Sexual Dimorphism in Primate Evolution

Transcription

1 YEARBOOK OF PHYSICAL ANTHROPOLOGY 44:25 53 (2001) Sexual Dimorphism in Primate Evolution J. Michael Plavcan Department of Anthropology, University of Arkansas, Fayetteville, Arkansas KEY WORDS sexual dimorphism; sexual selection; canine teeth; primates ABSTRACT Sexual dimorphism is a pervasive phenomenon among anthropoid primates. Comparative analyses over the past 30 years have greatly expanded our understanding of both variation in the expression of dimorphism among primates, and the underlying causes of sexual dimorphism. Dimorphism in body mass and canine tooth size is familiar, as is pelage and sex skin dimorphism. More recent analyses are documenting subtle differences in the pattern of skeletal dimorphism among primates. Comparative analyses have corroborated the sexual selection hypotheses, and have provided a more detailed understanding of the relationship between sexual selection, natural selection, and mating systems in primates. A clearer picture is emerging of the relative contribution of various selective and nonselective mechanisms in the evolution and expression of dimorphism. Most importantly, recent studies have shown that dimorphism is the product of changes in both male and female traits. Developmental studies demonstrate the variety of ontogenetic pathways that can lead to dimorphism, and provide additional insight into the selective mechanisms that influence dimorphism throughout the lifetime of an animal. Evidence from the fossil record suggests that dimorphism probably evolved in parallel twice, and the dimorphism in some extinct hominoids probably exceeded that of any living primate. Our advances in understanding the behavioral/ecological correlates of dimorphism in living primates have not improved our ability to reconstruct social systems in extinct species on the basis of dimorphism alone, beyond the inference of polygyny or intense male-male competition. However, our understanding of the behavioral/ecological correlates of growth and development, and of the expression of dimorphism as a function of separate changes in male and female traits, offers great potential for inferring evolutionary changes in behavior over time.yrbk Phys Anthropol 44:25 53, Wiley-Liss, Inc. TABLE OF CONTENTS Variation in Dimorphism in Primates Definitions Primary sex differences Secondary sex differences Body mass dimorphism The canine/premolar complex Dental and skeletal dimorphism Pelage and skin dimorphism Causal Models Sexual selection History of comparative studies of dimorphism sexual selection vs. other mechanisms Recent models of mate competition, intrasexual competition, and dimorphism Sperm competition Mate choice Models for other causes of dimorphism Body size Diet Terrestriality Correlated response Niche divergence Taxonomy Epiphenomenal variation Ontogeny Dimorphism in the Fossil Record Dimorphism in the primate fossil record WILEY-LISS, INC. DOI /ajpa.10011

2 26 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 44, 2001 Dimorphism in hominids Conclusions Acknowledgments Literature Cited Sexual dimorphism is a common phenomenon among living and extinct primates. The past several decades have seen a steady growth in studies of dimorphism, ranging from simple documentation of the phenomenon in various species to broad comparative analyses incorporating humans and a wide variety of primates. Consequently, our understanding of the expression, evolution, and development of sexual dimorphism in primates and humans has increased dramatically. This essay reviews recent advances in our understanding of the expression and evolution of dimorphism across primates. However, it should be kept in mind that anthropologists and primatologists are interested in sexual dimorphism for a variety of reasons. Forensic anthropologists seek characters that allow accurate sex identification on the basis of skeletal and dental remains. Human biologists are interested in sexual dimorphism as a component of population variation, and what it reveals about health and epiphenomenal adaptation in local populations. Systematists, morphologists, and paleontologists need to understand sexual dimorphism as a component of trait variation, either as a speciesspecific character, or as a component of variation that needs to be distinguished from potential interspecific character variation. Dimorphism can also offer evidence about the behavior of extinct species. Primate biologists are interested in understanding the causes of sexual dimorphism in primates, and how these relate to variation in behavior and ecology (e.g., Crook, 1972; Leutenegger and Kelly, 1977; Oxnard, 1987; Plavcan and van Schaik, 1994; Leigh and Shea, 1995). Most comparative analyses address two basic questions: what are the differences and similarities in the ways that dimorphism is expressed among species, and what are the causes of variation in the magnitude and pattern of dimorphism among species? In this essay, I review recent advances in answering these two questions. I shall first survey dimorphism among primates, focusing particularly on variation in the expression of dimorphism. Next, I shall evaluate comparative models of the causes of sexual dimorphism in primates, including a discussion of sexual selection and behavioral ecological models of dimorphism, and the relation between life-history, ontogeny, and dimorphism. Finally, I provide a brief survey of dimorphism in the fossil record and discuss whether it can be used to reconstruct behavior. VARIATION IN DIMORPHISM IN PRIMATES Definitions Biologists recognize two basic types of dimorphic characters: primary sex differences and secondary sex differences. Primary sex differences are those directly related to mating and reproduction, including obstetrically related differences in the pelvis. Secondary sex differences are any other differences not directly related to mating. Primary sex differences are not often referred to as dimorphic characters (though there is nothing wrong with doing so). Rather, the term sexual dimorphism is usually reserved for secondary sex differences, particularly if they are thought to arise from sexual selection (Crook, 1972). However, the term is sometimes used with reference to any sexual difference, including genetic and biochemical differences. For example, platyrrhine monkeys have sexually dimorphic alleles for color vision (Jacobs, 1994), while Gustafsson (1994) discusses metabolic liver dimorphism in rats. Primary sex differences Primary sex differences are not commonly considered in discussions of the evolution of primate sexual dimorphism, but it is worth noting that the genitalia of primates can differ substantially in form among species (Hershkovitz, 1977; Fooden, 1980; Dixson, 1987; Harcourt and Gardiner, 1994; Harcourt, 1995; Verrell, 1992). For example, most male primates have a bone (the baculum) in the penis which differs in form and size among species. The baculum is lacking in Ateles, Lagothrix, Brachyteles, Chiropotes, Tarsier, and Homo (Hershkovitz, 1977). Penile morphology is highly variable (Hershkovitz, 1977; Harcourt and Gardiner, 1994; Harcourt, 1995). Generally, strepsirrhines tend to have spines on the penis, while most anthropoids do not (Harcourt and Gardiner, 1994). Within these groups, though, even closely related species can differ substantially in penile morphology. Scrotal morphology is also variable, but to a lesser degree than penile morphology. Relative testicular volume also varies among species (Harcourt, 1995), as does the size of the seminal vesicles (Dixson, 1997). There is also a wide variation in the external appearance of the female genitalia: the labial folds, clitoris, and associated structures, even among closely related species (Le Gros Clark, 1971; Hershkovitz, 1977). Additionally, the skin in and around the genital area varies widely among primates in cyclical changes associated with the menstrual cycle (though this is usually considered a secondary sex characteristic). Such sex-skins have evolved at least three times among catarrhine primates, and are found in species of Pan, Colobus, Procolobus, Papio, Macaca, Cercocebus, Lophocebus, Mandrillus, Theropithecus (swellings located on the chest), Miopithecus, and Allenopithecus. (Crook, 1972; Pagel, 1994; van Schaik et al., 1999; Nunn, 1999).

