Ecological Research

, Volume 24, Issue 3, pp 521–531

Complexity, dynamics and diversity of sociality in group-living mammals


    • Department of Evolutionary Studies of BiosystemsThe Graduate University for Advanced Studies

DOI: 10.1007/s11284-008-0563-4

Cite this article as:
Kutsukake, N. Ecol Res (2009) 24: 521. doi:10.1007/s11284-008-0563-4


Numerous studies in group-living animals with stable compositions have demonstrated the complex and dynamic nature of social behaviour. Empirical studies occasionally provide principles that cannot be applied directly to other group-living species. Because of this, researchers are required to address fine-scaled conceptual questions and to incorporate species-specific characteristics of the study species. In this paper, I raise three key topics that will promote our understanding of animal sociality: the effects of heterogeneous social relationships on the pattern, distribution, and function of social interactions; conflict management for maintaining group living; and meta-dyad-level perspectives for understanding dyadic social relationships and behaviours. Through the discussion of these topics together with examples of group-living mammals, I emphasise the importance of direct behavioural observations and functional analyses in studies of species- or taxonomic-group-specific characteristics of social behaviour in a wide range of taxonomic groups. In addition to approaches focusing on specificity, another approach that examines the general principles or common characteristics found across different taxonomic groups could provide synthetic and reductive frameworks to understand divergent sociality. The complementary use of these two approaches will offer a comprehensive understanding of social evolution in group-living animals.


SocialitySocial relationshipGroup livingMammalsSocial diversity


Understanding the emergence and evolution of complex sociality is one of the central issues in ecology and evolutionary biology (Wilson 1975; Maynard Smith and Szmathmáry 1995). Theoretically, it is predicted that animals form groups when the benefits of group living exceed the costs. In general, the benefits of group living include a decreased probability of being preyed upon, the sharing of useful information, thermoregulation, and an increased probability of winning competitions with conspecific or competing species (Krause and Ruxton 2002). On the other hand, costs mainly result from competition for limited resources, such as food and reproductive opportunities. Additional costs include a high probability of parasitic infection and increased disease transmission (Krause and Ruxton 2002).

Aggregation is widely seen in animals of various taxonomic groups (Wilson 1975; Krause and Ruxton 2002). In contrast to simply aggregated species with temporally fluid group composition, some group-living species form groups that have relatively stable compositions of members with non-random structures (Whitehead 1997; Whitehead and Dufault 1999; individualized societies; de Waal and Tyack 2003). In such groups, group members repeatedly engage in social interactions, and as a result, the histories of the social interactions shape the stable social relationships (Fig. 1; Hinde 1976). Sociality within these groups is regarded as an autonomous distributed system in which each group member employs social strategies to maximise the (inclusive) fitness. In addition, the strategy of each group member is decided by a complex feedback system that is affected by the social strategies of other group members. Such social environments are regarded as complex adaptive systems that produce distinct selective pressures and promote the formation of complex and dynamic social relationships, which cannot be observed in species with unstable group membership. Due to the complex and dynamic nature of animal sociality, researchers occasionally reach conclusions that cannot be applied to other taxonomic groups because of differences in ecology, life history, and other biological characteristics (Fig. 1). In such situations, researchers are required to address fine-scaled conceptual questions regarding animal sociality by incorporating the specific characteristics of the study species.
Fig. 1

Scheme of sociality in an animal group with heterogeneous social relationships. Sociality is determined by the combination of selective pressures, including adaptation to ecological factors and constraints of phylogeny and life history. The social structure can be regarded as a social network (i.e., an entity of the social relationships within a group) composed of dyadic social relationships between group members. Dyadic relationships are largely determined by individual (e.g., age and sex) or relationship (e.g., relatedness) intrinsic factors. In the figure, dashed circles represent relatedness among group members. Differences in brightness and thickness of the lines represent social bonds with different magnitudes. Note that the characteristics of social relationships are not necessarily determined by the relationship intrinsic factors (see text). The characteristics of each social relationship are also affected by both (A) the history of dyadic social interactions between two group members and (B) other group members. A box (A) shows temporal sequences of social interactions between two individuals (t indicates the present social interaction, while the past interactions are indicated by t–n). A box (B) shows the influences of other group members on dyadic social relationships

