OUP user menu

Primate Mating Systems and the Evolution of Neocortex Size

Michael A. Schillaci
DOI: http://dx.doi.org/10.1644/06-MAMM-A-417.1 58-63 First published online: 19 February 2008

Abstract

The evolution of the neocortex in primates has been associated with social complexity, but the relationship between neocortex evolution and components of social complexity such as sexual selection and mating systems is not well studied. I examined the evolutionary relationship between relative neocortex size and the intensity of male–male competition for mates among various primate mating systems using bootstrapped estimates of least-squares regression parameters. A significant negative evolutionary relationship was found between relative neocortex size and the level of male–male competition for mates associated with various mating systems. The largest relative neocortex sizes among primate species were associated with monogamy This negative evolutionary relationship suggests that monogamy may require greater social acuity and abilities for deception and manipulation, and promote selection for larger brains.

Key words
  • brain evolution
  • monogamy
  • neocortex
  • primates
  • sexual dimorphism

Despite extensive research, the ultimate explanation for the evolution of brain size in primates remains largely unresolved (Walker et al. 2006). Research associating increasing brain size with larger social groups and social complexity in primates predicts that brain size, specifically the size of the neocortex relative to the rest of the brain, will coevolve with mating systems exhibiting social complexity (Dunbar 1995, 2003). In this context, larger brains are selected for because they confer greater reproductive fitness associated with greater social acuity or the ability to manipulate others within the group (Byrne and Whiten 1997; Dunbar 1995). Larger brains also reflect the cognitive demands associated with complex social systems requiring, for example, deception and the formation of social coalitions (Dunbar 1998; Whiten and Byrne 1988). These demands will be greater, in particular, for males competing in mating systems associated with larger groups, such as multimale–multifemale systems. Pawlowski et al. (1998) found that among polygamous primates the correlation between male rank and reproductive success was negatively correlated with relative neocortex size after accounting for shared ancestry. Pawlowski et al. (1998) interpreted their finding as indicating that lower-ranking males of a species with larger neocortices were circumventing dominance-based systems and capitalizing on mate choice by females through deployment of sophisti-cated social skills and coalition building (also see Dunbar 1998). Therefore, social complexity and the advantage of sophisticated social and cognitive ability associated with larger neocortices is applied not only to competition with other males for access to females, but also to manipulating female choice.

The social brain hypothesis predicts that sizes of social groups in primates are constrained by information-processing capacity, which is largely determined by the neocortex portion of the brain. Group sizes have, in fact, been demonstrated to covary with size of the neocortex relative to total brain size (see Dunbar 1992, 1995, 1998; Kudo and Dunbar 2001), lending considerable credence to the social brain hypothesis. Although there seems to be consensus that the evolution of the neocortex is associated with social complexity, the relationship between neocortex evolution and certain components of social complexity such as sexual selection and mating systems have not been well studied in primates (but see Pawlowski et al. 1998; Schillaci 2006).

In addition to social complexity, researchers have identified length of the period spent as a juvenile as an important contributor to the evolution of brain size in primates (e.g., Allman et al. 1993). Recent studies have identified life span, foraging complexity, and population density as the most significant predictors of length of the period spent as a juvenile in primates (e.g., Walker et al. 2006). It seems, therefore, that slower growth and development may facilitate social learning and promote foraging competency in primates. Larger neocortices would be adaptive in such a socioecological context.

I examined the evolutionary relationship between relative neocortex size and the intensity of male–male competition for mates among various mating systems in anthropoid primates. In addition, I tested the hypothesis that large relative neocortex size in male primates has coevolved with mating systems exhibiting social complexity. Generally speaking, social complexity is presumed to increase with larger sizes of social groups in primates because of the larger number of relationships involved (Dunbar 1998). I tested this hypothesis primarily by evaluating the linear relationship between group size and the relative size of the neocortex using least-squares regression. Regression analyses were conducted on species-level data, as well as on independent contrasts, which control for the effects of shared ancestry. Based on the social brain hypothesis we would expect the largest relative neocortex sizes to be associated with multimale–multifemale mating systems because these systems are typically associated with the largest social groups in primates.

Materials and Methods

I focused primarily on the relationships between neocortex ratio and the intensity of male–male competition for mates, sizes of social groups, and mating systems for male primates. The neocortex ratio is a measure of the size of the neocortex relative to the rest of the brain (Dunbar 1998). Data on neocortex ratios were gathered from the literature (Kudo and Dunbar 2001). The data presented in the literature do not distinguish between males and females; in other words, it has been assumed by researchers that the differences in brain sizes between males and females are minimal. Published values were not available for all species included in the analysis. For those species, neocortex ratios were estimated by regressing log10-transformed values of neocortex ratio presented in Kudo and Dunbar (2001) on log10-transformed brain weight published by Harvey et al. (1987; see Appendix I). The predicted values for neocortex ratios were then compared to the original published values using least-squares regression. The results of that comparison indicated that predicted values were very close approximations of the published values (r = 0.948, R2 = 0.969).