3 J.M. Plavcan] These vary among species in size, shape, and duration. The reasons for most interspecific variation in genital morphology are unclear (Harcourt and Gardiner, 1994; Harcourt, 1995). Sexual selection, the common explanation for sexual dimorphism in primates, probably plays little role in most variation of the scrotum, penis, baculum, clitoris, labia, and associated structures (Harcourt and Gardiner, 1994; Harcourt, 1995). Some genital differences are thought to function as a reproductive isolating mechanism: variation in genital morphology can create a lock-and-key morphology discouraging closely related species from interbreeding. Such species-specific variation can be useful for alpha taxonomy (e.g., Fooden, 1980). However, Le Gros Clark (1971) notes that genital variation is so great, even among closely related species, as to be useless for higher taxonomic studies. There has been a great deal of discussion of the evolutionary mechanisms underlying variation in testicular volume (Dixson, 1987; Verrell, 1992; Harcourt, 1995), and in the exaggerated perineal swellings of female catarrhines (Harcourt, 1995; van Schaik et al., 1999; Nunn, 1999; Stallmann and Froehlich, 2000). These are discussed below in the section on sexual selection. DIMORPHISM IN PRIMATES 27 Secondary sex differences Secondary sex differences are highly variable in expression among primates, and are sometimes spectacular. For the most part, secondary sex difference are confined to the anthropoid primates; strepsirrhine primates and tarsiers are, with a few exceptions, monomorphic. The most obvious secondary sex differences among primates are body mass dimorphism, and canine tooth size dimorphism. Skeletal dimorphism is often pronounced in primates, and is primarily a product of body mass dimorphism (Fig. 1). Pelage dimorphism, while not pervasive among primates, is present in several species and can be quite impressive. Table 1 summarizes the distribution of dimorphism among primates. Body mass dimorphism. In most anthropoid primates, males are larger than females. The most extremely size-dimorphic primates are gorillas, orangutans, mandrills, baboons, and proboscis monkeys, with males sometimes more than twice as large as females (Crook, 1972; Clutton-Brock et al., 1977; Leutenegger and Kelly, 1977; Plavcan and van Schaik, 1997b; Smith and Jungers, 1997). Most cercopithecoid male primates range from about 30 80% larger than females in body mass. Body mass differences between males and females are often characterized as species-specific traits, but there is good evidence for biologically meaningful subspecific variation in body mass dimorphism in some species (Turner et al., 1994, 1997; Smith and Jungers, 1997; Plavcan and van Schaik, 1997b; Jones et al., 2000). Fig. 1. Illustration of lateral views of male and female Macaca fascicularis skulls. This species shows typical dimorphism for a cercopithecoid primate. Note magnitude of size difference between male and female skulls, somewhat greater robusticity of the skull, longer male face, and much larger male canine. Reverse dimorphism 1 is only characteristic of some callitrichine primates and possibly some strepsirrhines (Kappeler, 1990, 1991; Godfrey et al., 1993; Plavcan and van Schaik, 1997b). In general, body mass dimorphism reaches its extremes in hominoids and papionines. Platyrrhines on the whole are characterized by lesser degrees of body mass dimorphism, though there are a few species (howler monkeys) that are comparable to cercopithecoids (Ford, 1994; Plavcan and van Schaik, 1997b). The canine/premolar complex. Canine tooth size dimorphism is also common in anthropoids (Crook, 1972; Leutenegger and Kelly, 1977; Harvey et al., 1978; Plavcan and van Schaik, 1992, 1994). The maxillary canine teeth tend to be more dimorphic than the mandibular canines. Cercopithecoid male canine teeth tend to be very long and daggerlike, and can be up to 400% taller than those of 1 The term reverse dimorphism is used to indicate that females are larger than males. Some object to this term as sexist, reflecting male bias in primatology. However, in primates almost all species have males larger than females. Dimorphism is usually expressed quantitatively as a ratio of the larger sex to the smaller (the opposite practice is statistically inadvisable), meaning that ratios reflect male values divided by female values. Reverse dimorphism is used to indicate that the opposite relation is true.