In this paper, I review the previous literature and selectively identify three key topics that I believe will promote our understanding of the complexity and dynamics of sociality in mammals. These topics include the following questions: (1) how do heterogeneous social relationships affect the pattern, distribution, and function of individual social behaviour, (2) how do group members regulate social relationships and maintain group living, and (3) how does the presence of other group members affect the relationships and interactions between two individual group members? Through the discussion of these questions, I point out that long-term comparative data of social behaviour are lacking in a wide range of taxonomic groups and emphasise the importance of direct behavioural observations and functional analyses of social interactions. In the latter part, I show the direction of future studies by discussing the contrasts between approaches focusing on specificity and generality for understanding animal sociality. Note that this paper does not aim to provide a comprehensive review of animal sociality. Rather, the cited examples are biased to long-lived mammals for the following two reasons. First, few attempts have been made to review social behaviour across different taxonomic groups of mammals, despite the fact that mammalian sociality has received attention in previous studies (Silk 2007). A lack of reflection on mammalian biological characteristics (i.e., long life history and the complexity of social structures) and research investigations might bias the comprehensive theorisation of social evolution in animals. Second, it is usually difficult to evaluate the fitness benefits of a single social interaction or social relationships in long-lived mammals (Silk 2002, 2007). Because of this problem, it has been difficult to estimate the importance of selection pressures caused by social behaviour (i.e., social selection) relative to the magnitude of natural selection and sexual selection. As a result, the investigations of evolutionary aspects of animal sociality remain one of the most challenging topics in current evolutionary ecology.

Effects of heterogeneous social relationships on social behaviour

Even in simply aggregated species with unstable compositions, the nature of temporal social relationships among individuals is heterogeneous because of differences in individual intrinsic factors, such as age and sex. In contrast to these species, the heterogeneity of social relationships is pronounced in group-living animals with structured societies, both because of the variation in relatedness between individuals and by the history of the repeated social interactions between group members. The characteristics of the social relationships shape the individual strategies and interactions between two individuals. Because the group structure is determined by the sum of the social relationships within a group, heterogeneity in social relationships ultimately results in heterogenic variation at the group level. The formation of stable and heterogeneous social relationships makes animal sociality increasingly complicated, such that principles found in aggregated species may not apply to group-living species with stable memberships and repeated social interactions. Therefore, it is vital to investigate how the heterogeneity of social relationships affects the pattern, distribution, and function of the social interactions. At the behavioural level, however, detailed analyses of social interactions have been conducted in limited species, such as primates, which prevents comparison of mammalian socialities.

Arguably the most influential hypothesis posited to explain heterogeneous social relationships is the kin selection hypothesis (Hamilton 1964; Maynard Smith 1964). The kin selection hypothesis predicts that related individuals engage in cooperative behaviour more often than unrelated individuals because of increased inclusive fitness (Hamilton 1964). Numerous studies have shown strong influences of relatedness on social behaviour, particularly on cooperation (reviewed in Dugatkin 1997; Korb and Heinze 2008). However, recent studies highlight that relatedness does not always guarantee valuable and cooperative relationships (Griffin and West 2002; West et al. 2002) and that kin selection benefits have been overestimated (Clutton-Brock 2002). For example, the costs of kin competition diminish the benefits of kin selection in viscous populations where individual dispersal is limited (Frank 1998; West et al. 2002). Cooperation among related individuals that superficially fits the kin selection hypothesis can often be explained by other ultimate mechanisms (Clutton-Brock 2002). Therefore, it is incorrect to assume a priori that relatedness is always tightly linked to cooperative social relationships.

Observations of social behaviour are one of the most informative and direct ways to investigate and quantify the effects of heterogeneous social relationships. Studies in the cooperatively breeding carnivore, the meerkat (Suricata suricatta), provide one such example. This species forms multi-male, multi-female groups with high reproductive skew. Subordinates, usually the offspring of dominant pairs, do not reproduce, but instead help in rearing the offspring of the dominant pairs. Like other cooperatively breeding species, help by subordinates enhances the reproductive output of dominants (Emlen 1997). At the same time, however, there is an intense intrasexual competition between same-sexed dominant and subordinate individuals, irrespective of the degree of relatedness between them. That is, intrasexual aggression is frequent among same-sexed group members in meerkats (Kutsukake and Clutton-Brock 2006a, b). This aggression is particularly pronounced among females (Clutton-Brock et al. 2006). Dominant females attack older subordinate females who are likely to reproduce and evict them from the group at the later stages of the dominant female’s pregnancy (Fig. 2; Kutsukake and Clutton-Brock 2006a). Evicted subordinate females show increased stress levels, decreased fertility, and abortion if pregnant (Young et al. 2006). Although the peaceful and cooperative aspects of cooperatively breeding species have received attention in previous studies, these results reveal the understudied competitive aspects of cooperatively breeding societies, i.e., relatedness does not guarantee peaceful relationships in cooperatively breeding species, and high reproductive skew is achieved as a result of intense intrasexual competition (Kutsukake and Clutton-Brock 2006a, b, 2008a, b; Clutton-Brock et al. 2006).
Fig. 2

