Animal Behaviour 80 (2010) 675e682

Contents lists available at ScienceDirect

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

Coalition formation among Barbary macaque males: the influence of scramble competition Andreas Berghänel a, b, Oliver Schülke a, b, Julia Ostner a, b, * a b

Courant Research Centre Evolution of Social Behaviour, University of Göttingen, Germany Integrative Primate Socio-Ecology Group, MPI Evolutionary Anthropology, Leipzig, Germany

a r t i c l e i n f o Article history: Received 23 April 2010 Initial acceptance 31 May 2010 Final acceptance 5 July 2010 Available online 14 August 2010 MS. number: 10-00285 Keywords: Affenberg Salem Barbary macaque contest competition Macaca sylvanus male coalition scramble competition

Owing to the nature of male competition over a nonshareable resource, cooperation and coalition formation are unexpected among mammalian males. Nevertheless, coalitions of unrelated males are widespread across mammals. Recently, a mathematical model for the evolution of male coalitions within groups was developed, predicting patterns and outcomes of coalitions depending on the degree of contest competition for mates. We tested this model in a species with presumably low-contest competition. Accordingly, under low contest, coalitions should be large, formed by mid- to low-ranking males targeting top-rankers and aimed at decreasing reproductive skew (all-up levelling coalitions). We also investigated the effect of scramble competition on coalitions, as its impact was not included in the model. Scramble competition is expected to be pronounced in our study population of Barbary macaques, Macaca sylvanus, and thus we predicted that coalitions should target mainly young immigrants, irrespective of rank, to isolate them and eventually reduce male group size and, consequently, scramble competition. Data on social interactions, including coalitions, were collected on 23 males of a group of Barbary macaques living at Affenberg Salem, Germany. Coalition size was small, coalitions rarely targeted top-ranking males and all-up levelling coalitions were rare. Overall, the model was not confirmed. In contrast, more than three-quarters of coalitions targeted young immigrants, which were isolated from the group. Thus, male coalition formation in our group seemed to be influenced by scramble rather than contest competition. Both modes of competition should thus be taken into account if payoffs of coalition formation are investigated. Ó 2010 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

In social carnivores, cetaceans and primates, male coalitions in multimaleemultifemale groups are a viable strategy in both between-group and within-group sexual competition (Wrangham 1979; Noë 1986; Connor et al. 1992, 2000; Zabel et al. 1992; Grinnell et al. 1995; Watts 1998; de Villiers et al. 2003; Watts et al. 2006). Several frameworks have been generated to explain different aspects of coalition formation, such as coalition size (Whitehead & Connor 2005), winner and loser effects (Dugatkin 1998a, b), fighting ability (Noë 1986; Mesterton-Gibbons & Sherratt 2007) and dynamics and outcome of coalitions (Chapais 1995). Among mammals, coalition formation has been studied most thoroughly in primates (Smith et al. 2010). Primate females form coalitions to aid relatives or to increase their access to food resources (Wrangham 1980; Chapais 1995). Primate males, however, compete over a nonshareable resource, that is, fertilizations, making coalitions ‘constant-sum’ and only to be expected under special circumstances (van Hooff & van Schaik 1992). * Correspondence: J. Ostner, Courant Research Centre Evolution of Social Behaviour, University of Göttingen, Kellnerweg 6, 37077 Göttingen, Germany. E-mail address: [email protected] (J. Ostner).

Recently, van Schaik and colleagues (Pandit & van Schaik 2003; van Schaik et al. 2004, 2006) provided a mathematical model predicting the occurrence of different types of within-group maleemale coalitions according to their feasibility in terms of defeating the target (combined strength of allies exceeds that of target, Noë 1994) and profitability in terms of increased reproductive success of the allies (Pandit/van Schaik model). Coalitions were classified as all-up, alldown or bridging according to the dominance rank positions of allies relative to the target (i.e. allies being all lower ranking than the target, all higher ranking, or a high- and a low-ranking ally targeting an intermediate-ranking individual; see also Chapais 1995) and as either rank changing or levelling, that is, flattening the payoff distribution. Variation in the occurrence of coalition types and sizes turned out to be driven by variation in male contest competition potential alone (van Schaik et al. 2004); for example, when contest potential is very high coalitions are always profitable, all-up coalitions are not feasible, and top-ranking males form small, bridging rank-changing coalitions with relatives. Contest potential or the intensity of direct mating competition largely depends on a single male’s ability to monopolize fertile females against rival males (Emlen & Oring 1977), which in turn is

0003-3472/$38.00 Ó 2010 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.anbehav.2010.07.002

