Comment on Amézquita et al. (2017) "Conspicuousness, color resemblance, and toxicity in geographically diverging mimicry: The pan-Amazonian frog Allobates femoralis".

Amézquita et al. (2017) recently concluded that species of the Allobates femoralis group are toxic to mice at levels equivalent to syntopic alkaloid-containing poison frogs, which they attributed to the presence of alkaloids in skin secretions. However, the chemical composition of skin secretions was not analyzed, and here we present additional data supporting the absence of alkaloids in skin secretions of the Allobates femoralis group. Instead, we suggest the observed toxicity was caused by the anesthetic benzocaine, which was applied to the buccal cavity to euthanize frogs prior to skin removal. We show that orally administered benzocaine is rapidly incorporated into the skin of species that sequester and do not sequester alkaloids, which casts doubt on the conclusion that Allobates femoralis group skin secretions are toxic and makes the results of experiments with alkaloid-containing species of Adelphobates and Ameerega uninterpretable. To prevent experimental errors and misinterpretations in studies of amphibian chemical defense, we encourage researchers to test the chemical composition of samples prior to experimentation, include all necessary controls to detect false positives, conduct small pilot studies for new methods, and consider the limitations of particular methods and their ability to address the intended research questions.