Animal Behaviour 81 (2011) 1109e1116

Contents lists available at ScienceDirect

Animal Behaviour journal homepage: www.elsevier.com/locate/anbehav

Social bonds predict future cooperation in male Barbary macaques, Macaca sylvanus Andreas Berghänel a, b, Julia Ostner a, b, Uta Schröder a, b, Oliver Schülke b, c, * a

Primate Social Evolution Group, Courant Research Centre Evolution of Social Behaviour, Georg August University Göttingen Integrative Primate Socio-Ecology Group, Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany c Courant Research Centre Evolution of Social Behaviour, Georg August University Göttingen b

a r t i c l e i n f o Article history: Received 7 July 2010 Initial acceptance 22 September 2010 Final acceptance 4 February 2011 Available online 29 March 2011 MS. number: 10-00472R Keywords: Affenberg Salem Barbary macaque coalition cooperation Macaca sylvanus social bond social relationship

Social bonds have been construed as mental representations mediating social interactions among individuals. It is problematic, however, to differentiate this mechanism from others that assume more direct exchanges and interchanges of behaviour such as reciprocity, market effects or mutualism. We used naturally occurring shifts in rates and patterns of social interactions among male Barbary macaques to test whether affiliation in the nonmating season predicts coalition formation in the mating season. We carried out 1377 h of observation of all 23 males living in a semifree-ranging bisexual group at Affenberg Salem, Germany. The mating season was characterized by significantly increased rates of aggression, male coalition formation and spontaneous submission as well as decreased rates of affiliation. Rates of coalition formation in the mating season were predicted by affiliation between these males several weeks earlier in the nonmating season after we controlled for affiliation in the mating season. We conclude that social bonding in the nonmating season may build reputations among partners that are crucial for cooperation in risk-prone agonistic coalitions against other males. From the pronounced time lag between bonding and cooperation and the fact that males largely ignored affiliation as well as agonistic patterns in the time-matched mating season, we conclude that short-term reciprocity, mutualism or market effects cannot explain these observations. Instead, we conclude that long-term reciprocity mediated by emotional book keeping may be the basic mechanism. Building on that, males may employ a mental representation of their social bonds when choosing partners for cooperative interactions. 2011 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd.

Social bonds, that is, lasting partner preferences in affiliative interactions, are a widespread phenomenon in gregarious mammals (e.g. bats: Wilkinson 1985; dolphins: Connor et al. 2001; carnivores: de Villiers et al. 2003; ungulates: Wasilewski 2003; Cameron et al. 2009; primates: Mitani 2009) and birds (Emery et al. 2007). The strength of the bond between two individuals is typically estimated from differential frequency or duration metrics for spatial proximity among individuals and grooming/preening behaviour (Silk 2002). While grooming may be exchanged for grooming or interchanged for other commodities in classical reciprocity (Hart & Hart 1992; Schino 2007) or in biological markets (Noë et al. 2005; Barrett & Henzi 2006), an alternative view proposes that grooming is an investment in a social bond (e.g. Dugatkin1997; Schülke et al. 2010; see also von Bayern et al. 2007; but see Henzi & Barrett 1999). * Correspondence: O. Schülke, Courant Research Centre Evolution of Social Behaviour, Georg August University Göttingen, Kellnerweg 6, 37077 Göttingen, Germany. E-mail address: [email protected] (O. Schülke).

Social bonds form through repeated affiliative interactions among partners. The evolutionary mechanisms proposed to maintain such bonding are kin selection, reciprocal altruism and mutualism (Clutton-Brock 2009). Intrasexual social bonds in particular are thought to be driven primarily by kin selection and are rarer between unrelated individuals because competition prevails among same-sex individuals in social groups (van Hooff & van Schaik 1994; Silk 2006; Clutton-Brock 2009; Smith et al. 2010; but see Schino & Aureli 2010). This applies especially for males, which, in contrast to females, compete for a nonshareable resource, that is, fertilizations (Trivers 1972; van Hooff & van Schaik 1994; van Schaik 1996). Nevertheless, affiliative behaviour between unrelated males can reduce agonistic tension (e.g. Judge & de Waal 1997; Paul et al. 2000), increase tolerance (e.g. Silk 2007) and is related to coalitionary support (Feh 1999; Schino 2007). Affiliation is based on reciprocal altruism if benefits for the actor are delayed, or mutualism if benefits inevitably accrue for both partners (e.g. Mitani 2006; Schülke et al. 2010). The benefits of male affiliation are assumed to accrue more or less immediately between nonkin, because temporal delays in returning

0003-3472/$38.00 2011 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. doi:10.1016/j.anbehav.2011.02.009

1110

A. Berghänel et al. / Animal Behaviour 81 (2011) 1109e1116

benefits (i.e. reciprocation) may provide cheating options to the receiver of the cooperative act (Trivers 1971; Connor 1995; CluttonBrock 2009). For the same reason, some authors have questioned the existence of social bonds as entities, in the sense of conscious or unconscious mental representations (Bachmann & Kummer 1980; Cheney & Seyfarth 1990; Dunbar & Shultz 2010), which causally mediate relations between temporally separated interactions (Hemelrijk 1996; Stevens et al. 2005; Barrett et al. 2007). The necessary delay between investment in a social bond and the return of benefits is also assumed to be constrained by the complex cognitive skills required, such as mnemonic time tracking and book keeping of interactions (Stevens et al. 2005; Barrett et al. 2007). Recent studies suggest that unrelated females may indeed delay reciprocation for a few hours or days in baboons, Papio hamadryas anubis and P. h. ursinus and mandrills, Mandrillus sphinx (Frank & Silk 2009; Schino & Pellegrini 2009; Cheney et al. 2010), macaques, Macaca fuscata and M. fascicularis (Schino et al. 2007; Massen 2010), chimpanzees, Pan troglodytes (Gomes et al. 2009), capuchins, Cebus apella (Schino et al. 2009) and rats, Rattus norvegicus (Rutte & Taborsky 2008; see also Schino & Aureli 2010). However, most of these studies were correlational, and did not investigate temporal contingency or include a time-specific component in the actual analyses, but pooled data across varying periods instead. Thus it remains unclear whether reciprocation may be mediated by social bonds over several weeks or months, especially among unrelated males. In this study, we used the pronounced mating seasonality of Barbary macaques to assess the mediating role of long-term social bonds between unrelated males in agonistic coalition formation, thus combining the correlational and the temporal contingency approach that takes the temporal sequence of events into account (e.g. Schino & Aureli 2009). The degree of despotism in social relationships, and thus the importance of the dominance hierarchy, and levels of aggressiveness are expected to be most pronounced in the mating season (e.g. Eaton et al. 1981; Sands & Creel 2004; Ostner et al. 2011) corresponding to the presence of the males’ key resource, fertile females (Trivers 1972; van Hooff & van Schaik 1994). Consequently, coalition formation should be most beneficial and frequent during the mating season, since mating success of males usually depends on direct coalitionary support in sexual contexts (e.g. Noë 1992; Feh 1999; Connor et al. 2001; Bissonnette et al. 2011) or rank position (de Ruiter & van Hoof 1993; Alberts et al. 2003), which again is often associated with the coalitionary support on which a male can rely (de Villiers et al. 2003; Schülke