676

A. Berghänel et al. / Animal Behaviour 80 (2010) 675e682

determined mainly by the distribution of fertile females in space and time (e.g. Nunn 1999; Ostner et al. 2008a). Where females live in relatively small groups and their fertile phases are asynchronous, a single male may be able to monopolize access to all females in the groups (Altmann 1962). Traits that are important in direct competition such as body size, strength, endurance and weaponry will be selected for (Clutton-Brock 1985; Plavcan 2004). But whenever male mating monopolies are broken and females mate with multiple males, males will compete also via indirect means in mating scrambles (Schwagmeyer 1988) and scramble and contest competition will coexist. When females are either spatially dispersed or temporally clumped, a single male will not be able to defend them and scramble competition will prevail (Emlen & Oring 1977; Ims 1988), promoting traits advantageous for locating fertile females, for example well-developed sensory organs and mobility, as well as for sperm competition (Schwagmeyer 1988; Birkhead & Møller 1998; Setchell & Kappeler 2003). Scramble competition may be enhanced by female promiscuity. If competition is mainly by scramble, male fitness will depend mainly on the number of male rivals, but not on rank, because each male has a similar chance of finding a female and fathering her offspring. We predict that scramble competition will have effects on male coalition formation. Specifically, we expect male coalitions to be aimed at reducing the number of male competitors within the group and coalitionary effort to be directed against new male immigrants (van Hooff & van Schaik 1994). The Pandit/van Schaik model proposes that in a pure scramble situation no coalitions will be formed (Pandit & van Schaik 2003; van Schaik et al. 2006). Barbary macaques, Macaca sylvanus, have been classified as a species with low to medium contest potential and have thus been predicted to show all-up levelling coalitions (van Schaik et al. 2004; Bissonnette 2009). Indeed, a study by Bissonnette et al. (2009a) on coalitions among Barbary macaque males confirmed this prediction. At the same time, Barbary macaques have been described as showing pronounced scramble competition over reproduction, owing to their large group size, as well as pronounced seasonality, making it difficult for single males to monopolize reproduction (Taub 1980; Paul 1989; Small 1990). Contest potential is additionally decreased by female synchronization of fertile periods (Brauch et al. 2008) and scramble competition is enhanced by the highly promiscuous mating pattern, in which females change partners in rapid succession and usually mate with a large proportion of a group’s resident males (Taub 1980; Paul 1989; Small 1990; Menard et al. 1992; Brauch et al. 2008). Thus, given the presumably low but nonzero contest potential combined with a strong scramble component, Barbary macaques provide a good model species to test the predictions regarding male coalitions forming under predominately sexual scramble competition. In this study, we first characterized the mode of competition in our study group of Barbary macaques at the Affenberg Salem, Germany to test the predictions laid out in the Pandit/van Schaik coalition model as well as our predictions for the effect of scramble competition on male coalition formation. According to the Pandit/ van Schaik model, in the case of low-contest potential, coalitions should be mainly all-up, coalition size should be large and the benefits of forming coalitions should be greatest for low- and middle-ranking males, by targeting top-ranking males, and for those with the highest mating success to decrease mating skew. Based on our considerations, under strong scramble competition all males will benefit from group size reduction irrespective of their rank position. The most likely targets should thus be recently immigrated males and the profit from coalitionary aggression will be to isolate these males, that is, force them to the periphery of the group and eventually expel them.

METHODS Study Species and Group Barbary macaques live in multimaleemultifemale groups with female philopatry and male dispersal (Paul & Kuester 1985; Scheffrahn et al. 1993; Kuester & Paul 1999). Males usually migrate into groups with fewer males than in their natal group (Kuester & Paul 1999). While secondary migration is rare in semifree-ranging situations (Kuester & Paul 1999), it is common in wild populations (Menard & Vallet 1993). Females reach sexual maturity within 3 years and males within 4 years, but males are not fully grown until the age of 7 years and show their maximum strength around the age of 14 years (Paul & Kuester 1988; Kuester & Paul 1989; Kuester et al. 1995; Bissonnette et al. 2009b). Data for this study were collected on a semifree-ranging group of Barbary macaques inhabiting a 14.5 ha enclosure at Affenberg Salem, Germany (for a history of the population and conditions, see de Turckheim & Merz 1984). Data were collected from October 2008 to January 2009 covering one mating season. 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 (unpublished data from Affenberg Salem). Five adult males were born in the study group (unpublished data from Affenberg Salem). Remaining in the natal group during adulthood is not an artefact of captivity, as has been shown in a long-term study on wild Barbary macaques where 50% of males became adult in their natal group and delayed their dispersal until after adolescence (Menard & Vallet 1993, 1996; Kuester & Paul 1999). Of the 31 females, 20 had been given hormonal contraceptive implants (Norplant; active ingredient: levonorgestrel) for birth control. Administered levonorgestrel allows contraception while neither preventing normal cyclicity of the ovaries nor suppressing the development of cyclic perineal swellings (Wallner et al. 2007). A previous study on mating behaviour in this group showed that males did not differentiate between implanted and nonimplanted females, mating with them at the same rate (Bissonnette 2009). Consequently, all adult females were included in the analysis. Behavioural Data Recording We recorded data using 45 min focal animal observations with continuous and instantaneous recording (Altmann 1974; Martin & Bateson 1993). All adult animals were individually known, and all 23 adult males were chosen as focal animals (906 focal hours, 39.4
0.6 h/male). The continuous protocol included sexual interactions (i.e. copulations) and agonistic interactions. Interactions are agonistic whenever they include 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. Every minute, we recorded whether or not the focal male was isolated from the group, that is, outside the known home range of the study group and at least 100 m away from the remaining male group members. Ad libitum data on agonistic interactions were recorded throughout the study period. Finally, 228 dyadic agonistic interactions were provoked opportunistically throughout the study period using the nut test method (Preuschoft et al. 1998; Bissonnette et al. 2009a). This entailed positioning a walnut at an equal distance between two males, sitting less than 10 m apart, who were attentive to the nut and without a third male present. All agonistic behavioural interactions during the test were recorded and treated as ad libitum data.

