Females drive primate
Patrik Lindenfors1*, Laila Fro ¨berg2
and Charles L. Nunn3
1Department of Biology, University of Virginia, Gilmer Hall,
Charlottesville, VA 22903, USA
2Department of Zoology, Stockholm University, 106 91 Stockholm,
3Section of Evolution and Ecology, University of California, Davis,
CA 95616, USA
*Author and address for correspondence: Department of Zoology,
Stockholm University, 106 91 Stockholm, Sweden
Recd 04.07.03; Accptd 01.10.03; Online 12.11.03
Within and across species of primates, the number
of males in primate groups is correlated with the
number of females. This correlation may arise
owing to ecological forces operating on females, with
subsequent competition among males for access to
between changes in male and female group member-
ship remains unexplored in primates and other
mammalian groups. We used a phylogenetic com-
parative method for detecting evolutionary lag to
test whether evolutionary change in the number of
males lags behind change in the number of females.
We found that change in male membership in pri-
mate groups is positively correlated with divergence
time in pairwise comparisons. This result is consist-
ent with male numbers adjusting to female group
size and highlights the importance of focusing on
females when studying primate social evolution.
Keywords: socioecological model; comparative methods;
evolutionary lag; primate behaviour
A fundamental question in social evolution concerns the
factors that influence group size and composition. What
factors influence variation in group size across species, and
how do these factors affect the numbers of adult males
and females within groups? These questions have been
addressed by focusing on the ecological variables that
determine the spatio-temporal distribution of females,
based on the expectation that resources and predation
account for variation in female reproductive success. By
comparison, access to females is generally the major factor
influencing male reproductive success (Trivers 1972;
Wilson 1975; Emlen & Oring 1977). Thus, after risks and
resources determine the spatio-temporal distribution of
females, the distribution of females is expected to influ-
ence the degree of male intrasexual competition (Emlen &
Oring 1977). The general framework is therefore one of
female-driven social evolution.
Comparative tests of this general framework have been
conducted across species of primates, with results showing
that the number of males in primate groups correlates with
the number of females (Andelman 1986; Ridley 1986;
Dunbar 1988; Altmann 1990; Mitani et al. 1996; Nunn
1999). A common interpretation of this pattern is that as
Proc. R. Soc. Lond. B (Suppl.) 271, S101–S103 (2004)
2003 The Royal Society
female group size increases owing to the ecological advan-
tages of group-living, it becomes more difficult for a single
male to defend access to the group of females, resulting
in multimale–multifemale groups. Recently, it has been
shown that additional variation in male membership can
be accounted for by female mating behaviour; species in
which females mate more synchronously exhibit relatively
more males in the group (Nunn 1999). This effect is
expected if it is more difficult for a single male to mono-
polize access when two or more females are mating simul-
taneously in the group.
These comparative tests, along with similar tests within
species (Altmann 2000), support one prediction of models
based on female-driven social evolution: the number of
males in primate groups varies according to the degree of
competition over females (sexual selection). Thus, males
are adjusting to variation in female spatial and temporal
availability, or, as stated by Altmann (1990), ‘primate
males go where the females are’ (p. 193).
In this paper, we test another prediction of female-
driven social evolution that has yet to be investigated
empirically and focuses on the causal links between males
and females. Specifically, female-driven social evolution
predicts sequential effects, with the number of females
influencing the number of males in primate groups. Thus,
evolutionary changes in the number of males are expected
to lag behind changes in the number of females. When
investigating patterns across species, at least three factors
may cause the number of males (male group size) to lag
behind the number of females (female group size). First,
changes in the number of males may be sensitive to female
sexual behaviour, such as oestrous synchrony, which may
not evolve instantaneously as female group size increases.
Selection for such adjustments in female sexual behaviour
may result from variation in the degree of predation or
infanticide risk (van Schaik & Ho ¨rstermann 1994; van
Schaik & Janson 2000). Second, individuals may require
time to evolve defences to infectious disease risk that are
expected to increase in multimale–multifemale groups,
including risks of acquiring socially or sexually transmitted
diseases (Freeland 1976; Møller et al. 1993; Nunn et al.
