Dunbar y Schultz - Bondedness and Socialty

Review

Bondedness and sociality

Robin I.M. Dunbar1) & Susanne Shultz(British Academy Centenary Research Project, Institute of Cognitive & Evolutionary

Anthropology, University of Oxford, 64 Banbury Road, Oxford OX2 6PN, UK)

(Accepted: 31 March 2010)

Summary

Approaches to sociality have, in the past, focused either on group typologies or on thefunctional aspects of relationships (mate choice, parental investment decisions). In contrast,the nature of the social relationships that scale from the individual-level behavioural decisionsto the emergent properties represented by group typology has received almost no attention atall. We argue that that there is now a need to refocus attention on the bonding processes thatgive rise to social groups. However, we lack any kind of language with which to describeor classify these operationally, in part perhaps because social bonding is emotional (and,hence, ‘felt’). One task for the future is, therefore, to identify suitable indices that can beused to compare the degree of bondedness between individual animals both between speciesand, within species, between individual dyads in such a way as to be able to test functionalquestions.

Keywords: social bonds, bondedness, social evolution, visual monitoring, behavioural syn-chrony.

1) Corresponding author’s e-mail address: [email protected]

© Koninklijke Brill NV, Leiden, 2010 Behaviour 147, 775-803DOI:10.1163/000579510X501151 Also available online - www.brill.nl/beh

776 Review: Dunbar & Shultz

“It is certain that associated animals [i.e., those living together in socialgroups] have a feeling of love for each other which is not felt by adultand non-social animals.”

Charles Darwin (1871). The Descent of Man, Vol. 1, p. 76

Introduction

Sociality has always been a focus of interest for both ethologists and behav-ioural ecologists. Nonetheless, despite considerable interest in the structuralaspects of social systems during the 1960s and 1970s (Crook & Gartlan,1966; Wilson, 1975; Crook et al., 1976), in subsequent decades behaviouralecology came to be dominated mainly by concerns with individual decision-making (mating strategies, parental investment decisions, alliance formation,cooperation, foraging decisions, etc., usually in contexts where an animal’sbehaviour directly influences its own fitness). This switch of focus during themid-1970s was, of course, both necessary and heuristically desirable: under-standing fitness from the individual’s perspective is an essential requirementfor a proper understanding of sociality and its evolution. Much of the weak-ness of the early studies lay precisely in the fact that they ignored the funda-mental place of the individual (and the gene) in the evolutionary process.

Nonetheless, it is also evident that few species are truly asocial. Animalsdo not make decisions about their behavioural tactics in a demographic vac-uum; for many (if not all) species, most of their decisions necessarily takeplace within a demographic context. In many ways, this was precisely thepoint of Hamilton’s (1964) argument about inclusive fitness: every decisionan animal makes about how to act has implications for every other individualin the local population because, whether or not they are related, the conse-quences ramify back onto the actor — which is why Hamilton structured hisanalysis in terms of neighbour-modulated fitnesses, offering the more famil-iar Hamilton’s Rule and inclusive fitness as a mathematically more tractableproxy for this. In Hamilton’s conception, animals can (and indeed do: seeCheney, 1982) eschew favouring relatives if they gain higher fitness by be-having altruistically towards unrelated individuals.

But for those species for whom sociality itself is part of the individual’sfitness strategy, being able to maintain the effective functionality of a groupthrough time may have very significant implications for the individual fit-nesses of its members. Familiar examples include cooperative hunting (as in

Bondedness and sociality 777

social carnivores), cooperative rearing (as in many birds) and group-living asa defence against predators (as in many primates). The point is that failure tomaintain coordination and collective focus may undermine the effectivenesswith which the group as such subserves these functions. This should not beconfused with group selection, since the benefits are exclusively at the levelof the individual. Rather, this is group-level selection (Wilson, 2007; in thecontext of (social) niche construction, see also Odling-Smee et al., 2003) ormultilevel selection: here, the benefit is created by the emergent propertiesof the group (e.g., more successful rearing of young, higher rates of resourceacquisition, or lower predator-derived mortality), but the payoff remains ex-plicitly at the level of the individual. Such a process commonly involves in-dividuals having to make tradeoffs between different levels of the biologicalsystem (Dunbar, 1983a).

The functional approach that characterises behavioural ecology has ten-ded to encourage the view that the issues of real interest are the observablebehavioural outcomes (actual choices of what to feed on or whom to matewith), for the (perhaps obvious) reason that these have directly measurableeffects on fitness. As a result, there has been no particular need to investigatethe mechanisms involved in maintaining groups as coherent entities. Instead,social groups tend to be seen as one of a kind irrespective of taxon, the pre-vailing view being what we might describe as the dungfly model of sociality— individuals distribute themselves in response to the shifting balance of themoment-by-moment costs and benefits of being in a group (Krause & Rux-ton, 2002). Indeed, an entire literature on ungulate sexual segregation (as oneoutcome of grouping processes) has been developed largely on this premise(for general reviews, see Ruckstuhl & Neuhaus, 2000, 2002). However, suchfluidity of grouping pattern is not found in all species; in many cases, indi-viduals have few options as to whom they can associate with (in philopatricspecies, for example, the only option may be members of the natal group).

Thus, an ‘ideal free’ approach to sociality overlooks a problem that is, al-though obvious, of considerable evolutionary significance: in highly socialspecies, groups consist of sets of individuals that associate together to vary-ing degrees, and the extent of that associativeness (in particular, how easilysuch relationships can be prized apart) has implications for the strategies in-dividuals use to maximise their fitnesses. The amount by which individualsare willing to compromise their immediate interests in order to maintain the


integrity of the group depends precisely on the benefits conferred by the char-acteristics of the group. As social cohesiveness in this sense increases, theoptions available to an individual are increasingly constrained by the needsof other group members. Individuals may then be locked into the effective-ness with which their social group functions, and that in turn may depend onhow closely integrated the individual life strategies of its members are. Pair-bonded monogamy is perhaps the archetypal case: two unrelated individualscooperate in an explicit fitness-related task (reproduction).

The issue for all such social solutions to the problems of fitness is thatthey are implicit social contracts, and are, thus, susceptible to freeriders whocan very easily destabilise the fragile balance of cooperative solutions (En-quist & Leimar, 1993; Orstom et al., 1994; Nettle & Dunbar, 1997; Dunbar,1999; Gürerk et al., 2006). As a result, mechanisms are required to minimisethis risk. Social insects effectively solve this problem by imposing roles oncolony members either through olfactory cues or developmental constraints.Such solutions are obviously less in evidence in mammals and birds, and al-ternative ways of dealing with the freerider problem may then be required.However, this is virtually unexplored territory outwith the social insects.

One solution to the free-rider problem is to develop more sophisticatedstrategies to track relationships and past interactions with other individuals.The Social Brain Hypothesis (SBH), as developed in the particular case ofprimates, argues that increases in brain size are associated with the need tomanage an increasingly complex web of social relationships as groups in-crease in size (Byrne & Whiten, 1988; Dunbar, 1992, 1998). However, itnow seems that the group size model is more or less peculiar to the primates;in most other mammalian taxa (including artiodactyl ungulates, carnivoresand bats) as well as birds, brain size appears to be associated with endur-ing social (principally pairbonded) relationships rather than the total numberof group members (Shultz & Dunbar, 2007). Shultz & Dunbar (2007) ar-gued that this suggests that brain enlargement was initially kicked off by thecognitive demands of (reproductive) pairbonding, but became generalisedto other group members (in the form of non-reproductive relationships orfriendships: cf., Silk, 2002) in primates (and perhaps a limited number ofother taxa). This explains why primates (and perhaps only anthropoid pri-mates) differ from the great majority of other birds and mammals in havingdisproportionately more (female-)bonded social systems (i.e., those in which


females remain together, often as natal sets who maintain close affiliative re-lationships by grooming together and/or resting in close proximity to eachother). The number of such relationships is limited only by the number of in-dividuals in a group, and this explains why the social brain hypothesis takesan explicitly quantitative form in primates, but a qualitative (i.e., categorical)form in almost all other birds and mammals.

While groups can be classified along a continuum of cohesiveness rang-ing from solitary individuals through unstable aggregations to stable socialstructures, the issue we want to highlight here is that we have neither thelanguage with which to describe this continuum of sociality nor any theoriesto account for the observed differences between taxa in what, for want of abetter term, we will refer to as bondedness (sensu Wrangham, 1980; see alsoSilk, 2002). We, therefore, begin by asking what we mean by bondednessand why it might be relevant to understanding sociality; we then address themuch more thorny issue of how we can describe it operationally in such away as to be able to compare species and individuals that differ in bonded-ness and, thus, perhaps social complexity.