A. Berghänel et al. / Animal Behaviour 80 (2010) 675e682

Female Attractiveness and Fertility To estimate female attractiveness and fertility, we recorded daily the size of sexual swellings for all adult females following a method established for the species (according to Möhle et al. 2005; Brauch et al. 2007). Anogenital swelling size was assessed on a 2-point scale (0 ¼ no or partial swelling, vulva not or partly visible, with wrinkles, ‘dull’ appearance; 1 ¼ maximal swelling, vulva clearly visible, protrusion of all genital structures, no wrinkles, tight and bright appearance) according to low and high likelihood of ovulation, respectively. To reduce estimation error in cases of unrecognizable swelling variation, we also considered obvious enhanced female attractiveness evident from sexual consorting behaviour or increased maleemale aggression around a certain female. These records, as well as the assessment of swelling size, were made daily by two or three observers independently and confirmed by majority (>50%) to ensure maximal conservatism. Enhanced female attractiveness according to male behaviour corresponded with maximal swelling sizes in 94.2% of individual daily records of female swellings; we thus included male behaviour in the analysis as an alternative criterion of attractiveness for females without clear swelling variation. The period of female attractiveness was defined by the period of maximal swelling size or, if necessary, by periods of obviously enhanced attractiveness itself. The fertile period, however, was defined as the first 6 of the last 7 days of the attractive phase using an established method that had been verified with hormonal data in Barbary macaques (Kuester & Paul 1984; Brauch et al. 2008; Heistermann et al. 2008; Bissonnette 2009). Data Analysis We constructed a dominance hierarchy on the basis of 755 dyadic decided agonistic interactions from focal, ad libitum and nut test data. Decided conflicts were either unprovoked submission or aggression followed by submission by only one individual. To assess linearity of the hierarchy, we applied Landau’s index h0 corrected for the number of unknown relationships (de Vries 1995). We calculated the directional consistency index as the mean of the proportions of decided interactions that were given in the main direction in all decided interactions per dyad. Dominance ranks used in further analyses were based on normalized David’s scores (Norm DS), corrected for variation in interaction frequencies (de Vries et al. 2006). Individual male mating success was based on the frequency of ejaculatory copulations per male and focal hour. Ejaculation was easily detected by a pre-ejaculatory pause and/or sperm on male or female genitalia. Two indices of male mating skew were calculated. First, the lambda index (l) measures the overall skew regardless of dominance rank positions (calculated using Skew Calculator 2003; http://www.eeb.ucla.edu/Faculty/Nonacs/shareware.htm) ranging from 0 (equal distribution) to 1 (one male takes all; Kokko & Lindström 1997). Second, we calculated the beta index (b) based on the alpha male’s mating success as a measure of the degree of despotism as used in the Pandit/van Schaik coalition model. Following de Waal & Harcourt (1992), we defined coalitions as joint aggression by two or more males against a common target. The analysis of coalitionary aggression was based on conflicts involving males only (N ¼ 189 coalitions of various sizes, N ¼ 155 coalitions of two partners against one target) and excluding scream fights (see Bissonnette 2009). If not stated otherwise, analyses were based on two-against-one coalitions. The context of a coalition was conservatively defined as sexual, whenever at least one animal participating in a coalition (as ally or target) was in close proximity (<1.5 m) to an adult female.

677

The relationship between rank (measured as dominance success, i.e. Norm DS values), mating success (with fertile females), age (continuous) and immigration status (categorical: natal or immigrant) as independent variables and the number of coalitionary attacks received, that is, how often an individual was a target of a coalition, as dependent variable was analysed in a generalized linear model (GLM). In another GLM we tested the relationship between dominance, age and immigration status as independent variable and active participation in a coalition as dependent variable. In both analyses the dependent variables were modelled as having a Poisson distribution. Overdispersion was not significant in either model (P ¼ 0.999). To assess the possibility for collinearity among predictor variables, we used variance inflation factors (VIFs, Petraitis et al. 1996). We found significant associations between the predictor variables (e.g. age and dominance). In all cases in both GLMs, however, VIFs were less than 4, thus well below the cut-off of >10 which indicates problems with collinearity (Petraitis et al. 1996). The association between dominance and mating success was tested with the distribution-free Spearman rank correlation. Differences between main and subgroup males with regard to their time spent away from the group were tested with a ManneWhitney U test and a Wilcoxon signed-ranks test was applied to compare the observed and expected mating frequency at the periphery of the group. The expected frequency was based on the proportion of time a male spent at the periphery, that is, if a male spent 10% of his activity time at the periphery, 10% of his matings should take place there. Analyses were carried out in Statistica 8.0 (StatSoft, Tulsa, OK, U.S.A.), S-Plus 6.1 (Insightful Corp., Seattle, WA, U.S.A.) and MatMan 1.1.4 (Noldus, Wageningen, The Netherlands). All statistical tests were two tailed with alpha set at 0.05. RESULTS Dominance Hierarchy Of 253 possible dyadic relationships among all 23 males, 88.5% were known, and 78.3% were one-way relationships with a clear dominant and a clear subordinate. The resulting hierarchy was significantly linear (h0 ¼ 0.78, N ¼ 755, P < 0.0001) and highly consistent in the direction of the interaction outcomes (DCI ¼ 0.92). Steepness of the hierarchy based on normalized David’s scores was 0.53 (P ¼ 0.0005) which is high considering the size of the hierarchy (Richter et al. 2009). Mode of Male Competition On average
SD, 5.2
2.2 females were attractive and 2.4
1.3 fertile on a given day (16.8
7.0% and 7.8
4.2% of all females, respectively). The observed mating skew across males was low (l ¼ 0.20, b ¼ 0.14, N ¼ 52 with fertile and l ¼ 0.16, b ¼ 0.10, N ¼ 84 with attractive females). Although low-ranking males achieved no mating success with fertile females, they nevertheless mated with attractive females (Fig. 1). Across all 23 males, mating success with fertile females depended significantly on dominance, indicating some degree of contest competition (Spearman rank correlation: rS ¼ 0.76, N ¼ 23, P < 0.0002). This relationship, however, disappeared if only the upper half of the hierarchy, that is, ranks 1e11, the males that actually mated with fertile females, was considered (Spearman rank correlation: rS ¼ 0.41, N ¼ 11, P ¼ 0.206). Hence, the contest effect was driven by the fact that lower-ranking males did not mate with fertile females at all, whereas matings were more or less evenly distributed across the higher-ranking males, indicating considerable scramble competition.