4 TABLE 1. Summary of body mass, canine tooth size, and pelage dimorphism in primates 1 Taxon Body mass dimorphism Canine dimorphism Pelage and other dimorphisms Lorisids Slight reverse to moderate Slight None Galagids Slight to moderate Slight to moderate None Lemurids Reverse to none Slight to moderate Most, none; Lemur macaco, dichromatic Cheirogalids Slight reverse to moderate None None Indriids Slight reverse None None Lepilemurids None Slight to moderate None Daubentonia None None None Tarsier None None None Callitrichines None, females sometimes slightly None None larger than males Aotus None None None Saimiri Moderate; males gain weight during Moderate None breeding season Cebus Slight Slight; female canines relatively None large Callicebus None None None Pithecia Slight Slight; female canines relatively P. pithecia dichromatic large Cacajao Slight Slight; female canines relatively None large Chiropotes Slight Slight; female canines relatively None large Ateles Moderate Moderate None Brachyteles None None None Lagothrix Moderate Moderate None Alouatta Strong Strong A. caraya dichromatic; males have enlarged hyoid bone Cercopithecus Moderate to strong Moderate to strong; female canines relatively large Most have none; some have different colors on genital skin Allenopithecus Strong Strong Females with exaggerated estrous swellings Miopithecus Strong Strong Females with exaggerated estrous swellings Erythrocebus Strong Extreme None Lophocebus Strong Strong Females with exaggerated estrous swellings Cercocebus Strong Strong Females with exaggerated estrous swellings Macaca Moderate to strong Strong to extreme Females with exaggerated estrous swellings in some species (M. fascicularis, nemestrina) Papio Strong to extreme Extreme Females with exaggerated estrous swellings; male P. hamadryas and P. papio have mane Mandrillus Extreme Extreme Females with exaggerated estrous swellings; male mandrills with brilliant facial and perianal colors; less brightly colored facial display in drills Theropithecus Extreme Extreme Males with large mane and lip display; females with brilliant chest and abdominal sexual skin; males have lesser developed homologous sex skins Colobus Slight to moderate Strong; guereza females have None relatively large canines Procolobus Moderate Strong Females with exaggerated estrous swellings Presbytis Slight Slight to strong None Trachypithecus Moderate Strong None Semnopithecus Moderate Strong None Kasi Moderate Strong None Nasalis Strong Strong Males with bulbous nose Simias Strong Strong None Rhinopithecus Strong Strong None Pygathrix Strong Strong None Hylobates Slight Slight H. concolor, hoolock, and pileatus dichromatic; male gibbons have enlarged hyoid bones Pan Moderate Moderate Females with exaggerated perianal swellings Gorilla Extreme Strong Older males develop silver back hair Pongo Extreme Strong Older males with gular pouches, cheek flanges Homo Moderate Slight Males with facial and body hair; females with enlarged breasts; different distribution of body fat 1 Anthropoid body mass from Plavcan and van Schaik (1997b). Strepsirrhine body mass dimorphism from Kappeler (1990). Anthropoid canine dimorphism from Plavcan and van Schaik (1992). Strepsirrhine canine dimorphism from Kappeler (1991). Information on pelage and other dimorphisms from Hershkovitz (1977), van Schaik et al. (1999), Crook (1972), and Fleagle (1999).

5 J.M. Plavcan] Fig. 2. Occlusal views of maxillary canine teeth of five primates, illustrating variation in canine size and shape among species, and within sexes. All teeth are oriented in same direction, with mesial at top, distal at bottom, lingual to left, and buchal to right. Note especially progression from strongly compressed in Varecia to rounded in Pongo, the heart-shaped outline of male Macaca and Presbytis teeth associated with a deeply excavated mesial groove, and differences in shape between males and females of Macaca and Presbytis. females. Among anthropoids, male canines vary in size and shape (Fig. 2), ranging from fat, relatively short teeth in the great apes (by comparison to cercopithecoids), to long, tall thin teeth in hylobatids (Greenfield and Washburn, 1991; Greenfield, 1992a; Plavcan, 1993a). Dimorphism is usually much greater in the crown height of the tooth than in the occlusal dimensions (the measurements of length and breadth at the base of the tooth). There are distinctive patterns of canine dimorphism that correspond roughly to differences in male canine shape. Cercopithecoids tend to show proportionally greater crown height dimorphism as compared to occlusal dimorphism, while in hominoids dimorphism tends to be more uniform among the tooth dimensions. Male and female canines in many species cannot be easily sorted when they are similar in size. However, Kelley (1995a,b) demonstrated that hominoid male and female canine teeth can be sorted on the basis of a few simple proportional differences, even where they do not differ substantially in size. Many male anthropoid canines are characterized by a strong mesial groove running down the anterior DIMORPHISM IN PRIMATES 29 surface of the tooth onto the root (Crook, 1972; Fleagle, 1999). This gives the tooth a heart-shaped occlusal outline which is easily recognized, especially in cercopithecoids. While most female anthropoids have smaller canine teeth than males, their canines nevertheless can be quite large (Plavcan and van Schaik, 1994; Plavcan et al., 1995). Female canine tooth size ranges from small, as in most colobines, to as large as those of males, as in the hylobatids. Females of some species also have a mesial groove, though not as exaggerated as that of males. Importantly, canine tooth size dimorphism is as much a function of variation in female canine tooth size as male canine tooth size. For example, both Callicebus and Hylobates show nearly monomorphic canine teeth. In Callicebus, canines in both sexes are relatively short and blunt, while in Hylobates, they are relatively tall. The mandibular premolar that lies adjacent to the mandibular canine (P2 in platyrrhines and P3 in catarrhines) can also be highly dimorphic (Zingeser, 1969; Crook, 1972; Greenfield, 1992a,b, 1996; Greenfield and Washburn, 1992). In most species, the male tooth is elongated to form a hone for sharpening the back edge of the maxillary canine tooth (Fig. 3). Consequently, the mandibular premolar is often considered with the canines as part of the canine/premolar complex. Females also can have an elongated premolar. Greenfield (1996) offers evidence that the hone of the female mandibular premolar is overdeveloped by comparison to the size of the maxillary canine. He proposes that the development of the female hone is a genetic correlatedresponse to the development of the male honing tooth. However, female canine crown height is strongly correlated with female honing premolar length, suggesting that the hone is functional in females as well as in males (though this does not rule out the effect of correlated response). Dental and skeletal dimorphism. Primates also show sexual dimorphism in the skeleton and teeth. Linear dimensions of the noncanine teeth can be up to about 10% larger in males than females (Garn et al., 1966; Plavcan, 1990; Cochard, 1985). It has been suggested that there is a field effect of sexual dimorphism in the dentition, centered on the canine teeth, such that teeth closer to the canines tend to be more dimorphic than those farther away (Garn et al., 1966). Others have suggested that different taxa share different patterns of sexual dimorphism in the teeth (Lieberman et al., 1985; Oxnard, 1987). It appears, however, that noncanine tooth size dimorphism is largely a correlate of body mass dimorphism (Cochard, 1985; Wood et al., 1991). While the pattern of tooth size dimorphism in the dentition tends to follow classic fields of tooth size, there are no substantial, taxonomically significant patterns of dimorphism in teeth other than the canines (Plavcan, unpublished findings).

