Abstract # 205:

Scheduled for Saturday, June 23, 2007 11:30 AM-11:45 AM: Session 20 (North Main Hall F/G) Oral Presentation

Capuchin Monkeys (Cebus apella) Demonstrate Knowledge of Social Group Members via an Oddity Task

J. J. Pokorny and F. B. de Waal
Emory University / Living Links / Yerkes National Primate Research Center, Atlanta, GA, USA
     The ability to distinguish individuals within one’s group from those outside is critical for survival, as outsiders pose a threat to a group’s food and mating resources. In humans, faces indicate the identity of an individual, allowing one to distinguish between familiar and unfamiliar individuals. The current study investigated whether brown capuchin monkeys (Cebus apella) use faces to identify conspecifics and categorize them based on group membership. Three monkeys from two socially housed groups were trained to perform an oddity task on a touch sensitive monitor using digital images of faces of conspecifics. Subjects were required to select the odd stimulus based on familiarity. In the familiar-odd condition, 3 images of individuals unfamiliar to the subject were presented along with 1 image of a familiar individual, this being the correct response. Likewise, in the unfamiliar-odd condition, subjects were to select the unfamiliar individual among 3 images of familiar individuals and 1 unfamiliar individual. The two conditions were initially presented together within a session, but due to less than expected performance, conditions were then presented separately. A Heterogeneity G-test compared performance with random chance, similar to a chi-square, but the G-test takes into account individual contributions. Separating the conditions resulted in performance significantly above chance for all three subjects [Fam-odd: Gp(1)=134.79, p<0.001; Unfam-odd: Gp(1)=6.17, p=0.01]. As a group, subjects performed above 25% chance when transferred to a new set of images under both conditions [40 trials each; Fam-odd: Gh=NS, Gp(1)=81.24, p<0.001; Unfam-odd: Gh(2)=7.05, p<0.05, Gp(1)=59.14, p<0.001]. Overall, results suggest that subjects recognize the familiar conspecifics in the images and use this knowledge, as well as a concept of familiarity, to complete the task. Results will be compared to a control subject familiar with the oddity task and rewarded for any response.