A. Berghänel et al. / Animal Behaviour 80 (2010) 675e682

20

(a)

18 16 14 10

Isolation of Coalitionary Targets

8 6 4 2 0

20 (b)

18 16 14 12 10

The six main targets of coalitions, that is, the young immigrants, were significantly more often isolated from the group (ManneWhitney U test: U ¼ 0.00, Nmain group males ¼ 17, Nsubgroup males ¼ 6, P ¼ 0.0004; Fig. 3). Although available as a target less often, the more time a male spent isolated from the main group the more often he was a target of a coalition (Spearman correlation by rank: rS ¼ 0.77, N ¼ 23, P ¼ 0.00002). The aim of isolating males may be to level the mating skew, but the proportion of matings young immigrants achieved at the periphery did not deviate from an expected value based on the proportion of time they spent there (Wilcoxon signed-ranks test: T ¼ 6.00, N ¼ 6, P ¼ 0.345; Fig. 4). Demographic Patterns of Coalition Formation

8 6 4 2 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Matings with attractive females (%)

immigrant male added to the likelihood of being a target of aggression. The second GLM with age, immigration status and dominance as predictors of the number of coalitions a male participated in as an ally revealed that only age and dominance rank had a significant effect (Table 2).

12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Matings with fertile females (%)

678

Dominance rank Figure 1. Dominance-related mating success measured as ejaculatory copulations with (a) fertile and (b) attractive females.

Patterns of Male Coalition Formation Based on the Pandit/van Schaik coalition model and given the low-contest potential, coalitions were expected to be large, predominately of the all-up levelling type and targeting highranking males and those with high mating success. However, coalition size was small with an average of 2.3
0.6 partners (N ¼ 189). Of the 155 two-against-one coalitions observed during the study period, 41% (N ¼ 64) were all-up, 40% (N ¼ 62) bridging and 19% (N ¼ 29) all-down coalitions. Overall, 22% of all coalitions occurred in the context of access to mating partners and, thus, may be classified as levelling; however, on only two or three occasions did one of the coalition partners obtain a female following the coalition. Only a third (31%) of the all-up coalitions had a sexual context. Also, coalitions were not primarily directed at males with a high mating success (see results below). Finally, coalitions did not primarily target the top-ranking males (Fig. 2). Instead, the three top-ranking males combined were targets in only five of all 189 coalitions. In accordance with our predictions for coalitions under strong scramble competition, more than three-quarters of all coalitions (78%; N ¼ 121) were directed against the six youngest immigrants, and 91% (N ¼ 141) of those coalitions were formed exclusively by the rest of the group consisting of either old males, thus long-term residents, or natal males (see Fig. 2 for age structure of the group). In a GLM we tested for the effects of immigration status, dominance, mating success with fertile females and age on coalitionary aggression received (Table 1). Age, dominance rank and immigration status were all predictors of the variation in the number of coalitionary attacks received, with younger, higher-ranking immigrant males being the main targets of coalitions. Mating success did not add to the explanatory value of the model. Thus, being an

Assuming constant breeding seasonality within a species, we predicted the number of males, as well as the number of females, that would have an effect on the frequency of coalition formation, as an increase in either number should lead to an increase in scramble competition. When we qualitatively compared data of nine studies, large male and female group sizes (which were strongly correlated: rS ¼ 0.89, P ¼ 0.002) were linked to the frequency of coalitions (categorized dichotomously as either ‘frequent’ or ‘rare’; see Table 3). Male group size in the five studies reporting frequent coalitions was on average
SD 23.9
8.3 and in the four studies with rare coalitions 6.3
1.5. Similarly, the mean number of females was 40.0
16.3 in groups forming male coalitions frequently and 12.5
3.5 in groups in which coalitions were rarely observed. DISCUSSION Testing the Pandit/van Schaik Model Females in the study group were highly seasonal and synchronous in their reproductive activity which has also been reported for other populations of Barbary macaques (Small 1990; Brauch et al. 2008). Oestrus synchrony in combination with a female strategy of pronounced promiscuity yields a competitive regime of lowcontest and high scramble competition. Consequently, male mating skew, expressed as the lambda index, was rather low, which is in accordance with results from previous work on the same population (Kuester & Paul 1992; Bissonnette 2009). According to the Pandit/van Schaik model, a competitive regime of low contest (b ¼ 0.14 in this study) should lead to large all-up levelling coalitions targeting high-ranking males, as has been reported, for example, for savannah baboons (Papio anubis: Smuts 1985; P. cynocephalus: Noë & Sluijter 1990, 1995; Alberts et al. 2003). Although all-up coalitions occurred in our study, a significant proportion of coalitions were of the all-down and of the bridging type. According to the model, bridging rank-changing coalitions are not profitable at low levels of contest; rather, they are expected at high contest competition where helping a relative may be beneficial in terms of an increase in indirect fitness (van Schaik et al. 2006). Instead, bridging coalitions are expected only when they mainly have a levelling function, which was not the case in our study where only 20% of coalitions occurred in a mating context.

A. Berghänel et al. / Animal Behaviour 80 (2010) 675e682

Rank

Age and immigration status

1

16

2

18

3

9

4

14

5

13

6

17

7

12

8

21

9

11

10

12

11

11

12

27

13

19

14

20

15

21

16

22

17

22

18

22

19

22

20

22

21

23

22

26

23

23

50

40

30

20

10

Number of coalitions given

0

0

10

20

679

30

40

50

Number of coalitions received

Figure 2. Number of coalitions a male was involved in as ally (coalitions given) or target (coalitions received). All coalitions (grey) are shown with their proportion of coalitions formed by two males against one target only (black) including those that were formed against the six subgroup members (shaded left side) at all and by main group members only (shaded right). >: natal; <: young immigrant; C old immigrant.

