A winner effect supports third-party intervention behaviour during fallow deer, Dama dama, fights

Animal Behaviour 77 (2009) 343–348

Contents lists available at ScienceDirect

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

A winner effect supports third-party intervention behaviour during fallow deer, Dama dama, fights Domhnall J. Jennings a, *, Caitriona M. Carlin b,1, Martin P. Gammell c, 2 a

Institute of Neuroscience, Newcastle University Natural England, Peterborough c Department of Life and Physical Sciences, Galway-Mayo Institute of Technology, Galway b

a r t i c l e i n f o Article history: Received 15 February 2008 Initial acceptance 26 March 2008 Final acceptance 15 October 2008 Published online 24 December 2008 MS. number: 08-00091R Keywords: Dama dama dominance fallow deer fight third-party intervention winner effect

Male ungulates engage in intense competition for access to females during the breeding season. Although fights are generally dyadic level encounters, they are on occasion disrupted by the intervention of third-party males. We investigated these third-party interventions using predictions derived from Dugatkin’s model (Dugatkin 1998, Proceedings of the Royal Society of London, Series B, 265, 433–437) of intervention behaviour. The model argues that when an individual successfully defeats an opponent there is an increase in the probability of winning a subsequent contest: a winner effect. Third-party intervention behaviour is predicted to occur as it serves to prevent either member of a competing dyad from successfully defeating his opponent, achieving a winner effect and subsequently becoming a threat to the intervener. Consistent with model predictions, our results show that intervening males held significantly higher rank than males that did not intervene and were also more likely to be dominant to both of the competing males. Intervening males did not selectively target competitors based on rank, nor did they target males based on overall dyadic rates of aggression between the intervener and competing males. Furthermore, interveners were more likely to have won their interaction immediately prior to intervention and were also likely to win their interaction subsequent to intervention when compared with contest success of the two competing males. Our results are consistent with predictions that support a winner effect for intervention behaviour in fallow deer fights. Ó 2008 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

Empirical and theoretical studies of animal competition have traditionally focused on dyadic level contests during which two competitors fight until one emerges as the victor. While these contests are undoubtedly frequent in nature, there exists a substantial body of work involving birds, carnivores, cetaceans and primates that have reported the existence of intervention behaviour by third-party group members (Widdig et al. 2006). During these third-party interventions, the ongoing contest is prematurely terminated by the intervener and, under such circumstances, no decisive outcome between the contestants is established. Perhaps reflecting the vast body of data collected by primatologists in this area, explanations to account for intervention behaviour include disturbance minimization, reduction of injuries

* Correspondence: D. Jennings, Centre for Behaviour and Evolution, Institute of Neuroscience, Newcastle University, Framlington Place, Newcastle upon Tyne NE2 4HH, U.K. E-mail address: [email protected] (D.J. Jennings). 1 C. Carlin is at Natural England, Northminster House, Northminster, Peterborough PE1 1UA, U.K. 2 M. Gammell is at the Department of Life and Physical Sciences, Galway-Mayo Institute of Technology, Dublin Road, Galway, Ireland.

between the contestants, promotion of group cohesion and to reduce tension or to strengthen alliances within the group (e.g. Ehardt & Bernstein 1992). To account for such behaviour in primate societies, it has been argued that intervention behaviour is indicative of some form of knowledge concerning third-party relationships between members of the group, a generally accepted metric of intelligence (Harcourt 1988). While our knowledge primate social behaviour is extensive, our current understanding of how ungulates structure aggressive interactions and their knowledge of third-party dominance relationships is far from comprehensive (Jennings & Gammell, in press). In general, published accounts of ungulate contest behaviour have centred on studies of dyadic level interactions that are used to assess dominance relations within the social group (e.g. CluttonBrock & Albon 1979; Appleby 1982; Jennings et al. 2006) and the extent to which this relates to phenotypic correlates of resourceholding potential (e.g. Bowyer 1986; Clutton-Brock et al. 1988; Pe´labon & Joly 2000) and mating success (e.g. Appleby 1982; FestaBianchet et al. 1990; Moore et al. 1995). While it has been shown that fighting is more likely to occur between similarly ranked individuals (Freeman et al. 1992; Jennings et al. 2006), it is unclear what decision-making process, opponent- or self-assessment, is

0003-3472/$38.00 Ó 2008 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.anbehav.2008.10.006

