimate dimorphism in crown height, this was not possible due to a paucity of unbroken canines in the samples. 2. Lower carnassial (first molar) size: maximum anteroposterior length of the lower first molar. The lower carnassial of carnivorans is typically the largest cheek tooth, and functions to cut through tough substances, such as skin (Van Valkenburgh, 1996). 3. Skull length: condylobasal length was measured parallel to the midsagittal plane from the caudal surface of the occipital condyles to the anterior tip of the premaxilla. Sexual dimorphism was calculated as the ratio of male mean to female mean for a given measure. For the Pleistocene species, sexual dimorphism was estimated using three methods. The first approach is based on the positive relationship between the extent of sexual dimorphism and overall sample variance. If males and females differ greatly in size, then the coefficient of variation ([standard deviation/mean] 3 100) of the combined sex sample is likely to be larger than when the sexes differ little in size (Fleagle et al., 1980; Kay, 1982a, b). Given this, a regression of sexual dimorphism on coefficients of variation (CV) for extant species was used to predict sexual dimorphism in the two Rancholabrean species. The second approach, called the ‘‘mean method’’ by Plavcan (1994) simply divides the total sample into two sub-samples about the mean. The ratio between the means of each of the subsamples is then used as estimated sexual dimorphism. Using known sex samples, this technique has proven to be reasonably accurate, especially when dimorphism is pronounced (Godfrey et al., 1993; Plavcan, 1994). The third approach taken here is finite mixture analysis (FMA). In this approach, the maximum separation of male and female mean values that could be contained within a unimodal distribution is estimated. It relies on the fact that within a given sample range, there is an expected number of standard deviations that will occur. Sample size affects this value, and Pearson (1932) provides a table of expected numbers of standard deviations according to sample size. This can be used to estimate the maximum variance allowable within two subsamples, and then to calculate the maximum possible dimorphism represented within the sample (Godfrey et al., 1993). When applied to extant species, FMA was found to yield sexual dimorphism values closer to actual values than those derived using other techniques when three conditions were met: (1) skulls exhibit no obvious indicators of sex such as horns; (2) distribution of skull lengths of species sampled at single localities lack visible bimodality, and (3) samples from single localities contain at least 10 individuals (Godfrey et al., 1993). However, in a simulation study, Plavcan (1994) found that FMA did not perform as well as the mean method when dimorphism was minimal and the sex ratio was balanced. RESULTS For the sample of extant felids and canids, the variance in upper canine dimensions in the mixed sex sample showed the best correlation with level of sexual dimorphism (Table 2, Fig. 1). The correlations with skull length (r2 5 0.477) and lower molar length (r2 5 0.331) were weaker but significant (Table 2). Regressions obtained using all 18 extant taxa were used to predict the level of sexual dimorphism for each variable in the

166

JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 22, NO. 1, 2002

TABLE 2. Statistics for the regressions of sexual dimorphism index on coefficient of variation (CV) for the four measures and using the sample of 18 extant species of felids and canids. Shown are the yintercepts, slopes, coefficients of determination (r2), and P values. For C1AP, the statistics shown are for a regression run without an outlier, the arctic for (Alopex lagopus). C1AP values are similar in male and female arctic foxes, but the mixed sex sample exhibits a moderate CV for the measure. For abbreviations see Table 1. Species included are listed in Table 4. Measurement

Intercept

Slope

r2

P

SKL C1AP C1ML m1L

0.988 0.872 0.898 1.01

0.015 0.029 0.024 0.008

0.477 0.630 0.743 0.331

,0.01 ,0.01 ,0.001 ,0.05

two Pleistocene species. In addition, regressions were run separately for the two families, and each was used to predict dimorphism in the appropriate extinct species. That is, canid regression equations were applied to C. dirus, and felid equations to S. fatalis. Dimorphism estimates obtained using the separate family regressions were very similar to those obtained using the combined felid and canid sample and are not presented here. Notably, all of the CV regressions provided values that differed by no more than eight percent from those produced using the mean method and FMA (Table 3). The highest estimates tended to be those made using the mean method and the lowest using the CV regression method. The dire wolf and S. fatalis appear to have exhibited similar levels and patterns of sexual dimorphism in their skulls and teeth (Table 3). Across both species, the range of dimorphism values for all four measures is not large, ranging from 1.04 (C. dirus, skull length) to 1.15 (C. dirus, mediolateral breadth of C1). Moreover, the level of dimorphism in both species was greatest in the canine teeth (1.06–1.15) and less in both skull length (1.04–1.07) and lower carnassial length (1.05–1.10). This was also true within the extant sample of canids and felids (Table 4). Comparison of estimated dimorphism values for the extinct