Also in contrast to the model, coalitions did not specifically target the top-ranking males, that is, alpha, beta and gamma males, or males with the highest mating success. Finally, the majority of coalitions did not have the predicted levelling function as they

occurred outside a mating context and even in the absence of a female. It may be argued that the observed coalitions may still serve a levelling function indirectly by intimidating subordinate males and thus keeping them from mating (see Kuester & Paul

Table 1 Independent variables (dominance rank measured as normalized David’s score; mating success measured as proportion of matings with fertile females; immigration status (natal or immigrant) and age) predicting the amount of individually received coalitionary aggression in a GLM

Table 2 Independent variables (dominance rank measured as normalized David’s score; immigration status (natal or immigrant) and age) predicting the frequency of participating in a coalition in a GLM

Dependent variable: received coalitionary attacks

Intercept Rank Mating success Immigration status Age

Estimate

SE

Z

Pð>jZjÞ

3.660 0.139 0.022 L3.167 L0.184

0.767 0.0425 0.027 0.541 0.030

4.769 3.281 0.808 L5.857 L6.074

0.0001 0.001 0.419 0.0001 0.0001

Coalitions: N ¼ 155. Bold values indicate significant results.

Dependent variable: participation in coalitionary attacks

Intercept Rank Age Immigration status

Estimate

SE

Z

Pð>jZjÞ

0.078 0.090 0.106 0.100

0.525 0.025 0.016 0.162

0.148 3.535 6.789 0.619

0.883 0.0001 0.0001 0.536

Coalitions: N ¼ 155. Bold values indicate significant results.

680

A. Berghänel et al. / Animal Behaviour 80 (2010) 675e682

24 Focal time spent isolated (%)

22 20 18 16 14 12 10 8 6 4 2 0

Main group males

Subgroup males

Figure 3. Time spent isolated from the group for males from the main group and for young immigrants (subgroup males). Box plots show medians, 25e75% percentiles and extreme values.

1992; Bissonnette 2009). However, the mating success of the main targets of coalitions was similar to that of higher-ranking males. Overall, the pattern of coalition formation found in this study did not confirm the predictions from the Pandit/van Schaik model.

Impact of Scramble Competition on Coalitions Scramble competition was pronounced in our study group. Mating success did not depend on rank across the upper half of the hierarchy. In accordance with our predictions for coalitions under strong scramble competition, immigrant males were more likely to be targets of coalitionary aggression. These coalitions usually occurred independent of a sexual context, thus do not qualify as ‘levelling’ and were effective in isolating the young immigrants from the group (see also Paul 1989; Kuester & Paul 1992; Bissonnette 2009). Isolation of competitors via coalitions has been reported for this species and interpreted as an alternative male reproductive tactic to lower the mating success of the target

60 50

Percentage

40 30 20 10 0

Expected Observed Proportion of matings at periphery

Figure 4. Expected and observed proportion of matings achieved by young immigrants at the periphery. The expected value is based on the proportion of time these males spent at the periphery. Box plots show medians, 25e75% percentiles and extreme values.

males (Paul et al. 1993; van Schaik et al. 2006; Bissonnette 2009). However, our results indicate that isolating young immigrants did not function to keep those males from mating, that is, it had no indirect levelling function because mating frequency outside the group was not different from mating success when associated with the main group. van Schaik et al. (2006) did not include coalitions against immigrants (‘unranked males’, sensu van Schaik et al. 2006) in the model. They proposed, however, that the benefit of those coalitions may be rank maintenance and thus maintenance of the current payoff associated with a given rank position. In our study, the degree of contest competition, that is, rank-dependent mating success, was low, and consequently rank maintenance should not be the crucial driver of coalitions. Also, low-ranking males were often involved in coalition formation, again arguing against a possible rank-maintaining function of coalitions. Instead, we propose that the aim of coalitions was to reduce scramble competition by reducing male group size as a result of isolating and eventually ousting immigrant males. Whether our findings can be generalized to Barbary macaques living under natural conditions cannot be decided at this point. Owing to the semicaptive setting of the study with only three groups in the population, the options for immigrant males were constrained and secondary dispersal events are rare (Paul & Kuester 1988; Kuester & Paul 1999). The only long-term study of wild populations of Barbary macaques showed that males migrate frequently and continue to disperse throughout their lifetime (Menard & Vallet 1993). Therefore, it may be assumed that the aggressive isolation we see in a semicaptive setting may, in natural habitats without the limitation of dispersal options, often lead to the emigration of target males and thus to successful group size regulation. This assumption, however, still needs to be tested in a wild population of Barbary macaques or other species with high levels of scramble competition, that is, living in large groups or exhibiting pronounced female synchrony. van Schaik et al. (2006) proposed that an increase in male group size will lead to an increase in the frequency of coalitions as well as an increase in coalition size. This means that at a given constant number of females, coalitions will be more frequent and larger at larger male numbers as well as at a higher sex ratio. Our qualitative comparison of available data on Barbary macaques showed that coalitions were indeed more frequent in groups with more males, but also in groups with more females. Sex ratio, however, did not point in the direction predicted by the Pandit/van Schaik model as coalitions were actually more frequent at lower sex ratios, that is, fewer males to females. Thus, it seems that it is rather the number of females, and thus a possible scramble effect, that drives the frequency of coalitions, at least based on this qualitative and small data set. Peripheral, recently immigrated males were also the main target of male coalitions among wild Assamese macaques, Macaca assamensis, in a population experiencing low-contest potential (Ostner et al. 2008b; O. Schülke & J. Ostner, unpublished data). In this study, it was not clear whether the isolation served a contest effect, that is, keeping the males from mating, or a scramble effect, that is, driving the males out of the group, or possibly both. Notably, however, the two most peripheral males and main targets of coalitionary attacks left the group soon after the study, whereas similar-aged but more integrated males remained (O. Schülke & J. Ostner, unpublished data). The aims of coalitionary activity could not be determined in a captive group of bonnet macaques, Macaca radiata, where coalitions were formed at a high frequency but virtually none of them involved direct competition over access to receptive females or had a measurable effect on dominance ranks (see Silk 1992a, b). In a population of Tibetan macaques, Macaca thibetana, where strong