The nature of sociality

It seems likely that the ancestral state for most vertebrate lineages was soli-tary (Pérez-Barbería et al., 2007). In mammals, this is most commonly in-stantiated in the lifestyles of many of the smaller insectivores, although mostother orders offer comparable examples (the galagines among the primates,most of the smaller forest cervids and antelopes among the ungulates, andsome carnivores). Sociality (meaning permanent stable social groups) hasarisen independently across most vertebrate (and many invertebrate) orders,because individuals gain benefits from aggregating, including the reductionin predation risk (e.g., selfish herd, dilution, or improved vigilance benefits),more effective exploitation of patchy resources, mate finding or defence, andcooperative foraging or rearing of young.

The early socioecology literature implicitly recognised that there are de-grees of sociality by drawing a distinction between aggregations and con-gregations (labelled ‘casual’ and ‘demographic’ societies, respectively, byWilson, 1975). This distinction focussed, in a qualitative way, on the con-trast between loose herds of the kind found in many grazing ungulates (e.g.,


reindeer, wildebeest) on the one hand and, on the other hand, more formalgroups of the kind found in many primates, equids and social carnivores. Theloose aggregations of many ungulates, for example, often lack any kind ofstable membership and, hence, lack stable relationships; they contrast withthe kinds of intensely focussed relationships found in some Old World mon-keys. Similarly, the stable pairbonds of species like gibbons or, among theantelope, klipspringer seem to contrast starkly both with species that haveephemeral encounters with reproductive partners and with the looser socialrelationships that characterise not only some group-living monkeys (e.g.,squirrel monkeys and colobines) but also many cervids and bovids whoseinterest in social interaction might be described as, at best, casual.

Among mammals, the two most intense relationships are those betweenmother and offspring and those between reproductive partners, and it is likelythat bonded relationships arose from one or other (or both) of these (Curly& Keverne, 2005) (a similar claim could be made about other vertebrates,especially birds, but we restrict our comments here to the less controversialcase of mammals). Although these kinds of relationships are widespread,their duration and stability are highly variable across species. The generalmammalian reproductive strategy involves females nursing young until theyreach independence (and much the same could be said of altricial birds). Insome species, weaning (i.e., nutritional independence) coincides with dis-persal, at which point the mother-offspring bond is terminated; but in others,the bond between mothers and offspring extends well into maturity (and, insome cases, may be lifelong). Cervids and bovids show a somewhat extendedmother-offspring bond, such that females and their offspring from severalprevious litters may continue to associate (e.g., Guinness et al., 1979). How-ever, these relationships are often fluid, changing over seasons and graduallybecoming less intense over successive years. In contrast, by far the mostintense social relationships seem to occur in primates (and specifically in‘female-bonded’ primates sensu Wrangham, 1980). In these species, moth-ers and daughters (and/or sisters) often maintain close relationships for theirentire lives (in terms of both spatial and behavioural coordination), resultingin multi-generational matrilines (Silk et al., 2006a,b) whose coherence overtime impacts directly on the fitnesses of their members (Silk et al., 2003).

As with mother-offspring relations, relationships between reproductivepartners can also be highly variable. In some species, the relationship only


persists during courtship and copulation. Alternatively, a relationship be-tween reproductive partners can be much more stable and can result in anextended pairbond that, in some cases, may be lifelong. Although pairbondsmay have evolved in order to protect or ensure paternity, mate guarding cangive rise to either monogamy or to harem-holding depending on the distribu-tion of females and the synchrony of oestrus (Dunbar, 1988, 1995, 2003; vanSchaik & Dunbar, 1990; Komers & Brotheron, 1997; Brotherton & Komers,2003). Lifelong maintenance of pairbonds is commonly seen in (among othergroups) corvid and psittaciform birds, most canids, some small-bodied un-gulates, and some (mainly New World) primates.

In most of these cases, our evidence for claiming a relationship betweentwo individuals derives from observing that they interact closely or spenda lot of time together (sensu Hinde, 1976). However, this is purely a be-havioural description: it does not tell us anything about the nature of therelationship that underpins this association, or necessarily anything about itsstrength. Exactly what that ‘bonding’ process is remains unclear, but what-ever it is, it almost certainly has something to do with the individual’s inter-nal emotional state (Damasio, 1994; LeDoux, 1998; Niedenthal, 2007), andis probably mediated by an array of neuroendocrines that appear to play acrucial role in facilitating bonding (Keverne et al., 1989; Bales et al., 2004;Curly & Keverne, 2005; Dunbar, 2010a).

Defining bondedness operationally

In recent years, a number of seminal attempts have been made to under-stand how animals perceive their social environment by looking at individualrecognition (sheep: Tate et al., 2006; fish: Bshary et al., 2007), recognitionof third-party and kin relationships (primates: Parr & de Waal, 1999; Ch-eney & Seyfarth, 1999, 2007; Bergman et al., 2003) or the understanding ofother individuals’ perspectives (primates: Call, 2001; Hare et al., 2001, 2006;Bugnyar & Heinrich, 2005; ungulates: Kaminski et al., 2006; birds: Emery& Clayton, 2001; Dally et al., 2006; Emery et al., 2007; Clayton et al., 2007).Both primates and corvids, for example, appear to take into account relativesocial status and ‘knowledge’ state when making feeding decisions.

Such studies remind us that animals recognise other individuals’ iden-tities, status and perspectives, as well as remember past actions, but it is


not clear what they actually tell us about how animals perceive social rela-tionships as such. Nor does it tell us much about what, in proximate terms,makes an individual likely to behave generously towards one individual butnot another — unless, of course, such relationships involve nothing but cog-nition (i.e., are essentially explicitly rational). Studies of this kind undoubt-edly identify some of the mechanisms involved at the cognitive level, butthis is not what is at issue in respect of bondedness. Bondedness is not just acognitive process, even though cognition may be a necessary component;rather, bonded relationships appear to involve two parallel and quite dis-tinct mechanisms — a cognitive mechanism (derivative of the social brain)and an emotionally based form of attachment (often involving a psycho-pharmacological mechanism) (Dunbar, 2010b).

This is perhaps most explicitly so in anthropoid primates, where socialgrooming seems to be used to create relationships that demonstrably havefitness consequences (e.g., in terms of protective coalitions (gelada: Dunbar,1980, 1989), fertility (gelada: Dunbar, 1989) or rearing success (baboons:Silk et al., 2003)). Although we do not fully understand the processes in-volved, it seems that grooming, through the release of endorphins (Keverneet al., 1989; Dunbar, 2010a), provides a psycho-pharmacological mechanismthat enables two individuals to build a bonded relationship with some kindof deeply emotional basis. We suggest that this is a necessary prerequisitefor such individuals to build a more cognitively-based relationship, one thatis akin to trust in the sense that individuals are willing to behave altruis-tically towards each other. In the absence of some such explanation, it isdifficult to explain why such intense (and otherwise functionless) groom-ing should apparently play so important a role in primate sociality (Dunbar,1991; Lehmann et al., 2007) — and perhaps even in birds (Radford & DuPlessis, 2006) and other mammals that allo-groom regularly (e.g., equids).

One reason why defining the nature of bondedness is so problematicseems to be that it is intrinsically an emotional process, and we have no ad-equate language with which to describe such relationships even for our ownspecies. In this sense, the nature of our relationships is hidden even fromour own capacities to comment on them. We know when we have such rela-tionships, and we recognise them when we see them in others, but we do nothave the vocabulary to describe them. One plausible reason for this is that re-lationships are ‘felt’ (in the ‘embodied cognition’ sense of Damasio (1994))rather than being cognised. In other words, relationships are the product


of an internal emotional state that is created by a complex neuroendocrineenvironment (Young & Wang, 2004; Curly & Keverne, 2005; Niedenthal,2007) rather than being a strictly cognitive one that we can reflect on analyt-ically. In effect, these relationships are what cognitive psychologists wouldrefer to as ‘hot cognition’, as opposed to those kinds of ‘cold cognition’ weuse when we describe and reflect on our motives and experiences (Abelson,1963; Damasio, 1994).

In this respect, work on human relationships might provide some insights.Although there is an extensive social psychological literature on human re-lationships dating back several decades (Kelley et al., 1983), there has beenalmost no intellectual exchange between ethologists and social psycholo-gists, in part perhaps because the former have focussed on function whereasthe latter have emphasised mechanisms (often in the form of psychologi-cal intervening variables). The fact that early psychological approaches torelationships were exclusively qualitative has only served to exaggerate thegulf between the two disciplines. However, since Rubin’s (1970) seminalattempts to introduce a quantitative approach into the psychological descrip-tion of relationships, a number of studies have explored their underlying psy-chological dimensions (albeit, in most cases, using questionnaires). Withinthis literature, two approaches merit attention.

Sternberg’s (1997) rather unfortunately named ‘triangular theory of love’focusses mainly on romantic relationships, and argues that such relationshipsinvolve three key dimensions (intimacy, passion and commitment). Passionreflects those aspects of romantic relationships commonly associated withthe phenomenon of ‘falling in love’ (intense focus on the object of one’s de-sire, a sense of exhilaration, with or without a sexual component: Fisher etal., 2006), and may be of less immediate interest to us in the present context.In contrast, intimacy (associated with feelings of closeness, connectednessand bondedness) and commitment (associated with a desire to support, andremain in the physical presence of, another individual) apply to all types ofrelationships, and may, thus, have greater heuristic purchase for understand-ing animal relationships.

