Space to choose: network analysis of social preferences in a captive chimpanzee community, and implications for management

Abstract

Social network analysis (SNA) is rapidly gaining popularity in primatology, but its application to the management of zoo-housed primates has been largely overlooked. Here I use SNA techniques to explore the social structure of chimpanzees (Pan troglodytes) housed in the new "Budongo Trail" exhibit at Edinburgh Zoo, UK. Given that individuals have extensive space (2332 m(2)), and access to several interconnected exhibit sections, I test the hypothesis that individuals are able to choose to interact with specific social partners. Spatial association and social interaction data were recorded from 400 focal watches on 11 individuals, and association, affiliative, and agonistic networks were constructed. Matrix correlations showed that individuals who spent time in close proximity were likely to affiliate with one another, but spatial association did not predict the frequency of agonistic encounters. Cluster analysis revealed significantly distinct sub-groups in the affiliative network (but not association or agonistic networks) in line with maternal kinship. Overall my findings support the hypothesis that the Budongo Trail exhibit facilitates the expression of social preferences, and suggests that SNA can be a useful tool to study zoo primates when proximity between individuals is not forced (i.e. in large, modern exhibits). Now that I have validated a set of SNA methods for this community, they can be used to trace changes in social dynamics over a longer time period, and ultimately assist zoo staff in their management decisions.

How can social network analysis improve the study of primate behavior?

“…The other two examples in the special issue concern captive chimpanzees. Composite indices are often used as a proxy for the strength of social relationships in primates [Silk et al, 2006a,b], and Clark [2011] shows that the grooming network of a captive chimpanzee community was positively correlated with the association network, whereas Kanngiesser et al's [2011] study found that grooming networks were correlated to both age and kinship.…”

“…When defining a group, it could then be argued that one should use only interaction data, and not association data. Nevertheless, association data can be extremely informative, especially where these associations correlate to interaction data [Clark, 2011;Silk et al, 2002a,b], since association data are typically easier to collect. How can SNA of relationship data be used to define a group?…”

“…Similarly, where opportunity for immigration in wild populations is limited due to physical barriers created by human populations, natural fission events may be prevented, potentially escalating human-primate conflicts. Using cluster analysis on grooming interactions, researchers could predict the best group division that would result in preservation of the most affiliated individuals (see Clark [2011] for use of cluster analysis on captive chimpanzees). Using centrality measures on aggressive interactions, one can also identify the individual most disruptive to the network and then remove it in order to potentially decrease aggression rates and enhance group stability.…”

“…Using centrality measures on aggressive interactions, one can also identify the individual most disruptive to the network and then remove it in order to potentially decrease aggression rates and enhance group stability. Clark [2011] provides an informative discussion of how SNA can be used from a captive management perspective, and two empirical examples are presented in this special issue. Dufour et al [2011] used SNA to study the effect of relocation on stress and proximities in a captive capuchin and squirrel monkeys, and Beisner et al [2011] used SNA to demonstrate the degree of integration of highranking natal adult males into social network of rhesus macaques.…”

How can social network analysis contribute to social behavior research in applied ethology?

“…Numerous studies have used this measure of community (e.g. Lusseau and Newman, 2004; Kasper and Voelkl, 2009; Clark, 2011). One recent study looked at the effect of degree of relatedness in captive rhesus macaque matrilines, called matriline fragmentation, on grooming community structure (Beisner et al, 2011a).…”