TABLE 3. Estimated sexual dimorphism (male mean/female mean) in Canis dirus and Smilodon fatalis for four measures using three different methods: mean method (MM), finite mixture analysis (FMA), and coefficient of variation regression (CV). The average of the three estimates is also shown (AVERAGE). Measurement abbreviations as defined in Table 1. Species Canis dirus SKL C1AP C1ML m1L Smilodon fatalis SKL C1AP C1ML m1L

MM

FMA

CV

Average

1.05 1.12 1.15 1.09

1.04 1.11 1.12 1.08

1.04 1.08 1.09 1.06

1.05 1.10 1.12 1.08

1.07 1.12 1.13 1.10

1.06 1.09 1.13 1.07

1.05 1.06 1.09 1.05

1.06 1.09 1.12 1.07

species with their living relatives reveals that C. dirus might have been somewhat more dimorphic than extant canids, and S. fatalis somewhat less than extant felids (Table 4). The upper canine teeth of male C. dirus are estimated to have been about 11–12 percent larger than those of females, whereas the average difference between the sexes in the extant canid sample was only six percent. Both the mean method and FMA are known to overestimate dimorphism when the level of dimorphism is less than or equal to 1.10, and thus the estimates for the extinct species must be considered as an upper limit for the true dimorphism value (Plavcan, 1994). Consequently, it is difficult to know whether the dire wolf was significantly more dimorphic than its closest living analog among canids, the gray wolf, which exhibits a maximum dimorphism level of 1.08 for upper canine length in our sample. Given that the methods tend to overestimate dimorphism, S. fatalis does appear to have been slightly less dimorphic in skull length and canine dimensions than a typical felid (Table 4). It certainly seems to have been significantly less dimorphic in its upper canine teeth than the most social of felids, the lion, and

FIGURE 1. Least squares linear regression of sexual dimorphism for upper canine tooth mediolateral diameter (C1ML) against coefficient of variation for the same measure for 11 extant felids (1–11) and 7 extant canids (12–18). For regression statistics, see Table 2. 1, Caracal caracal; 2, Acinonyx jubatus; 3, Puma concolor; 4, Felis silvestris; 5, Lynx rufus; 6, Felis chaus; 7, Panthera tigris; 8, Panthera onca; 9, Leptailurus serval; 10, Panthera leo; 11, Panthera pardus; 12, Urocyon cinereoargentus; 13, Canis lupus; 14, Lycaon pictus; 15, Alopex lagopus; 16, Canis latrans; 17, Vulpes vulpes; 18, Vulpes velox.

VAN VALKENBURGH AND SACCO—SEXUAL DIMORPHISM IN PLEISTOCENE CARNIVORANS TABLE 4. Levels of sexual dimorphism in skull length, canine dimensions and lower molar length in extant canids and felids compared with the average values of dimorphism predicted for Canis dirus and Smilodon fatalis. Data for extant species are from Gittleman and Van Valkenburgh (1997) except for those species marked with an asterisk (*) which were measured for this paper. For abbreviations, see Table 1. Species

SKL

Canis lupus, gray wolf C. latrans, coyote Vulpes velox, swift fox V. vulpes, red fox Alopex lagopus, arctic fox Urocyon cinereoargenteus, gray fox *Lycaon pictus, African wild dog Extant canid sample mean Canis dirus Panthera leo, lion P. tigris, tiger P. pardus, leopard *P. onca, jaguar *Puma concolor, puma Caracal caracal, caracal Lynx rufus, bobcat Felis silvestris, wild cat F. chaus, jungle cat Leptailurus serval, serval *Acinonyx jubatus, cheetah Extant felid sample mean Smilodon fatalis

