Meeting Abstract
The mouse-eared bats (Myotis) are found on every continent except Antarctica and comprise three primary ecomorphs with different feeding strategies (i.e. aerial hawking, gleaning, and trawling for aquatic prey). Despite striking morphological similarities within ecomorphs, molecular phylogenies have shown that these groups are not monophyletic. In this study, we use phylogenetic comparative methods, machine learning tools and morphological data to understand which traits are linked to the convergent phenotypes and foraging strategies of Myotis. We evaluated 15 traits hypothesized to be of significance for predicting foraging strategies in an analysis of over 300 specimens representing 54 species from 6 continents. No phylogenetic signal was found for any of the traits, with each trait displaying a low K value and lacking statistical significance, suggesting that similarities among different species in these traits is due to convergence rather than shared ancestry. Convergence analyses using comparative methods revealed significant changes in the mean values for each trait at particular nodes of our tree, and detected a lower number of regimes than shifts for each trait, which together are indicative of convergent evolution. A machine-learning analysis to predict feeding type resulted in an accuracy rate equal to or greater than 75%. The three most important traits for prediction of foraging strategy in Myotis are ear length, tibia length, and foot length. Ear length may play a role in the ability of bats to hear prey-generated sounds, and tibia and foot length may play a role in prey capture using the uropatagium and feet.