344

D.J. Jennings et al. / Animal Behaviour 77 (2009) 343–348

used during dyadic level fights (Clutton-Brock & Albon 1979; Jennings et al. 2004, 2005a; Bartosˇ et al. 2007). Based on published accounts of contest behaviour in ungulates, it has been asserted that, at least in the wild, ungulates do not engage in triadic level interactions (Freeman et al. 1992; Ceacero et al. 2007). Furthermore, it has been argued that cognitively complex behaviour such as knowledge of rank relations within the social group or the ability to form alliances is beyond deer (Freeman et al. 1992; Appleby 1993; Shultz & Dunbar 2007). However, contrary to assertions made by Freeman et al. (1992) and Ceacero et al. (2007), we have noted that fights are frequently terminated by third-party intervention in rutting fallow deer. Based on studies of triadic level encounters in primates and other species, this observation suggests that social complexity in this species might be greater than hitherto reported. While predictions for such third-party behaviour in fallow deer might be derived from primate models of intervention, alternative explanations that are somewhat less Machiavellian in outlook have been put forward to account for intervention behaviour. A mathematical model developed by Dugatkin (1998) proposes that intervention behaviour serves to break up fights between group members to prevent rival males from achieving a winner effect. The model is based on the observation that prior fighting experience can affect subsequent fight performance (Hsu et al. 2006); winning a fight can lead to increased success in later contests whereas a loss can lead to an increased probability of losing subsequent contests (Dugatkin 1997; Rutte et al. 2006). Where a winner or mixed winner–loser effect exists, Dugatkin’s model predicts that intervention and disruption of a fight serves simply to prevent the contestants from achieving a winner effect and consequently advancing up any hierarchy that exists. Since intervention serves simply to prevent either competitor from experiencing a winner effect, the intervening individual is expected to target one of the competitors at random. Therefore, although the model does not specifically rule out the possibility that individuals have knowledge of the social status of group members, an explanation of intervention behaviour does not rest on this concept. Consistent with this prediction, empirical studies that have described intervention behaviour have frequently shown that neither contestant was favoured in both primate (e.g. de Waal & van Hoof 1981; Watts 1991) and nonprimate species (e.g. Nelissen 1985; Keil & Sambraus 1998). The model also predicts that intervention behaviour is related to dominance rank because high-ranking individuals are more successful in terms of the outcome of competitive interactions. Thus, they are more likely to experience a winner effect. The European fallow deer is a polygynous, seasonally breeding species with the annual rut taking place during October in the northern hemisphere (Apollonio et al. 1989; Moore et al. 1995). The breeding system of our study population is described as a follower system: most matings are recorded while males consort with females as they move over their day range (Moore et al. 1995). With the exception of the rut, mature males in this population reside in a large, stable, multigenerational group composed of upwards of 60 individuals. Furthermore, they have a stable and transitive dominance hierarchy which is rapidly (re-) established following cleaning of the velvet from the antlers in late August in the northern hemisphere (Jennings 2007). Aggressive behaviour recorded during the prerut and rut is highly correlated with dominance rank; the most dominant individuals interact with the greatest number of opponents, have the greatest overall interaction rate (Jennings 2000) and achieve most matings (Moore et al. 1995). The present study represents an opportunity to investigate thirdparty intervention behaviour within the framework of a model that hypothesizes a winner effect in intervention behaviour (Dugatkin 1998). Dugatkin’s model hypothesizes that interveners will not favour either competing individual and, therefore, we predicted

that interveners will not target a competing individual based on relative social rank. In addition, we predicted that intervention behaviour is positively related to dominance rank and that intervening males should hold a higher dominance rank than nonintervening males. Furthermore, for intervention to be favoured within the population, the intervening male should target and defeat one of the competing males. METHODS Study Site and Population This study was conducted on a large herd of free-ranging European fallow deer resident in the Phoenix Park: a large enclosed city park consisting of 709 ha located in Dublin, Ireland (53 220 N, 6 210 W). Most of the park (80%) is open grassland with the remaining 20% covered by mixed woodland. Annual tagging of the herd began in 1971, and the entire herd was rounded up and tagged between 1991 and 1992. The majority of fawns are tagged in each ear with unique colour/numbered tags shortly after birth in June and July each year. At the time of this study, most mature males (1996: 97%; 1997: 93%  4 years old) still retained their eartags. The relatively few males that were not tagged were easily identifiable from a combination of antler form and coat colour. Procedures From about the time that velvet was cleaned from the antlers in late August, the bachelor herd was observed 5–6 days each week by two to three observers. Following cleaning, males remain in an aggregated bachelor herd and there is a marked increase in the number of agonistic interactions. We recorded agonistic interactions using all-event recording procedures (Altmann 1974); data recorded included the identities of the two animals involved in the interaction and the time, date, location and outcome of the interaction. We placed these interactions in two categories: noncontact interactions which do not involve physical contact, and fights in which competing males lock their antlers and engage in vigorous pushing contests (Jennings et al. 2003, 2004). From mid to late September, the bachelor herd disperses onto the females’ range and males become increasingly intolerant of each other, begin to display typical rutting behaviour (e.g. vocalizing, scent marking) and heightened sexual interest in females (e.g. herding, sniffing; Chapman & Chapman 1975). The tendency for males to escalate to fighting increases in frequency as the time of the annual rut approaches. To monitor males’ rutting behaviour, we increased the observation schedule from the end of September. We observed the deer from dawn to dusk 7 days a week and the number of observers increased to approximately 10 on any given day. Fight interventions occurred when one male moved towards a fighting dyad and disrupted the fight by directly interacting with the dyad. Intervention resulted in the end of the fight with the males from the fighting dyad moving away from each other. A second interaction usually occurred immediately between the intervener and one of the fighting males; however, in a minority of cases the intervening male did not go on to interact with either member of the fighting dyad. Interventions occurred during slightly fewer than 10% of all fights recorded in this population (1996: 159 (9.5%); 1997: 196 (8.8%)). The majority of disruptions were the result of mature males interacting with dyads consisting of mature males (133 and 151 interventions in each year, respectively) and these interventions are the subject of the present study. We recorded the identity of the intervening male and the category and outcome of the subsequent interaction as described above. When the intervening male interacted with one of the fighting males we denoted this male as the ‘target’. The second male in the