6 30 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 44, 2001 Fig. 3. Lateral views of mandibular canines and honing premolars of four anthropoid species. A: Pan troglodytes. B: Mandrillus leucophaeus. C: Alouatta seniculus. D: Macaca mulatta. Mesial is to left, distal to right. Note strong elongation of the hone in Macaca and Mandrillus, in both males and females. In Pan, the mesial enamel face of the tooth is extended, but not clearly derived as in the cercopithecoids. Cochard (1985) noted that dimorphism of the molar and premolar teeth is less than expected on the basis of body mass dimorphism. He suggested that males of dimorphic primates tend to have relatively small teeth for their body size because size dimorphism is achieved largely through an extension of male growth, continuing well after tooth size has been determined. Cranial and skeletal dimorphism in primates is less well-studied than canine or body mass dimorphism (Wood, 1976; Anderson, 1981, 1982; Leutenegger and Larson, 1985; Richtsmeier and Cheverud, 1989; Corner and Richtsmeier, 1991, 1992, 1993; Ravosa, 1991; Richtsmeier et al., 1993; Ravosa and Ross, 1994; Lague, 2000). As a rule, skeletal dimorphism tends to be pronounced in more sizedimorphic primates: larger males have proportionally larger skeletons (Wood, 1976). Hence, skeletal dimorphism clearly arises primarily as a consequence of size dimorphism, and not as a consequence of selection for different male and female adaptations (even though there is good evidence that morphological dimorphism in some other vertebrates and invertebrates is a direct consequence of niche Fig. 4. Range of dimorphism of 38 measurements of skull and jaws in Macaca nemestrina (28 males, 18 females). Dimorphism is estimated as natural log of the ratio of average male value divided by average female value for each dimension. Measurements are organized loosely into regions: N, neurocranium; O, orbits; B, basicranium; P, palate; F, face; M, muscle attachments; J, jaws. Note that dimorphism is strongest in the palate, face, muscle attachment surfaces, and jaws, and least in the basicranium, orbits, and neurocranium. However, there is a substantial range of variation within these regions. divergence between males and females; Andersson, 1994). Dimorphism of linear skeletal dimensions tends to be proportional to the cube root of body mass dimorphism (Gingerich, 1981). For example, a species showing a body mass dimorphism ratio of 1.5 (males 50% larger than females) can be expected show dimorphism of about 1.14 (males 14% larger than females) in linear skeletal dimensions. This proportionality between mass and skeletal dimorphism is only a generalization. Skeletal dimorphism is not uniformly expressed as a simple function of body mass dimorphism in primates (Oxnard, 1987; O Higgins et al., 1990; Wood et al., 1991; Plavcan, in press), and in fact appears to be very complex in the pattern of its expression both within and among species. In most species, measurements of the brain case, basicranium, and orbits tend to show less dimorphism than measurements of the face and jaws (Fig. 4; Leigh and Cheverud, 1991; Ravosa, 1991; Ravosa and Ross, 1994; Corner and Richtsmeier, 1991, 1992, 1993; Masterson, 1997; Masterson and Hartwig, 1998; Lockwood, 1999; Plavcan, in press). However, measurements of a single bone, such as the mandible, can show substantially different degrees of dimorphism within and among species. For example, in a sample of 80 Gorilla specimens, males are on average 8% larger than females

7 J.M. Plavcan] Fig. 5. Profile of dimorphism in 38 craniofacial dimensions for three species of Alouatta. Note the general similarity of profiles, indicating that all three species share similar patterns of craniofacial dimorphism. Dimorphism is measured as natural log of the ratio of average male measurement divided by average female measurement for each dimension. DIMORPHISM IN PRIMATES 31 in mandibular depth, but 29% larger than females in mandibular symphyseal thickness. In Colobus angolensis, males are about 14% larger than females in both dimensions. In fact, there is a hierarchy of variation in patterns of dimorphism: there is a basic pattern of dimorphism among skeletal components, but species vary in patterns that are superimposed on this general pattern (Plavcan, in press). Patterns of craniofacial dimorphism tend to be conserved among closely related species (Plavcan, in press; but see Jones et al., 2000, for evidence of variation in craniofacial dimorphism among subspecies of Alouatta paliatta). Thus, species of Alouatta (Fig. 5) tend to show much more similar patterns of craniofacial dimorphism to each other, than to, for example, Macaca (Plavcan, in press). Great apes are unusual in that, proportional to their body mass dimorphism, they tend to show less craniofacial dimorphism than other primates (Plavcan, in press). This observation is intriguing for those interested in understanding hominid dimorphism, because it implies that the intense craniofacial dimorphism of some hominids is associated with even greater body mass dimorphism than formerly suspected (cf. Lockwood, 1999). In contrast, no such taxonomic patterns are found in the dentition, apart from those seen in the canine/ premolar complex (Wood, 1976; Wood et al., 1991; Plavcan, unpublished findings). Lague (2000) recently found little evidence for significant pattern differences in dimorphism of the elbow and knee joint surfaces in a series of catarrhine primates. Likewise, Wood (1976) found little evidence for substantial pattern differences in the postcranial skeletons of seven primate species. Beyond this, comparative studies of pattern differences in dimorphism are relatively few, so it is no surprise that the causes of such differences are poorly known. The complexity of the results of the above studies suggests a variety of factors underlying pattern differences in skeletal and cranial dimorphism. Some basic pattern differences can be easily explained. For example, colobine primates show little body mass dimorphism, but rather intense canine tooth size dimorphism (Plavcan and van Schaik, 1997b). Consequently, only those craniofacial dimensions near the canine teeth show substantial dimorphism (Plavcan, in press). Conversely, cercopithecine primates all show intense body mass and canine tooth size dimorphism and intense craniofacial dimorphism (Plavcan and van Schaik, 1997b; Plavcan, in