Intrasexual reproductive conflicts among female meerkats. a Aggression by dominant females in relation to the age of subordinate females. Aggression frequency indicates the proportion of encounters with aggression divided by the total number of encounters for each dyad of the same-sexed dominant and subordinate females. b The proportion of females evicted by a dominant female from a group for each category of subordinate female age. The figures are redrawn from Kutsukake and Clutton-Brock (2006a). Individual mean ± SE is shown

The study of heterogeneous social relationships occasionally provides an opportunity to test general ecological principles and the function of social behaviour in sophisticated ways. One such example is the reconsideration of group size effects in vigilance behaviour. Vigilance is used mainly for anti-predator purposes in group-living animals (Caro 2005). Individual vigilance rates are known to decrease as group size or local gregariousness increases due to dilution and the presence of many ‘eyes’ (reviewed in Krause and Ruxton 2002; Caro 2005). However, recent careful studies have begun to notice social influences on vigilance and have clarified that group size effects are not universal (reviewed in Treves 2000). For example, in the wild chimpanzee (Pan troglodytes) the presence of other group members increases, not decreases, individuals’ levels of vigilance (Fig. 3a, b; Kutsukake 2006, 2007). Given that chimpanzees are a highly social species; this result indicates that vigilance may be directed toward competitive group members. If vigilance functions as a way to monitor other group members, it is predicted that the characteristics of social relationships affect vigilance level, with individuals being more vigilant toward competitive group members. The above prediction is supported by studies indicating that chimpanzees are more vigilant when non-association group members are nearby than when other categories of group members are nearby (Fig.  3c; Kutsukake 2006). Likewise, behavioural observations on wild chimpanzees show that the rate of self-directed behaviour, a behavioural indicator of emotional state and individual stress level (Maestripieri et al. 1992), shows a similar pattern (Kutsukake 2003). Specifically, females increased the rate of self-directed behaviour more when other group members were nearby than when a non-affiliative group member was in their proximity (Kutsukake 2003). Influences of social context and relationship quality have been reported in other mammals (e.g., brown capuchin monkeys Cebus apella, Hirsch 2002; giraffe Giraffa camelopardalis Cameron and du Toit 2005; elk Cervus elaphus Lung and Childress 2007; European rabbits Oryctolagus cuniculus; Monclus and Roedel 2008), suggesting that social influences of vigilance are common in group-living mammals.
Fig. 3

Influences of group members nearby (within 3 m) on individual vigilance level in wild chimpanzees. a Vigilance duration measured by a 2-min focal observation method in relation to the number of group members nearby. b Vigilance level measured by point sampling during a 5-min interval in relation to the number of group members nearby. c Effects of social relationships with proximate group members on individual vigilance level measured by point sampling during a 5-min interval. The figures are redrawn from Kutsukake (2006, 2007). Individual mean ± SE is shown

These studies suggest that intrinsic factors of relationships, such as relatedness, do not always determine the characteristics of social relationships. At the same time, the consideration of heterogeneous social relationships may facilitate a more fine-scaled understanding of ecological principles. The quantification of heterogeneous social relationships cannot be performed from long-term demographic data or life history data. Therefore, detailed functional analyses of social behaviour in individual animals are indispensable for understanding animal sociality.

Conflict management for the regulation of social relationships and maintenance of group living

Although animals form groups when the benefits exceed the costs, it is too simplistic to assume that this payoff is passively determined with no room for group members to affect its outcome. Some animals actively maximise the net benefits by reducing the costs in order to maintain group living. Social behaviours that reduce the costs of aggression (i.e., consumption of energy and time and damage of social relationships between opponents) are referred to as conflict management (Aureli and de Waal 2000). Here, I exemplify three behavioural options of conflict management (dominance, greeting, and reconciliation) and discuss what is known and unknown for each behaviour.