1.04 1.01 1.01 1.10 1.00 1.04 1.01 1.03 1.05 1.12 1.16 1.13 1.06 1.08 1.08 1.10 1.10 1.07 1.08 1.10 1.10 1.06

C1AP C1ML 1.06 1.07 1.13 1.03 1.00 1.05 1.09 1.06 1.10 1.25 1.16 1.24 1.12 1.14 1.06 1.16 1.15 1.16 1.10 1.15 1.15 1.09

1.08 1.01 1.16 1.08 1.01 1.02 1.04 1.06 1.12 1.23 1.08 1.26 1.11 1.09 1.03 1.11 1.13 1.15 1.14 1.11 1.13 1.12

m1L 1.05 1.00 1.01 1.03 1.08 1.02 1.05 1.03 1.08 1.13 1.12 1.12 1.06 1.04 1.08 1.06 1.08 1.10 0.97 1.03 1.07 1.07

the solitary leopard, where males of both species display canines that are about 25 percent larger than those of females. Its canine dimorphism level is similar to that of a number of solitary felid species, including the tiger, jaguar, puma, and wild cat. The predicted level of dimorphism in skull length for S. fatalis is also low relative to the mean value and range for the 11 sampled felids (Table 4). Carnassial dimorphism is similar in the extinct cat and the living felids. DISCUSSION The data on living felids and canids presented here and in Gittleman and Van Valkenburgh (1997) indicate that level of sexual dimorphism is related to breeding system. Canids are generally less dimorphic than felids (Fig. 1). Canids tend to live as monogamous pairs, along with their offspring in some instances (e.g., wolves, dholes, jackals) (Kleiman and Eisenberg, 1973; Gittleman, 1989; Geffen et al., 1996). Notably, the monogamy that is assumed in several pair-bonded canid species has been found to be false; extra-pair copulations are not uncommon (Lehman et al., 1992; Sillero-Zubiri et al., 1996; Gompper and Wayne, 1996). Nevertheless, battles between males for access to females are observed relatively rarely. Canids vary little in levels of sexual dimorphism, and there is no clear association between dimorphism and group size (e.g., pack-living vs. mated pair vs. solitary). Felids tend to be solitary, with the sexes associating only briefly during the mating period (Kleiman and Eisenberg, 1973; Sandell, 1989; Kitchener, 1991). Felids are polygynous, with males defending large territories that encompass several smaller female territories (Sandell, 1989; Kitchener, 1991). Territories are actively marked and patrolled by single dominant males that exclude rival males by force if necessary. Among felids, the lion is the most social and one of the most dimorphic. In addition to having a mane and larger body size, male lions on average have upper canine teeth that are about 25 percent larger than those of females (Table 4; Smuts et al., 1978). Lions compete intensely for access to females; males that are not members of a pride often form coalitions of two or three that attempt to