Data on body masses of males and females; group size; social, or grooming, clique size; and type of mating system were taken from the literature (Harcourt et al. 1981; Harvey et al. 1987; Kudo and Dunbar 2001). Within the order Primates, single-male mating systems typically comprise 1 adult male controlling the reproduction of multiple adult females. Multimale–multifemale mating systems are composed of multiple adult females copulating promiscuously with multiple adult males. Generally speaking, monogamous mating systems are composed of 1 adult female and her pair-bonded male. However, some species of the subfamily Callitricinae form polyandrous social groupings but mate monogamously. For the purpose of this study, humans were considered to be largely monogamous, although there is likely considerable variability among human populations in mating practices. Even within human populations for which mating practices are nominally monogamous, extrapair copulations may be common.

Sizes of social groups and social cliques were used as proxy measures of social complexity. Social cliques are composed of primary grooming partners within a larger social group (Kudo and Dunbar 2001). The availability of published data on sizes of social groups and cliques was limited. Estimates of sexual dimorphism, used as a measure of the intensity of male–male competition for mates (Plavcan and Van Schaik 1997), were calculated as the log10-transformed ratio of the body mass of males to that of females. In addition to sexual dimorphism, published estimates of the operational sex ratio (Mitani et al. 1996) also were used as a measure of the intensity of male-male competition for mates.

Like all mammals, brain size scales allometrically with body size in primates (Rilling 2006). Therefore, it was necessary to determine if a similar scaling relationship exists between neocortex ratio and body size in anthropoid primates. Similar to total brain size, the neocortex ratio of primates does appear to scale closely with body mass of males, with 89% of the variation in neocortex ratio explained by variation in body weight of males (Fig. 1a). The same condition is observed for females, with approximately 88.9% of the variation in neocortex ratio attributable to variation in body mass. The confounding effects of body mass on the evaluation of the relationship between neocortex size and social complexity also were apparent when body mass and group size were included as independent variables in a multiple regression model predicting neocortex size. Combined, these independent variables represented a significant predictor of log10-transformed neocortex ratios in primates (F = 22.455, d.f. = 3, P = 0.0003, R2 = 0.833). Separately, however, although body mass was a significant predictor (F = 43.344, d.f. = 1, P = 0.0001), explaining 80.4% of the variation in neocortex ratios, group size was not (F = 0.232, d.f. = 1,P = 0.642), explaining only 0.43% of the variation in neocortex ratios.

Fig. 1

Bivariate plots describing a) regression of the log10 values of neocortex ratio on the log10 values of male body mass, b) regression of neocortex residuals on log10 values of male body mass, c) regression of neocortex residuals on log10 values of sexual dimorphism, d) regression of neocortex residuals on log10 values of the operational sex ratio, e) regression of neocortex residuals on log10 values of group size, and f) regression of neocortex residuals on log10 values of social clique size. Species with multimale–multifemale mating systems are represented by gray-filled circles, whereas species with single-male systems are represented by open circles. Monogamous species are represented by black-filled circles and are labeled.

Clearly, body mass is a significant confounder of any assessment of the relationship between neocortex ratio and size of the social group. Consequently, to account for the confounding effects of variation in body size, relative neocortex ratio was estimated using the residuals from the least-squares regression of raw log10-transformed neocortex ratio on log10-transformed the body mass of males.

Statistical analyses.—Analyses of the relationships between relative neocortex ratio and the intensity of male–male competition and social complexity among primate taxa comprised both phylogenetically based and nonphylogenetically based comparative methods. Nonphylogenetically based procedures examining the relationships between relative neocortex size and sizes of groups and cliques, as well as the relationships between relative neocortex ratio and measures of male–male competition for mates including the level of sexual dimorphism and the operational sex ratio, consisted of least-squares linear regression. The relationships between neocortex ratio and body mass and group size were assessed using multiple regression. The model residuals from the least-squares regression of neocortex ratio on male body mass were compared among mating systems using a standard analysis of variance (ANOVA). Because humans are unique relative to other primates by having a very large neocortex, some tests were conducted with humans both included and excluded from the analysis. This allowed for a more generalized picture of the relationship between mating system and relative neocortex size within the order Primates. All least-squares regression parameters including slope, intercept, and the coefficient of determination were estimated by resampling with replacement (bootstrapping) using the NCSS statistical software package (Hintze 2004). Variables were log10-transformed before analysis, and an arbitrary 2-tailed significance level of α = 0.05 was used for all tests.