(a)

et al. 2010). At the same time, increased rates of aggression during the mating season lead to increased spatial avoidance and stronger rank effects in behaviour and thus to reduced possibilities for affiliative interactions between males, which restricts investments in the social bond largely to the nonmating season (e.g. Horiuchi 2005; Cordoni 2009). Consequently, affiliative interactions that establish or maintain male social bonds in the nonmating season have been predicted to result in coalition formation only several weeks or months later during the mating season (Paul et al. 1996, 2000). In one scenario, the pronounced affiliative behaviour of the nonmating season, but not the constrained affiliation of the mating season, is predicted to drive coalitionary support during the mating season (Fig. 1a), which would render strong support for the idea that bonds form to establish reputations for future cooperation. If affiliative interactions are viewed as cooperative acts then partners can gain reputations as cooperators over several iterations and these reputations may bias future partner choice. Alternatively, coalition formation in the mating season may be related ‘just-in-time’ (Barrett et al. 2007) to this season’s affiliative activity despite the suboptimal conditions for affiliation (Fig. 1b). In the latter case, coalition formation will just reflect short-term mutualism involving immediate direct benefits or symmetry-based reciprocity (de Waal & Luttrell 1988; CluttonBrock 2009). Long-term social bonds will then emerge as an epiphenomenon of persistent associations over time without any need for higher cognitive skills such as time tracking, mental representations or mediator roles of these bonds (de Waal & Luttrell 1988; Hemelrijk 1996; Stevens et al. 2005; Barrett et al. 2007; Clutton-Brock 2009). In this study, we aimed to distinguish between these two possibilities by investigating social bonding and coalition formation in male Barbary macaques. Unlike other primates, adult Barbary macaque males groom each other only rarely but frequently engage in triadic interactions involving two males and one infant (for details see Deag 1980; Paul et al. 1996) that may be characterized as affiliative (Paul et al. 1996). Almost all body contacts, grooming interactions and friendly approaches between males are preceded by triadic interactions (Paul et al. 1996, 2000). Triadic interactions have been proposed (1) to function as agonistic buffering mechanisms facilitating male affiliation (Deag 1980; see also Zhao 1996 for Tibetan macaques, Macaca thibetana), (2) to represent male mating effort and to increase access to mothers in the future (Ménard et al. 2001; but see Paul et al. 1996), and (3) to serve to establish coalitionary

A

(b)

Affiliative Interactions MS

Affiliative M Interactions NMS

A

Coalitionary support MS B

M Affiliative Interactions MS

Affiliative Interactions NMS

Coalitionary support MS

B

Figure 1. Causal step approach to assess mediatorship (Baron & Kenny 1986). Two distinct causal pathways are conceivable regarding the three variables. (a) Affiliative interactions during the nonmating season (NMS) predict both affiliative interactions and coalition formation during the mating season (MS) and thus mediate a possible positive correlation between the latter two. (b) Affiliative interactions of the nonmating season predict only affiliative interactions of the mating season by continual association, which in turn predicts coalition formation of the mating season immediately and thus mediates a possible positive correlation between affiliative interactions during the nonmating season and coalition formation during the mating season. Variables are A independent, B dependent and M mediating according to the causal step approach (see Methods)

A. Berghänel et al. / Animal Behaviour 81 (2011) 1109e1116

relationships between males (Paul et al. 1996, 2000). We investigated the previously untested third explanation. The overall aim of the study was to investigate the mediating role of social bonds for high-risk cooperation in agonistic coalitions among males. Specifically, we aimed to evaluate the basic assumptions that rates of dyadic and coalitionary aggression are increased during the mating season compared to the nonmating season. As a consequence of increased conflict avoidance among males, the rate of affiliative behaviour (namely triads, body contact, grooming and spatial proximity) was predicted to decrease during the mating season, hence restricting affiliation largely to the nonmating season. Building on that, we proposed that patterns of affiliative interactions (i.e. triadic interactions) during the nonmating season predicted patterns of coalition formation during the subsequent mating season on the dyadic level (see Fig. 1). Assuming that males eavesdrop on others’ social interactions (e.g. Earley 2010), we predicted that if cooperation was based on reputation, males with high affiliation rates in the nonmating season would have more coalition partners, form more coalitions and receive more agonistic support in the mating season. METHODS Study Group We collected data on one of three semifree-ranging groups of Barbary macaques which share a 14.5 ha enclosure at Affenberg Salem, Germany (for a history of the population see de Turckheim & Merz 1984). Data were collected during one nonmating season (4 June 2006 to 27 September 2008) and one mating season (27 October 2008 to 23 January 2009). During the study period, the study group (H) consisted of 67 animals (23 adult males, 31 adult females, three subadult males, 10 immatures) of known ages. Only five adult males were born into the study group (information on age and group history: Affenberg Salem, unpublished data). As a result of male dispersal (Scheffrahn et al. 1993), only 3.6% of all possible male dyads (N ¼ 9 of 253) show a coefficient of maternal relatedness 0.25 (r ¼ 0.25: N ¼ 7; r ¼ 0.5: N ¼ 2; Affenberg Salem, unpublished data,). Paternal kinship is not recognized by Barbary macaques (Kuester et al. 1994). Hence, possible effects of kinship in our study were negligible. Behavioural Data Recording We recorded data using focal animal observations each 45 min long with continuous recording (Altmann 1974). All adult animals were individually known, and all 23 adult males were chosen as focal animals (nonmating season: 471 h, 20.5  1.8 h/male; mating season: 906 h, 39.4  0.6 h/male). The continuous protocol included sexual interactions, that is, copulations, affiliative and agonistic interactions. Affiliative interactions included close proximity (<1.5 m), body contact, grooming and triads (i.e. interactions including two males and an infant in body contact; see Introduction, Deag 1980). Interactions were agonistic whenever they included at least one aggressive (i.e. bite, chase, ground slap, lunge, open mouth, point, slap, stare, pushepull) or one submissive (crouch, flee, fear scream, give ground, make room) behavioural element. Additionally, ad libitum data on agonistic and sexual interactions were recorded. Data Analysis Interaction rates were calculated as the individual number of interactions/h of focal protocol observation and male. Rates were calculated for triadic interactions (nonmating season: N ¼ 450; mating season: N ¼ 354), sitting in body contact for more than 10 s