D.J. Jennings et al. / Animal Behaviour 77 (2009) 343–348

fighting dyad was labelled the ‘ignored’ male. If the intervening male did not interact with either member of the fighting dyad we labelled the next male that he interacted with as the ‘fourth male’. Dominance Ranking Dominance ranks were calculated using David’s score (DS: David 1987, 1988), a consistent and accurate ranking measure (Gammell et al. 2003). To calculate the dominance rank, we used all decisively resolved noncontact interactions that were recorded prior to the first mating in the population in that year. We used this dominance rank as an index of male quality as it correlated positively with subsequent mating success in both years (Spearman correlations: 1996: rS ¼ 0.51, N ¼ 70, P < 0.001; 1997: rS ¼ 0.63, N ¼ 63, P < 0.001). During both years a number of males were chemically immobilized under the supervision of a veterinary surgeon; however, this had no effect on individual males’ ability to compete with each other (Jennings 2007).

345

dominance ranks of the fighting dyad and the intervening male (calculated as the average of the difference between the target– intervener and ignored–intervener; 1996: X ¼ 21:77  1:3; 1997: X ¼ 22:05  1:2), and this difference was significant (ANOVA: F1,198 ¼ 12.65, P < 0.001). There was no main effect of year (F1,198 ¼ 0.08, P ¼ 0.78) and no interaction between year and difference in rank (F1,198 ¼ 0.46, P ¼ 0.5). Males were more likely to intervene in fights where they were dominant to both the fighting males (1996: N ¼ 38; 1997: N ¼ 40 fights) than when they were dominant to only one (1996: N ¼ 14; 1997: N ¼ 25) or neither buck (1996: N ¼ 7, c22 ¼ 26:9, P < 0.001, ø ¼ 0.68; 1997: N ¼ 21, c22 ¼ 7, P < 0.03, ø ¼ 0.29). There was no evidence that intervening males were electing to interact with the lower-ranked male in a fighting dyad as indexed by prerut dominance rank (high versus low-ranked buck in dyad: 1996: 33 versus 26; 1997: 41 versus 45). This was the case irrespective of whether the interrupting buck was dominant to both (sign test (data pooled over both years): Z ¼ 0.23, N ¼ 78, NS), one (Z ¼ 0.74, N ¼ 39, NS) or neither of the fighting males (Z ¼ 0.76, N ¼ 28, NS).

Statistical Tests Where the same male intervened in a fighting dyad more than once, we used only a single interaction from the data set to avoid pseudoreplication (1996: 22 males; 1997: 30 males engaged in more than one intervention). We used analysis of covariance (ANCOVA) and a general linear model (GLM) with year included as a factor to investigate the effect of year. To facilitate interpretation of categorical data between years we include phi (ø) as a measure of effect size (Cohen 1988). To investigate whether intervening bucks elected to engage with a competing male following intervention based on overall interaction rate, we constructed two square matrices. Each matrix contained all males that were involved in interventions whether as intervener, target or ignored (1996: N ¼ 49; 1997: N ¼ 50). One of the two matrices contained the overall number of dyadic interactions while the second matrix contained the overall number of interventions. These matrices were analysed and the association using the Mantel Z statistic computed (Mantel 1967). Descriptive statistics are given as X  SE. All P values presented are two tailed. RESULTS Dominance and Intervention Behaviour Dominance rank appeared to be related to the number of interventions by individual males in both years (Fig. 1a, b). The tendency to engage in third-party intervention behaviour ranged from 48.6% (N ¼ 34) of males in 1996 to 66.6% (N ¼ 42) of males in 1997. An ANCOVA using tendency to intervene as a categorical variable with year as a covariate indicated that males that held high dominance rank in both years (1996: X ¼ 289  34; 1997: X ¼ 33:5  3:1) were significantly more likely to intervene than low-ranked males (1996: X ¼ 41:7  3:2; 1997: X ¼ 37:4  4; ANCOVA: F1,130 ¼ 5.41, P ¼ 0.02) and this was not influenced by year (F1,130 ¼ 0.49, P ¼ 0.49). The tendency to intervene also appeared to be related to individual aggression rates; the number of interventions increased with tendency to fight (Fig. 1c, d). As with dominance rank, individuals that were observed to engage in third-party intervention behaviour fought significantly more during the observation period (1996: X ¼ 7:2  0:7; 1997: X ¼ 4:8  0:6) than individuals that did not intervene (1996: X ¼ 2:4  0:5; 1997: X ¼ 3:3  0:7; ANOVA: F1,130 ¼ 23.87, P < 0.001) and this effect was not influenced by year (F1,130 ¼ 2.49, P ¼ 0.12). After we randomly removed duplicate dyads, the difference in dominance ranks of the fighting males (1996: X ¼ 16:76  1:5; 1997: X ¼ 17:76  1:3) was smaller than the difference in