Berscheid et al. (1989) developed a particularly influential approach thattries to unpack this a little more. Their questionnaire-based instrument (theRelationship Closeness Inventory, or RCI) treats closeness as a multidimen-sional psychological construct composed of three key dimensions: frequency(time spent together), diversity (the variety of activities performed together)


and strength (influence that one person has on the other). Taken together, thisthree-dimensional index is robust across relationship types (romantic versusfriendship), reliable across time (high test-retest validity) and predictive ofrelationship duration. A subsequent factor analysis of the RCI and a numberof other verbal and pictorial relationship indices concluded that they clus-ter on just two key dimensions, labelled, respectively, ‘Behaving close’ and‘Feeling close’ (Aron et al., 1992). This suggests that intense relationshipscan be described in terms of a combination of physical proximity (i.e., fre-quency of interaction?) and emotional engagement.

To the extent that all these findings might hold across animals as wellas humans, the first of these (proximity) provides no difficulty, since it canbe operationalised in terms of measures of spatial proximity or affinitiveinteraction. In contrast, emotional involvement (or intimacy in Sternberg’sconceptualisation) is more difficult to characterise. Nonetheless, it clearlyresonates with the problem that is exercising us here. The thorn in the side forus is the fact that these social psychological approaches require subjects tosay something about their inner states. While it is one thing to ask language-competent humans to say something about their emotional states, it is anotherto find practical indices that can be used with non-verbal organisms (andthose, of course, include very young children).

One possible solution might be to take a leaf out of the social psycholo-gist’s book and examine correlations between different behavioural measuresand use factor analysis to extract a set of dimensions that might reflect theseinner states. Cords & Aureli (2000) used this approach with primates andidentified three key dimensions, which they termed value, compatibility andsecurity. Their schema has been used with some success in a number of stud-ies (Koski et al., 2007; Fraser et al., 2008; Majolo et al., 2009). Compatibil-ity reflects tolerance, affiliation and spatial proximity, and is easy to identifywith the psychologists’ concept of ‘behaving close’. However, the other twodimensions (value and security) are essentially functional criteria: value re-flects willingness to share food or provide coalitionary support, and securityhas to do with the predictability and consistency of a relationship over time.The problem is that both of these are functional outcomes, rather than simpledescriptors of relationship quality: we are still missing that ‘inner something’that makes the relationship work and gives rise to tolerance and relationshipstability. More worryingly, the factor analysis suggests that these are threeorthogonal dimensions, implying that the two functional dimensions do not


correlate with the only strictly descriptive dimension (compatibility), whichin turn implies that the functional value of relationships may have nothing todo with their strength (or at least this one aspect of their strength).

There has been considerable interest in the possibility of a neuroendocrinedescription of bonding. The evidence that oxytocin (and/or vasopressin inmales) plays a role in facilitating pairbonding in voles (Young & Wang,2004) has raised the possibility that the emotional component of bondingmay have a relatively simple neuroendocrine basis. While it seems unlikelythat anything as psychologically complex as a relationship could in fact be sosimply determined, there remains nonetheless an encouraging prospect thatsome quantitative measure of neuroendocrine receptor gene up-regulationmight provide an objective index of social bonding. However, despite a grow-ing interest in emotion in the neurosciences (Rolls, 1995; Le Doux, 1998),our current knowledge of the neuroendocrine basis of relationships is at bestpoor — not least because it is often based on a simple dichotomy betweenmonogamous (i.e., pairbonded) and promiscuous mating systems in threevole species (Bales et al., 2004; Young & Wang, 2004). Even with the bestwill in the world, monogamous voles can hardly be considered archetypal ex-emplars of bondedness in mammals, never mind primates. Although there areindications that other neuroendocrines, including serotonin and dopamine,are associated with more general measures of sociability, and a stronger casestill for endorphins, we are still a long way from a model linking species typ-ical social behaviour to genetic or hormonal profiles. Indeed, considerablecare is needed to avoid naïve overinterpretation. In one study of humans,for example, the presence or absence of a specific allele (RS3) for the va-sopressin receptor was found to differentiate between males with stable andunstable marital relationships, with the perhaps inevitable conclusion thatthis was the gene functionally involved in (male) pairbonding (Wallum etal., 2008). A more likely explanation is that this allele is related to an in-ability to control one’s behaviour (an ‘act first, think after’ syndrome) andhas nothing at all to do with any gene for pairbonding capacity as such (eventhough the ability to control one’s behaviour may be important for successfulpairbonding) (Knafo et al., 2007).

More importantly, even if we can specify exactly what is happening inneurobiological terms, this may not help us with exploring the functionalconsequences of bonding in freeliving (or even captive) animals. Assays for


neuroendocrines invariably require intervention studies, and while experi-mental studies provide an appropriate means of studying how a mechanismworks, they are usually too disruptive to help with understanding functionalconsequences, especially in free-living animals (which is what behaviouralecologists and ethologists will usually want to know).

This might lead us to be content (as many have been) with using indicesof physical proximity as a measure of bond strength. However, much de-pends on whether Sternberg and the psychologists are right in claiming thatbehaving close and feeling close are independent dimensions. A case couldbe made for claiming that humans (and perhaps great apes) have a uniquecapacity to keep other individuals in mind even when they are not physicallypresent (Barrett et al., 2003; Dunbar, 2003); because of this, humans mightbe able to separate out these two dimensions of bondedness in a way thatother species cannot. If so, then indices of proximity might reasonably sub-stitute for indices of bondedness for most (if not all) mammals and birds.But, if humans are not unique in this respect, then the one will not be anadequate index for the other: using purely behavioural indices of proximitywill then give us only a partial assessment of bond strength. This is a purelyempirical issue, but to resolve it we will need better indices of the emotionalcomponent of bondedness.

How to measure bonding?

Our problem, then, lies in finding suitable operational indices for the emo-tional component of bonded relationships that we can use both in compara-tive studies across a wide range of taxonomic groups and, within species, incomparing individual relationships and their functional outcomes. Our aimin this section is not to provide a comprehensive review, but rather to con-sider a range of possibilities and the limitations that they impose and, in sodoing, to provide further insights into the definition of bondedness. One ofthe central difficulties is that, in both the social psychological and the etho-logical literatures, relationships have been defined mainly in terms of instru-mental criteria (see Cords & Aureli, 2000; Silk, 2003). A great deal has beenmade of the instrumental reciprocities that underpin friendship, but in real-ity these are the outcomes of well-bonded relationships, not the relationshipitself. For this reason, we would not want to include phenomena like length


of pairbond/group membership, or the probability that one individual goesto another’s support in agonistic interactions, selectively shares nest sites orother resources with it or seeks to prevent third parties from gaining accessto the partner (see, for example, Silk et al., 2006a,b; Emery et al., 2007) asdefining criteria for a pairbond, since these are all instrumental outcomes ofthe bonded relationship rather than descriptors of the relationship itself: theyare the phenomena we want to explain by the strength of the pairbond.

In primates, spatial proximity (Kummer, 1968; Cords, 1997) and socialgrooming (Dunbar, 1991; Lehmann et al., 2007) are well documented andwidely used proxies for social bondedness, and this may likewise be trueof at least some other taxa (e.g., equids, some bats, some birds) that arefrequent social groomers (Wilkinson, 1986; Radford & Du Plessis, 2006).Neither proximity nor grooming are random in these species, but focussedpredominantly on a very select subset of group members (Kudo & Dunbar,2001; Lehmann & Dunbar, 2009). However, social grooming is not espe-cially widespread outside this handful of taxa, so it may not necessarily bethat useful as a general index of bondedness unless we can first show thatonly those species with bonded relationships engage in social grooming andthat all species that do not are one of a kind socially. While spatial proxim-ity is a more widely applicable index, it suffers from one key disadvantage:two individuals may associate with each other simply because of a commoninterest in a third party or because predation or patchy foraging conditionsinduce them to form aggregations (Cords, 1997; Silk, 2002).

On a broader level, the amount of time devoted to social interaction mayprovide an index of how demanding different relationships are for individuals(see Shultz & Dunbar, 2007). However, there are a number of issues that needto be resolved in this case. Firstly, individuals in large groups may spendmore total time socialising, yet have only relatively loose associations withmany individuals. Conversely, pairbonded species such as klipspringer mightspend less time in social activity, but devote it all to a single partner. Which ofthese species is the more bonded? A second issue is whether data for differenttaxa have been recorded in comparable form. In primates, most behaviouralresearchers record social activity as a separate class of behaviour (in additionto feeding, resting, moving, etc.) because these species devote so much timeto social grooming. However, in less intensely social species, researchersoften include social behaviour in a nondescript category of miscellaneousbehaviours or even simply subsume it under a more general category of


resting time. As a result, it can be quite difficult to obtain truly comparablemeasures of social engagement across taxonomic groups. Nonetheless, on afirst pass analysis of data of this kind, Shultz & Dunbar (2007) found thatpairbonded primates and ungulates devote significantly more time per groupmember to social activities than primate species living in bonded groups,who in turn spend more time than ungulate species that live in unbondedgroups. However, while useful proxies, all these measures involve the samekind of circularity as other instrumental behaviours. These proxies are asmuch outcomes of being bonded as they are indices of bonding and we willoften want to know why individuals who groom or forage together do so.