167

take over existing prides (Packer and Pusey, 1982). This high level of intrasexual competition takes its toll, resulting in shorter average life spans for male lions (Schaller, 1972). However, the leopard also is very dimorphic in canine dimensions and yet has the typical felid breeding system described above. Thus, enhanced dimorphism cannot be used to predict a lion-like breeding structure in an extinct felid with confidence, but reduced dimorphism strongly suggests low levels of male–male competition relative to lions. Based on the sample from Pit 13, C. dirus and S. fatalis exhibit similar levels of dimorphism. In the case of C. dirus, it appears to have been similar to gray wolves in skull and carnassial size dimorphism, but slightly more dimorphic than gray wolves in canine dimorphism. However, the difference in canine dimorphism could be an artifact of the estimation techniques as both the mean method and FMA tend to overestimate dimorphism slightly when intersex differences are small (less than 10 percent) (Plavcan, 1994). As noted above, dimorphism levels vary little among canids despite differences in group size, and thus the fact that C. dirus had typical canid levels of dimorphism tells us only that they lived as monogamous pairs. If we had found that dire wolves were much more dimorphic than their extant counterparts, then this would have suggested a breeding system distinct from that of extant canids. Are there other lines of evidence that can be used to infer the social system of dire wolves? Yes; the substantial numbers of dire wolves preserved at Rancho La Brea suggest that they lived in larger groups. Moreover, their large body size and highly carnivorous dentition are consistent with a predator that takes relatively large prey (Van Valkenburgh and Koepfli, 1993; Van Valkenburgh and Hertel, 1998). Extant canids cannot use their limbs to grapple with prey; instead they rely on their jaws and must work together to kill ungulates much larger than themselves. Among canids, those that take prey larger than themselves do so in packs that consist of an alpha pair and their offspring from current and previous years (e.g., African wild dog, dhole, gray wolf) (Mech, 1970; Fox, 1984; Girman et al., 1997). Thus, it seems reasonable to assume that C. dirus lived in packs of related individuals led by an alpha pair. Inferring the breeding structure of S. fatalis is more problematic. Its level of dimorphism in both skull length and canine dimensions is much less than that of the lion and similar to that observed among most solitary felids (excluding the aberrant leopard). In fact, it shows very weak dimorphism in skull length (1.06), suggesting that males and females differed little in body size. If S. fatalis lived in prides that regularly experienced takeover battles by rival males, it seems probable that selection would have favored larger body size at least in male S. fatalis. Thus, it seems likely that male–male competition for mates was not as intense in S. fatalis as the lion. The level of dimorphism predicted for S. fatalis is consistent with a typical felid social structure in which both sexes are usually solitary, and males defend a territory that includes females. If S. fatalis had the typical felid breeding structure, then how can we explain the large numbers of individuals preserved in the Rancho La Brea asphalt deposits? As stated in the introduction, the prevalence of large carnivorans at Rancho La Brea is consistent with a predator trap scenario in which numerous individuals were lured to their death by a single dying or dead herbivore. Among living felids, it has been observed in a few species (tiger, puma, bobcat, lynx) that when food is plentiful and the habitat is fairly open, typically solitary females in neighboring territories will share kills, although not peacefully (Kitchener, 1991; Caro, 1994). In some instances, the resident male will join them as well. For example, in Rathambhore Park, India, where visibility is good, multiple tigers may be attracted to kills by circling vultures and will reluctantly share their meal (Thapar, 1986). The deposits at Rancho La Brea have yielded

168

JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 22, NO. 1, 2002

numerous scavenging birds, such as vultures and condors, and formed in a fairly open setting (Howard, 1962; Stock, 1992). It is likely that carnivorans, such as S. fatalis, observed them and congregated at carcasses. Among felids, females have been observed to feed together more often and in larger numbers than males, whether they exist as solitary individuals or as members of a pride or coalition (Kitchener, 1991). Consequently, it might be expected that the felid samples at Rancho La Brea would be biased towards females. This in turn might lead one to question our estimates of sexual dimorphism; perhaps the female bias results in an inaccurate and much reduced estimate of actual dimorphism. However, the same problem should exist for the Rancholabrean American lion (Panthera atrox) and yet marked sexual dimorphism is apparent in this species at Rancho La Brea despite a smaller sample size than is available for S. fatalis (G. Jefferson, pers. comm.). The methods we used set an upper bound on dimorphism and have been demonstrated to be robust to deviations from a 50/50 sex ratio as great as 4:1 (Plavcan, 1994). Consequently, it seems unlikely that dimorphism at the level of extant lions would be completely obscured by sampling biases in Rancho La Brea S. fatalis. It is of course possible that S. fatalis displayed a social structure and breeding system that differed from that of any extant felid. After all, the evolutionary divergence between the extant felids and the branch leading to S. fatalis is ancient, occurring at least 17–20 Ma (Turner and Anton, 1997). Moreover, S. fatalis was a different sort of felid, a sabertooth, with a rather specialized dentition. Given the low level of dimorphism, it may be that S. fatalis was more canid-like in its social structure, existing as monogamous pairs and their offspring from the current year as well as perhaps previous years. There are several reasons why some sort of group living might have been favored in S. fatalis. First, because of their bizarre dentition, juvenile S. fatalis may have required a prolonged period of parental care and delayed weaning until the adult dentition was functional. Among living carnivorans, spotted and brown hyenas, both of which exhibit massive teeth and skulls, have unusually late weaning times that coincide with the appearance of the adult dentition (Binder and Van Valkenburgh, 2000). The presence of helpers (non-reproductive group members) in S. fatalis could have promoted cub survivorship through guarding and food provisioning during the prolonged period of juvenile dependence. In addition, groups of S. fatalis would have been better able to defend or steal carcasses from the putatively social species with which they coexisted, American lions and dire wolves. There is some reason to believe that interspecific encounters over carcasses were frequent in the late Pleistocene. The habitat at Rancho La Brea was fairly open in the late Pleistocene, facilitating the discovery of kills (Stock, 1992). Moreover, a high incidence of teeth broken in life among Rancho La Brea carnivorans suggests that interspecific competition for food was significant (Van Valkenburgh and Hertel, 1993). Today, the greatest diversity of large social carnivoran species occurs in open habitats in Africa where interspecific interactions over kills occur with regularity. There, the lion, spotted hyena, wild dog, and, to a much lesser extent, cheetah live in groups that, within each species, hunt and feed together. Their ability to either defend a kill or steal a kill is enhanced by having greater numbers (Eaton, 1979). Consequently, it seems likely that social groups of S. fatalis would have been favored in the late Pleistocene. Finally, it has been suggested that S. fatalis existed in social groups because healed skeletal injuries are present in some individuals (Shaw, 1992). The injuries are argued to have been so severe as to have impaired hunting success; consequently, survival of the individuals must have required that they be pro-