Association Between Intervening and Target/ignored Male The analysis of patterns of interaction between the intervening male and the target male showed a significant relationship between the dyadic and intervention matrices (Mantel t: 1996: t ¼ 6, P < 0.001, r ¼ 0.17; 1997: t ¼ 7, P < 0.001, r ¼ 0.21). A second analysis comparing patterns of interaction between the intervening male and the ignored male also indicated a significant relationship between the dyadic and intervention matrices (Mantel t: 1996: t ¼ 4.5, P ¼ 0.002, r ¼ 0.15; 1997: t ¼ 6.5, P < 0.001, r ¼ 0.19). According to these analyses, intervening males had similar patterns of agonistic contact with the target and ignored males and were not deliberately selecting the target male based on superior knowledge of its competitive ability.

Before the Intervention To investigate whether the intervening male was simply continuing a prior interaction with one of the fighting males, we examined the last recorded interaction (noncontact or fight) of both the target and the ignored males prior to the intervention. Interactions recorded prior to fight intervention were more likely to be with a fourth male rather than the target (target male versus fourth male: 1996: 5 versus 53; 1997: 4 versus 82) or ignored males (ignored male versus fourth male: 1996: 7 versus 52; 1997: 4 versus 82 interactions). Therefore, males did not intervene as a continuation of a prior interaction. The outcome (win, lose or draw) of the previous interaction was compared across the intervener, the target and the ignored males (Table 1). Males that intervened in fights were more likely to have won their previous fight than either the target or ignored males in both years of this study (1996: c24 ¼ 21:76, P < 0.001, ø ¼ 0.35; 1997: c24 ¼ 15:27, P ¼ 0.004, ø ¼ 0.27). Following the Intervention Intervention during a fight led to a higher probability that the intervening male would then interact with one member (the target male) of the fighting dyad than with a different (fourth) male (1996: 59 versus 18; 1997: 86 versus 26). Interveners were more likely to win their next interaction if it was against a member of the fighting dyad rather than against a fourth male (Table 2, data were pooled over both years to eliminate the effects of low expected values; c22 ¼ 21:43, P < 0.001).

346

D.J. Jennings et al. / Animal Behaviour 77 (2009) 343–348

12

(a)

10

8

Number of interventions

6

6

4

4

2

2

0

10

10

20

40 50 30 Dominance rank

60

70

0

20

10

30 40 50 Dominance rank

60

70

14 (c)

12

(d)

10

8

8

6

6

4

4

2 0

(b)

12 10

8

12

14

2 20

40 60 Number of fights

80

100

0

20

40 60 Number of fights

80

100

Figure 1. The distribution of third-party interventions and dominance rank held by individual males in (a) 1996 and (b) 1997 and the distribution of third-party interventions and overall fight rate of individual males in (c) 1996 and (d) 1997. Low numbers indicate high rank.

DISCUSSION Our data support a model of intervention behaviour where a winner or mixed winner–loser effect operates within the population (Dugatkin 1998). Under the model, interveners are not expected to target selectively one of the competing individuals. This prediction has been supported in a variety of species (de Waal & van Hoof 1981; Nelissen 1985; Watts 1991; Keil & Sambraus 1998; Perry 2003) and, in this respect, the present study adds to these findings. However, this study is the first that we are aware of that has investigated intervention behaviour in a free-ranging ungulate. Interveners did not elect to target one particular competitor because they had greater experience of dyadic contests with that male. The level of association between interveners and target males was equivalent to that between interveners and ignored males. Dugatkin’s model predicts that for intervention behaviour to be maintained in a population, winning an aggressive interaction should lead to further victories: a winner effect (Dugatkin 1997; Rutte et al. 2006). The specific mechanism by which experience of contest outcome mediates fighting behaviour still remains to be established (Hsu et al. 2006), although it has been argued that winning alters an individual’s resource-holding potential (Parker 1974). Where such an effect is evident in a population, interveners should act to prevent other group members

Table 1 The distribution of outcomes for interactions occurring in prior to fight intervention by the intervening, target and ignored males Male