1111

(nonmating season: N ¼ 309; mating season: N ¼ 53), grooming (nonmating season: N ¼ 95; mating season: N ¼ 15) and staying in close proximity of less than 1.5 m for more than 10 s (nonmating season: N ¼ 616; mating season: N ¼ 456). Only one-third of triadic interactions were initiated directly by the infant-handling male. In the remaining cases, the approaching male was the initiator of the triadic interaction. Rates of agonistic interactions (individual aggression, i.e. attacks or threats to another individual) were calculated as the individual number of interaction bouts/h. An interaction bout was defined as a series of behavioural acts that were not separated by more than 10 s, and were either dyadic (i.e. involving two partners only; aggressive interaction bouts: nonmating season: N ¼ 119; mating season: N ¼ 202) or polyadic (see below). Cumulative aggression rate accounts for both dyadic and polyadic aggressive interactions. Unprovoked submission indicating spontaneous avoidance was defined as submission without any preceding or simultaneous aggression by the other male within the same bout (nonmating season: N ¼ 49; mating season: N ¼ 460). Following de Waal & Harcourt (1992), we defined coalitions as joint aggression of two or more males against a common target. The analysis of coalitionary aggression was based on conflicts involving males only (mating season: N ¼ 189 coalitions of various sizes; N ¼ 155 coalitions of two partners against one target) and excluding scream fights (see Bissonnette 2009). Coalitions during the nonmating season were ignored in the correlation analysis owing to their low number (N ¼ 5). If not stated otherwise, analyses were based on two-against-one coalitions. Directional coalitionary support was defined conservatively as supportive third-party interventions in ongoing conflicts (N ¼ 70 of the 155 two-againstone coalitions). To investigate the relationships between affiliative behaviour during the nonmating season and the mating season and coalition formation through the mating season and to assess mediatorship (see Fig. 1), we carried out a ‘causal step approach’ following Baron & Kenny (1986; reviewed in MacKinnon 2008) using row-wise and partial row-wise matrix correlations of symmetric (i.e. nondirectional) interaction matrices (de Vries 1993; tests for significance based on 10 000 permutations). Accordingly, mediation requires positive correlations of all of the three variables with each other (MacKinnon 2008) as well as that the correlation between the mediator variable M and the dependent variable B prevails after controlling for the independent variable A. Furthermore, partial mediation by M is indicated by a weakened correlation between A and B after controlling for M, while a total loss of correlation indicates complete mediation by M, and unmodified or increased correlation indicates no mediation by M (A, B and M refer to Fig. 1; for details see MacKinnon 2008). Since grooming events were extremely rare between males (see above) we based our analysis of affiliative behaviour on the frequencies of triadic interactions and time spent in close proximity, two important behaviours regarding discussions of shortterm mutualism (Hemelrijk 1996). Despite more focal hours during the mating season (see above), the total number of observed triadic interactions, as well as the total time spent in close proximity, was higher in the nonmating season than the mating season. In nonsaturated matrices (i.e. including false zeros) increased total counts potentially increase variance and total covariance and could therefore influence outcomes of the partial matrix correlations. To rule out such effects conclusively, we balanced the total sample sizes of the nonmating season and mating season by subsampling. We successively omitted protocols from the end of the nonmating season until the total number of triadic interactions (N ¼ 352 out of 450) matched the sample from the mating season and the total duration of time spent in close proximity in the mating-season

1112

A. Berghänel et al. / Animal Behaviour 81 (2011) 1109e1116

sample matched that in the nonmating-season sample (total time: 14.5 h versus 14.7 h, N ¼ 518 out of 616). This procedure extended the temporal gap between the nonmating season and mating season samples to 8 weeks in both cases (triads: 58; time spent in close proximity: 53 days). Coalitions in the study group were primarily formed against young immigrant males to isolate them from the group (Berghänel et al. 2010). Most coalitions were formed by older immigrant males (mean  SD ¼ 21.9  1.7 years) and natal males and mainly targeted young immigrants (mean  SD ¼ 12.2  1.2 years), yielding what seemed to be two subgroups (Berghänel et al. 2010). Consequently, we predicted that coalitionary support and affiliative interactions would be primarily concentrated among partners of the same subgroup too. As this special pattern would strongly bias dyadic partner choice options and may drive positive row-wise matrix correlations, we tested for this assumption and controlled for it as a covariable. Average-linkage cluster analyses (Whitehead 2009) revealed the presumed subgroup formation regarding coalitionary support of the mating season. Maximal modularity of this analysis was 0.33 and thus indicated meaningful division, and the cophenetic correlation coefficient was 0.73, indicating that the basal clustering of the analysis into the two subgroups matched the interaction matrix well (see Whitehead 2009). All interaction matrices were tested by row-wise matrix correlations (Spearman rank; de Vries 1993) for correlation with a control matrix coding between-subgroup dyads as 0 and withinsubgroup dyads as 1. All interaction matrices (i.e. coalition formation in the mating season, and triads and time spent in close proximity in the nonmating season or mating season) were significantly correlated with the subgroup control matrix (all: P  0.002) and were thus corrected by mean scaling (as usually applied in general linear models, GLMs, see e.g. Kiebel & Holmes 2003). For this purpose we divided all between- and withinsubgroup values by their respective mean for each row separately, so that for each male the means of the between- and of the withinsubgroup interactions equalled one. Thus, all further analyses using these mean-scaled interaction matrices were controlled for differences between within- and between-subgroup dyads. Analyses were carried out using Statistica 8.0 (Statsoft, Tulsa, OK, U.S.A.), SPSS 18.0.2 (Polar Engineering and Consulting, Nikiski, AK, U.S.A.), MatMan 1.1.4 (Noldus Information Technology, Wageningen, The Netherlands; de Vries et al. 1993) and SocProg 2.4 (Whitehead 2009). All statistical tests were two tailed with alpha set to 0.05, except matrix correlations, which used one-tailed alpha estimations according to Monte Carlo permutation tests with alpha set to 0.05. RESULTS Interaction Rates in Nonmating and Mating Seasons While dyadic aggression rate did not differ between the two seasons, coalitionary aggression rate increased significantly from the nonmating season to the mating season, as did the cumulative aggression rate (Fig. 2a, b, Table 1). Similarly, conflict avoidance, that is, the rate of unprovoked submission was also elevated in the mating season compared to the nonmating season (Fig. 2c, Table 1). In contrast, rates of affiliative behaviours, that is, proximity, body contact, grooming and triadic interactions, all decreased significantly from the nonmating season to the mating season (Fig. 2d-f, Table 1), indicating that affiliative activities were restricted largely to the nonmating season. This was not caused by reduced infant handling times (Table 1). Males affiliated less often with almost all (97%) their closer partners, that is, those with above-average individual affiliation rates. Reduction in affiliation was even stronger for more closely bonded males than for others in which only 67% of dyads affiliated

