Cross-site comparison studies of behavioral variation can provide evidence for traditions in wild species once ecological and genetic factors are excluded as causes for cross-site differences. These studies ensure behavior variants are considered within the context of a species' ecology and evolutionary adaptations. We examined wide-scale geographic variation in the behavior of spider monkeys (Ateles geoffroyi) across five long-term field sites in Central America using a well established ethnographic cross-site survey method. Spider monkeys possess a relatively rare social system with a high degree of fission-fusion dynamics, also typical of chimpanzees (Pan troglodytes) and humans (Homo sapiens). From the initial 62 behaviors surveyed 65% failed to meet the necessary criteria for traditions. The remaining 22 behaviors showed cross-site variation in occurrence ranging from absent through to customary, representing to our knowledge, the first documented cases of traditions in this taxon and only the second case of multiple traditions in a New World monkey species. Of the 22 behavioral variants recorded across all sites, on average 57% occurred in the social domain, 19% in food-related domains and 24% in other domains. This social bias contrasts with the food-related bias reported in great ape cross-site comparison studies and has implications for the evolution of human culture. No pattern of geographical radiation was found in relation to distance across sites. Our findings promote A. geoffroyi as a model species to investigate traditions with field and captive based experiments and emphasize the importance of the social domain for the study of animal traditions.

Candidate traditions were documented across three communities of wild spider monkeys (Ateles geoffroyi) using an a priori approach to identify behavioral variants and a statistical approach to examine differences in their proportional use. This methodology differs from previous studies of animal traditions, which used retrospective data and relied on the ‘exclusion method’ to identify candidate traditions. Our a priori approach increased the likelihood that behavior variants with equivalent functions were considered and our statistical approach enabled the proportional use of ‘universal’ behaviors, i.e., used across all communities, to be examined for the first time in any animal species as candidate traditions. Among universal behaviors we found 14 ‘community preferred’ variants. After considering the extent to which community preferred variants were due to ecological and, to a lesser degree, genetic differences, we concluded that at least six were likely maintained through social learning. Our findings have two main implications: (i) tradition repertoires could be larger than assumed from previous studies using the exclusion method; (ii) the relative use of universal behavior variants can reinforce community membership.

No abstract

No abstract