Win

Lose

Draw

1996 Intervener Target Ignored

33 14 11

17 27 29

9 17 19

1997 Intervener Target Ignored

44 21 30

24 40 29

17 25 27

achieving a winner effect (Dugatkin 1998). It has been proposed that winner effects are dependent on the presence of a linear and, transitive hierarchy (Dugatkin 1997). Such conditions are met in our population (Jennings 2007), and, consistent with theoretical predictions, a previous study that we conducted indicated that a winner effect operates between mature males in this population (Jennings et al. 2004). The present study has extended this earlier finding by demonstrating that interveners were more likely to have won their interaction prior to intervention when compared with either of the competing males. Furthermore, interveners were also more likely to win the interaction following intervention if they interacted with a member of the competing dyad, rather than a different, fourth male. Aggressive interactions are costly to the participants in terms of time, energy, injury and/or death (e.g. Wilkenson & Shank 1976; Clutton-Brock et al. 1979; Briffa & Elwood 2002). Following intervention in our study, it was highly likely that the intervener would immediately engage in a contest. Intervention should, therefore, be a costly form of behaviour particularly where there is a high probability of retaliation on the part of one of the dyadic competitors (e.g. Kaplan 1978). The tendency for individual males to intervene in fights and their tendency to fight is related to dominance rank (Jennings 2000; this study). Furthermore, the difference in dominance rank between interveners and the competing males

Table 2 The outcome of interactions that directly followed a fight intervention based on whether the intervening male interacted with a male from the fighting dyad (target) or a different (fourth) male Interaction

Win

Lose

Draw

1996 Target male Fourth male

44 5

5 8

10 5

1997 Target male Fourth male

65 16

3 5

18 5

Raw data from both years were combined for statistical analysis.

D.J. Jennings et al. / Animal Behaviour 77 (2009) 343–348

in the present study was significantly greater than the difference in rank between the competing males themselves. While this indicates that interveners behaved in a manner consistent with a form of risk or cost minimization, dominant males do engage in riskier behaviour than subordinates during fights. For example, dominant males are more likely to terminate displays during fights in favour of a renewed bout of fighting relative to a subordinate (Jennings et al. 2002, 2003). In addition, dominant males engage in more risky behaviour than subordinate males during fights (Jennings et al. 2005a). We also predicted that for intervention to be selectively favoured in a population, interveners should benefit by defeating the target male in any subsequent interaction. This prediction was also supported; where interveners disrupted a fight and continued to interact with one member of the fighting dyad, the intervener was more likely to win than when the intervener engaged in a contest with a fourth male. Support for our predictions suggests that not only do fallow deer seek to deny a winner effect to rival males but they also seek to maintain their own winner effect relative to the males in the competing dyad. While our data indicate that interveners show no preference for either member of the fighting dyad, there is evidence from other species that group members preferentially interact with dominant group members (e.g. Seyfarth 1980; Packer & Pusey 1982; de Waal 1991; Feh 1999; Engh et al. 2005; Connor & Mann 2006). It has been claimed that knowledge of third-party relationships is an important proxy indicator of social intelligence, a premise based on the idea that the ability to process third-party relationships is cognitively challenging (Tomasello & Call 1997). Whether cervids can process social information derived from dyadic and triadic interactions between group members has been questioned (e.g. Freeman et al. 1992; Appleby 1993). Consistent with this line of argument, it has been proposed that mammals of the order Artiodactyla should not be capable of such complex cognitive tasks based on their small relative cortex size and the low cognitive demands of ungulate social systems (Shultz & Dunbar 2007). Nevertheless, it could be argued that the capacity to understand rank relations within the wider social group would be advantageous in terms of risk reduction. Males in our study population form a large multigenerational group for most of the year. Furthermore, dominance relations in this population are both stable and transitive (Jennings 2007). Given the high levels of fighting observed in this population, it is therefore possible that males might benefit from monitoring and (re-) assessing each other’s social rank (Bond et al. 2003; Croney & Newberry 2007). However, in deer, whether males monitor each other for competitive ability is, at present, open to debate (e.g. Jennings et al. 2003, 2004, 2005a; Bartosˇ et al. 2007). In the present study, intervening males showed no preference for either of the fighting males, consistent with the proposal that deer lack the ability to discriminate the rank position of group members relative to each other (Appleby 1993; Jennings et al. 2006). Certain predictions derived from Dugatkin’s (1998) model are shared with other models of third-party intervention, specifically, models that describe intervention behaviour within primate social groups. Primate social systems are considered to be highly complex, involving strategic social behaviours (Whiten & Byrne 1988; Harcourt 1989; Pawlowski et al. 1998; Kudo & Dunbar 2001; Cheney & Seyfarth 2007). These complex forms of social behaviour suggest that some level of monitoring behaviour takes place within groups and it is within this context that models of policing behaviour have been developed. Policing models predict that intervention is an impartial method of breaking up potentially costly interactions and that intervention is conducted by high-status individuals (Flack et al. 2005). Similar to Dugatkin’s model, therefore, policing models predict that intervention will not favour one individual over the other and that dominant

