Evolutionary and ecological dynamics of aposematism and mimicry in poison frogs

In aposematism, defended prey advertise their aversive qualities to predators using a warning signal. Predators that have learned the warning signal are thus the selective agent that can promote similarity in warning signals between distantly related species (i.e. defensive mimicry). Here, I use poison frogs to study the evolutionary and ecological dynamics of aposematism and mimicry. An incredible diversity of color patterns exists among poison frogs and this bright coloration has evolved multiple times from cryptic ancestors, making them a unique system for investigating the complexities and controversies of warning signals. In chapter 1, I explore the phylogenetic relationships of poison frogs (Dendrobatidae) to other frogs, testing the assumption that lack of bright coloration and alkaloid skin toxins is ancestral for dendrobatids. Most skin toxins of poison frogs are sequestered from their diet. In chapter 2, I examine a critical prediction of the diet- toxicity hypothesis, which states that independent origins of dietary specialization will be correlated with independent origins of chemical defense. Using comparative methods, I found a recurring association of dietary specialization and alkaloid sequestration, suggesting parallel evolutionary trends in the origins of aposematism. In chapter 3, I investigate the relative importance of aposematic signal components, conspicuousness and unpalatability, for anti-predator defense using natural signal variation among poison frogs of Ecuadorian Amazonia. I found equally effective predator avoidance strategies with differential investment in conspicuous coloration and toxicity across species, demonstrating a mechanistic explanation for natural diversity in warning signals and providing empirical evidence for Batesian mimicry in dendrobatid frogs. In chapter 4, I examine the dynamics of Batesian mimicry where mutiple model species co-occur and, therefore, more than one warning signal could be exploited by the mimic. I demonstrate that mimics resemble the less toxic and less abundant model species, and that this counter-intuitive mimicry pattern is selectively advantageous due to the psychological phenomenon of toxicity dependent stimulus-generalization in educated predators. Finally, in chapter 5, I explore how evidence from experimental psychology can further improve our conceptualization of predator learning and memory and therefore enhance our predictions for mimicry dynamics.