The effects of shared evolutionary history on the relationships among measures of relative neocortex size, the intensity of male–male competition, and social complexity were assessed using bootstrapped regression of phylogenetically independent contrasts (Felsenstein 1985). Independent contrasts were estimated using the primate phylogeny presented in Purvis (1995) with all branch lengths set to 1. The evolutionary relationship among variables was assessed using Pearson correlation coefficients and least-squares regression through the origin. Analyses of independent contrasts were conducted with the PDAP:PDTREE (Garland et al. 1992) module of Mesquite 1.06 (Maddison and Maddison 2005).

Results

A comparison of residuals from the regression of neocortex ratio on body mass indicated that humans fell outside the 95% confidence intervals of the variation among primates in relative neocortex size (Fig. 1b). The hominoids that are considered to be largely monogamous, that is, humans and gibbons (family Hylobatidae) exhibited the largest relative neocortex sizes ( = 0.0639), whereas the primate species with multimale-multifemale mating systems exhibited the next highest level of relative neocortex size ( = 0.0018). However, the remaining monogamous primates species, that is, the owl monkey (Aotus trivirgatus) and 2 callitrichine species (Callithrix jacchus and Saguinus oedipis), exhibited comparatively low relative neo-cortex ratios ( = −0.0169). Primate species exhibiting single-male mating systems had the lowest mean residual among mating systems (−0.0167).

Relative neocortex ratio exhibited a significant negative linear relationship (P = 0.039) with the intensity of male–male competition for mates as measured by the level of sexual dimorphism (Fig. 1c). A similar result was obtained when humans were excluded from the analysis (P = 0.026). The relationship between relative neocortex size and the operational sex ratio was not statistically significant (cf. Fig. 1d). Similarly, relationships between relative neocortex size and size of the social group, and between relative neocortex ratio and size of the social clique, also were not statistically significant (Figs, 1e and 1f).

The results from the ANOVA indicated that there was a significant difference in relative neocortex ratios among mating systems, with monogamous species (n = 6) exhibiting a greater average relative neocortex ratio than species exhibiting multimale–multifemale mating systems (n = 15), as well as species with single-male mating systems (n = 10). However, the ANOVA was not significant when humans were removed from the analysis (Figs. 2a and 2b).

Fig. 2

Error-bar charts describing the distribution of neocortex residuals across mating systems including a) and excluding b) humans. Error bars equal 1 standard error of the mean. Mating systems are labeled as SM = single-male; M = multimale–multifemale; and MG = monogamous.

The results of the phylogenetically based comparative method were largely consistent with the results from the previous comparisons. The least-squares regression analysis of independent contrasts indicated a significant negative relationship between sexual dimorphism and relative neocortex ratio (Table 1). All other relationships based on independent contrasts were not statistically significant, including the relationship between the measures of social complexity and relative neocortex ratio. However, the lack of statistical significance may be a product of small sample sizes for some tests.

View this table:
Table 1

Least-squares regression analysis of independent contrasts with relative neocortex size as the dependent variable. Significant probabilities are given in bold type.

VariablesNo. contrastsSlopeR2RP
Humans included in the analysis
Dimorphism30−0.13310.1934−0.42540.0166
Operational sexratio90.08320.28000.30310.1784
Group size12−0.01150.1946−0.30090.2957
Clique size10−0.00260.1357−0.16710.5210
Humans excluded from the analysis
Dimorphism29−0.09600.2212−0.44940.0115
Operational sexratio90.04670.23700.32090.2383
Group size120.00810.12600.04710.9793
Clique size10−0.00260.1357−0.16710.5210

Discussion

The lack of a positive linear relationship between relative neocortex ratio and measures of social complexity, and the observation that the largest relative neocortex sizes are found among monogamous primates, would initially seem to be in-consistent with the social brain hypothesis. The finding of a significant inverse relationship between relative neocortex ratio and the intensity of male–male competition as reflected in the level of sexual dimorphism is difficult to explain. Research on sexual selection and brain evolution in bats has demonstrated that high levels of sperm competition constrain brain-size evolution because of an energetic trade-off between metabolically costly testes and brain tissues (Pitnick et al. 2005). Therefore, the intensity of sexual selection resulted in smaller brain size in some bat species. However, recent research on primates found no relationship between the level of sperm competition and brain size (Schillaci 2006). These studies of sexual selection and evolution of brain size in bats and primates do not support a sexual conflict hypothesis that states that both males and females are under selection to subvert the reproductive investment made by the other sex. The sexual conflict hypothesis predicts species with promiscuous matings with have larger relative brain sizes than monogamous species (see discussion in Pitnick et al. 2005). In primates at least, overall brain size relative to body size seems to be larger in species with lower levels of male–male competition, particularly in monogamous species.