The most common form of conflict management is the mediation of dominance relationships by dominant and submissive interactions and signals (Pusey and Packer 1997; Kutsukake 2000; Preuschoft and van Schaik 2000; Searcy and Nowicki 2005). Although dominance relationships reduce unnecessary conflicts over limited resources (Pusey and Packer 1997), relatively little is known regarding the function of submission as a mechanism of conflict management. Submissive signals are used to convey relative dominance relationships among group members and act to reduce the probability of aggression by dominant individuals (Preuschoft and van Schaik 2000). However, closer examination of these functions has been rare (Flack and de Waal 2007). Some studies have found that social signals that were previously regarded as submissive are not (Stevens et al. 2005) and that submission does not function as a form of conflict management (Fig.  4a; Kutsukake and Clutton-Brock 2008b). Mathematical theories of submission have been formulated only recently (Matsumura and Hayden 2006). Therefore, it is necessary to functionally test submissive signals.
Fig. 4

Effects of a submission, b greeting, and c reconciliation on the occurrence of aggression or affiliation after each type of social behaviour. a The effects of submission during aggression on the probability of aggression reoccurrence in the next social interaction in meerkats. The figure is redrawn from Kutsukake and Clutton-Brock (2008b). Individual mean + SE are shown. b Effects of greeting behaviour on the probability of affiliation between interactants in black and white colobus. The figure is redrawn from Kutsukake et al. (2006). The mean probability based on pooled data is shown. c Effects of reconciliation on the probability of aggression reoccurrence in Japanese macaques. The figure is redrawn from Kutsukake and Castles (2001). Individual means are shown

To facilitate social harmony, animals engage in a ritualised pattern of non-aggressive behaviour that usually occurs during a reunion. This is referred to as greeting behaviour (Colmenares et al. 2000). For example, in the black-and-white colobus (Colobus guereza) individuals engage in ‘over-head mounting’ and other types of contact behaviour. Functional and contextual analyses have shown that this behaviour is performed mainly by subordinates to dominant individuals and facilitates the occurrence of affiliation between interactants (Fig.  4b; Kutsukake et al. 2006). This suggests that greeting behaviour functions as conflict management for this species. In other species, greeting behaviour may have different functions, including the reaffirmation of social bonds (guinea baboons Papio papio, Whitham and Maestripieri 2003), suggesting that ritualised contact among group members has evolved for divergent functions in different species.

The most explicit form of conflict management is ‘reconciliation’. Reconciliation is the affiliation of opponents following an aggressive interaction (de Waal and van Roosmalen 1979). Reconciliation functions to reduce the probability that the victim will suffer further attack by the aggressor (Fig. 4c; Aureli et al. 2002). In addition, the behaviour reduces the post-aggression stress of the opponents (Aureli et al. 2002). These observations indicate that this behaviour reduces the costs associated with aggression and resolves conflict among individuals (conflict resolution). Reconciliation does not occur after all cases of aggression, and occurrence rates show inter-species, intra-species, and within-group variation (Arnold and Aureli 2006). Reconciliation is particularly effective in repairing damaged social relationships between group members with strong social bonds because these dyads experience higher levels of post-aggression stress than ones with weak social bonds (long-tailed macaques Macaca fascicularis: Aureli 1997; Japanese macaques Macaca fuscata: Kutsukake and Castles 2001; chimpanzees: Koski et al. 2007). Aureli et al. (2002) predicted that reconciliation should be common among species that live in stable social groups, have individualised relationships, and experience hostility after aggression, particularly among species in which aggressive interactions disturb biologically valuable social relationships. Reconciliation is widely observed in group-living primates (Arnold and Aureli 2006) and in other group-living mammals (reviewed in Schino 2000; domestic goats Capra hircus: Schino 1998; spotted hyenas Crocuta crocuta: Wahaj et al. 2001; bottlenose dolphins Tursiops truncatus: Samuels and Flaherty 2000; Weaver 2003; Tamaki et al. 2006; domestic dogs Canis familiaris: Cools et al. 2008; wolves Canis lupus: Cordoni and Palagi 2008), suggesting that conflict resolution is a common behavioural option for group-living animals. However, reconciliation is not always demonstrated in group-living animals that fit the aforementioned criteria set by Aureli et al. (2002) (red-bellied tamarins Saguinus labiatus: Schaffner and Caine 2000; Schaffner et al. 2005; common marmosets Callithrix jacchusjacchus: Westlund et al. 2000; black lemurs Eulemur macaco: Roeder et al. 2002; meerkats: Kutsukake and Clutton-Brock 2008b; naked mole-rats Heterocephalus glaber: Kutsukake, personal observation; see also Kappeler 1993 and Rolland and Roeder 2000 for ring-tailed lemurs Lemur catta: van den Bos 1998 for domestic cats Felis catus). Social factors associated with the evolution of reconciliation require further discussion (Aureli et al. 2002; Kutsukake and Clutton-Brock 2008b), but it is likely that no single factor can explain the absence of reconciliation in all species.