A more promising index might be behavioural responses to separation,an index that was widely used in captive studies of primates (Mason, 1965;Mason & Mendoza, 1998; Aureli & Schaffner, 2002). However, while thisprobably does provide a good index of emotional bondedness, the fact that itrequires experimental manipulation rather obviates its value as a tool for fieldstudies. Even in laboratory studies, disturbing one’s study animals to such adegree will usually be unhelpful for functional studies of the value of socialbonds. A refinement of this approach would be to use the cost an animal isprepared to incur in order to be with specific individuals — an index thathas been used to very good effect to assess the demand characteristics ofanimals’ welfare needs (Dawkins, 1983). However, ‘welfare trade-off ratios’of this kind suffer much the same disadvantages as separation studies: theyare not easy to effect without experimental interventions.

An alternative approach might be to use behavioural responses when an-imals meet after a natural period of separation or following a natural chal-lenge (e.g., a territorial dispute or an attempted group takeover). In elephants,for example, members of the same bond group greet each other after an ab-sence with levels of excitement that are never seen when members of dif-ferent clans (the next social layer up) meet (Moss & Poole, 1983; McCombet al., 2000). Spider monkeys (Ateles geoffroyi) use ritualised greeting be-haviours when sub-groups reconvene (Aureli & Schaffner, 2007). Similarly,blue monkey (Cercopithecus mitis) females groom each other intensely afterengaging in collective defence of their group territory (Cords, 2002). How-ever, detailed descriptions of such behaviour are rare, and quantitative de-scriptions that would permit comparative analyses are virtually non-existent.More importantly, we can only use behaviours displayed after a period ofseparation in those species that have natural fission–fusion social systems; it


cannot solve the problem of how to measure bond strength in those specieswhere bonded individuals are rarely separated (e.g., pairbonded ungulatesand many primates).

Another possibility might be the frequency with which reconciliation isused to ameliorate the disruptive (and hence dispersive) effects of within-group aggression (see Aureli & de Waal, 2000). There is, after all, no point inreconciling after aggression if there is no long term relationship that requiresrepair. Indeed, the question of how reconciliation is affected by the qualityof a relationship (which might be interpreted in terms of bond strength) isone that has exercised those who work on this topic (van Schaik & Aureli,2000). However, some key exceptions aside (e.g., dogs, hyaenas, goats anddolphins: Schino, 1998; Samuels & Flaherty, 2000; Wahaj et al., 2001; Coolset al., 2008), there is little evidence to suggest that reconciliation is widelycommon outside the primates (Aureli & de Waal, 2000) and, thus, its valuemay be as limited as grooming. More importantly, perhaps, reconciliation isan instrumental response and risks circularity: one of the core questions inthe study of reconciliation is why some pairs reconcile more than others, and‘good relationships’ (i.e., bondedness) is one common explanation.

One aspect of bonded relationships that would seem to be germane to theirsurvival through time is the need to coordinate and synchronise behaviour.Pairs of individuals, whether mates or allies, would not be able to achievetheir fitness objectives (whether these be shared reproduction or functionalalliances) if the absence of behavioural coordination resulted in their driftingapart. Physical togetherness not only constitutes one of the core dimensionsthat social psychologists have identified as defining friendships (Sternberg,1977; Berscheid et al., 1989), but it also highlights a second, namely the needto share activities and interests. The broader issue, however, is not so muchthe sharing of common interests (though that is how this might be manifestedin humans), but in coordinating and synchronising one’s essential activitieswith those of one’s partner(s). Failure to do so inevitably results in the pairmembers failing to be in the right place at the right time. In species thatoperate a fission–fusion form of sociality (such as many ungulates), lack ofbehavioural synchrony inexorably leads to group fission and the spatial sep-aration of individuals (Conradt, 1998a; Calhim et al., 2006; Dunbar & Shi,2008). Studies of ungulates, for example, have identified activity synchronyand preferences for the certain kinds of habitat as major factors giving riseto sexual segregation (Ruckstuhl & Kokko, 2000; Ruckstuhl & Neumann,


2002; Calhim et al., 2006). Similarly, in monogamous species of antelope(e.g., klipspringer, dikdik) and primates (e.g., gibbons), pairs that do not co-ordinate their behaviour are likely to become separated as their individualactivity cycles become desynchronised; physical separation may then resultin animals being exposed to increased risks of predation, infanticide, extra-pair matings, harassment or reduced foraging efficiency (Dunbar & Dunbar,1980; van Schaik & Dunbar, 1990; Komers & Brotherton, 1997; Brotherton& Komers, 2003). At least in humans, synchronised action yields a signif-icant upsurge in endorphin production (Cohen et al., 2010) and has a dra-matically positive effect on generosity in public good games (Wiltermuth &Heath, 2009).

Behavioural coordination is clearly a significant issue for most pairbondedspecies that rely on biparental care for the successful rearing of offspring.Monogamous birds that do not coordinate their foraging schedules risk leav-ing eggs or nestlings exposed to cooling and/or nest predation if both adultsforage at the same time. Similarly, in klipspringer (perhaps the most intenselypairbonded of all the ungulates), one member of the pair is required to standguard and keep watch for predators while the other feeds (Dunbar & Dunbar,1980); failure to ensure a balanced sharing of this duty is likely to result inone or both members of the pair being forced to feed alone, with increasedrisk of being caught unawares by a predator. Thus, anti-synchrony (as op-posed to the asynchronised behaviour that gives rise to social segregation inungulates) may be as much a hallmark of a deeply bonded relationship asclose synchrony. The problem is more one of finding a suitable index of syn-chrony that reflects the extent to which pairs of individuals coordinate theirbehaviour. Some variant on the synchrony indices developed to study ungu-late social segregation (e.g., Conradt, 1998b) might, thus, prove profitable.

Vocal interactions might be relevant here. In many species, communica-tion between individuals serves to coordinate group activity and may, thus,indicate the amount of negotiation necessary to maintain synchrony. For ex-ample, contact calls including clear calls (Uster & Zuberbuhler, 2001), grunts(Stewart & Harcourt, 1994) and trills (Boinski, 1993) have been linked togroup movement and coordination in primates. In gelada, females engagein contact call bouts with their core grooming partners during feeding thathave all the hallmarks of ‘grooming at a distance’ (Dunbar, 1988: p. 251).Boubou shrikes (Lanarius spp.) are similarly renowned for the closely timed


vocal duets of pairs (Thorpe & North, 1966). Other forms of non-vocal syn-chronised display behaviour (such as that seen in dolphins: Connor, 2007)might also be relevant. It is worth noting that in both primates (McComb &Semple, 2005) and chickadees (Freeberg, 2006) vocal complexity increaseswith group size, within as well as between species, suggesting that theremay be selection pressure to increase both the amount and complexity ofvocalisations when there are more relationships to be managed. However,the possibility that vocalisations might be involved in servicing social rela-tionships in this way is probably one of the most under-researched topics inbehavioural ecology.

Another little explored aspect of behavioural synchrony that might pro-vide some valuable insights is social monitoring. Paying close attention toa partner is likely to be an essential requirement for behavioural synchrony,in that behavioural coordination requires that at least one member of thedyad adjust its behaviour to fit that of its partner — something it can onlydo this only if it attends closely to the partner’s behaviour. Kummer et al.(1974), for example, noted that hamadryas males seem to use the attentive-ness of a female to her harem male as a cue of the strength of the relationshipbetween them (and hence of the chances the male might have of capturingthe female) (see also Dunbar, 1984; Smuts, 1985). In humans, fixated gaze iswidely recognised as a feature of romantic engagement (Fisher, 2006), and iscertainly an important cue indicating relational interest during conversations(Vertegaal et al., 2001; Vertegaal & Ding, 2002). Hence, social monitoringof other group members (and the extent to which this is reciprocated) maybe a particularly useful index of the extent to which individuals are bonded.

That within-group social vigilance may be an important dimension of so-cial dynamics is not an entirely new suggestion: Chance & Jolly (1970) sug-gested many years ago that primates could be differentiated into those with‘centripetal’ and those with ‘centrifugal’ patterns of visual attention. The for-mer category would identify species with more intense social relationships inwhich individuals constantly monitor other members of the group. Given theintellectual climate of the time, Chance & Jolly emphasised the overridingimportance of dominance rank as the key organising principle in centripetalgroups where individuals were anxious to keep track of rivals and avoid thethreat of attack by more dominant animals. In contrast, we would want toplace a greater emphasis on affiliative relationships.