The results presented here indicate, similar to overall brain size (Schillaci 2006), that relative neocortex size also exhibits a negative linear relationship with reproductive competition among males as measured by the level of sexual dimorphism in body mass. A closer evaluation of the regression plots reveals that monogamous gibbons are driving much of the observed negative relationship between relative neocortex ratios and sexual dimorphism, as well as the similarly negative relationship between relative neocortex ratio and size of the social group. It is important to note that there are no available data published on sizes of social cliques for gibbons, which would be predicted to be very small. The relationship between sexual dimorphism and relative neocortex ratios may be driven to some extent by collinearity between dimorphism and some unidentified covariate of monogamy, such as length of the period spent as a juvenile (see below), body mass of females, substrate, or reproductive rates (see Plavcan and Van Schaik 1997).

There are several potential explanations for the observed distribution of relative neocortex size across mating systems. First, apes, including monogamous apes, have longer periods of maturation with possibly more biparental investment than primates with multimale–multifemale mating systems. As mentioned earlier, a longer period of growth may contribute to the evolution of larger brain size, including size of parts of the brain such as the neocortex. Recent research has substantiated the notion that longer periods spent as a juvenile are associated with increased brain sizes and brain-part sizes in primates (Barton 1999; Deaner et al. 2003; Kaplan et al. 2003; Leigh 2004; Walker et al. 2006). Second, it is also possible that monogamous mating systems require an intense and intricate social relationship between mates. Such relationships in primates may require greater social acuity and abilities for deception and manipulation. This requirement would promote selection for a larger neocortex in both sexes. If this explanation is accurate then the evolution of brain size in primates is, in fact, tied to social complexity, and the observed distribution of relative neocortex size across mating systems is consistent with the social brain hypothesis. However, sizes of social groups may not be the best measure of social complexity when body mass is accounted for.

Acknowledgments

The study was supported by a grant from the Connaught Fund of the University of Toronto. This research benefited from the comments of 2 anonymous reviewers, and discussions with R. Walker. Responsibility for any errors, omissions, and interpretations lies solely with the author.

Appendix I

View this table:

Data on mating system, body mass, estimated neocortex ratio, and mass dimorphism.

TaxonMating systemaBody massbNeocortexcDimorphism
Callithrix jacchusMG0.320.160268280.0290
Saguinus oedipusMG0.520.17163380−0.0174
Saimiri sciureusM0.780.258587950.1116
Aotus trivirgatusMG1.020.23302882−0.0362
Lagothrix lagotrichaM5.220.378372670.0691
Alouatta palliataM7.260.329605490.1134
Ateles geoffroyiM7.940.390589190.0290
Chlorocebus aethiopsM4.950.336742050.1204
Macaca fascicularisM4.420.349470560.1581
Macaca radiataM8.650.358555510.2514
Macaca mulattaM9.200.377188950.3153
Macaca nemestrinaM9.980.386649340.1249
Macaca arctoidesM10.510.385072430.0607
Papio hamadryasSM20.170.412447470.3593
Papio cynocephalusM24.320.427368940.1249
Papio anubisM26.400.430408800.2430
Papio ursinusM31.750.448062440.0843
Papio papioM31.980.425387390.3010
Theropithecus geladaSM20.400.405708300.1782
Trachypithecus rubicundaSM6.230.374960470.0000
Trachypithecus cristataSM6.580.342659910.0260
Trachypithecus obscuraSM7.450.347431070.1062
Trachypithecus entellusSM17.000.407862720.2079
Colobus polykomosSM10.250.358441920.0928
Nasalis larvatusSM20.640.376359930.3119
Hylobates molochMG5.440.392763090.0223
Hylobates larMG5.500.388036490.0316
Pan troglodytesM44.340.504649080.1263
Pongo pygmaeusSM74.640.505284230.2706
Gorilla gorillaSM169.000.522909900.2356
Homo sapiensMG65.000.601773450.0772
  • a MG = monogamous; M = multimale–multifemale; SM = single maie.

  • b Data on body mass (kg) of males taken from Plavcan and Van Schaik (1997).

  • c Log10 value of neocortex ratio.

Footnotes

  • Associate Editor was Craig L. Frank.

Literature Cited

View Abstract