visioned in some way. While we agree that this does suggest that injured individuals were able to feed on kills made by others, we do not feel it necessarily indicates a lion type of pride system for S. fatalis. Such a system in which males fight vigorously for the opportunity to mate with several females would be expected to result in males that greatly exceed females in body size and canine dimensions. Alternatively, injured relatives might have been allowed to feed if S. fatalis had a monogamous breeding system in which individuals lived in extended family groups, such as can be seen in wolves (Mech, 1970). Unfortunately, we are unlikely ever to be certain of the social structure of S. fatalis or of any extinct species for that matter. As molecular techniques continue to improve, it eventually may be possible to sex individuals from the tar pits and quantify kinship as has been done for living species (Lehman et al., 1992; Girman et al., 1997). These data certainly would help answer questions concerning group structure and social systems. However, it will likely remain difficult to decide between a monogamous canid-like pack structure and a typical solitary felid pattern, in which neighboring (and often related) females occasionally shared kills. Moreover, it is likely that, as is true of many extant social carnivorans, the social system of S. fatalis would have varied according to resource levels, ranging from living in larger groups when resources are plentiful to living as pairs or individuals when resources are limited (Macdonald, 1983; Caro, 1994). However, when in groups, the analysis of sexual dimorphism in S. fatalis does seem to exclude a lion type of pride system and favors a less polygynous, if not monogamous breeding structure. ACKNOWLEDGMENTS We thank J. Harris, C. Shaw and S. Cox at the George C. Page Museum for access to the Rancho La Brea collection. P. Adam, A. Friscia, F. Hertel, K. Koepfli, and C. Shaw provided helpful comments. This study was partially supported by a University of California Academic Senate grant (to BVV) and NSF EAR 98047-42 (to BVV). LITERATURE CITED Bertram, B. C. B. 1979. Serengeti predators and their social systems; pp. 221–248 in A. R. E. Sinclair and M. Norton-Griffiths (eds.), Serengeti, Dynamics of an Ecosystem. University of Chicago Press, Chicago. Binder, W. J., and B. Van Valkenburgh. 2000. Development of bite strength and feeding behaviour in juvenile spotted hyenas (Crocuta crocuta). Journal of Zoology 252:273–283. Caro, T. M. 1994. Cheetahs of the Serengeti Plains: Group Living in an Asocial Species. University of Chicago Press, Chicago, 478 pp. Coombs, M. C. 1975. Sexual dimorphism in chalicotheres (Mammalia, Perissodactyla). Systematic Zoology 24:55–62. Eaton, R. L. 1979. Interference competition among carnivores: a model for the evolution of social behavior. Carnivore 2:9–16. Ewer, R. F. 1973. The Carnivores. Cornell University Press, Ithaca, 494 pp. Fleagle, J. G., R. F. Kay, and E. L. Simons. 1980. Sexual dimorphism in early anthropoids. Nature 287:328–330. Fox, M. W. 1984. The Whistling Hunters: Field Studies of the Asiatic Wild Dog (Cuon alpinus). State University of New York Press, Albany, 150 pp. Geffen, E., M. E. Gompper, J. L. Gittleman, H.-K. Luh, D. W. Macdonald, and R. K. Wayne. 1996. Size, life history traits, and social organization in the Canidae: a reevaluation. American Naturalist 147:140–160. Girman, D. J., M. G. L. Mills, E. Geffen, and R. K. Wayne. 1997. A molecular genetic analysis of social structure, dispersal, and interpack relationships of the African wild dog (Lycaon pictus). Behavioral Ecology and Sociobiology 40:187–198. Gittleman, J. L. 1989. Carnivore group-living: comparative trends; pp.

