DIFFERENTIAL AVOIDANCE OF CORAL SNAKE BANDED PATTERNS BY FREE-RANGING AVIAN PREDATORS IN COSTA RICA.

Empirical studies of mimicry have rarely been conducted under natural conditions. Field investigations of some lepidopteran systems have provided a bridge between experiments examining artificial situations and the mimicry process in nature, but these systems do not include all types of mimicry. The presence of dangerous or deadly models is thought to alter the usual rules for mimicry complexes. In particular, a deadly model is expected to protect a wide variety of mimics. Avoidance of different types of mimics should vary according to how closely they resemble the model. Coral snake mimicry complexes in the neotropics may provide natural systems in which these ideas can be examined, but there is no direct evidence that the patterns of venomous coral snakes or potential mimics are avoided in the wild. Plasticine replicas of snakes were used to assess the frequency of avian predation attempts as a function of color pattern. Avian predators left identifiable marks on the replicas, the position of which indicated that replicas were perceived as potentially dangerous prey items by birds. The number of attacks on unmarked brown replicas was greater than that on tricolor coral snake banded replicas. This result was true whether replicas were placed on natural or plain white backgrounds, suggesting that coral snake banded patterns function aposematically. In a separate experiment, replicas representing all six patterns of proposed coral mimics at the study site were attacked less often than unmarked brown replicas. Within these six banded patterns, some were attacked significantly more often than others. This study provides direct field evidence that coral snake banded patterns are avoided by free-ranging avian predators and supports theoretical predictions about mimicry systems involving deadly models.