347

animals are expected to engage in policing more often than subordinate animals. Our results show that interveners were more likely to be dominant to both members of the competing dyad, consistent with behaviour shown in cercopithecine primates (reviewed in Harcourt 1989; also Chapais 1983; Silk 1993; Roeder et al. 2002). It is, however, the underlying rationale for intervention that differs between these models. In the case of Dugatkin’s model, intervention serves to prevent either animal from achieving a winner effect, thereby advancing up the hierarchy and becoming a threat to the intervener. Policing, on the other hand, serves to preserve social cohesion within the group, thereby maintaining stable resource networks: a significant benefit of group living (Flack et al. 2006). In terms of the present study, Dugatkin’s model would appear to be more appropriate for two reasons. (1) Any sense of group cohesion enjoyed by males in this population ends with the dispersion of the bachelor group in mid-September when mature males begin the rut. (2) In the majority of instances where intervention was noted, the intervening male went on to interact aggressively with one of the fighting males. In the context of the present study, this result is more than likely to be a consequence of the relationship between the tendency for individuals to intervene and dominance rank. In more general terms, Flack et al. (2005) have noted that policing does not appear to operate in multimale groupings, although it is unknown why this is the case. In the context of the present study, we speculate that where there is competition for limited, nondivisible resources such as mating opportunities, policing as a strategy should fail. Although it has been argued that agonistic behaviour in deer is restricted to dyadic level contests, the present study has shown that this is not the case. Approximately 10% of all fights in this population were terminated by third-party intervention. In lieu of comparable data from other ungulate species, our results indicate a moderate level of intervention behaviour compared with intervention rates in primate species such as boboons, Papio cynocephalus (4–6%, Silk et al. 2004) and bonnet macaques, Macaca radiata (22%, Silk 1992). The data collected for the present study do not lend themselves to an explanation as to why intervention behaviour falls consistently at about 10%. However, we suggest that intervention rates are not only mediated by a winner effect in this population; our results suggest that interveners must also be both dominant to the fighting males and in physical proximity to the fight to intervene. Nevertheless our results do support a winner effect in promoting intervention behaviour. Among different adaptive hypotheses of winner effects, intervening has been linked with the idea that it prevents rival males from experiencing a winner effect and thus serves to benefit the intervener directly (Dugatkin 1998). We further speculate that intervention also benefits the intervener in other ways, specifically that the dominance relationship between the fighting males remains in an ambiguous state (Jennings et al. 2005b). We have shown that fallow deer intervene in fights comprising subordinate males and suggest that this could serve to regulate risks and costs associated with intervention behaviour. This permits an already high-status individual to (re)assert his dominance status over subordinate males with a high probability of retaining the benefits of any winner effect. Acknowledgments We extend our thanks to Professor Tom Hayden and our colleagues at the Mammal Research Group for their help in the field. We also thank Drs Daniel Nettle, Mike Cox and Ben Brilot and the anonymous referees for their helpful comments on the manuscript. M.P.G. and C.M.C. received financial support from Enterprise Ireland.

