Abstract
BACKGROUND: Predator avoidance can have immense impacts on fitness, yet individual variation in the expression of anti-predator behaviour remains largely unexplained. Existing research investigating learning of novel predators has focused either on individuals or groups, but not both. Testing in individual settings allows evaluations of learning or personality differences, while testing in group settings makes it impossible to distinguish any such individual differences from social dynamics. In this study, we investigate the effect of social dynamics on individual anti-predator behaviour. We trained 15 captive ravens to recognize and respond to a novel experimental predator and then exposed them to this predator in both group and isolation settings across 1.5 years to tease apart individual differences from social effects and evaluate two hypotheses: (1) weaker anti-predator responses of some individuals in the group occurred, because they failed to recognize the experimental predator as a threat, leading to weak responses when separated, or (2) some individuals had learned the new threat, but their scolding intensity was repressed in the group trials due to social dynamics (such as dominance rank), leading to increased scolding intensity when alone. RESULTS: We found that dominance significantly influences scolding behaviour in the group trials; top-ranked individuals scold more and earlier than lower ranking ones. However, in the separation trials scolding duration is no longer affected by rank. CONCLUSIONS: We speculate that, while top-ranked individuals use their anti-predator responses to signal status in the group, lower-ranking ravens may be suppressed from, or are less capable of, performing intense anti-predator behaviour while in the group. This suggests that, in addition to its recruitment or predator-deterrent effects, alarm calling may serve as a marker of individual quality to conspecifics.