Unfortunately, almost without exception, the work on vigilance hithertohas focussed on monitoring either for predators (Hart & Lendrem, 1984;Elgar, 1989; Hunter & Skinner, 1998; Treves, 2000) or for social competi-tors (Keverne et al., 1978; Caine, 1984; Aureli & Schaffner, 2002), ratherthan monitoring social partners; most such studies, thus, likely reflect anxietyrather than bondedness (see also Castles et al., 1999). In contrast, monitoringof one’s key social partner(s) may reflect precisely the kind of attentivenessthat one might expect of a bonded relationship. Figure 1 suggests that thereare marked differences in females’ social monitoring rates (i.e., looks di-rected specifically at other members of the group) between a pairbonded un-gulate (klipspringer, Oreotragus oreotragus), an ungulate that has a typicalfission–fusion form of sociality (in this case, feral goats, Capra hircus), anda species of well-bonded primate (represented here by the harem-forminggelada baboon, Theropithecus gelada). The differences between taxa arehighly significant (F2,41 = 37.94, p < 0.001), and post hoc tests confirmthat all three pairwise differences are significant (p � 0.002). On this sim-ple index of social intensity (essentially an index of attentiveness to othergroup members), monogamous klipspringer are significantly more attentiveto their mates than is a polygamous primate, who in turn are significantlymore attentive to other members of their social group than a loosely bondedungulate. While we would obviously need a larger sample of species to con-firm these findings, Figure 1 does at least provide prima facie evidence tosuggest that social monitoring might provide us with a useful behaviouralindex of bondedness that could be used both within and between species. Weemphasise that, functionally, the issue here is the extent to which attentionto others facilitates coordination of activity. But it may, at the same time,also function as a declaration of interest (and probably does so in humans, atleast: Dunbar et al., 2002).

While it is certainly true that social monitoring occurs in order to mon-itor the whereabouts of resource competitors or to avoid harassment fromhigher ranking individuals, it is worth pointing out that there is only limitedempirical evidence to support the claim that social monitoring is actually re-lated to the avoidance of within-group aggression, and that only from cagedgroups (Keverne et al., 1978; McNelis & Boatright-Horowitz, 1998) whereaggression rates are commonly an order of magnitude higher than in wildpopulations (Gartlan, 1968). Even then, the correlations between dominancerank and monitoring received and given are mixed, and tend to be higher


Figure 1. Mean±SE frequency of social monitoring by adult females for three mammaliantaxa (gelada, Theropithecus gelada; klipspringer, Oreotragus oreotragus; feral goats, Caprahircus). The plotted value is the mean rate per minute with which individual females lookedat other adult group members. Gelada are intensely social Old World monkeys that live in acomplex multi-level social system; klipspringer are intensely monogamous ungulates; goatslive in loose matrilineal social groups (heft groups). Mean group size (including the focalsubject) is 6.9 for gelada, 2 for klipspringer and, 19.4 for goats. Vigilance patterns wererecorded from natural, habituated populations using 15-min focal samples for gelada andklipspringer and 5-min focal samples for goats. Gelada were sampled in the Simen MountainsNational Park, Ethiopia (1975); klipspringer at Kekopey Ranch, Gilgil, Kenya (1981–1982);and goats on the Great Orme Nature Reserve, North Wales (2006–2007); with all samplingby R.I.M.D. and Patsy Dunbar. Sample sizes were: 14 adult female gelada (mean sample150 min/female), 12 adult female klipspringer (mean sample 229 min/female) and 15 adultfemale goats (mean sample 57 min/female). Inter-observer reliability was tested for geladaand was 87.5% (Cohen’s κ = 0.74, τ = 6.21, p = 0.001). Although ungulates have 120◦vision, looks at both objects in the environment and other group members are invariablyaccompanied by a snap of the head, which we used as the defining criterion. Kaminski etal. (2006) provide evidence that not only is it obvious to human observers where a goat islooking (and that human observers can agree), but so it also is for other goats, as well as for

the fact that goats do not have all-round attentional vision.

for males than females (Keverne et al., 1978); moreover, neither Keverne etal. (1978) nor McNelis & Boatright-Horowitz (1998) controlled for alterna-tive social factors such as grooming partnerships. That high-rank individualsattract more social monitoring may simply reflect their social attractiveness(cf., Seyfarth’s (1977) seminal explanation for why social grooming oftenaccumulates ‘up the hierarchy’). Although we would be surprised if animals


never monitored each other out of fear of attack (Keverne et al. (1978) notedthat vigilance rates rose significantly during the period immediately after anagonistic conflict in their talapoins, Miopithecus talapoin), it is clear thatsocial monitoring commonly occurs for other reasons. Altmann (1980), forexample, found that, among baboon mothers, glance rate (which includedboth predator vigilance and social monitoring) was a good predictor of ma-ternal style, but a poor predictor of dominance rank. In gelada, rank did notcorrelate with the frequency with which individuals visually monitored orwere monitored by other members of their reproductive unit; instead, adultfemales preferentially both glanced at and faced their core grooming part-ners (Dunbar, 1983b). Similarly, a detailed critical tests analysis of vigilancepatterns in humans showed that visual monitoring is driven predominantlyby mate searching considerations rather than predation risk (Dunbar et al.,2002), and Keverne et al. (1978) likewise noted that, when a female was inoestrus, monitoring rates of the opposite sex increased significantly amongtalapoins. That avoidance of conflict is, in general, unlikely to be the wholestory is, in any case, indicated by the fact that, even if this is the key factordriving the vigilance rates of the goats in Figure 1, it cannot be in the casefor klipspringer where within-pair aggression and displacement hardly everoccurs — yet vigilance rates are around eight times higher than they are inthe socially more promiscuous goats.

Thus, while we surely need to take care not to confound vigilance drivenby different motivations, social monitoring may nonetheless be as valuable atool for the non-invasive assay of social bondedness as anything else on offer.It is simple to use and reliable (inter-observer reliability rates are commonlyvery high not only in primates, but also in ungulates and birds: Dunbar,1983b; McNelis & Boatright-Horowitz, 1998; Treves, 2000; Dawkins, 2002;Kaminski et al., 2005); in addition, it can be implemented in both captivityand the wild, and can be used to assay individual relationships as well asoverall levels of group bondedness.

At a more cognitive level, the current interest in social network analy-sis might offer an index of social complexity that at least taps directly intothe cognitive demands of bonding. Although network analysis is a well de-veloped area within human sociology (Pattison, 1993; Watts, 1999; Palla etal., 2007), it is only very recently that attempts have been made to applythese analytical techniques to non-human species (see, for example, Lusseauet al., 2004; Lehmann & Dunbar, 2009). So far, however, network analysis


has been limited to relatively simple descriptions of the structure of socialgroups, and offers little real insight into the nature of the relationships be-tween network members; there are few, if any, formal predictions about hownetwork structure should vary between different kinds of sociality. However,two possible aspects of network structure might recommend themselves asindices of social engagement.

First, in an analysis of social networks in Old World monkeys, Lehmann& Dunbar (2009) have been able to show that several indices of networkdensity and connectedness correlate independently with relative neocortexsize and group size. In effect, as group size increases, animals may spendmore time grooming, but focus this increasingly on a more select subset ofindividuals. In effect, their relationships become more intense and focussed.While network indices may provide us with a way ahead, this result shouldcaution us against naïve interpretations of these indices: highly connectedanimals (in the network sense) may not be those with the most intense so-cial bonds. Second, it has been shown, for both humans (Zhou et al., 2005;Hamilton et al., 2007) and some animal taxa (Moss & Poole, 1983; Kudo &Dunbar, 2001; Wittemyer et al., 2005; Hill et al., 2008) that social networksin complexly organised societies have a distinctive hierarchical structure: anindividual’s network of relationships can be described as a hierarchically in-clusive series of circles of acquaintanceship which have a consistent scalingratio very close to 3. Species may differ in the size of the base layer (thenumber of individuals in the innermost circle of relationships, with the sizeof successive layers in the network a constant multiple of this number) orthey may all have the same base layer (but differ in the number of layersthey can support above that). We suggest that the level of social complexity(and hence the bondedness) characteristic of a species might be indexed as acombination of the size of the base unit and the number of layers of acquain-tanceship that emerge from it (Kudo & Dunbar, 2001). More development ofthese ideas is, therefore, needed to identify network indices that might reflectboth social complexity and social bondedness.

Conclusion

We have argued that at present we lack the conceptual framework as wellas the operational indices to compare either social systems or the affiliative


processes (bondedness) that give social systems their form and temporal sta-bility. Instead, we have tended to categorise social systems in terms of simpledemographic or mating criteria, and as a result have often pooled social sys-tems that are of very different kinds. In comparative studies of the evolutionof sexual advertisement in primates, for example, Sillén-Tullberg & Møller(1993) and Pagel & Meade (2006) were able to dichotomise mating systemsin radically contrasting ways: the first did so on the basis of number of fe-males (pooling multimale and harem groups to contrast with monogamy), theother did so in terms of number of males (contrasting multimale with singlemale groups by pooling harem systems with monogamy). This of necessityforced them into testing very different hypotheses for exactly the same phe-nomenon.