348

D.J. Jennings et al. / Animal Behaviour 77 (2009) 343–348

References Altmann, J. 1974. Observational study of behavior: sampling methods. Behaviour, 49, 227–267. Appleby, M. C. 1982. The consequences and causes of high social rank in red deer stags. Behaviour, 80, 259–273. Apollonio, M., Festa-Bianchet, M. & Mari, F. 1989. Correlates of copulatory success in a fallow deer lek. Behavioral Ecology and Sociobiology, 25, 89–98. Appleby, M. C. 1993. How animals perceive a hierarchy: reactions to Freeman, et al. Animal Behaviour, 46, 1232–1233. Bartosˇ, L., Fricˇova´, B., Bartosˇova´, J., Panama´, J., Sˇustr, P. & Sˇmidova´, E. 2007. Estimation of probability of fighting in fallow deer (Dama dama) during the rut. Aggressive Behavior, 33, 7–13. Bond, A. B., Kamil, A. C. & Balda, R. P. 2003. Social complexity and transitive inference in corvids. Animal Behaviour, 65, 479–487. Bowyer, R. T. 1986. Antler characteristics as related to social status of male southern mule deer. Southwest Naturalist, 31, 289–298. Briffa, M. & Elwood, R. W. 2002. Power of shell-rapping signals influences physiological costs and subsequent decisions during hermit crab fights. Proceedings of the Royal Society of London, Series B, 269, 2331–2336. Ceacero, F., Landete-Castillejos, T., Garcı´a, A. J., Este´vez, J. A. & Gallego, L. 2007. Kinship discrimination and effects on social rank and aggressiveness levels in Iberian red deer hinds. Ethology, 113, 1133–1140. Chapais, B. 1983. Dominance, relatedness and the structure of female relationships in rhesus monkeys. In: Primate Social Relationships (Ed. by R. A. Hinde), pp. 208–219. Oxford: Blackwell Scientific. Chapman, D. I. & Chapman, N. G. 1975. Fallow Deer: Their History, Distribution and Biology. Lavenham: T. Dalton. Cheney, D. L. & Seyfarth, R. M. 2007. Baboon Metaphysics: The Evolution of a Social Mind. Chicago: University of Chicago Press. Clutton-Brock, T. H. & Albon, S. D. 1979. The roaring of red deer and the evolution of honest advertisement. Behaviour, 69, 145–170. Clutton-Brock, T. H., Albon, S. D., Gibson, R. M. & Guinness, F. E. 1979. The logical stag: adaptive aspects of fighting in red deer (Cervus elaphus L.). Animal Behaviour, 27, 211–225. Clutton-Brock, T. H., Green, D., Hiraiwa-Hasegawa, M. & Albon, S. D. 1988. Passing the buck: resource defence, lek breeding and mate choice in fallow deer. Behavioral Ecology and Sociobiology, 23, 281–296. Cohen, J. 1988. Statistical Power Analysis for the Behavioral Sciences. Hillsdale, New Jersey: L. Erlbaum. Connor, R. C. & Mann, J. 2006. Social cognition in the wild: Machiavellian dolphins? In: Rational Animals? (Ed. by S. Hurley & M. Nudds), pp. 329–367 Oxford: Oxford University Press. Croney, C. C. & Newberry, R. C. 2007. Group size and cognitive processes. Applied Animal Behaviour Science, 103, 215–228. David, H. A. 1987. Ranking from unbalanced paired-comparison data. Biometrika, 74, 432–436. David, H. A. 1988. The Method of Paired Comparisons. London: C. Griffin. Dugatkin, L. A. 1997. Winner and loser effects and the structure of dominance hierarchies. Behavioral Ecology, 8, 583–587. Dugatkin, L. A. 1998. Breaking up fights between others: a model of intervention behaviour. Proceedings of the Royal Society of London, Series B, 265, 433–437. Ehardt, C. L. & Bernstein, I. S. 1992. Conflict intervention behaviour by adult male macaques: structural and functional aspects. In: Coalitions and Alliances in Humans and Other Animals (Ed. by A. H. Harcourt & F. B. M. de Waal), pp. 83–111. Oxford: Oxford University Press. Engh, A. L., Siebert, E. R., Greenberg, D. A. & Holekamp, K. E. 2005. Patterns of alliance formation and postconflict aggression indicate spotted hyaenas recognize third-party relationships. Animal Behaviour, 69, 209–217. Feh, C. 1999. Alliances and reproductive success in Camargue stallions. Animal Behaviour, 57, 705–713. Festa-Bianchet, M., Appllonio, M., Mari, F. & Rasola, G. 1990. Aggression among lekking male fallow deer (Dama dama): territory effects and relationship with copulatory success. Ethology, 85, 236–246. Flack, J. C., de Waal, F. B. M. & Krakauer, D. C. 2005. Social structure, robustness and policing cost in a cognitively sophisticated species. American Naturalist, 165, 126–139. Flack, J. C., Girvan, M., de Waal, F. B. M. & Krakauer, D. C. 2006. Policing stabilizes construction of social niches in primates. Nature, 439, 426–429. Freeman, L. C., Freeman, S. & Romney, A. K. 1992. The implications of social structure for dominance hierarchies on red deer (Cervus elaphus L.). Animal Behaviour, 44, 239–245. Gammell, M. P., de Vries, H., Jennings, D. J., Carlin, C. M. & Hayden, T. J. 2003. David’s score: a more appropriate dominance ranking method than CluttonBrock, et al.’s index. Animal Behaviour, 66, 601–605. Harcourt, A. H. 1988. Alliances in contests and social intelligence. In: Machiavellian Intelligence (Ed. by R. W. Byrne & A. Whiten), pp. 132–152. Oxford: Clarendon. Harcourt, A. H. 1989. Social influences on competitive ability: alliances and their consequences. In: Comparative Socioecology (Ed. by V. Standen & R. A. Foley), pp. 223–242. Oxford: Blackwell Scientific.