VAN VALKENBURGH AND SACCO—SEXUAL DIMORPHISM IN PLEISTOCENE CARNIVORANS 183–207 in J. L. Gittleman (ed.), Carnivore Behavior, Ecology, and Evolution. Cornell University Press, Ithaca. ———, and B. Van Valkenburgh. 1997. Sexual dimorphism in the canines and skulls of carnivores: effects of size, phylogeny, and behavioural ecology. Journal of Zoology 242:97–117. Godfrey, L. R., S. K. Lyon, and M. R. Sutherland. 1993. Sexual dimorphism in large-bodied primates: the case of the subfossil lemurs. American Journal of Physical Anthropology 90:315–334. Gompper, M. E., and R. K. Wayne. 1996. Genetic relatedness among individuals within carnivore societies; pp. 429–452 in J. L. Gittleman (ed.), Carnivore Behavior, Ecology, and Evolution, Vol. 2. Cornell University Press, Ithaca. Harvey, P. H., M. Kavanagh, and T. H. Clutton-Brock. 1978. Sexual dimorphism in primate teeth. Journal of Zoology (London) 186: 475–487. Howard, H. 1962. A comparison of avian assemblages from individual pits at Rancho La Brea, California. Contributions in Science 58:1– 24. Kay, R. F. 1982a. Sexual dimorphism in the Ramapithecinae. Proceedings of the National Academy of Science, U.S.A. 79:209–212. ——— 1982b. Sivapithecus simonsi, a new species of Miocene hominoid, with comments on the phylogenetic status of Ramapithecinae. International Journal of Primatology 3:113–173. ———, J. M. Plavcan, G. E. Glander, and P. C. Wright. 1988. Sexual selection and canine dimorphism in New World monkeys. American Journal of Physical Anthropology 77:385–397. Kitchener, A. 1991. The Natural History of the Wild Cats. Comstock, Ithaca, 280 pp. Kleiman, D. G., and J. F. Eisenberg. 1973. Comparisons of canid and felid social systems from an evolutionary perspective. Animal Behaviour 21:637–659. Krishtalka, L., R. K. Stucky, and K. C. Beard. 1990. The earliest fossil evidence for sexual dimorphism in Primates. Proceedings of the National Academy of Science, U.S.A. 87:5223–5226. Lehman, N., P. Clarkson, L. D. Mech, T. J. Meier, and R. K. Wayne. 1992. A study of the genetic relationships within and among wolf packs using DNA fingerprinting and mitochondrial DNA. Behavioral Ecology and Sociobiology 30:83–94. Macdonald, D. W. 1983. The ecology of carnivore social behaviour. Nature 301:379–383. Marcus, L. F., and R. Berger. 1984. The significance of radiocarbon dates for Rancho La Brea; pp. 159–188 in P. S. Martin and R. G. Klein (eds.), Quaternary Extinctions. University of Arizona Press, Tucson. Martin, R. D., L. A. Willner, and A. Dettling. 1994. The evolution of sexual dimorphism in primates; pp. 159–200 in R. V. Short and E. Balaban (eds.), The Differences Between the Sexes. Cambridge University Press, Cambridge. McHenry, H. M. 1991. Sexual dimorphism in Australopithecus afarensis. Journal of Human Evolution 20:21–32. Mech, L. D. 1970. The Wolf: The Ecology and Behavior of an Endangered Species. University of Minnesota Press, Minneapolis, 384 pp. Menard, H. W., Jr. 1947. Analysis of measurements in length of the metapodials of Smilodon. Bulletin of the Southern California Academy of Sciences 46:127–135. Nigra, J. O., and J. F. Lance. 1947. A statistical study of the metapodials of the dire wolf group from the Pleistocene of Rancho La Brea.