press). Hence the contrasting patterns of craniofacial dimorphism of colobines and cercopithecines seem to directly reflect the contrasting patterns of canine and body mass dimorphism. More subtle differences in patterns of dimorphism probably reflect interactions between growth patterns and functional and biomechanical constraints, though there has been little comparative work on this topic. Leigh (personal communication) noted that patterns of sutural closure and epipheseal fusion throughout the skeleton vary among primates. Such developmental variation is likely to influence the expression of sexual dimorphism in the skeleton by varying the relative duration of growth among skeletal components. Studies of craniofacial development generally point to patterns of ontogenetic scaling between males and females, i.e., male growth curves tend to be extensions of the female growth curves (Leutenegger and Masterson, 1989a,b; Richtsmeier and Cheverud, 1989; Masterson and Leutenegger, 1990; Ravosa, 1991; Corner and Richtsmeier, 1991, 1992, 1993; Richtsmeier et al., 1993; Ravosa and Ross, 1994). Heterogeneity in the rate and duration of growth among different skeletal regions accounts for different patterns of dimorphism. For example, in baboons, facial dimensions grow faster than neurocranial dimensions. With extended male growth, this means that facial dimensions tend to be more dimorphic than neurocranial dimensions. Potential biomechanical or functional reasons for variation in patterns of dimorphism are not well understood at this time. In the craniofacial complex, simply supporting large canine teeth requires large jaws and maxillae. In order to effectively use large canine teeth, an animal needs to open its jaws wide enough for the canines to clear the mandibular dentition (Lucas et al., 1986; Lucas, 1981). The long faces of many male papionine primates, and the attendant facial dimorphism, may reflect selection to increase gape in order to do this (Lucas, 1981; but see also Ravosa, 1990). On the other hand, recent studies found no evidence of shape differences in humeral and femoral joint surfaces (Lague, 2000) or

8 32 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 44, 2001 the craniofacial complex (Spencer and Hogard, 2001) that might reflect selection for different biomechanical properties. The postcranial data are particularly interesting in light of the hypothesis of Stern and Susman (1983) that Australopithecus afarensis males and females may have shown different locomotor repertoires based on postcranial variation. Some pattern differences can be explained with reference to static adult allometric relations between skeletal size and muscle and/or organ size. For example, larger male body mass is usually accompanied by more robust skeletal components, more intense development of muscle scarring, and stronger cresting of bones in comparison to females. Such differences are manifest even in humans, which are only modestly dimorphic in body mass (males less than 20% larger than females), making it possible to identify the sex of individuals with between 80 90% accuracy (White and Folkens, 1991). It is well known that the large sagittal and nuchal crests of male gorillas are a function of the disproportionate increase in the size of the temporalis muscles and nuchal musculature as compared to the brain. Brain size in male and female gorillas is on average quite similar, even though male gorillas can be twice the body mass of females (Smith and Jungers, 1997; Plavcan and van Schaik, 1997b). This means that the brain case in males and females is very similar in size. The larger male body mass, though, means that the muscles attaching to the brain case require substantially more attachment area in males than in females. Consequently, males develop large crests and flanges to provide this additional bony surface area. In a comparative context, however, the relative development of sagittal and nuchal cresting in primates is itself a function of overall body size. Because muscle force is a function of cross-sectional muscle area, rather than mass, muscle force increases by the 2/3 power of body mass. This means that larger primates need proportionally larger muscles in order to maintain an isometric relationship between body size and muscle force. This in turn requires proportionally larger muscular attachment areas for larger primates. As a result, smaller primates tend to have less cresting on the skull than larger primates, regardless of degree of body mass dimorphism or brain size. Finally, one other unusual skeletal dimorphism is found in male gibbons and howler monkeys, which possess enlarged, hollow hyoid bones that serve as resonators for the voice. Much more research needs to be done to understand the underlying causes of variation in patterns of skeletal dimorphism in primates. Such research is likely to increase our understanding not only of the causes of skeletal dimorphism, but also of the growth, development, constraints, and adaptations of the skeleton. Skeletal dimorphism often represents a significant degree of morphological variation within a species that needs to be accommodated within any model of the evolution of morphological form. Pelage and skin dimorphism. Pelage and skin can be strikingly dimorphic in primates. The most familiar example is the brightly colored face of male mandrills as compared to that of females (Setchell and Dixson, 2001). Skin and pelage differences are not limited to mandrills. As already noted above, there are dramatic female estrous swellings in many anthropoid species. Proboscis monkeys get their name from the enormous, pendulous nose of the male. Guenons are famous for the red-white-andblue display of males. The male penis is bright red, the scrotum is bright blue, and the genital region is surrounded by white fur. This makes for a conspicuous penile display that is used as a threat gesture by males (Fedigan and Fedigan, 1988). Male gelada baboons are characterized by a large mane which probably serves to make males appear larger (Crook, 1972), and might serve as a signal to females of male health. Orangutan males possess large cheek flanges and laryngeal pouches. In several species (Lemur macaco, Pithecia pithecia, Alouatta caraya, Hylobates concolor, H. hoolock, and H. pileatus), males and females differ dramatically in pelage color (Crook, 1972; Hershkovitz, 1977). A number of other species show