A. Berghänel et al. / Animal Behaviour 80 (2010) 675e682

681

Table 3 Demographic data and coalition pattern across nine studies of Barbary macaques Number of adult males

Number of adult females

Sex ratio

Coalitions

Comment

Source

23 24 31 31 10.5

31 27 46 66 30

1.3 1.1 1.5 2.1 2.9

Frequent Frequent Frequent Frequent Frequent

This study Bissonnette et al. 2009a Widdig et al. 2000 Kuester & Paul 1992 Paul 1989

8 7 5 5

16 9 10 15

2.0 1.3 2.0 3.0

Rare Rare Rare Rare

189 coalitions, 906 focal hours 111 coalitions, 279 focal hours 106 coalitions, 196 focal hours 52 coalitions, 263 focal hours 2 mating seasons, coalitions as drivers of mating success 8 coalitions, 250 focal hours 15 months, ‘supports rare’ <20 coalitions, 325 focal hours 2 mating seasons, coalitions rare (K. Brauch, personal communication)

contest competition should override scramble effects on coalition formation, most coalitions were protective all-down and a small proportion all-up rank-changing (Berman et al. 2007). Our results contrast with those of a study on the same group in Salem but 2 years beforehand, which revealed predominantly allup coalitions (58%) often over access to females (Bissonnette et al. 2009a). However, group structure differed between the two studies. While in our study the top three dominance positions were all held by natal males, in the study by Bissonnette et al. (2009a) young immigrant males were highest ranking. Natal males may delay dispersal past adolescence also in the wild (Menard & Vallet 1993, 1996; Kuester & Paul 1999), indicating that the presence of natal males in our study group is not purely an effect of limited dispersal options. Consequently, although the constellation of coalitions may differ between the two studies, the targets are in both cases the same, namely young immigrant males. In conclusion, our results indicate that the main intention of coalition formation in our study group experiencing unusually high levels of indirect male competition was to isolate males and thus reduce male group size. We could not detect a primary levelling function of coalitions as predicted by the Pandit/van Schaik model. This model makes clear predictions for coalitions under various degrees of contest and for the lack of coalitions under pure scramble. However, among primates scramble and contest competition will often coexist (e.g. Eberle & Kappeler 2004) making it impossible to test the coalition model thoroughly. Under certain circumstances scramble and contest will lead to similar patterns of coalition formation (see above) making it difficult to determine the relative importance of either factor. 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 and Maude Erasmy for help with data collection, Annie Bissonnette for valuable discussion and Mathias Franz for help with the statistics. The study was funded by the Max Planck Society and the German Initiative of Excellence. References Alberts, S., Watts, H. & Altmann, J. 2003. Queuing and queue-jumping: long-term 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. Altmann, S. A. 1962. A field study of the sociobiology of the rhesus monkey, Macaca mulatta. Annals of the New York Academy of Sciences, 102, 338e435. Berman, C. M., Ionica, C. & Li, J. 2007. Supportive and tolerant relationships among male Tibetan macaques at Huangshan, China. Behaviour, 144, 631e661.