169

Bulletin of the Southern California Academy of Sciences 46:26– 35. O’Leary, M. A., S. G. Lucas, and T. E. Williamson. 2000. A news specimen of Ankalagon (Mammalia, Mesonychia) and evidence of sexual dimorphism in mesonychians. Journal of Vertebrate Paleontology 20:387–393. Packer, C., and A. E. Pusey. 1982. Cooperation and competition within coalitions of male lions: kin selection or game theory? Nature 296: 740–742. Pearson, E. 1932. The percentage limits for the distribution of range in samples from a normal population. Biometrika 24:404–417. Plavcan, J. M. 1994. Comparison of four simple methods for estimating sexual dimorphism. American Journal of Physical Anthropology 94:465–476. ———, and C. P. van Schaik. 1992. Intrasexual competition and canine dimorphism in anthropoid primates. American Journal of Physical Anthropology 87:461–477. Sandell, M. 1989. The mating tactics and spacing patterns of solitary carnivores; pp. 164–182 in J. L. Gittleman (ed.), Carnivore Behavior, Ecology, and Evolution. Cornell University Press, Ithaca. Schaller, G. B. 1972. The Serengeti Lion: A Study of Predator-Prey Relations. University of Chicago Press, Chicago, 480 pp. Shaw, C. A. 1992. Old wounds: the paleopathology of Rancho La Brea. Terra 31:17. Short, R. V., and E. Balaban. 1994. The Differences Between the Sexes. University Press, Cambridge, 479 pp. Sillero-Zubiri, C., D. Gottelli, and D. W. Macdonald. 1996. Male philopatry, extra-pack copulations and inbreeding avoidance in Ethiopian wolves (Canis simensis). Behavioral Ecology and Sociobiology 38:331–340. Smuts, G. L., J. L. Anderson, and J. C. Austin. 1978. Age determination of the African lion (Panthera leo). Journal of Vertebrate Zoology 185:115–146. Stock, C. 1992. Rancho La Brea: A Record of Pleistocene Life in California. (7th ed., revised by J. M. Harris). Natural History Museum of Los Angeles County, Science Series, 113 pp. Thapar, V. 1986. Tiger: Portrait of a Predator. Facts on File, New York, 200 pp. Turner, A., and M. Anton. 1997. The Big Cats and their Fossil Relatives. Columbia University Press, New York, 234 pp. Van Valkenburgh, B. 1996. Feeding behavior in free-ranging, large African carnivores. Journal of Mammalogy 77:240–254. ———, and F. Hertel. 1993. Tough times at La Brea: tooth breakage in large carnivores of the late Pleistocene. Science 261:456–459. ———, and ——— 1998. The decline of North American predators during the late Pleistocene; pp. 357–374 in J. J. Saunders, B. W. Styles, and G. F. Baryshnikov (eds.), Quaternary Paleozoology in the Northern Hemisphere. Illinois State Museum Scientific Papers, Springfield. ———, and K.-P. Koepfli. 1993. Cranial and dental adaptations to predation in canids; pp. 15–37 in N. Dunston and J. L. Gorman (eds.), Mammals as Predators. Oxford University Press, Oxford. Weckerly, F. 1998. Sexual-size dimorphism: influence of body mass and mating systems in the most dimorphic mammals. Journal of Mammalogy 79:33–52. Received 15 September 2000; accepted 1 March 2001.