less often (or stayed zero) in the mating season (Pearson c2 ¼ 17.4, P < 0.0001). Moreover, the increase in unprovoked submission rate per dyad between the nonmating and mating seasons was not negatively correlated with triadic interaction rate in the nonmating season (Spearman rank correlation: rS ¼ 0.035, P ¼ 0.22; partial rowwise matrix correlation controlled for subgroup control matrix). Thus, preferred dyads were similarly characterized by conflict avoidance and limited affiliation during the mating season. Social Bonds For triadic interactions, as well as time spent in close proximity, dyadic patterns of coalition formation during the mating season and affiliative interactions during both the mating and the nonmating seasons were significantly positively correlated with each other (Fig. 3). The lack of a correlation between the first two variables after we controlled for the latter one and, in addition, the just slightly modified correlation between affiliative interactions in the nonmating season and coalition formation in the mating season when we controlled for affiliative interactions in the mating season suggest complete mediation by affiliative interactions in the nonmating season (see Figs 1, 3). Accordingly, not only rates but also patterns of affiliative behaviour changed between the nonmating season and the mating season, since the remaining variance between these two variables (about 70%; see bivariate correlations, Fig. 3) could not simply be explained by random noise but turned out to be rather meaningful and systematic regarding coalition formation (see partial correlations, Fig. 3). Thus, coalition formation during the mating season was exclusively predicted by affiliative interactions through the nonmating season. The apparent effect of affiliative interactions during the mating season on coalition formation was mediated by its correlation with the affiliative interactions of the nonmating season (Fig. 1a), and changes in affiliative behaviour between the nonmating season and the mating season did not account for discrepancies between affiliative behaviour in the nonmating season and coalition formation in the mating season. Reputation Independent of their own interactions with other males, males may base their partner choice on the latter’s reputation, that is, on the partner’s observable propensity to cooperate in affiliative interactions with third parties. However, the frequency of triadic interactions in which a male took part during the nonmating season was not correlated with the frequency of coalitions it formed in the mating season (Spearman rank correlation: rS ¼ 0.134, P ¼ 0.54; partial correlation controlled for subgroup pattern: r ¼ 0.290, P ¼ 0.19), the frequency of coalitionary support it received (Spearman rank correlation: rS ¼ 0.006, P ¼ 0.98; partial correlation controlled for subgroup pattern: r ¼ 0.276, P ¼ 0.21) or the number of coalition partners over the mating season (Spearman rank correlation: rS ¼ 0.155, P ¼ 0.48; partial correlation controlled for subgroup pattern: r ¼ 0.335, P ¼ 0.13). DISCUSSION The mating season, compared to the preceding months, was characterized by a steep increase in aggression rate, indicating a rise in competition and tension among males. This change was accompanied by increased rates of unprovoked submission and decreased rates of affiliative behaviour. Although not explicitly tested, this indicates that males may have avoided contact (see also Horiuchi 2005; van Wolkenten et al. 2006) rather than managing conflict through affiliative tension reduction (e.g. Judge & de Waal

A. Berghänel et al. / Animal Behaviour 81 (2011) 1109e1116

0.7

0.8

1.6 (a)

(c)

(b)

0.6 0.5 0.4

1.4

0.7

1.2

0.6

1

0.5

0.8

0.4

0.6

0.3

0.4

0.2

0.2

0.1

0.3 0.2

Interactions/h

0.1

0

0 Coalitionary aggression NMS

Coalitionary aggression MS

4.5 4

Cumulative aggression NMS

Cumulative aggression MS

0

Unprovoked Unprovoked submission NMS submission MS

4.5

3 (d)

(e)

4

2.5

(f)

3.5

3.5 2

3

3 2.5

2.5

1.5 2

2

1.5

1

1.5

1

1 0.5

0.5

0.5 0

1113

Proximity NMS

Proximity MS

0

Body contact NMS

Body contact MS

0

Triads NMS

Triads MS

Figure 2. Differences in individual interaction rates between the nonmating (NMS) and the mating season (MS). (a) Polyadic aggression, (b) cumulative aggression, (c) unprovoked submission, (d) spatial proximity (<1.5 m), (e) body contact, (f) triadic interactions.

1997; Videan & Fritz 2007). Affiliation rates dropped and spontaneous submission rates increased in all dyads irrespective of their overall affiliation rate (i.e. bond strength). Thus, tension caused by reproductive competition among males was not simply reduced by affiliative and cooperative social bonds, and social bonds may not have been formed to compensate for, but only to deal with, competition. Table 1 Differences in maleemale interaction rates between the nonmating (NMS) and the mating (MS) season Interaction

Median Median P* (interquartile range) (interquartile range) NMS MS

Rates (interactions/h and Dyadic aggression Coalitionary aggression Cumulative aggression Unprovoked submission Proximity (<1.5 m) Body contact Grooming Triads Percentage of time/male Proximity (<1.5 m) Body contact Grooming Body contact with infants

male) 0.16 (0.09e0.32) 0 (0e0.05) 0.17 (0.09e0.33) 0.05 (0.02e0.09) 0.99 (0.68e1.55) 0.48 (0.23e0.80) 0.05 (0.00e0.23) 0.67 (0.37e1.12)

0.15 (0.08e0.36) 0.28 (0.20e0.44) 0.51 (0.32e0.68) 0.41 (0.23e0.54) 0.40 (0.28e0.53) 0.02 (0.00e0.05) 0 (0e0.01) 0.28 (0.15e0.37)

0.99 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001

1.96 1.58 0.05 2.21

1.04 (0.62e1.67) 0.02 (0e0.18) 0 (0e0.09) 1.18 (0.71e3.60)

0.001 <0.001 0.003 0.75

*

Wilcoxon signed-ranks test.