Bissonnette 2009, Daun group Taub 1980 Kohpahl 1986 Brauch et al. 2008

Birkhead, T. R. & Møller, A. P. 1998. Sperm Competition and Sexual Selection. London: Academic Press. Bissonnette, A. 2009. Testing a model on coalitions in Barbary macaque males (Macaca sylvanus). Ph.D. thesis, University Zurich. Bissonnette, A., de Vries, H. & van Schaik, C. P. 2009a. Coalitions in male Barbary macaques, Macaca sylvanus: strength, success and rules of thumb. Animal Behaviour, 78, 329e335. Bissonnette, A., Lange, E. & van Schaik, C. P. 2009b. A cardinal measure of competitive ability in Barbary macaque males (Macaca sylvanus). Ethology, 115, 671e681. Brauch, K., Pfefferle, D., Hodges, K., Möhle, U., Fischer, J. & Heistermann, M. 2007. Female sexual behavior and sexual swelling size as potential cues for males to discern the female fertile phase in free-ranging Barbary macaques (Macaca sylvanus) of Gibraltar. Hormones and Behavior, 52, 375e383. Brauch, K., Hodges, K., Engelhardt, A., Fuhrmann, K., Shaw, E. & Heistermann, M. 2008. Sex-specific reproductive behaviours and paternity in free-ranging Barbary macaques (Macaca sylvanus). Behavioral Ecology and Sociobiology, 62, 1453e1466. Chapais, B. 1995. Alliances as means of competition in primates: evolutionary, developmental and cognitive aspects. Yearbook of Physical Anthropology, 38, 115e136. Clutton-Brock, T. H. 1985. Size, sexual dimorphism and polygyny in primates. In: Size and Scaling in Primate Biology (Ed. by W. L. Jungers), pp. 51e60. New York: Plenum. Connor, R. C., Smolker, R. & Richards, A. 1992. Dolphin alliances and coalitions. In: Coalitions and Alliances in Humans and other Animals (Ed. by A. H. Harcourt & F. B. M. de Waal), pp. 415e444. Oxford: Oxford University Press. Connor, R. C., Read, A. J. & Wrangham, R. 2000. Male reproductive strategies and social bonds. In: Cetacean Societies (Ed. by J. Mann, R. C. Connor, P. L. Tyack & H. Whitehead), pp. 247e269. Chicago: University of Chicago Press. Dugatkin, L. A. 1998a. Breaking up fights between others: a model of intervention behaviour. Proceedings of the Royal Society B, 265, 433e437. Dugatkin, L. A. 1998b. A model of coalition formation in animals. Proceedings of the Royal Society B, 265, 2121e2125. Eberle, M. & Kappeler, P. M. 2004. Sex in the dark: determinants and consequences of mixed male mating tactics in Microcebus murinus, a small solitary nocturnal primate. Behavioral Ecology and Sociobiology, 57, 77e90. Emlen, S. T. & Oring, L. W. 1977. Ecology, sexual selection, and the evolution of mating systems. Science, 197, 215e223. Grinnell, J., Packer, C. & Pusey, A. E. 1995. Cooperation in male lions: kinship, reciprocity or mutualism? Animal Behaviour, 49, 95e105. Heistermann, M., Brauch, K., Möhle, U., Pfefferle, D., Dittami, J. & Hodges, K. 2008. Female ovarian cycle phase affects the timing of male sexual activity in free-ranging Barbary macaques (Macaca sylvanus) of Gibraltar. American Journal of Primatology, 70, 44e53. van Hooff, J. A. R. A. M. & van Schaik, C. P. 1992. Cooperation in competition: the ecology of primate bonds. In: Coalitions and Alliances in Humans and other Animals (Ed. by A. H. Harcourt & F. B. M. de Waal), pp. 357e390. Oxford: Oxford University Press. van Hooff, J. A. R. A. M. & van Schaik, C. P. 1994. Male bonds: affiliative relationships among nonhuman primate males. Behaviour, 130, 309e337. Ims, R. A. 1988. Spatial clumping of sexually receptive females induces space sharing among male voles. Nature, 335, 541e543. Kohpahl, G. 1986. Saisonabhängige Veränderungen in den Beziehungen zwischen männlichen Berberaffen (Macaca sylvanus). Eine Semi-Freilanduntersuchung. Diplomarbeit, Universität Göttingen. Kokko, H. & Lindström, J. 1997. Measuring the mating skew. American Naturalist, 149, 794e799. Kuester, J. & Paul, A. 1984. Female reproductive characteristics in semifree-ranging Barbary macaques (Macaca sylvanus L. 1758). Folia Primatologica, 43, 69e83. Kuester, J. & Paul, A. 1989. Reproductive strategies of subadult Barbary macaque males at Affenberg Salem. In: The Sociobiology of Sexual and Reproductive Strategies (Ed. by V. V. Rasa), pp. 93e109. New York: Routledge, Chapman & Hall.

682

A. Berghänel et al. / Animal Behaviour 80 (2010) 675e682

Kuester, J. & Paul, A. 1992. Influence of male competition and female mate choice on male mating success in Barbary macaques (Macaca sylvanus). Behaviour, 120, 192e217. Kuester, J. & Paul, A. 1999. Male migration in Barbary macaques (Macaca sylvanus) at Affenberg Salem. International Journal of Primatology, 20, 85e106. Kuester, J., Paul, A. & Arnemann, J. 1995. Age-related and individual differences of reproductive success in male and female Barbary macaques (Macaca sylvanus). Primates, 36, 461e476. Martin, P. & Bateson, P. 1993. Measuring Behaviour. 2nd edn. Cambridge: Cambridge University Press. Menard, N., Scheffrahn, W., Vallet, D., Zidane, C. & Reber, C. 1992. Application of blood protein electrophoresis and DNA fingerprinting to the analysis of paternity and social characteristics of wild Barbary macaques. In: Paternity in Primates: Genetic Tests and Theories: Implications of Human DNA Fingerprints (Ed. by R. Martin, A. Dixson & E. Wickings), pp. 155e174. Basel: Karger. Menard, N. & Vallet, D. 1993. Population dynamics of Macaca sylvanus in Algeria: a 8-years study. American Journal of Primatology, 30, 101e118. Menard, N. & Vallet, D. 1996. Demography and ecology of Barbary macaques (Macaca sylvanus) in two different habitats. In: Evolution and Ecology of Macaque Societies (Ed. by J. E. Fa & D. G. Lindburg), pp. 106e131. Cambridge: Cambridge University Press. Mesterton-Gibbons, M. & Sherratt, T. 2007. Coalition formation: a game-theoretic analysis. Behavioral Ecology, 18, 277e286. Möhle, U., Heistermann, M., Dittami, J., Reinberg, V. & Hodges, J. K. 2005. Patterns of anogenital swelling size and their endocrine correlates during ovulatory cycles and early pregnancy in free-ranging Barbary macaques (Macaca sylvanus) of Gibraltar. American Journal of Primatology, 66, 351e368. Noë, R. 1986. Lasting alliances among adult male savannah baboons. In: Primate Ontogeny, Cognition and Social Behaviour (Ed. by J. G. Else & P. C. Lee), pp. 381e392. Cambridge: Cambridge University Press. Noë, R. 1994. A model of coalition formation among male baboons with fighting ability as the crucial parameter. Animal Behaviour, 47, 211e213. Noë, R. & Sluijter, A. A. 1990. Reproductive tactics of male savannah baboons. Behaviour, 113, 117e170. Noë, R. & Sluijter, A. A. 1995. Which adult male savanna baboons form coalitions? International Journal of Primatology, 16, 77e105. Nunn, C. L. 1999. The number of males in primate groups: a comparative test of the socioecological model. Behavioral Ecology and Sociobiology, 46, 1e13. Ostner, J., Nunn, C. L. & Schülke, O. 2008a. Female reproductive synchrony predicts skewed paternity across primates. Behavioral Ecology, 19, 1150e1158. Ostner, J., Heistermann, M. & Schülke, O. 2008b. Dominance, aggression and physiological stress in wild male Assamese macaques (Macaca assamensis). Hormones and Behavior, 54, 613e619. Pandit, S. & van Schaik, C. 2003. A model for leveling coalitions among primate males: toward a theory of egalitarianism. Behavioral Ecology and Sociobiology, 55, 161e168. Paul, A. 1989. Determinants of male mating success in a large group of Barbary macaques (Macaca sylvanus) at Affenberg Salem. Primates, 30, 461e476. Paul, A. & Kuester, J. 1985. Intergroup transfer and incest avoidance in semifreeranging Barbary macaques (Macaca sylvanus) at Salem (FRG). American Journal of Primatology, 8, 317e322. Paul, A. & Kuester, J. 1988. Life-history patterns of Barbary macaques (Macaca sylvanus) at Affenberg Salem. In: Ecology and Behavior of Food-enhanced Primate Groups (Ed. by J. E. Fa & C. H. Southwick), pp. 199e228. New York: A.R. Liss. Paul, A., Kuester, J., Timme, A. & Arnemann, J. 1993. The association between rank, mating effort, and reproductive success in male Barbary macaques (Macaca sylvanus). Primates, 34, 491e502. Petraitis, P., Dunham, A. & Niewlarowski, P. 1996. Inferring multiple causality: the limitations of path analysis. Functional Ecology, 10, 421e431. Plavcan, J. M. 2004. Sexual selection, measures of sexual selection, and sexual dimorphism in primates. In: Sexual Selection in Primates: New and Comparative Perspectives (Ed. by P. M. Kappeler & C. van Schaik), pp. 230e252. New York: Cambridge University Press. Preuschoft, S., Paul, A. & Kuester, J. 1998. Dominance styles of female and male Barbary macaques (Macaca sylvanus). Behaviour, 135, 731e755.