minor sex differences in pelage color or patterning (Hershkovitz, 1977; Napier, 1981, 1985). The reasons for this are not clear, though it could be related to mate recognition (as a function of speciation), mating preferences, or female choice (Crook, 1972; Hershkovitz, 1977; Martin et al., 1994). CAUSAL MODELS Sexual selection Sexual dimorphism in body mass, canine tooth size, and display features is commonly attributed to sexual selection. However, there are other factors including developmental, phylogenetic, and genetic constraints, natural selection, and epiphenomenal factors that can influence the expression of dimorphism in species (Crook, 1972; Clutton-Brock et al., 1977; Leutenegger and Kelly, 1977; Harvey et al., 1978; Leutenegger and Cheverud, 1982, 1985; Gaulin and Sailer, 1984; Cheverud et al., 1985; Clutton- Brock, 1985; Gautier-Hion and Gautier, 1985; Leutenegger and Lubach, 1987; Shea, 1986; Oxnard, 1987; Kay et al., 1988; Ely and Kurland, 1989; Kappeler, 1990, 1991, 1996; Greenfield, 1992a,b, 1996; Leigh, 1992, 1995a,b; Plavcan and van Schaik, 1992, 1994, 1997b; Ford, 1994; Martin et al., 1994; Hayes et al., 1995; Leigh and Shea, 1995; Mitani et al., 1996; Plavcan, 1998, 1999). The relative importance of these factors to the expression of dimorphism among primates has been at times a topic of heated debate. In order to reconcile the sometimes confusing literature on this topic, a review of mechanisms

9 J.M. Plavcan] of sexual selection, and how they have been quantified in studies of dimorphism, should help. Following is a brief, simplified description of mate choice and mate competition. There is substantial variation on the themes outlined below. The literature on primates alone describing different male and female mating strategies is enormous. The most comprehensive general review of sexual selection to date is that of Andersson (1994). Sexual selection encompasses a diverse array of mechanisms (Andersson, 1994). Sexual selection theory is broadly divided into two components: mate competition and mate choice. Individuals can increase their reproductive fitness relative to conspecific rivals by either excluding rivals from mating (mate competition), or by selectively choosing mates (mate choice). Both of these mechanisms generate selection by increasing the fitness of offspring (if mates who win fights or are chosen pass on better genes), and/or by creating differential mating opportunities (by excluding some individuals from mating, either by force or by active choice). Neither mechanism is limited to one or the other sex. However, in primates mate competition is largely associated with males, while mate choice is associated with females. Mate choice and mate competition theoretically arise ultimately from anisogamy (differential size and production of gametes between sexes), which in turn favors different levels of parental investment in individual offspring (Trivers, 1972; Reynolds and Harvey, 1994; Andersson, 1994). In primates, males are limited in their reproductive output by the number of females that they can inseminate. Sperm supply for a male can be temporarily reduced through repeated ejaculation (Dixson, 1997), but the sperm count rebounds much more quickly than the supply of females that a male can effectively mate with over a period of time (cf. van Schaik et al., 1999). Consequently a male can dramatically increase his fitness relative to other males by excluding them from access to a group of females. This provides a powerful incentive for males to compete intensely with one another for access to females (Pagel, 1994). Any weaponry or other morphological feature such as size, displays, agility, speed, coordination, or other behaviors that help males win fights should be powerfully selected for (Crook, 1972; Clutton-Brock, 1985; Kappeler, 1991; Martin et al., 1994; Reynolds and Harvey, 1994; Plavcan, 1999). The other half of the equation is mate choice. Generally speaking, a female s reproductive output is limited by the number of offspring she can produce and raise (Andersson, 1994; Reynolds and Harvey, 1994). While a female s egg production theoretically can be very large, the costs of pregnancy, lactation, and child care can be quite high, limiting the number of offspring that can be produced in a lifetime. Presumably, the survivorship and reproductive success of the female s offspring will be at least partly a function of the quality of the genes DIMORPHISM IN PRIMATES 33 contributed by the male. Male parental investment also can improve offspring survival (Andersson, 1994; Reynolds and Harvey, 1994). The result is selective pressure for females to choose their mates, either to control genetic quality, or to select for males that will in some way be better providers. Female choice is not necessarily limited to characters that signal male fitness. For example, if females have, for some reason, a preexisting bias to mate with males bearing a particular trait, then selection might favor the development of that trait, even if it is not associated with any fitness benefit (Fisher, 1930; Ryan, 1990; Andersson, 1994; Stallmann and Froehlich, 2000). Another mechanism that may produce dimorphism is mate recognition. Females or males may prefer traits that act as signals to ensure conspecificity, avoiding cross-breeding with closely related species that may produce less viable offspring. While these mechanisms are known among biologists interested in sexual selection (Andersson, 1994), they have not received a great deal of scrutiny among primatologists interested in dimorphism. History of comparative studies of dimorphism sexual selection vs. other mechanisms. The hypothesis that sexual dimorphism in primates is a consequence of mate competition has had a long historical development, during which various alternative explanations for the evolution of dimorphism were put forward. The way that mate competition is measured and analyzed, and views on its relative importance in explaining the evolution of dimorphism in primates, are intertwined with the development of these alternative hypotheses. Hence, before presenting a detailed review of recent comparative studies of the sexual selection hypothesis and dimorphism in primates, it will be helpful to review comparative studies of dimorphism over the past 30 years. Following this, I shall present a detailed review of recent models of mate competition, mate choice, and dimorphism. This is followed by a more detailed discussion of evolutionary models for other behavioral, ecological, and morphological correlates of dimorphism. Mate competition has long been thought to be the primary mechanism favoring the development of sexual dimorphism in primates (Darwin, 1871; Crook, 1972; Leutenegger and Kelly, 1977; Clutton- Brock, 1985; Kay et al., 1988; Kappeler, 1990, 1991; Plavcan and van Schaik, 1992, 1994, 1997b; Ford, 1994; Martin et al., 1994; Lindenfors and Tullberg, 1998). In easily observed species, such as baboons, males are largely intolerant of one another, often engaging in spectacular fights and threat displays. Such displays and fighting seem to determine access to females. Species like baboons are intensely dimorphic. Males have enormous canine teeth and are much larger than females, fitting well with the concept that sexual selection has favored these traits because of the advantages they confer in winning

