The use of cave models to assess potential predation on cave salamanders

Clay models are frequently used in ecological studies to estimate potential predation rates on small and cryptic vertebrates (Bateman et al. 2016). These models are easy to shape and they retain predator marks, allowing the identification of the body part attacked and of the predator. In herpetological studies, prey models are used to evaluate colour polymorphism, aposematism, mimicry and to assess differential predation rates in different habitats. Clay models have been applied in studies on both tailed and tailless amphibians in different geographic regions, with a first application in Europe by Velo-Antón & Cordero Rivera (2011). The present research aimed to evaluate by means of realistic clay models: i) if different potential predation rates were observed on the plethodontid salamander Speleomantes strinatii living in different habitats, and ii) if the presence of a brooding female reduced the potential predation of attended egg clutches, in comparison to unattended clutches. The results show that predators attacked the models significantly more (P < 0.01) in woodlands (43 attacked models out of 90) than inside caves (16 attacked model out of 94). Therefore, salamander populations inhabiting underground environments may benefit from reduced predation risk in comparison to forest-dwelling populations (Salvidio et al. 2017). Concerning the second experiment, similar predation rates on isolated eggs and on eggs attended by females were observed. This may suggest that all kind of models (i.e. eggs and/or salamanders) were not discriminated by the predators and attacked with similar frequencies