Are the alarm calls of Geoffroy's marmosets (Callithrix geoffroyi) functionally referential?

The current study aimed to determine whether or not Geoffroy's marmosets have referential alarm calls by analyzing both the acoustic properties of the calls given in response to snake and perched raptor models (production criterion), as well as the marmosets' behaviors in response to playbacks of those calls (perception criterion). I predicted that the production criterion would be met, and that the marmosets would demonstrate the perception criterion by modifying their behavior in ways that are appropriate to the particular threats posed by raptors versus snakes. Discriminant function analysis showed that the vocal responses to the predator models were, in fact, different, therefore meeting the production criterion. Binomial tests confirmed that the marmosets would look down in response to snake alarm calls and would approach the speaker. Also as predicted, the monkeys did not approach the speaker following raptor alarm calls, and they did tend to look up, but not significantly more than was predicted by chance. All other hypotheses were not supported. The data suggest that the marmosets' most immediate reactions to snake and perched raptor alarm calls meet the perception criterion for referential communication, but alarm calls in the absence of repeated announcement or confirmatory visual evidence of a predator's presence may be dismissed as a false alarm, an interpretation that is consistent with the marmosets' tendency to reduce the costs of unnecessary vigilance. KEY WORDS: referential communication, Callithrix geoffroyi, alarm calls, captive marmosets