10 34 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 44, 2001 fights. In contrast, monomorphic gibbons are very obviously monogamous, showing no apparent differential male reproductive success that is associated with male-male competition. Sexual selection is not the only explanation for sexual dimorphism in primates. One of the oldest alternative hypotheses is that large male body size and canine tooth size evolved for predation defense (DeVore and Washburn, 1963). In fact, the popularity of the sexual selection hypothesis as an explanation for the evolution of primate sexual dimorphism has actually waxed and waned over the years. This has come about partly as a result of the failure of many studies to adequately demonstrate a relation between variation in dimorphism and variation in estimates of sexual selection, and partly through the elaboration of alternative mechanisms that might explain variation in dimorphism. Debates in the comparative literature of the past 30 years were stoked by the observation that sexual dimorphism in body mass and canine tooth size varies considerably in magnitude across primates (Clutton-Brock et al., 1977; Leutenegger and Kelly, 1977; Harvey et al., 1978; Pickford, 1986; Martin et al., 1994; Plavcan and van Schaik, 1994). At the same time, as more became known about primate social systems, behavior, and ecology, it became apparent that sexual selection must vary in intensity among species. This led to attempts to associate variation in sexual dimorphism with variation in estimates of sexual selection. The first such attempts were anecdotal. Crook (1972) observed the association between polygyny and dimorphism in primates, and provided detailed hypotheses for the evolution of dimorphism in pelage, body size, canine size, and other morphological traits in relation to various mating systems and habitats. Leutenegger and Kelly (1977) presented a comparison of dimorphism in canine size and body mass in primates. They concluded that dimorphism in the two characters is moderately correlated, suggesting that factors other than sexual selection may play a role in constraining or driving the evolution of dimorphism in different characters in different species. For example, in addition to sexual selection, they suggested that predation defense may play a role in the evolution of large body size in male proboscis monkeys, but that an arboreal mode of locomotion may constrain the evolution of large male size in other colobines, thereby constraining the expression of body mass dimorphism. Clutton-Brock et al. (1977) attempted the first formal statistical test of the sexual selection hypothesis in primates. They measured sexual selection using the socionomic sex ratio the number of adult females in a typical group divided by the number of adult males. The reasoning behind this ratio is that the fewer males per group, the greater the differential male reproductive success associated with malemale competition. Hence, the greater the ratio, the more intense sexual selection must be. They found a weak positive correlation between the socionomic sex ratio and body mass dimorphism. However, they noted that when monogamous species were removed from the analysis, there was no correlation between the socionomic sex ratio and dimorphism, even though there is a large range of variation in both variables. They noted that this could reflect either the effect of other factors on the evolution and expression of sexual dimorphism, or the inadequacy of the socionomic sex ratio to capture variation in sexual selection (now widely accepted; cf. Clutton- Brock, 1985; Andersson, 1994; Plavcan, 1999). They also noted that terrestrial species tend to be more dimorphic than arboreal species, and that body mass dimorphism is significantly positively correlated with overall body size. They suggested that the distinction between terrestrial and arboreal species might reflect a constraint on larger male size associated with the use of terminal branches in arboreal species. Harvey et al. (1978) provided the first statistical comparative analysis of canine tooth size dimorphism. To estimate sexual selection, they used a tripartite categorization of primate mating systems: monogamy, multimale/multifemale social groups, and single-male/multifemale harems. They reasoned that sexual dimorphism should be minimal in monogamous primates because there is little male differential reproductive success. Dimorphism should be intermediate in multi-male groups, and greatest in single-male groups, reflecting the ability of males to either partially or totally exclude other males from access to a group of females. They found that monogamous species are indeed less dimorphic than polygynous species, but they found no significant difference in dimorphism between the multimale and single-male social systems. They too noted that terrestrial species are more dimorphic than arboreal species, supporting the predation-defense hypothesis. Up to this point, studies confirmed that sexual selection has something to do with the evolution of sexual dimorphism in primates, but they also suggested that much of the variation in dimorphism among species must reflect other factors. Subsequently, two influential studies suggested that the evolution of larger body size alone leads to the evolution of greater sexual dimorphism (Leutenegger and Cheverud, 1982, 1985), and that phylogenetic inertia explains a great deal of variation in dimorphism (Cheverud et al., 1985). Leutenegger and Cheverud (1982, 1985) developed a quantitative genetic model which suggested that sexual dimorphism will increase over time if male traits are more variable than those of females, or if male traits are less heritable. This hypothesis was put forward to explain the well-known correlation between sexual dimorphism and body mass in primates (Crook, 1972; Clutton-Brock et al., 1977; Harvey et al., 1978; Pickford, 1986; Plavcan and van Schaik, 1992, 1994, 1997b; Plavcan, 1999). However,