(1.19e3.06) (0.4e2.97) (0e1.72) (0.68e5.04)

Although affiliative behaviour and coalition formation were most frequent in temporally separated mating and nonmating seasons, we found that triadic interactions in general predicted coalitionary relationships between males as suggested by Paul et al. (1996, 2000). Thus, our results present a further demonstration of affiliation between unrelated males for the benefit of agonistic coalition formation (Silk 1994; Feh 1999; Schülke et al. 2010). Currently, Tibetan macaques are the only primates known to match Barbary macaques in their rates, pattern and relationship-regulatory function of triadic interactions (Zhao 1996; Paul et al. 2000). But in contrast to our results, coalition formation was not related to affiliation in Tibetan macaques (Berman et al. 2007). Comparing seasons, we found that not only rates but also patterns of affiliative interactions changed from the nonmating to the mating season. Affiliation patterns during the mating season did not predict coalition formation. Instead, males seemed to recognize their social bonds from the nonmating season also during the mating season, a period of restricted affiliation and increased tension. Here we used nonmating season affiliation patterns observed at least 8 weeks before the onset of the mating season to predict coalition formation in the mating season. The directional causality between affiliation in the nonmating season and coalition formation in the mating season is supported by the sequence of the temporally well-separated interactions. There was a delay between investment in the social bond and the high-risk cooperation in a coalition against another male, which opens up opportunities for cheating. Cooperation during

1114

A. Berghänel et al. / Animal Behaviour 81 (2011) 1109e1116

(a)

(b) Triads MS

***

19

30

Proximity NMS

0.0

42

0.1

10

**

0.054

0.065

0.110*

0.1

0.084*

0.3

0.3

Triads NMS

Proximity MS

***

99

** 0.0

Coalitionary support MS

74

*

*

Coalitionary support MS

Figure 3. Results of the causal step analysis relating (a) triadic interactions or (b) time spent in close proximity in the mating and the nonmating seasons with coalitionary support in the mating season, using bivariate and partial row-wise matrix correlations (Spearman rank correlations) of symmetric (i.e. nondirectional) interaction matrices. Grey: bivariate; black: partial correlations; dashed line and numbers in italics: nonsignificant correlations; size of the lines and numbers: row-wise matrix correlation coefficients; *P < 0.05; **P < 0.01; ***P < 0.001.

agonistic coalitions always carries the risk of injury, especially among males that often carry weapons for direct conflict such as the long canine teeth of many male primates (Preuschoft et al. 1998; Plavcan 1999). Moreover, male Barbary macaques often form all-up coalitions in which both allies rank below the target (Bissonnette 2009; Berghänel et al. 2010); hence, neither of the allies can win alone, making defection by one partner very risky for the other. Our results suggest that males choose their allies according to their nonpublic reputation as reciprocator in affiliative interactions to reduce the risk of defection in coalitions (see also for nonpublic reputation in rats and finches: Rutte & Taborsky 2008; St-Pierre et al. 2009). Accordingly, reputations were built during the nonmating season when competition is reduced. The costs associated with defection in triadic interactions or grooming are small but significant. It has been shown that the overall duration of infant carrying by male Barbary macaques is associated with increased levels of faecal glucocorticoid excretion (Henkel et al. 2010). Thus, males that invest in a triadic interaction with an infant may signal to their partner their willingness to invest in more risky cooperative behaviours (see also Whitham & Maestripieri 2003). Reducing the risk of cheating in the way described requires some mechanisms for tracking relationships through time, for example a mental representation of the bond. Several authors have questioned the existence of such abstract conceptual notions as bonds and time-tracked, mediating social relationships in social animals altogether (e.g. Stevens et al. 2005; Barrett et al. 2007). Instead, they have postulated direct and immediate behavioural adaptation to social dynamics by ‘just-in-time learning’, based on the spatial and affiliative structure among animals in their environment and thus ‘the world as its own best model’ suggesting a mechanism of ‘short-term mutualism’ (see also Hemelrijk 1996; Stevens et al. 2005; Barrett et al. 2007; Henzi et al. 2009). But the spatial and affiliative structure of our study group was explicitly not used ‘just in time’ to adapt coalition formation. Instead, the history of the respective relationships was taken into account. Likewise, the relationship between affiliation and coalition formation cannot be explained by symmetry-based reciprocity (see de Waal & Luttrell 1988) or short-term market forces (Noë et al. 2005; Barrett & Henzi 2006) because neither kinship nor proximity nor affiliative behaviour accounted for coalition formation immediately within the mating season. The view that animal sociality works on the

basis of short-term reciprocity or mutualism (Barrett et al. 2007) has also been challenged by recent studies revealing considerable delays between giving and returning benefits from several minutes or hours (Schino et al. 2007, 2009; Frank & Silk 2009; Schino & Pellegrini 2009; Cheney et al. 2010) up to several days (Rutte & Taborsky 2008; Gomes et al. 2009; Massen 2010). Our results suggest that Barbary macaque males may keep track of events that are separated by several weeks or months. Since the abovementioned studies involved taxa as different as rats, macaques and chimpanzees, it seems unlikely that the required time tracking involves advanced cognitive processes (Stevens et al. 2005; Barrett et al. 2007). Instead, emotional book keeping (Schino & Aureli 2009) has been proposed as a mechanism linking events separated in time, a mechanism that is similar to attitudinal reciprocity (de Waal 2000) or liking (Trivers 1971). Accordingly, emotions are construed as intervening variables that are affected by previous social interactions and in turn affect subsequent interactions, thereby allowing individuals to keep track of social interactions over a longer time and to convert different behavioural commodities into a common currency (Schino & Aureli 2009). This mechanism may have the potential to explain our results if several conditions are met. First, both affiliative and coalitionary behaviours need to be converted into one emotional currency, because affiliation in the mating season alone does not predict coalition formation. At the onset of the mating season, males would then base their choice on previous affiliation only and update their emotional state with both affiliation and coalitions as the mating season goes on. Second, emotional state would need to represent interactions from a longer previous time period. This is because as soon as the mating season starts, even bonded partners show changed affiliation rates and patterns. If emotional state reflects past interactions from only a day or two it would be difficult to maintain the nonmating season affiliation pattern with the relatively rare coalitions alone. Third, negative emotions would need to be registered independently from positive ones because keeping a positive attitude is hampered by increased rates of aggression and spontaneous submission among closely bonded partners. We conclude that a strictly noncognitive emotional book-keeping mechanism is unlikely to explain our results. Instead, the current emotional state may emerge from contextual reactivation and integration of different (and potentially contradictory) previous emotional states. This would suggest that some mechanism for