2000; Nunn & Altizer 2004). Finally, cognitive constraints
may limit the number of individuals in groups (Dunbar
1988). For example, male alliances in multimale primate
groups require a degree of cooperation that is unlikely to
exist in strictly polygynous systems (Sommer 1988).
2. MATERIAL AND METHODS
To test the hypothesis that male group size lags behind female
group size in primate groups, we acquired data on the number of
males and the number of females in groups, and primate phylogeny.
Data on group composition were taken from Nunn & Barton (2000),
to which we added data on solitary strepsirhines and tarsiers from
Kappeler & Heymann (1996). Group sizes were log-transformed
prior to analyses. Analyses were carried out using both the full dataset
and a dataset with solitary species removed, as evolutionary analyses
of social animals should provide the strongest tests of the hypothesis.
Phylogenetic information was taken from Purvis (1995), which is a
dated supertree synthesized using both morphological and molecu-
To test the lag hypothesis, we used a phylogenetic method
described in Deaner & Nunn (1999), which was originally used to
test whether primate brain size lags behind body size. The method
is implemented by first calculating unstandardized independent con-
trasts (Felsenstein 1985) in female number (?Xn) and male number
(?Yn) using only the tips of the phylogeny. Contrasts are constructed
such that ?Xnis forced to be positive, based on the fact that the
direction of subtraction is arbitrary for independent contrasts
(Garland et al. 1992). The second step calculates residuals from a
regression through the origin of ?Ynagainst ?Xn. The final step
involves regressing these residuals on time since the taxa diverged.
S102P. Lindenfors and others
Females drive primate social evolution
log (female group size) contrasts
log (male group size) contrasts
0.9 0.8 0.7 0.6 0.50.4 0.3 0.20.10
log (female group size) contrasts
Figure 1. The relationship between male group size and female group size in primates. Plots show independent contrasts
calculated at the tips of the tree only, with the least-squares regression line forced through the origin. (a) Using the full
dataset, (b) solitary species excluded. See § 3 for statistical results.
Evolutionary lag is indicated by larger residuals on contrasts
characterized by longer branches, which indicates that relative to
?Xn, ?Ynincreases when there is more time available for change to
take place. In other words, we are testing whether male group size
is larger than expected on longer branches and smaller than expected
on shorter branches. We used only pairwise contrasts calculated using
extant trait values because the method relies on accurate estimates
of time since divergence and the amount of evolutionary change since
two species last shared a common ancestor. Following Deaner &
Nunn (1999), we also checked whether more change occurs on
longer branches, but found no significant relationship between pair-
wise contrasts in female group size and divergence dates (b = 0.009,
R2= 0.020, n = 37, p = 0.407; after excluding solitary species:
b = 0.012, R2= 0.012, n = 27, p = 0.593). Further information on the
method and its application is available in Deaner & Nunn (1999).
All statistical results reported are from two-tailed tests.
Groups sizes of males and females were significantly
correlated, both when using the full dataset (b = 0.506,
R2= 0.506, n = 37, p ? 0.001) and when excluding soli-
tary species (b = 0.508, R2= 0.359, n = 28, p ? 0.001)
(figure 1). To test the lag hypothesis, we calculated
residuals from the regression of male number on female
number. Residual male group size scaled positively with
the divergence dates of pairwise comparisons, both when
using all available data ( y = 0.017x ? 0.064, R2= 0.167,
n = 37, p = 0.012) and when excluding solitary species
( y = 0.025x ? 0.069,
These results are consistent with the hypothesis that male
group size lags behind female group size (figure 2).
The results can be summarized as occurring due to
reduced variation in male group size on shorter branches
of the primate tree, leading to negative residuals on the
left-hand sides of figure 2a,b and positive residuals on the
right-hand sides of these figures. This pattern could arise
if comparisons among two unimale–multifemale primate
species, as compared with pair-living, unifemale–multi-
male or multimale–multifemale species, are characterized
by shorter branch lengths. Specifically, contrasts involving
male group size among two unimale–multifemale sister
species would be expected to be small, owing to the fact
that these species by definition have a similar number of
males, whereas contrasts in female group size would exhi-
bit a greater degree of variation and would be larger. On
the contrasts plot (figure 1), unimale–multifemale species
n = 28,
p = 0.025).