This situation has arisen, in part, because ethologists have generallyducked the question of what social bonds actually are, and as a result wehave no descriptive indices that we can use to quantify differences either be-tween species or, within species, between individuals. It may now be timeto engage more directly with the nature of social relationships and the phe-nomenon of bondedness in animals. We argue that, at least in respect ofour understanding of the more intensely social species of animals, progresshas been impeded by an over-dependence on the behaviourist stance — theclaim made by the early twentieth century behaviourists that we can onlyever study behaviour and not the mind behind the behaviour (a perspectivethat was largely abandoned in cognitive psychology during the 1980s). Thisis not merely a matter of parochial interest to those who study the handful ofspecies that are socially more complex. It raises important issues about thenature of sociality, and the role of group-level processes in the way groupcohesion is maintained that may have important implications for multilevelselection.

Unfortunately, our understanding of what is involved in bonded relation-ships is so poor that we are probably going to have to go back to basicsand do a great deal more intensive natural history. We simply do not knowenough about these aspects of animal behaviour, and in the absence of suchknowledge experiments may have limited heuristic value. We will also needto come up with much better behavioural indices of bondedness. We havesuggested three novel ones (behavioural synchrony, directed vocal exchangesand social monitoring), but these are neither exhaustive nor necessarily ideal.Our plea is rather for more and better data of these kinds on a wider range ofspecies at all taxonomic levels.


References

Abelson, R. (1963). Computer simulation of ‘hot’ cognition. — In: Computer simulation ofpersonality (Tomkins, S.S. & Messick, S., eds). Wiley, New York, NY, p. 277-298.

Altmann, J. (1980). Baboon mothers and infants. — Harvard University Press, Cambridge,MA.

Aron, A., Aron, E.N. & Smollan, D. (1992). Inclusion of other in the self scale and thestructure of interpersonal closeness. — J. Pers. Soc. Psychol. 63: 596-612.

Aureli, F. & Schaffner, C.M. (2007). Aggression and conflict management at fusion in spidermonkeys. — Biol. Lett. 3: 147-149.

Aureli, F. & Schaffner, C.M. (2002). Relationship assessment through emotional mediation.— Behaviour 139: 393-420.

Aureli, F. & de Waal, F.B.M. (eds) (2000). Natural conflict resolution. — University ofCalifornia Press, Berkeley, CA.

Bales, K.L., Kim, A.J., Lewis-Reese, A.D. & Carter, C.S. (2004). Both oxytocin and va-sopressin may influence alloparental care in male prairie voles. — Horm. Behav. 45:454-361.

Barrett, L., Henzi, S.P. & Dunbar, R.I.M. (2003). Primate cognition: from ‘what now?’ to‘what if?’ — Trends Cogn. Sci. 7: 494-497.

Bergman, T.J., Beehner, J.C., Cheney, D.L. & Seyfarth, R.M. (2003). Hierarchical classifica-tion by rank and kinship in baboons. — Science 302: 1234-1236.

Berscheid, E., Snyder, M. & Omoto, A.M. (1989). The relationship closeness inventory:assessing the closeness of interpersonal relationships. — J. Pers. Soc. Psychol. 57:792-807.

Boinski, S. (1993). Vocal coordination of group movement among white-faced capuchinmonkeys, Cebus capuchinus. — Am. J. Primat. 30: 85-100.

Bshary, R., Salwiczek, L.H. & Wickler, W. (2007). Social cognition in non-primates. —In: Oxford handbook of evolutionary psychology (Dunbar, R.I.M. & Barrett, L., eds).Oxford University Press, Oxford, p. 83-102.

Brotherton, P.N.M. & Komers, P.E. (2003). Mate guarding and the evolution of socialmonogamy in mammals. — In: Monogamy: mating strategies and partnerships in birds,humans and other mammals (Reichard, U.H. & Boesch, C., eds). Cambridge UniversityPress, Cambridge, p. 59-80.

Bugnyar, T. & Heinrich, B. (2007). Ravens, Corvus corax, differentiate between knowledge-able and ignorant competitors. — Proc. Roy. Soc. Lond. B: Biol. Sci. 272: 1641-1646.

Byrne, R. & Whiten, A. (eds) (1988). Machiavellian intelligence. — Oxford University Press,Oxford.

Caine, N.G. (1984). Visual scanning by tamarins: a description and test of two derived hy-potheses. — Folia Primatol. 43: 59-67.

Calhim, S., Shi, J. & Dunbar, R.I.M. (2006). Sexual segregation among feral goats: testingbetween alternative hypotheses. — Anim. Behav. 72: 31-41.

Call, J. (2001). Chimpanzee social cognition. — Trends Cogn. Sci. 5: 388-393.Castles, D.L., Whiten, A. & Aureli, F. (1999). Social anxiety, relationships and self-directed

behaviour among wild female olive baboons. — Anim. Behav. 58: 1207-1215.Cheney, D.L. (1982). Extrafamilial alliances among vervet monkeys. — In: Primate social

relationships (Hinde, R.A., ed.). Blackwell, Oxford, p. 278-285.


Cheney, D.L. & Seyfarth, R.M. (1999). Recognition of other individuals’ social relationshipsby female baboons. — Anim. Behav. 58: 67-75.

Cheney, D.L. & Seyfarth, R.M. (2007). Baboon metaphysics: the evolution of a social mind.— University of Chicago Press, Chicago, IL.

Chance, M.R.A. & Jolly, C.J. (1970). Social groups of monkeys, apes and men. — Cape,London.

Clayton, N.S., Dally, J.M. & Emery, N.J. (2007). Social cognition by food-caching birds:the western scrub-jay as a natural psychologist. — Phil. Trans. Roy. Soc. Lond. B 362:507-522.

Cohen, E., Ejsmond-Frey, R., Knight, N. & Dunbar, R.I.M. (2010). Rowers’ high: behaviouralsynchrony is correlated with elevated pain thresholds. — Biol. Lett. 6: 106-108.

Connor, R.C. (2007). Complex alliance relationships in bottlenose dolphins and a consider-ation of selective environments for extreme brain size evolution in mammals. — Phil.Trans. Roy. Soc. Lond. B 362: 587-602.

Conradt, L. (1998a). Could asynchrony in activity between the sexes cause intersexual socialsegregation in ruminants? — Proc. Roy. Soc. Lond. B: Biol. Sci. 265: 1359-1363.

Conradt, L. (1998b). Measuring the degree of sexual segregation in group-living animals. —J. Anim. Ecol. 67: 217-226.

Cools, A.K.A., van Hout, A.J.M. & Nelissen, M.H.J. (2008). Canine reconciliation and third-party-initiated postconflict affiliation: do peacemaking social mechanisms in dogs rivalthose of higher primates? — Ethology 114: 53-63.

Cords, M. (1997). Friendships, alliances, reciprocity and repair. — In: Machiavellian in-telligence II (Whiten, A. & Byrne, R., eds). Cambridge University Press, Cambridge,p. 24-49.

Cords, M. (2002). Friendship among adult female blue monkeys (Cercopithecus mitis). —Behaviour 139: 291-314.

Cords, M. & Aureli, F. (2000). Reconciliation and relationship qualities. — In: Naturalconflict resolution (Aureli, F. & de Waal, F.B.M., eds). University of California Press,Berkeley, CA, p. 177-198.

Crook, J.H., Ellis, J.E. & Goss-Custard, J.D. (1976). Mammalian social systems: structureand function. — Anim. Behav. 24: 261-274.

Crook, J.H. & Gartlan, J.S. (1966). Evolution of primate societies. — Nature 210: 1200-1203.Curly, J.P. & Keverne, E.B. (2005). Genes, brains and mammalian social bonds. — TREE 20:

581-587.Dally, J.M., Emery, N.J. & Clayton, N.S. (2006). Food-caching western scrub-jays keep track

of who was watching when. — Science 312: 1662-1665.Damasio, A. (1994). Descartes’ error. — Putnam: New York, NY.Dawkins, M.S. (1983). Battery hens name their price: consumer demand theory and the

measurement of ethological ‘needs’. — Anim. Behav. 31: 1195-1205.Dawkins, M.S. (2002). What are birds looking at? Head movements and eye use in chickens.

— Anim. Behav. 63: 991-998.Dunbar, R.I.M. (1980). Determinants and evolutionary consequences of dominance among

female gelada baboons. — Behav. Ecol. Sociobiol. 7: 253-265.Dunbar, R.I.M. (1983a). Lifehistory tactics and alternative strategies of reproduction. — In:

Mate choice (Bateson, P., ed.). Cambridge University Press, Cambridge, p. 423-433.Dunbar, R.I.M. (1983b). Structure of gelada baboon reproductive units. IV. Organisation at

group level. — Z. Tierpsychol. 63: 265-282.


Dunbar, R.I.M. (1984). Reproductive decisions: an economic analysis of gelada baboon socialstrategies. — Princeton University Press, Princeton, NJ.