A. Berghänel et al. / Animal Behaviour 81 (2011) 1109e1116

contextual emotional association is needed which could be provided by mental memory representation. According to Owren & Rendall (1997), this could be facilitated by classical conditioning (i.e. learning) on a partner’s visual and vocal attributes as conditional stimuli. In this view, the perception of the conditioned attributes of a certain male triggers emotional states, which in turn trigger motivational states of approach or avoidance. This conditioned relationship could further be modulated by contextual trigger variables such as sensory inputs, or even seasonal changes in the neurophysiological environment (e.g. Wingfield et al. 1990; see also Schino & Aureli 2009). This classical conditioning mechanism would present an implicit mental memory representation of the social relationships and bonds beyond perceptual mental representation (i.e. representations that are provided by the perception itself), mediating between interactions over long time periods. Although not usually mentioned (e.g. Schino & Aureli 2009), this mechanism seems to be compatible with the original concept of emotional book keeping (see Aureli & Schaffner 2002). This suggests that different behavioural measures should be taken into account when analysing temporal patterns of social relationships, since they may represent different manifestations of the same mental representation of a social bond that can be expressed and modified largely independently. In line with this view, mammals as different as primates, dolphins and sealions, as well as birds such as parrots and pigeons, may possess conceptual representations (Katz et al. 2007; Newen & Bartels 2007; but see Penn et al. 2008). Several mammal and bird species seem to adjust their behaviour in relation to third-party relationships (e.g. Silk 1999; Perry et al. 2004; Holekamp et al. 2007; Weiss et al. 2010). In zebra finches, Taeniopygia guttata, long-term social bonds mediate cooperation in the iterated Prisoner’s Dilemma (St-Pierre et al. 2009) and male hyaenas, Crocuta crocuta, returning to their natal clan territory after several months (up to 1.6 years) maintained their natal social status in interactions with members of their natal clan (Höner et al. 2010). We emphasize that these implicit representations of social bonds do not require episodic memory or anticipation of future rewards (see also Schino & Aureli 2009; but see Barrett et al. 2007). Altogether, the social relationships between unrelated Barbary macaque males were long term and time-tracked, and integrated information from many social interactions in at least implicit mental representations of social bonds. Thus, social bonds can be viewed as mental entities that mediate between different kinds of interactions separated by long time periods. Moreover, they are characterized by bilateral initiation of affiliative interactions and withstand some degree of tension and conflict and periods of infrequent affiliation. Therefore, male social bonds are phenomenologically strikingly similar to close human friendships (see Silk 2002, 2003) indicating that the basic mechanisms underlying human friendships may be evolutionarily old and not unique to humans (see also Massen et al. 2010). Acknowledgments We thank Roland Hilgartner, Gilbert de Turckheim and Ellen Merz for permission to work at the Affenberg Salem. We are grateful to Roland Hilgartner and the staff of Affenberg for assistance. We thank Christin Minge, Maude Erasmy and Miriam Waldmann for help with data collection, Mathias Franz for help with the statistics, and Annie Bissonnette and Claire Cunningham for comments on the manuscript. The study was funded by the Max Planck Society and the German Initiative of Excellence to University of Göttingen.

1115

References Alberts, S. C., Watts, H. E. & Altmann, J. 2003. Queuing and queue-jumping: longterm patterns of reproductive skew in male savannah baboons, Papio cynocephalus. Animal Behaviour, 65, 821e840. Altmann, J. 1974. Observational study of behavior: sampling methods. Behaviour, 49, 227e267. Aureli, F. & Schaffner, C. M. 2002. Relationship assessment through emotional mediation. Behaviour, 139, 393e420. Bachmann, C. & Kummer, H. 1980. Male assessment of female choice in hamadryas baboons. Behavioral Ecology and Sociobiology, 6, 315e321. Baron, R. M. & Kenny, D. A. 1986. The moderatoremediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173e1182. Barrett, L. & Henzi, S. P. 2006. Monkeys, markets and minds: biological markets and primate sociality. In: Cooperation in Primates and Humans (Ed. by P. M. Kappeler & C. P. van Schaik), pp. 209e232. Heidelberg: Springer. Barrett, L., Henzi, P. & Rendall, D. 2007. Social brains, simple minds: does social complexity really require cognitive complexity? Philosophical Transactions of the Royal Society B, 362, 561e575. von Bayern, A. M. P., de Kort, S. R., Clayton, N. S. & Emery, N. J. 2007. The role of food- and object-sharing in the development of social bonds in juvenile jackdaws (Corvus monedula). Behaviour, 144, 711e733. Berghänel, A., Schülke, O. & Ostner, J. 2010. Coalition formation among Barbary macaque males: the influence of scramble competition. Animal Behaviour, 80, 675e682. Berman, C. M., Ionica, C. & Li, J. H. 2007. Supportive and tolerant relationships among male Tibetan macaques at Huangshan, China. Behaviour, 144, 631e661. Bissonnette, A. 2009. Testing a model on coalitions in Barbary macaque males (Macaca sylvanus). Ph.D. thesis, University Zurich. Bissonnette, A., Bischofberger, N. & van Schaik, C. P. 2011. Mating skew in Barbary macaque males: the role of female mating synchrony, female behavior, and maleemale coalitions. Behavioral Ecology and Sociobiology, 65, 176e182. doi:10.1007/s00265-010-1023-z. Cameron, E. Z., Setsaasa, T. H. & Linklater, W. L. 2009. Social bonds between unrelated females increase reproductive success in feral horses. Proceedings of the National Academy of Sciences, U.S.A., 106, 13850e13853. Cheney, D. L. & Seyfarth, R. M. 1990. The representation of social relations by monkeys. Cognition, 37, 167e196. Cheney, D. L., Moscovice, L. R., Heesen, M., Mundry, R. & Seyfarth, R. M. 2010. Contingent cooperation between wild female baboons. Proceedings of the National Academy of Sciences, U.S.A., 107, 9562e9566. Clutton-Brock, T. 2009. Cooperation between non-kin in animal societies. Nature, 462, 51e57. Connor, R. C. 1995. Impala allogrooming and the parcelling model of reciprocity. Animal Behaviour, 49, 528e530. Connor, R. C., Heithaus, M. R. & Barre, L. M. 2001. Complex social structure, alliance stability and mating access in a bottlenose dolphin ‘super-alliance’. Philosophical Transactions of the Royal Society B, 268, 263e267. Cordoni, G. 2009. Social play in captive wolves (Canis lupus): not only an immature affair. Behaviour, 146, 1363e1385. Deag, J. M. 1980. Interactions between males and unweaned Barbary macaques: testing the agonistic buffering hypothesis. Behaviour, 75, 54e81. Dugatkin, L. A. 1997. Cooperation among Animals: an Evolutionary Perspective. New York: Oxford University Press. Dunbar, R. I. M. & Shultz, S. 2010. Bondedness and sociality. Behaviour, 147, 775e803. Earley, R. L. 2010. Social eavesdropping and the evolution of conditional cooperation and cheating strategies. Philosophical Transactions of the Royal Society B, 365, 2675e2686. Eaton, G. G., Modahl, K. B. & Johnson, D. F. 1981. Aggressive behavior in a confined troop of Japanese macaques: effects of density, season, and gender. Aggressive Behavior, 7, 145e164. Emery, N. J., Seed, A. M., von Bayern, A. M. P. & Clayton, N. S. 2007. Cognitive adaptations of social bonding in birds. Philosophical Transactions of the Royal Society B, 362, 489e505. Feh, C. 1999. Alliances and reproductive success in Camargue stallions. Animal Behaviour, 57, 705e713. Frank, R. E. & Silk, J. B. 2009. Impatient traders or contingent reciprocators? Evidence for the extended time-course of grooming exchanges in baboons. Behaviour, 146, 1123e1135. Gomes, M. G., Mundry, R. & Boesch, C. 2009. Long-term reciprocation of grooming in wild West African chimpanzees. Philosophical Transactions of the Royal Society B, 276, 699e706. Hart, B. L. & Hart, L. A. 1992. Reciprocal allogrooming in impala, Aepyceros melampus. Animal Behaviour, 44, 1073e1083. Hemelrijk, C. K. 1996. Reciprocation in apes: from complex cognition to selfstructuring. In: Great Apes Societies (Ed. by W. C. McGrew, L. F. Marchant & T. Nishida), pp. 185e195. Cambridge: Cambridge University Press. Henkel, S., Heistermann, M. & Fischer, J. 2010. Infants as costly social tools in male Barbary macaque networks. Animal Behaviour, 79, 1199e1204. Henzi, S. P. & Barrett, L. 1999. The value of grooming to female primates. Primates, 40, 47e59. Henzi, S. P., Lusseau, D., Weingrill, T., van Schaik, C. P. & Barrett, L. 2009. Cyclicity in the structure of female baboon social networks. Behavioral Ecology and Sociobiology, 63, 1015e1021.

