Wheeler et al_Associative learning of alarm signals

Many vertebrate taxa respond to heterospecific alarm calls with appropriate anti¬predator behaviours, although it is unclear how apparent recognition is achieved. Such responses are widely thought to be based on learned associations between the occurrence of the call and the presence of a predator. Conclusive evidence that this behaviour is indeed underpinned by learning, however, is scarce. This study tested whether wild black capuchin monkeys (Sapajus nigritus) learn to associate novel sounds with the presence of predators through a two¬-phase field experiment. During an initial training phase, three study groups were each presented with a playback of one of three novel sounds together with a simulated felid predator on four occasions over a period of 8 to 12 weeks. This was followed by a test phase, wherein each of the three sounds was played back to individuals in all three groups, allowing each sound to serve as both a test stimulus for individuals trained with that sound, and a control stimulus for individuals trained with another sound. Antipredator responses were significantly stronger in response to test sounds than to controls. Antipredator responses continued to be strong for at least two to three years without reinforcement of predator-presence, although noisier novel sounds were more effective at eliciting responses than were more tonal sounds. This study provides the strongest evidence to date that the responses of primates to sounds such as heterospecific alarm calls can be shaped by associative learning, and supports the contention that signals provide receivers with information.