Proc. R. Soc. Lond. B (Suppl.)
therefore would be more likely to have negative residuals.
If this alternative explanation accounts for the results in
figure 2, then contrasts linking unimale–multifemale spec-
ies should have shorter branches than other contrasts used
in our analyses. Out of the 37 contrasts available in the
complete dataset, there were only nine contrasts linking
two unimale–multifemale species. We found, however, no
statistically significant differences between the branch
lengths involving unimale–multifemale species as related
to other contrasts in the dataset. In addition, we repeated
the analyses after excluding these nine pairwise compari-
sons. Residual male group size scaled positively with diver-
gence date when using this subset of all available data
( y = 0.018x ? 0.060, R2= 0.193, n = 28, p = 0.019), with
results approaching significance after further excluding the
solitary species ( y = 0.021x ? 0.057, R2= 0.128, n = 25,
p = 0.078).
In this study we confirmed previous results (Andelman
1986; Dunbar 1988; Altmann 1990; Mitani et al. 1996;
Nunn 1999) that male and female group sizes are tightly
correlated across species of primates. Our results also sup-
port predictions of the lag hypothesis: change in the num-
ber of males lags behind change in the number of females.
These results point to females as the driving sex in primate
social evolution, with female group size changing first and
male group size subsequently adjusting to female number.
Alternatives to the female-driven model, or variations
on this general model, have been proposed. For example,
females may seek protection from predation, sexual
coercion or infanticide by forming associations with males
(Wrangham 1979; Smuts & Smuts 1993; van Schaik
1996; van Schaik & Kappeler 1997). Such hypotheses
require the opposite pattern to that found here, with
female group size predicted to lag behind male group size.
We cannot rule out such effects in individual species and
there is much variation left unexplained (figure 2). Our
results nevertheless indicate that, in general, female-driven
social evolution accounts for significant variation in pat-
terns of group membership across primates.
The ecological model proposed by Emlen & Oring
(1977) is expected to account for variation both across
Females drive primate social evolution
P. Lindenfors and othersS103
divergence date (MYA)
residual male group size
16 14 1210 86420–2
divergence date (MYA)
16 1412 1086420–2
Figure 2. Evolutionary lag in male group size. Residual male group size (see figure 1) correlates with divergence date,
indicating that male group size lags behind female group size evolution. (a) Using the full dataset, (b) solitary species
excluded. Plots show regression lines with 95% confidence intervals (dashed lines). See § 3 for statistical results.
and within species, and previous work on primates has
shown that intraspecific variation also supports the general
prediction of the female-driven model, namely that male
and female numbers are correlated (Altmann 2000).
Because the processes that produce these patterns operate
within species, it may be surprising that lag would exist
across species. As noted in § 1, however, several factors
may account for this pattern at the interspecific level,
including the time required for males and females to
evolve sexual, behavioural and physiological adaptations
to living in groups of different size and composition.
In conclusion, these results highlight the importance of
focusing on females when studying primate social evol-
ution. It is female group size that apparently responds to
natural selection, and it is consequently their group size
that holds the key to understanding what drives social
evolution in the first place. Future research should con-
tinue to integrate the patterns linking male and female
group size to the ecological variation that ultimately
underlies cross-species variation in primate sociality.
The authors thank B. S. Tullberg and two anonymous reviewers for
comments on the manuscript. This work was financed by the tax-
payers of Sweden (through P.L. and L.F.).
Altmann, J. 1990 Primate males go where the females are. Anim.
Behav. 39, 193–195.
Altmann, J. 2000 Models of outcome and process: predicting the
number of males in primate groups. In Primate males: causes and
consequences of variation in group composition (ed. P. M. Kappeler),
pp. 236–247. Cambridge University Press.
Andelman, S. 1986 Ecological and social determinants of cercopi-
thecine mating patterns. In Ecological aspects of social evolution (ed.
D. I. Rubenstein & R. W. Wrangham), pp. 201–216. Princeton
Deaner, R. O. & Nunn, C. L. 1999 How quickly do brains catch up
with bodies? A comparative method for detecting evolutionary lag.