Dunbar, R.I.M. (1988). Primate social systems. — Chapman & Hall, London.Dunbar, R.I.M. (1989). Reproductive strategies of female gelada baboons. — In: Sociobi-

ology of sexual and reproductive strategies (Rasa, A., Vogel, C. & Voland, E., eds).Chapman & Hall, London, p. 74-92.

Dunbar, R.I.M. (1991). Functional significance of social grooming in primates. — FoliaPrimatol. 57: 121-131.

Dunbar, R.I.M. (1992). Neocortex size as a constraint on group size in primates. — J. HumanEvol. 22: 469-493.

Dunbar, R.I.M. (1995). The mating system of callitrichid primates: I. Conditions for thecoevolution of pair bonding and twinning. — Anim. Behav. 50: 1057-1070.

Dunbar, R.I.M. (1998). The social brain hypothesis. — Evol. Anthropol. 6: 178-190.Dunbar, R.I.M. (1999). Culture, honesty and the freerider problem. — In: The evolution

of culture (Dunbar, R., Knight, C. & Power, C., eds). Edinburgh University Press,Edinburgh, p. 194-213.

Dunbar, R.I.M. (2003). Why are apes so smart? — In: Primate life histories and socioecology(Kappeler, P.H. & Pereira, M., eds). Chicago University Press, Chicago, IL, p. 285-298.

Dunbar, R.I.M. (2010a). The social role of touch in humans and primates: behavioural func-tion and neurobiological mechanisms. — Neurosci. Biobehav. Rev. 34: 260-270.

Dunbar, R.I.M. (2010b). Brain and behaviour in primate evolution. — In: Mind the gap:tracing the origins of human universals (Appeler, P.H. & Silk, J., eds). Springer, Berlin,p. 315-330.

Dunbar, R.I.M. & Dunbar, P. (1980). The pairbond in klipspringer. — Anim. Behav. 28:251-263.

Dunbar, R.I.M., Cornah, L., Daly, F. & Bowyer, K. (2002). Vigilance in humans: a test ofalternative hypotheses. — Behaviour 139: 695-711.

Dunbar, R.I.M. & Shi, J. (2008). Sex differences in feeding activity results in sexual segrega-tion of feral goats. — Ethology 114: 444-451.

Elgar, M.A. (1989). Predator vigilance and group size in mammals and birds: a critical reviewof the empirical evidence. — Biol. Rev. 64: 13-33.

Emery, N.J. & Clayton, N.S. (2001). Effects of experience and social context on prospectivecaching strategies by scrub jays. — Nature 414: 443-446.

Emery, N.J., Seed, A.M., von Bayern, A.M.P. & Clayton, N.S. (2007). Cognitive adaptationsof social bonding in birds. — Phil. Trans. Roy. Soc. Lond. B 362: 489-506.

Enquist, M. & Leimar, O. (1993). The evolution of cooperation in mobile organisms. —Anim. Behav. 45: 747-757.

Fisher, H.E., Aron, A. & Brown, L.L. (2006). Romantic love: a mammalian brain system formate choice. — Phil. Trans. Roy. Soc. Lond. B 361: 2173-2186.

Fraser, O.N., Schino, G. & Aureli, F. (2008). Components of relationship quality in chim-panzees. — Ethology 114: 834-843.

Freeberg, T.M. (2006). Social complexity can drive vocal complexity: group size influencesvocal information in Carolina chickadees. — Psychol. Sci. 17: 557-561.

Gartlan, J.S. (1968). Structure and function in primate society. — Folia Primatol. 8: 89-120.Guinness, F.E., Hall, M.J. & Cockerill, R.A. (1979). Mother-offspring association in red deer.

— Anim. Behav. 27: 536-544.


Gürerk, Ö., Irlenbusch, B. & Rockenbach, B. (2006). The competitive advantage of sanction-ing institutions. — Science 312: 108-111.

Hamilton, M.J., Milne, B.T., Walker, R.S., Burger, O. & Brown, J.H. (2007). The complexstructure of hunter-gatherer social networks. — Proc. Roy. Soc. Lond. B: Biol. Sci. 274:2195-2002.

Hamilton, W.D. (1964). The genetical evolution of social behaviour. — J. Theor. Biol. 7:1-51.Hare, B., Call, J. & Tomasello, M. (2001). Do chimpanzees know what conspecifics know?

— Anim. Behav. 61: 139-151.Hare, B., Call, J. & Tomasello, M. (2006). Chimpanzees deceive a human competitor by

hiding. — Cognition 101: 495-514.Hart, A. & Lendrem, D. (1984). Vigilance and scanning patterns in birds. — Anim. Behav.

32: 1216-1224.Hill, R.A., Bentley, A. & Dunbar, R.I.M. (2008). Network scaling reveals consistent fractal

pattern in hierarchical mammalian societies. — Biol. Lett. 4: 748-751.Hinde, R.A. (1976). Interactions, relationships, and social structure. — Man 11: 1-17.Hunter, L.T.B. & Skinner, J.D. (1998). Vigilance behaviour in African ungulates: the role of

predation pressure. — Behaviour 135: 195-211.Kaminski, J., Call, J. & Tomasello, M. (2006). Goats’ behaviour in a competitive food para-

digm: evidence for perspective taking? — Behaviour 143: 1341-1356.Kaminski, J., Reidel, J., Call, J. & Tomasello, M. (2005). Domestic goat, Capra hircus, follow

gaze direction and use social cues in an object choice task. — Anim. Behav. 69: 11-18.Kelley, H.H., Berscheid, E., Christensen, A., Harvey, J.H., Huston, T.L., Levinger, G., Mc-

Clintock, E., Peplau, L.A. & Peterson, D.R. (1983). Close relationships. — Freeman,New York, NY.

Keverne, E.B., Leonard, R.A., Scruton, D.M. & Young, S.K. (1978). Visual monitoring ionsocial groups of talapoin monkeys (Miopithecus talapoin). — Anim. Behav. 26: 933-944.

Keverne, E.B., Martensz, N. & Tuite, B. (1989). Beta-endorphin concentrations in cere-brospinal fluid of monkeys are influenced by grooming relationships. — Psychoneu-roendocrinology 14: 155-161.

Knafo, A., Israel, S., Darvasi, A., Bachner-Melman, R., Uzefovsky, F., Cohen, L., Feldman,E., Lerer, E., Laiba, E., Raz, Y., Nemanov, L., Gritsenko, I., Dina, C., Agam, G., Dean,B., Bornstein, G. & Ebstein, R.P. (2007). Individual differences in allocation of fundsin the dictator game associated with length of the arginine vasopressin 1a receptorRS3 promoter region and correlation between RS3 length and hippocampal mRNA.— Genes Brain Behav. 7: 266-275.

Komers, P.E. & Brotherton, P.N.M. (1997). Female space use is the best predictor ofmonogamy in mammals. — Proc. Roy. Soc. Lond. B: Biol. Sci. 264: 1261-1270.

Koski, S.E., Koops, K. & Sterck, E.H.M. (2007). Reconciliation, relationship quality, andpostconflict anxiety: testing the integrated hypothesis in captive chimpanzees. — Am.J. Primatol. 69: 158-172.

Krause, J. & Ruxton, G.D. (2002). Living in groups. — Oxford University Press, Oxford.Kudo, H. & Dunbar, R.I.M. (2001). Neocortex size and social network size in primates. —

Anim. Behav. 62: 711-722.Kummer, H. (1968). Social dynamics of hamadryas baboons. — Karger, Basel.Kummer, H., Goetz, W. & Angst, W. (1974). Triadic differentiation: an inhibitory mechanism

protecting pair bonds in baboons. — Behaviour 49: 62-87.


Le Doux, J. (1998). The emotional brain. — Weidenfeld & Nicolson, London.Lehmann, J. & Dunbar, R.I.M. (2009). Network cohesion, group size and neocortex size in

female-bonded Old World primates. — Proc. Roy. Soc. Lond. B: Biol. Sci. 276: 4417-4422.

Lehmann, J., Korstjens, A.H. & Dunbar, R.I.M. (2007). Group size, grooming and socialcohesion in primates. — Anim. Behav. 74: 1617-1629.

Lusseau, D., Schneider, K., Boisseau, O.J., Haase, P., Slooten, E. & Dawson, S.M. (2004).The bottlenose dolphin community of Doubtful Sound features a large proportion oflong-lasting associations: can geographic isolation explain this unique trait? — Behav.Ecol. Sociobiol. 54: 396-405.

Majolo, B., Ventura, R. & Koyama, N.F., 2009. A statistical modelling approach to theoccurrence and timing of reconciliation in wild Japanese macaques. — Ethology 115:152-166.

Mason, W.A. (1965). The social development of monkeys and apes. — In: Primate behaviour(DeVore, I., ed.). Holt, Rinehart & Winston, New York, NY, p. 514-543.

Mason, W.A. & Mendoza, S.P. (1998). Generic aspects of primate attachments: parents,offspring and mates. — Psychoneuroendocrinology 23: 756-778.

McComb, K., Moss, C., Sayialel, S. & Baker, L. (2000). Unusually extensive networks ofvocal recognition in African elephants. — Anim. Behav. 59: 1103-1109.