1116

A. Berghänel et al. / Animal Behaviour 81 (2011) 1109e1116

Holekamp, K. E., Sakai, S. T. & Lundrigan, B. L. 2007. Social intelligence in the spotted hyena (Crocuta crocuta). Philosophical Transactions of the Royal Society B, 362, 523e538. Höner, O. P., Wachter, B., Hofer, H., Wilhelm, K., Thierer, D., Trillmich, F., Burke, T. & East, M. L. 2010. The fitness of dispersing spotted hyaena sons is influenced by maternal social status. Nature Communications, 1, 1e7. van Hooff, J. A. R. A. M. & van Schaik, C. P. 1994. Male bonds: affiliative relationships among nonhuman primate males. Behaviour, 130, 309e337. Horiuchi, S. 2005. Affiliative relations among male Japanese macaques (Macaca fuscata yakui) within and outside a troop on Yakushima Island. Primates, 46, 191e197. Judge, P. G. & de Waal, F. B. M. 1997. Rhesus monkey behaviour under diverse population densities: coping with long-term crowding. Animal Behaviour, 54, 643e662. Katz, J. S., Wright, A. A. & Bodily, K. D. 2007. Issues in the comparative cognition of abstract-concept learning. Comparative Cognition and Behavior Reviews, 2, 79e92. Kiebel, S. & Holmes, A. P. 2003. The general linear model. In: Human Brain Function (Ed. by R. S. Frackowiak, K. J. Friston, C. D. Frith, R. J. Dolan, C. J. Price, J. Ashburner, W. D. Penny & S. Zeki), pp. 725e760. Oxford: Oxford Academic Press. Kuester, J., Paul, A. & Arnemann, J. 1994. Kinship, familiarity and mating avoidance in Barbary macaques, Macaca sylvanus. Animal Behaviour, 48, 1183e1194. MacKinnon, D. P. 2008. Introduction to Statistical Mediation Analysis. New York: L. Erlbaum. Massen, J. J. M. 2010. ‘Friendship’ in macaques: economics and emotions. Ph.D. thesis, Utrecht University. Massen, J. J. M., Sterck, E. H. M. & de Vos, H. 2010. Close social associations in animals and humans: functions and mechanisms of friendship. Behaviour, 147, 1379e1412. Ménard, N., von Segesser, F., Scheffrahn, W., Pastorini, J., Valleta, D., Gacic, B., Martin, R. D. & Gautier-Hion, A. 2001. Is maleeinfant caretaking related to paternity and/or mating activities in wild Barbary macaques (Macaca sylvanus)? Life Sciences, 324, 601e610. Mitani, J. C. 2006. Reciprocal exchange in chimpanzees and other primates. In: Cooperation in Primates and Humans (Ed. by P. M. Kappeler & C. P. van Schaik), pp. 107e120. Heidelberg: Springer. Mitani, J. 2009. Male chimpanzees form enduring and equitable social bonds. Animal Behaviour, 77, 633e640. Newen, A. & Bartels, A. 2007. Animal minds and the possession of concepts. Philosophical Psychology, 20, 283e308. Noë, R. 1992. Alliance formation among male baboons: shopping for profitable partners. In: Coalitions and Alliances in Humans and other Animals (Ed. by A. H. Harcourt & F. B. M. de Waal), pp. 285e322. Oxford: Oxford University Press. Noë, R., van Hooff, J. A. R. A. M. & Hammerstein, P. 2005. Economics in Nature: Social Dilemmas, Mate Choice and Biological Markets. Cambridge: Cambridge University Press. Ostner, J., Heistermann, M. & Schülke, O. 2011. Male competition and its hormonal correlates in Assamese macaques (Macaca assamensis). Hormones and Behavior, 59, 105e113. Owren, M. J. & Rendall, D. 1997. An affect-conditioning model of nonhuman primate vocal signaling. In: Perspectives in Ethology: Communication (Ed. by D. H. Owings, M. D. Beecher & N. S. Thompson), pp. 299e346. New York: Plenum. Paul, A., Kuester, J. & Arnemann, J. 1996. The sociobiology of maleeinfant interactions in Barbary macaques, Macaca sylvanus. Animal Behaviour, 51, 155e170. Paul, A., Preuschoft, S. & van Schaik, C. P. 2000. The other side of the coin: infanticide and the evolution of affiliative maleeinfant interaction in Old World primates. In: Infanticide by Males and its Implications (Ed. by C. P. van Schaik & C. H. Janson), pp. 269e292. Cambridge: Cambridge University Press. Penn, D. C., Holyoak, K. J. & Povinelli, D. J. 2008. Darwin’s mistake: explaining the discontinuity between human and nonhuman minds. Behavioral and Brain Sciences, 31, 109e178. Perry, S., Barrett, H. C. & Manson, J. H. 2004. White-faced capuchin monkeys show triadic awareness in their choice of allies. Animal Behaviour, 67, 165e170. Plavcan, J. M. 1999. Mating systems, intrasexual competition and sexual dimorphism in primates. In: Comparative Primate Socioecology (Ed. by P. C. Lee), pp. 241e260. Cambridge: Cambridge University Press. Preuschoft, S., Paul, A. & Kuester, J. 1998. Dominance styles of female and male Barbary macaques (Macaca sylvanus). Behaviour, 135, 731e755. de Ruiter, J. R. & van Hoof, J. A. R. A. M. 1993. Male dominance rank and reproductive success in primate groups. Primates, 34, 513e523. Rutte, C. & Taborsky, M. 2008. The influence of social experience on cooperative behaviour of rats (Rattus norvegicus): direct vs generalised reciprocity. Behavioral Ecology and Sociobiology, 62, 499e505. Sands, J. & Creel, S. 2004. Social dominance, aggression and faecal glucocorticoid levels in a wild population of wolves, Canis lupus. Animal Behaviour, 67, 387e396. van Schaik, C. P. 1996. Social evolution in primates: the role of ecological factors and male behaviour. Proceedings of the British Academy, 88, 9e31. Scheffrahn, W., Menard, N., Vallet, D. & Gaci, B. 1993. Ecology, demography, and population genetics of Barbary macaques in Algeria. Primates, 34, 381e394. Schino, G. 2007. Grooming and agonistic support: a meta-analysis of primate reciprocal altruism. Behavioral Ecology, 18, 115e120.

