Experimental evidence for predator learning and Müllerian mimicry in Peruvian poison frogs (Ranitomeya, Dendrobatidae)

摘要

Poison frogs are characterized by bright coloration, striking patterns, and toxicity; they have thus become a classic example of aposematism. Ranitomeya imitator mimics three congeneric model species (R. fantastica, R. summersi, and two morphs of R. variabilis), creating geographically distinct populations of the species, including four allopatric mimetic morphs. These complexes are thought to represent a case of Müllerian mimicry, but no empirical data exist on learned avoidance by predators to support this claim. In this study we used young chickens (Gallus domesticus) as naïve predators to determine if a mimetic morph of R. imitator and R. variabilis contribute to reciprocal learned avoidance by predators--a key component of Müllerian mimicry. Chickens exposed to either stimulus species demonstrated learned avoidance of both species; thus our results indicate that this complex functions as a Müllerian mimicry system. Our study shows no difference between learned avoidance in stimuli frogs and a "novel" morph of R. imitator which differed in both colour and pattern--indicating that predator learned avoidance may be generalized in this system. This study provides empirical evidence demonstrating the first known case of Müllerian mimicry in anurans. Further, it demonstrates generalized learning which provides a plausible mechanism for the maintenance of both polymorphic mimicry and the maintenance of intrapopulation phenotypic heterogeneity.