As discussed, group-living animals have behavioural mechanisms that manage and resolve conflicts in order to cope with the temporal disturbance of social relationships. These types of behaviours are important in controlling the dynamics of social relationships and act to maximise the net benefits of group living. At the same time, the forms of behaviour and their relative importance vary among different societies. Furthermore, behavioural distributions across species are currently unknown. Further empirical studies in species with divergent socialities will facilitate our understanding of the selective pressures shaping conflict management.

Understanding of social behaviour from the meta-dyad-level perspective

One important lesson from previous studies of animal sociality is that animals live in complex social networks within which group members are embedded (Krause et al. 2007). Since a single social interaction between two individuals is affected by the social behaviours and relationships among other individuals in their social network, understanding the effects of social interactions should not be restricted to a local scale (i.e., specifically between the interactants), but should instead be examined at a meta-dyad or more global level (i.e., between third parties and within a social network; Cheney and Seyfarth 1990, 2007; Harcourt and de Waal 1992; McGregor 2005; Conradt and Roper 2005). Here, I outline two examples of why meta-dyad-level perspectives are necessary and how the consideration of meta-dyad-level perspectives promotes our understanding of animal sociality.

Among group-living mammals, social interactions involving more than three individuals are commonly seen. Such polyadic interactions can be frequently observed within the context of aggression. Group-living mammals, such as primates, spotted hyenas, bottlenose dolphins, wild dogs (Lycaon pictus), and coatis (Nasua nasua), use coalitionary aggression, i.e., joint aggression toward a third party, to gain social benefits such as dominance and reproductive advantages (de Waal 1982; Nishida 1983; Harcourt and de Waal 1992; Kutsukake and Hasegawa 2005; Engh et al. 2005; Romero and Aureli 2008). Third parties who are not involved in the aggressive bout may interact with the opponents following the behaviour (e.g., consolation, appeasement, solicited consolation; Watts et al. 2000; Das 2000; Wittig and Boesch 2003; Kutsukake and Castles 2004; Koski et al. 2007; Fraser et al. 2008), or dominant individuals can intervene in ongoing aggression, thus terminating the aggression (Petit and Thierry 2000; Flack et al. 2006). Social interactions in a given dyad influence the nature and occurrence of social interactions among bystanders: the third-party individuals interact with each other following aggression (Cheney and Seyfarth 1986, 1989; Aureli et al. 1992; Judge and Mullen 2005). The function of these social behaviours following aggression have been regarded as conflict management because they reduce the probability of escalated aggression and regulate the social relationships; however, only a few empirical studies have investigated the function of these types of social behaviour (Palagi et al. 2006; Koski and Sterck 2007). These polyadic interactions increase the complexity and dynamics of animal sociality because the possible combinations of interactants should increase in species with polyadic interactions relative to ones without these interactions.