Richter, C., Mevis, L., Malaivijitnond, S., Schülke, O. & Ostner, J. 2009. Social relationships in free-ranging male Macaca arctoides. International Journal of Primatology, 30, 625e642. van Schaik, C., Pandit, S. & Vogel, E. 2004. A model for within-group coalitionary aggression among males. Behavioral Ecology and Sociobiology, 57, 101e109. van Schaik, C., Pandit, S. & Vogel, E. 2006. Toward a general model for maleemale coalitions in primate groups. In: Cooperation in Primates and Humans (Ed. by P. Kappeler & C. van Schaik), pp. 151e172. Heidelberg: Springer. Scheffrahn, W., Menard, N., Vallet, D. & Gaci, B. 1993. Ecology, demography, and population genetics of Barbary macaques in Algeria. Primates, 34, 381e394. Schwagmeyer, P. L. 1988. Scramble competition polygyny in an asocial mammal: male mobility and mating success. American Naturalist, 131, 885e892. Setchell, J. & Kappeler, P. M. 2003. Selection in relation to sex in primates. Advances in the Study of Behavior, 33, 87e173. Silk, J. 1992a. The patterning of intervention among male bonnet macaques: reciprocity, revenge, and loyalty. Current Anthropology, 33, 318e325. Silk, J. 1992b. Patterns of intervention in agonistic contests among male bonnet macaques. In: Coalitions and Alliances in Humans and other Animals (Ed. by A. H. Harcourt & F. B. M. de Waal), pp. 215e232. Oxford: Oxford University Press. Small, M. 1990. Promiscuity in Barbary macaques (Macaca sylvanus). American Journal of Primatology, 20, 267e282. 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, 21, 284e303. Smuts, B. B. 1985. Sex and Friendship in Baboons. New York: Aldine. Taub, D. 1980. Female choice and mating strategies among wild Barbary macaques (Macaca sylvanus L. In: The Macaques: Studies in Ecology Behavior and Evolution (Ed. by D. Lindburg), pp. 287e344. New York: Van Nostrand Reinhold. 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. 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. 1995. An improved test of linearity in dominance hierarchies containing unknown or tied relationships. Animal Behaviour, 50, 1375e1389. de Vries, H., Stevens, J. & Vervaecke, H. 2006. Measuring and testing the steepness of dominance hierarchies. Animal Behaviour, 71, 585e592. 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. Wallner, B., Möstl, E. & Dittami, J. 2007. Chronic levonorgestrel treatment in Macaca sylvanus: effects on perineal swelling size and fecal sex steroid excretion. Neuroendocrinology Letters, 28, 326e332. Watts, D. P. 1998. Coalitionary mate guarding by male chimpanzees at Ngogo, Kibale National Park, Uganda. Behavioral Ecology and Sociobiology, 44, 43e55. Watts, D. P., Muller, M., Amsler, S., Mbabazi, G. & Mitani, J. C. 2006. Lethal intergroup aggression by chimpanzees in Kibale National Park, Uganda. American Journal of Primatology, 68, 161e180. Whitehead, H. & Connor, R. 2005. Alliances I: how large should alliances be? Animal Behaviour, 69, 117e126. Widdig, A., Streich, W. & Tembrock, G. 2000. Coalition formation among male Barbary macaques (Macaca sylvanus). American Journal of Primatology, 50, 37e51. Wrangham, R. W. 1979. On the evolution of ape social systems. Social Sciences Information, 18, 335e368. Wrangham, R. W. 1980. An ecological model of female-bonded primate groups. Behaviour, 75, 262e300. Zabel, C., Glickman, S., Frank, L., Woodmansee, K. & Keppel, G. 1992. Coalition formation in a colony of prepubertal spotted hyenas. In: Coalitions and Alliances in Humans and other Animals (Ed. by A. H. Harcourt & F. B. M. de Waal), pp. 113e136. Oxford: Oxford University Press.