Hsu, Y., Early, R. L. & Wolf, L. L. 2006. Modulation of aggressive behaviour by fighting experience: mechanisms and contest outcomes. Biological Review, 81, 33–74. Jennings, D. J. 2000. Contest structure, dyad composition, duration and outcome of fights between male fallow deer (Dama dama L.). Ph.D. thesis, National University of Ireland, Dublin. Jennings, D. J. 2007. The effect of chemical immobilization on dominance rank in the fallow deer. Animal Behaviour, 74, 1107–1110. Jennings, D. J. & Gammell, M. P. In press. Ungulate contest behaviour. In: Animal Contests (Ed. by I.C.W. Hardy & M. Briffa). Cambridge: Cambridge University Press. Jennings, D. J., Gammell, M. P., Carlin, C. M. & Hayden, T. J. 2002. Does lateral presentation of the palmate antlers during fights by fallow deer (Dama dama) signify dominance or submission? Ethology, 108, 389–401. Jennings, D. J., Gammell, M. P., Carlin, C. M. & Hayden, T. J. 2003. Is the parallel walk between competing male fallow deer (Dama dama) a lateral display of individual quality? Animal Behaviour, 65, 1005–1012. Jennings, D. J., Gammell, M. P., Carlin, C. M. & Hayden, T. J. 2004. Effect of body weight, antler length, resource value and experience on fight duration and intensity of fallow deer. Animal Behaviour, 68, 213–221. Jennings, D. J., Gammell, M. P., Payne, R. J. H. & Hayden, T. J. 2005a. An investigation of assessment games during fallow deer fights. Ethology, 111, 511–525. Jennings, D. J., Gammell, M. P., Carlin, C. M. & Hayden, T. J. 2005b. Win, lose or draw: a comparison of fight structure based on fight conclusion in the fallow deer. Behaviour, 142, 423–439. Jennings, D. J., Gammell, M. P., Carlin, C. M. & Hayden, T. J. 2006. Is difference in body weight, antler length, age or dominance rank related to the number of fights between fallow deer (Dama dama). Ethology, 112, 258–269. Kaplan, J. R. 1978. Fight interference and altruism in rhesus monkeys. American Journal of Physical Anthropology, 49, 241–250. Keil, N. M. & Sambraus, H. H. 1998. ‘Interveners’ in agonistic interactions amongst domesticated goats. International Journal of Mammalian Biology, 63, 266–272. Kudo, H. & Dunbar, R. I. M. 2001. Neocortex size and social network size in primates. Animal Behaviour, 62, 711–722. Mantel, N. 1967. The detection of disease clustering and a generalized regression approach. Cancer Research, 27, 209–220. Moore, N. P., Kelly, P. F., Cahill, J. P. & Hayden, T. J. 1995. Mating strategies and mating success of fallow (Dama dama) bucks in a non-lekking population. Behavioral Ecology and Sociobiology, 36, 91–100. Nelissen, M. 1985. Structure of the dominance hierarchy and dominance determining ‘group factors’ in Melanochromis auratus. Behaviour, 94, 85–107. Packer, C. & Pusey, A. E. 1982. Cooperation and competition within coalitions of male lions: kin selection or game theory? Nature, 296, 740–742. Parker, G. A. 1974. Assessment strategy and the evolution of fighting behaviour. Journal of Theoretical Biology, 47, 223–243. Pawlowski, B., Lowen, C. B. & Dunbar, R. I. M. 1998. Neocortex size, social skills and mating success in primates. Behaviour, 135, 357–368. Pe´labon, C. & Joly, P. 2000. What, if anything, does visual asymmetry in fallow deer antlers reveal? Animal Behaviour, 59, 193–199. Perry, S. 2003. Coalitionary aggression in white-faced capuchins. In: Animal Social Complexity (Ed. by F. B. M. de Waal & P. L. Tyack), pp. 111–114. Cambridge, Massachusetts: Harvard University Press. Roeder, J. J., Duval, L. & Gosset, D. 2002. Aggressive and neutral interventions in conflicts in captive groups of brown lemurs (Eulemur fulvus fulvus). American Journal of Physical Anthropology, 118, 253–258. Rutte, C., Taborsky, M. & Brinkhof, M. W. G. 2006. What sets the odds of winning and losing? Trends in Ecology & Evolution, 21, 16–21. Seyfarth, R. M. 1980. The distribution of grooming and related behaviours among adult female vervet monkeys. Animal Behaviour, 28, 798–813. Shultz, S. & Dunbar, R. I. M. 2007. The evolution of the social brain: anthropoid primates contrast with other vertebrates. Proceedings of the Royal Society of London, Series B, 274, 2429–2436. Silk, J. B. 1992. 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. 212–232. Oxford: Oxford University Press. Silk, J. B. 1993. Does participation in coalitions influence dominance relationships among male bonnet macaques? Behaviour, 126, 171–189. Silk, J. B., Alberts, S. C. & Altmann, J. 2004. Patterns of coalition formation by adult female baboons in Amboseli, Kenya. Animal Behaviour, 67, 573–582. Tomasello, M. & Call, J. 1997. Primate Cognition. Oxford: Oxford University Press. de Waal, F. B. M. 1991. Rank distance as a central feature of rhesus monkey social organization: a sociometric analysis. Animal Behaviour, 41, 383–395. de Waal, F. B. M. & van Hoof, J. A. R. A. M. 1981. Side-directed communication and agonistic interactions in chimpanzees. Behaviour, 77, 164–198. Watts, D. P. 1991. Harassment of immigrant female mountain gorillas by resident females. Ethology, 89, 135–153. Whiten, A. & Byrne, R. W. 1988. Tactical deception in primates. Behavioral and Brain Sciences, 11, 233–273. } rnberg, P., Croucher, P. J. P., Bercovitch, F. R. & Widdig, A., Streich, W. J., Nu Krawczak, M. 2006. Paternal kin bias in the agonistic interventions of adult female rhesus macaques (Macaca mulatta). Behavioral Ecology and Sociobiology, 61, 205–214. Wilkenson, P. F. & Shank, C. C. 1976. Rutting fight mortality among musk oxen on Banks Island, Northwest Territories, Canada. Animal Behaviour, 24, 756–758.