The presence of group members affects social behaviour and relationships in animals even if these individuals are not directly involved in the social interactions. The biological market effect (Noë et al. 1991; Noë and Hammerstein 1994, 1995) is one theory that formulates how other group members affect the relationships between two individuals. This theory proposes that exchanges of social behaviour (e.g., cooperation) are affected by the ratio of asymmetric actor and receiver individuals. Previous studies have shown that the market effect operates in various ecological conditions (e.g., interspecific mutualism, Bshary 2001; Hoeksema and Schwartz 2001), including social behaviour (the economy of coalition formation, Noë 1992; allogrooming exchanges, Barrett and Henzi 2006; obtaining female sexual information by males, Stopka and Macdonald 1999; attitude toward ex-group members, Schaffner and French 1997). Studies in meerkats have shown that similar principles might operate in cooperatively breeding species. In meerkats, a low offspring-to-helper ratio reduces the per capita work burden of group members and is positively related to the reproductive success of dominant individuals (Clutton-Brock et al. 2001). Due to the Allee effect, groups with few numbers of helpers do not succeed in reproduction (Clutton-Brock et al. 2001). Therefore, it is predicted that the value of each subordinate is high for dominant males in small groups compared to large groups. Accordingly, it is predicted that the aggressiveness of a dominant individual is positively correlated with the number of subordinates within a group. Indeed, analyses of aggression by dominant males in meerkats have demonstrated that the number of subordinates in a group is positively related to the rate of intrasexual aggression by a dominant male (Fig. 5; Kutsukake and Clutton-Brock 2008a). At the same time, dominant males attack non-natal males who can be reproductive within a group more frequently than natal males, suggesting that reproductive conflict is reflected in social behaviour. Furthermore, the difference in aggression toward non-natal and natal males is slight in small groups, but exaggerated in large groups (Fig. 5; Kutsukake and Clutton-Brock 2008a). This result suggests that aggressiveness by dominant males is moderated by the number of subordinates within a group, including reproductive rivals, which is in agreement with the predictions of the biological market effect.
Fig. 5

Intrasexual reproductive conflicts reflected in aggression between dominant and subordinate male meerkats. Aggression frequency indicates the proportion of aggressive encounters divided by the total number of encounters for each dyad of same-sexed dominant and subordinate males. Subordinate males are categorised according to their relationship with dominant males. Non-natal males, both unrelated and non-sibling, can be reproductive rivals for dominant males. Figures are redrawn from Kutsukake and Clutton-Brock (2008a). Individual mean ± SE is shown

These examples show that social interactions and relationships between two individuals are under the influence of other group members. These examples further posit the necessity of meta-dyad-level perspectives in predicting and correctly interpreting animal social behaviour. Furthermore, proper statistics that would enable researchers to analyse social interactions at the meta-dyad level without dividing groups into dyads have not been fully developed. Recent advances in analytical techniques (e.g., Krause et al. 2007; Wey et al. 2008; Whitehead 2008) are expected to overcome this problem in future studies. For example, social network analysis (Scott 2000), which is based on mathematical ideas of graph theory, is a powerful analytic method that quantifies the metric of the social structure; a graph (i.e., network) is simply a set of lines (i.e., dyadic relationship between group members) connecting nodes (i.e., individuals), and social network analysis calculates the properties of the patterns formed by the lines, such as the density of the network or centrality of each node (Scott 2000). Social network analysis has recently been applied to investigate the dynamics of the social structure and relationships of animals (e.g., Krause et al. 2007; Wey et al. 2008; Whitehead 2008). By analyzing the structure of social networks without breaking them down to individual relationships, social network analysis could successfully quantify sociality across taxonomic groups and could provide unique insights that cannot be gained from analyses of dyadic relationships alone.

Conclusions and future directions: toward a comprehensive understanding of social diversity

In this paper, I have highlighted three conceptual questions that promote our understanding of sociality in animals: (1) the effects of heterogeneous social relationships on the pattern, distribution, and function of social behaviour, (2) conflict management for the maintenance of group living, and (3) meta-dyad-level perspectives in analysing social behaviour. Unfortunately, detailed analyses on sociality, as well as contextual and functional analyses on social behaviour, have been mainly conducted in limited taxonomic groups (e.g., primates and eusocial insects). Additional data on the structure and functions of sociality in various taxonomic groups is necessary. Although this paper mainly focused on mammals, the questions addressed here will be useful in analysing sociality in other vertebrates (de Waal and Tyack 2003; Korb and Heinze 2008; e.g., birds: Marler 1996; Emery et al. 2007; Seed et al. 2007). Some invertebrates also show social behaviour similar to that discussed in this paper (e.g., the individual recognition in the paper wasp Polistes fuscatus: Tibbetts 2002; ritualized greeting in crayfish Procambarus clarkii: Issa and Edwards 2006; coalition formation in fiddler crab Uca mjoebergi: Backwell and Jennions 2004; aggression and submission in paper wasp Polistes dominulus: Cant et al. 2006; allocleaning in the ocypodid crab Macrophthalmus banzai: Ueda and Wada 1996). Investigations on the functional and fitness consequences of these interactions will aid in clarifying whether these social behaviours have similar functions to those found in phylogenetically distant mammals and will help us understand the selective forces that promote the evolution of particular social behaviours. The similarities may be superficial because of the critical differences in biological characteristics between invertebrates and vertebrates. For example, the cognitive ability of study species may differentiate the function of the social behaviour, and researchers must consider the function of social behaviour separately according to taxonomic group. Although we do not have enough long-term and comparative data to discuss this idea, it seems that there is no a priori reason to employ such classification. More importantly, such a classification may not allow us to test the interesting hypothesis that the exhibition of socially complex behaviour requires sophisticated cognitive ability.