McComb, K. & Semple, S. (2005). Coevolution of vocal communication and sociality inprimates. — Biol. Lett. 1: 381-385.

McNelis, N.L. & Boatright-Horowitz, S.L. (1998). Social monitoring in a primate group: therelationship between visual attention and hierarchical ranks. — Anim. Cogn. 1: 65-69.

Moss, C. & Poole, J. (1983). Relationships and social structure of African elephants. — In:Primate social relationships (Hinde, R.A., ed.). Blackwell, Oxford, p. 314-325.

Nettle, D. & Dunbar, R.I.M. (1997). Social markers and the evolution of reciprocal exchange.— Curr. Anthropol. 38: 93-98.

Niedenthal, P.M. (2007). Embodying emotion. — Science 316: 1002-1005.Odling-Smee, F.J., Laland, K.N. & Feldman, M.W. (2003). Niche construction: the neglected

process in evolution. — Princeton University Press, Princeton, NJ.Orstrom, E., Gardner, R. & Walker, J. (1994). Rules, games and common-pool resources. —

University of Michigan Press, Ann Arbor, MI.Pagel, M. & Meade, A. (2006). Bayesian analysis of correlated evolution of discrete charac-

ters by reversible-jump Markov chain Monte Carlo. — Am. Nat. 167: 808-825.Palla, G., Barabasi, A.L. & Vicsek, T. (2007). Quantifying social group evolution. — Nature

446: 664-667.Parr, L.A. & de Waal, F.B.M. (1999). Visual kin recognition in chimpanzees. — Nature 399:

647-648.Pattison, P. (1993). Algebraic models for social networks. — Cambridge University Press,

Cambridge.Pérez-Barbería, J., Shultz, S. & Dunbar, R.I.M. (2007). Evidence for intense coevolution of

sociality and brain size in three orders of mammals. — Evolution 61: 2811-2821.Radford, A. & Du Plessis, M.A. (2006). Dual function of allopreening in the cooperatively

breeding green woodhoopoe, Phoeniculus purpureus. — Behav. Ecol. Sociobiol. 61:221-230.

Rolls, E.T. (1995). A theory of emotion and consciousness, and its application to understand-ing the neural basis of emotion. — In: The cognitive neurosciences (Gazzaniga, M.S.,ed.). MIT Press, Cambridge, MA, p. 1091-1106.


Rubin, Z. (1970). Measurement of romantic love. — J. Pers. Soc. Psychol. 16: 265-273.Ruckstuhl, K.E. & Kokko, H. (2002). Modelling sexual segregation in ungulates: effects of

group size, activity budgets and synchrony. — Anim. Behav. 64: 909-914.Ruckstuhl, K.E. & Neuhaus, P. (2000). Sexual segregation in ungulates: A new approach. —

Behaviour 137: 361-377.Ruckstuhl, K.E. & Neuhaus, P. (2002). Sexual segregation in ungulates: a comparative test of

three hypotheses. — Biol. Rev. 77: 77-96.Samuels, A. & Flaherty, C. (2000). Peaceful conflict resolution in the sea? — In: Natural

conflict resolution (Aureli, F. & de Waal, F.B.M., eds). University of California Press,Berkeley, CA, p. 229-231.

van Schaik, C.P. & Aureli, F. (2000). The natural history of valuable relationships in primates.— In: Natural conflict resolution (Aureli, F. & de Waal, F.B.M., eds). University ofCalifornia Press, Berkeley, CA, p. 307-333.

van Schaik, C.P. & Dunbar, R.I.M. (1990). The evolution of monogamy in large primates: anew hypothesis and some critical tests. — Behaviour 115: 30-62.

Schino, G. (1998). Reconciliation in domestic goats. — Behaviour 135: 343-356.Seyfarth, R.M. (1977). A model of social grooming among adult female monkeys. — J.

Theoret. Biol. 65: 671-698.Shultz, S. & Dunbar, R.I.M. (2007). The evolution of the social brain: anthropoid primates

contrast with other vertebrates. — Proc. Roy. Soc. Lond. B: Biol. Sci. 274: 2429-2436.Silk, J.B. (2002). The ‘F’-word in primatology. — Behaviour 139: 421-446.Silk, J.B. (2003). Cooperation without counting: the puzzle of friendship. In: Genetic and

cultural evolution of cooperation (Hammerstein, P., ed.). — MIT Press, Cambridge,MA, p. 37-54.

Silk, J.B., Alberts, S.C. & Altmann, J. (2003). Social bonds of female baboons enhance infantsurvival. — Science 302: 1232-1234.

Silk, J.B., Altmann, J. & Alberts, S.C. (2006a). Social relationships among adult femalebaboons (Papio cynocephalus). I. Variation in the strength of social bonds. — Behav.Ecol. Sociobiol. 61: 183-195.

Silk, J.B., Altmann, J. & Alberts, S.C. (2006b). Social relationships among adult femalebaboons (Papio cynocephalus). II. Sources of variation in the quality and stability ofsocial bonds. — Behav. Ecol. Sociobiol. 61: 197-204.

Sillén-Tullberg, B. & Møller, A.P. (1993). The relationship between concealed ovulation andmating systems in anthropoid primates: a phylogenetic analysis. — Am. Nat. 141: 1-25.

Smuts, B.B. (1985). Sex and friendship in baboons. — Aldine, New York, NY.Sternberg, R.J. (1997). Construct validation of a triangular love scale. — Eur. J. Soc. Psychol.

27: 313-335.Stewart, K.J. & Harcourt, A.H. (1994). Gorilla vocalisations during rest periods — signals of

impending departure. — Behaviour 130: 29-40.Tate, A.J., Fisher, H., Leigh, A.E. & Kendrick, K.M. (2006). Behavioural and neurophys-

iological evidence for face identity and face emotion processing in animals. — Phil.Trans. Roy. Soc. Lond. B 361: 2155-2172.

Thorpe, W.H. & North, M.E.W. (1966). Vocal imitation in the tropical boubou shrike Lenariusaethiopicus Major as a means of establishing and maintaining social bonds. — Ibis 108:432-435.

Treves, A. (2000). Theory and method in studies of vigilance and aggregation. — Anim.Behav. 60: 711-722.


Uster, D. & Zuberbulher, K. (2001). The functional significance of Diana monkey ‘clear’calls. — Behaviour 138: 741-756.

Vertegaal, R. & Ding, Y. (2002). Explaining effects of eye gaze on mediated group conversa-tions: amount or synchronization? — In: Proc. ACM conference on computer supportedcooperative work (CSCW’02), New Orleans, LA. ACM Press, New York, NY, p. 41-48.

Vertegaal, R., Slagter, R., van der Veer, G. & Nijholt, A. (2001). Eye gaze patterns in conver-sations: there is more to conversational agents than meets the eyes. — In: Proc. SIGCHIconference on human factors in computing systems (CHI’01), Seattle, WA. ACM Press,New York, NY, p. 301-308.

Wahaj, S.A., Guse, K. & Holekamp, K.E. (2001). Reconciliation in the spotted hyena (Cro-cuta crocuta). — Ethology 107: 1057-1074.

Walum, H., Westberg, L., Henningsson, S., Neiderhiser, J.M., Reiss, D., Igl, W., Ganiban,J.M., Spotts, E.M., Pedersen, N.L., Eriksson, E. & Lichtenstein, P. (2008). Geneticvariation in the vasopressin receptor 1a gene (AVPR1A) associates with pair-bondingbehaviour in humans. — Proc. Natl. Acad. Sci. USA 105: 14153-14156.

Watts, D.J. (1999). Small worlds: the dynamics of networks between order and randomness.— Princeton University Press, Princeton, NJ.

Wilkinson, G.S. (1986). Social grooming in the common vampire bat, Desmodus rotundus.— Anim. Behav. 34: 1880-1889.

Wilson, D.S. (2007). Group-level evolutionary processes. — In: Oxford handbook of evolu-tionary psychology (Dunbar, R.I.M. & Barrett, L., eds). Oxford University Press, Ox-ford, p. 49-56.

Wilson, E.O. (1975). Sociobiology. — Harvard University Press, Cambridge, MA.Wiltermuth, S.S. & Heath, C. (2009). Synchrony and cooperation. — Psychol. Sci. 20: 1-5.Wittemyer, G., Douglas-Hamilton, I. & Getz, W.M. (2005). The socioecology of elephants:

analysis of the processes creating multitiered social structures. — Anim. Behav. 69:1357-1371.

Wrangham, R.W. (1980). An ecological model of female-bonded primate groups. — Behav-iour 75: 262-300.

Young, L.J. & Wang, Z.X. (2004). The neurobiology of pair bonding. — Nature Neurosci. 7:1048-1054.

Zhou, W.-X., Sornette, D., Hill, R.A. & Dunbar, R.I.M. (2005). Discrete hierarchical organi-zation of social group sizes. — Proc. Roy. Soc. Lond. B: Biol. Sci. 272: 439-444.

Documents

Dunbar y Schultz - Bondedness and Socialty