Schino, G. & Aureli, F. 2009. Reciprocal altruism in primates: partner choice, cognition, and emotions. Advances in the Study of Behavior, 39, 45e69. Schino, G. & Aureli, F. 2010. The relative roles of kinship and reciprocity in explaining primate altruism. Ecology Letters, 13, 45e50. Schino, G. & Pellegrini, B. 2009. Grooming in mandrills and the time frame of reciprocal partner choice. American Journal of Primatology, 71, 884e888. Schino, G., di Sorrentino, E. P. & Tiddi, B. 2007. Grooming and coalitions in Japanese macaques (Macaca fuscata): partner choice and the time frame of reciprocation. Journal of Comparative Psychology, 121, 181e188. Schino, G., di Giuseppe, F. & Visalberghi, E. 2009. The time frame of partner choice in the grooming reciprocation of Cebus apella. Ethology, 115. Schülke, O., Bhagavatula, J., Vigilant, L. & Ostner, J. 2010. Social bonds enhance reproductive success in male macaques. Current Biology, 20, 2207e2210. Silk, J. B. 1994. Social relationships of male bonnet macaques: male bonding in a matrilineal society. Behaviour, 130, 271e291. Silk, J. B. 1999. Male bonnet macaques use information about third-party rank relationships to recruit allies. Animal Behaviour, 58, 45e51. Silk, J. B. 2002. Using the ‘F’-word in primatology. Behaviour, 139, 421e446. Silk, J. B. 2003. Cooperation without counting: the puzzle of friendship. In: Genetic and Cultural Evolution of Cooperation (Ed. by P. Hammerstein), pp. 37e54. Cambridge, Massachusetts/London: MIT Press/Dahlem University Press. Silk, J. B. 2006. Practicing Hamilton’s rule: kin selection in primate groups. In: Cooperation in Primates and Humans (Ed. by P. M. Kappeler & C. P. van Schaik), pp. 25e46. Heidelberg: Springer. Silk, J. B. 2007. The adaptive value of sociality in mammalian groups. Philosophical Transactions of the Royal Society B, 362, 539e559. Smith, J. E., van Horn, R. C., Powning, K. S., Cole, A. R., Graham, K. E., Memenis, S. K. & Holekamp, K. E. 2010. Evolutionary forces favoring intragroup coalitions among spotted hyenas and other animals. Behavioral Ecology and Sociobiology, 21, 284e303. St-Pierre, A., Larose, K. & Dubois, F. 2009. Long-term social bonds promote cooperation in the iterated Prisoner’s Dilemma. Proceedings of the Royal Society B, 276, 4223e4228. Stevens, J. R., Cushman, F. A. & Hauser, M. D. 2005. Evolving the psychological mechanisms for cooperation. Annual Review of Ecology, Evolution, and Systematics, 36, 499e518. Trivers, R. L. 1971. The evolution of reciprocal altruism. Quarterly Review of Biology, 46, 35e57. Trivers, R. L. 1972. Parental investment and sexual selection. In: Sexual Selection and the Descent of Man, 1871e1971 (Ed. by B. Campbell), pp. 136e179. London: Heinemann. de Turckheim, G. & Merz, E. 1984. Breeding Barbary macaques in outdoor open enclosures. In: The Barbary Macaque: a Case Study in Conservation (Ed. by J. E. Fa), pp. 241e261. New York: Plenum. Videan, E. N. & Fritz, J. 2007. Effects of short- and long-term changes in spatial density on the social behavior of captive chimpanzees (Pan troglodytes). Applied Animal Behaviour Science, 102, 95e105. de Villiers, M. S., Richardson, P. R. K. & van Jaarsveld, A. S. 2003. Patterns of coalition formation and spatial association in a social carnivore, the African wild dog (Lycaon pictus). Journal of Zoology, 260, 377e389. de Vries, H. 1993. The rowwise correlation between two proximity matrices and the partial rowwise correlation. Psychometrika, 58, 53e69. de Vries, H., Netto, W. J. & Hanegraaf, P. L. H. 1993. MATMAN: a program for the analysis of sociometric matrices and behavioural transition matrices. Behaviour, 125, 157e175. de Waal, F. B. M. 2000. Attitudinal reciprocity in food sharing among brown capuchin monkeys. Animal Behaviour, 60, 253e261. de Waal, F. B. M. & Harcourt, A. H. 1992. Coalitions and alliances: a history of ethological research. In: Coalitions and Alliances in Humans and other Animals (Ed. by A. H. Harcourt & F. B. M. de Waal), pp. 1e19. Oxford: Oxford University Press. de Waal, F. B. M. & Luttrell, L. M. 1988. Mechanisms of social reciprocity in three primate species: symmetrical relationship characteristics or cognition. Ethology and Sociobiology, 9, 101e118. Wasilewski, A. 2003. ‘Friendship’ in ungulates? Sociopositive relationships between non-related herd members of the same species. Ph.D. thesis, PhilippsUniversität Marburg. Weiss, B. M., Kehmeier, S. & Schloegl, C. 2010. Transitive inference in free-living greylag geese, Anser anser. Animal Behaviour, 79, 1277e1283. Whitehead, H. 2009. SOCPROG programs: analysing animal social structures. Behavioral Ecology and Sociobiology, 63, 765e778. Whitham, J. C. & Maestripieri, D. 2003. Primate rituals: the function of greetings between male Guinea baboons. Ethology, 109, 847e859. Wilkinson, G. S. 1985. The social organization of the common vampire bat I. Pattern and cause of association. Behavioral Ecology and Sociobiology, 17, 111e121. Wingfield, J. C., Hegner, R. E., Dufty, A. M. J. & Ball, G. F. 1990. The ‘challenge hypothesis’: theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. American Naturalist, 136, 829e846. van Wolkenten, M. L., Davis, J. M., Gong, M. L. & de Waal, F. B. M. 2006. Coping with acute crowding by Cebus apella. International Journal of Primatology, 27, 1241e1256. Zhao, Q. 1996. Maleeinfantemale interactions in Tibetan macaques. Primates, 37, 135e143.