Proc. R. Soc. Lond. B 266, 687–694. (DOI 10.1098/rspb.1999.0690.)
Dunbar, R. I. M. 1988 Primate social systems. Ithaca, NY: Cornell
Emlen, S. T. & Oring, L. W. 1977 Ecology, sexual selection, and the
evolution of mating systems. Science 197, 215–223.
Felsenstein, J. 1985 Phylogenies and the comparative method. Am.
Nat. 125, 1–15.
Freeland, W. J. 1976 Pathogens and the evolution of primate social-
ity. Biotropica 8, 12–24.
Proc. R. Soc. Lond. B (Suppl.)
Garland Jr, T., Harvey, P. H. & Ives, A. R. 1992 Procedures for the
analysis of comparative data using phylogenetically independent
contrasts. Syst. Biol. 41, 18–31.
Kappeler, P. M. & Heymann, E. W. 1996 Nonconvergence in the
evolution of primate life history and socio-ecology. Biol. J. Linn.
Soc. 59, 297–326.
Mitani, J. C., Gros-Louis, J. & Manson, J. H. 1996 Number of males
in primate groups: comparative tests of competing hypotheses. Am.
J. Primatol. 38, 315–332.
Møller, A. P., Dufva, R. & Allander, K. 1993 Parasites and the evol-
ution of host social behavior. Adv. Study Behav. 22, 65–102.
Nunn, C. L. 1999 The number of males in primate social groups: a
comparative test of the socioecological model. Behav. Ecol.
Sociobiol. 46, 1–13.
Nunn, C. L. & Altizer, S. 2004 Sexual selection, behavior and sex-
ually transmitted diseases. In Sexual selection in primates: new and
comparative perspectives (ed. P. M. Kappeler & C. P. van Schaik).
Cambridge University Press. (In the press.).
Nunn, C. L. & Barton, R. A. 2000 Allometric slopes and inde-
pendent contrasts: a comparative test of Kleiber’s law in primate
ranging patterns. Am. Nat. 156, 519–533.
Nunn, C. L., Gittleman, J. L. & Antonovics, J. 2000 Promiscuity and
the primate immune system. Science 290, 1168–1170.
Purvis, A. 1995 A composite estimate of the primate phylogeny. Phil.
Trans. R. Soc. Lond. B 348, 405–421.
Ridley, M. 1986 The number of males in a primate troop. Anim.
Behav. 34, 1848–1858.
Smuts, B. B. & Smuts, R. W. 1993 Male aggression and sexual
coercion of females in nonhuman primates and other mammals:
evidence and theoretical implications. Adv. Stud. Behav. 22, 1–63.
Sommer, V. 1988 Male competition and coalitions in langurs
(Presbytis entellus) at Jodhpur, Rajasthan, India. Hum. Evol. 3,
Trivers, R. L. 1972 Parental investment and sexual selection. In
Sexual selection and the descent of man 1871–1971 (ed. B. Campbell),
pp. 136–179. Chicago, IL: Aldine.
van Schaik, C. P. 1996 Social evolution in primates: the role of eco-
logical factors and male behaviour. In Evolution of social behaviour
patterns in primates and man (ed. W. G. Runciman, J. Maynard
Smith & R. I. M. Dunbar), pp. 9–31. Oxford University Press.
van Schaik, C. P. & Ho ¨rstermann, M. 1994 Predation risk and the
number of adult males in a primate group: a comparative test.
Behav. Ecol. Sociobiol. 35, 261–272.
van Schaik, C. P. & Janson, C. H. (eds) 2000 Infanticide by males
and its implications. Cambridge University Press.
van Schaik, C. P. & Kappeler, P. M. 1997 Infanticide risk and the
evolution of male-female association in primates. Proc. R. Soc.
Lond. B 264, 1687–1694. (DOI 10.1098/rspb.1997.0234.)
Wilson, E. O. 1975 Sociobiology, the new synthesis. Cambridge, MA:
Wrangham, R. W. 1979 On the evolution of ape social systems. Soc.
Sci. Information 18, 334–368.