11 J.M. Plavcan] it also suggested that sexual selection may play a relatively minor role in the evolution of sexual dimorphism. The analysis received a variety of intense criticism. Gaulin and Sailer (1984) pointed out that the measure of sexual dimorphism used in the study was biologically unrealistic, de facto creating a strong correlation between size and dimorphism. Leutenegger and Cheverud (1982, 1985) treated a 5-kg difference in mass the same between galagoes and gorillas. For a galago to show proportionally as much dimorphism as a gorilla by this measure, the male galago would have to weigh over 1,000 times as much as the female. Plavcan and Kay (1988) and Plavcan (1990, 2000b) demonstrated that male primates are not relatively more variable than females for dental and cranial traits. Kappeler (1990, 1991) and Godfrey et al. (1993) pointed out that there is no correlation between body size and dimorphism in strepsirrhine primates. Andersson (1994) noted that several studies likewise found little such correlation in other groups of animals, and that the model only works as long as body size continually increases in a lineage over time. It is interesting to note that in the human lineage, dimorphism apparently has decreased through time while body size has increased (Frayer, 1980; McHenry, 1992). In spite of all this criticism, it is important to emphasize that the model of Leutenegger and Cheverud (1982, 1985) has not been disproved. The above criticisms primarily targeted the analysis of Leutenegger and Cheverud (1982, 1985), but not the model itself, which incorporates selective and nonselective components, and is based on established quantitative genetic work by Lande (1980). All of the above papers have critiqued only the suggestion that body size increase is the primary factor driving the evolution of dimorphism in primates, by critiquing the claims of a strong correlation between body size and dimorphism, and greater male character variation. The critiques of the correlation between size and dimorphism do not deny the widely acknowledged fact that such a correlation exists in anthropoid primates they only point out that Leutenegger and Cheverud (1982, 1985) overemphasized the correlation through the way that they measured dimorphism, and that such a correlation does not exist in other groups of mammals. Even arguments about male and female trait variability are not conclusive. Plavcan and Kay (1988) and Plavcan (1990, 2000b) demonstrated that male traits are not relatively more variable than female traits. But in terms of absolute variation, there is little doubt that male primates tend to be more variable than females. The key is deciding which is more important: relative or absolute variation. Given that absolute variation is fundamentally linked to size itself, this particular question about male and female variability is difficult to resolve. In another important paper, Cheverud et al. (1985) suggested that much variation in dimorphism DIMORPHISM IN PRIMATES 35 among primates can be explained as a correlate of taxonomy, and not sexual selection. They used an autocorrelation model to partition variation in dimorphism correlated with taxonomy from that which is correlated with body size, mating system, habitat, and diet, and found that taxonomy explains 50% of the variation, body size explains 36%, and habitat, mating system, and diet explain the rest. Ely and Kurland (1989) reanalyzed the data of Cheverud et al. (1985), and challenged their conclusions. Regardless of the validity of the model or the interpretation of the results, Cheverud et al. (1985) brought the concept of phylogenetic inertia to the forefront of analyses of sexual dimorphism. While taxonomic differences in dimorphism have been known for a long time (cf. Crook, 1972), they had been largely ignored as a potential explanatory or confounding factor in understanding dimorphism among primates. Almost all comparative analyses of dimorphism in primates now acknowledge the existence of taxonomic correlations, and the potential importance of phylogenetic inertia as an explanation for variation in the magnitude and expression of dimorphism among species. Other hypotheses received support in various papers over the years. Rowell and Chism (1986) presented evidence of multimale influxes and matings in supposedly single-male species of guenons. They suggested that the role of sexual selection and malemale competition had been overemphasized. Phillips-Conroy and Jolly (1981) likewise questioned the role of sexual selection, noting that the magnitude of dimorphism in hamadryas and olive baboons did not appear to follow the pattern predicted by mating systems. A number of papers emphasized the role that predation defense must play in the evolution of sexual dimorphism (Crook, 1972; Leutenegger and Kelly, 1977; Gautier-Hion and Gautier, 1985; Rowell and Chism, 1986; Plavcan and van Schaik, 1992, 1997b), based on the observation that terrestrial species tend to be more dimorphic than arboreal species. Leutenegger and Lubach (1987) pointed out that Cercopithecus neglectus is apparently monogamous, yet is among the most dimorphic of guenons. They suggested that dimorphism is retained in this species as a consequence of phylogenetic inertia. Several papers entertained the concept of niche divergence in primates (e.g., Clutton-Brock et al., 1977; Harvey et al., 1978), though none has ever provided compelling field data for this hypothesis. Some suggested that arboreal locomotion places an upper limit on male body size, constraining the expression of body size dimorphism (Clutton-Brock et al., 1977; Leutenegger and Kelly, 1977). Lucas (1981) suggested that maximum jaw gape may constrain male canine crown height, constraining canine dimorphism. He also suggested that enlarged male canines evolved in an anthropoid ancestor, and that variation in canine dimorphism is a function of gape constraints.