Thus far, I have stressed the importance of enhancing long-term behavioural observations of identified individuals in a wide range of species for the detailed analyses of social behaviour. In each species, sociality is determined by the combined effects of different ecological (e.g., predation pressure and food resources) and evolutionary (e.g., constraints of phylogeny and life history) factors (Fig. 1). In addition, social traits are shaped by complex frequency-dependent interactions and feedback loop systems among social strategies by conspecific individuals, which makes the social behaviour peculiar to a particular species. Based on these backgrounds, the uniformed fitting of ecological or evolutionary principles found in the majority of species may overlook species or taxonomic group-specific characteristics. In the examples cited in this paper, polyadic interactions and reconciliation do not occur in all group-living species. The investigation of such specific social behaviour is useful for elucidating the magnitude of the social complexity in each study species, and it is vital to address fine-scaled questions that incorporate species-specific or taxonomic-group-specific characteristics of social complexity and dynamics. At the same time, however, increasing numbers of empirical studies show divergent patterns of animal sociality and provide a synthetic view of social diversity. Previous studies have focused on key biological parameters that can be commonly confirmed in animals of different taxonomic groups and formulate the synthetic view of animal social diversity (e.g., reproductive skew: Vehrencamp 1983a, b; Emlen 1997; Johnstone 2000; Kutsukake and Nunn 2006, 2009; social network: Krause et al. 2007; Wey et al. 2008; Whitehead 2008; genetic structure: Ross 2001; life history: Arnold and Owens 1998, 1999; Hatchwell and Komdeur 2000; Ligon and Burt 2004; Blumstein and Armitage 1998; Helms Cahan et al. 2002; food type and distribution or socioecology: van Schaik 1983, 1989; Sterck et al. 1997; Isbell and Young 2002). An advantage of such a synthetic approach is that researchers can directly compare social characteristics across taxonomic groups in a quantitatively similar manner, which could reveal the broad evolutionary pattern in animal sociality and provide powerful predications that can be applied to broad ranges of taxonomic groups. Attempts to synthesise social diversity are still in the early stages, and it is questionable whether these attempts are sufficient to explain social diversity in animals. Therefore, more emphasis must be placed on the building and testing of synthetic theories of social evolution. It should be also noted that synthetic theories occasionally miss species- or taxon-specific traits. Too much emphasis on general principles will prevent our understanding of species-specific traits (“the devil is in the details”), while too much emphasis on specific questions will miss the broad perspectives of social diversity (“cannot see the forest for the trees”); thus, I emphasise the importance of the complementary use of both approaches for a comprehensive understanding of animal socialities.


This paper is based on the presentation for the Miyadi Award at the 55th Annual Meeting of the Ecological Society of Japan, March 2008. I would like to thank all colleagues who supported my previous studies. Special thanks go to my supervisors, mentors, and colleagues, particularly Toshikazu Hasegawa, Duncan L. Castles, Toshisada Nishida, Noyuri Suetsugu, Takafumi Ishida, Tim H. Clutton-Brock, Charlie L. Nunn, Kazuo Okanoya, Mariko Hasegawa, and Keiko K. Fujisawa. Masayo Soma, Masakado Kawata, Kazuhiro Eguchi, and Dan Blumstein gave critical comments on the manuscript. My studies were supported by JSPS Research Fellowships, RIKEN Special Postdoctoral Researchers Program, Hayama Center for Advanced Studies and Grant-in-Aid for Young Scientists B and Start-up (no. 18870025 and 20770023).

Copyright information

© The Ecological Society of Japan 2009