Molecular phylogeny of the horse flies: A framework for renewing tabanid taxonomy

Introduction: Horse flies (Diptera: Tabanidae) are an economically, medically, and ecologically important group containing bloodfeeding pests and long-proboscid pollinators. This group, with over 4,300 described species, has experienced a severe loss of taxonomic expertise. The TabanidPEET project will fill this gap by training at least four postgraduates. As a framework for student taxonomic revisions, we present here a new higher-level molecular phylogeny of the horse flies. This work will guide morphological character choice and focus student projects on evolutionary units. We also ask the following questions: 1) Are the three current subfamilies monophyletic? 2) Is pollinator morphology ancestral in Tabanidae? 3) When did the major tabanid clades diverge and radiate? 4) What are the morphological and life history features supporting tabanid clades? METHODS: Using exemplars from 10 of 11 tribes, we sampled two nuclear (CAD, AATS), and one mitochondrial (COI) gene. Phylogenetic inference was performed using Parsimony and Maximum Likelihood in Paup*, and using Bayesian Inference in Beast. Character reconstruction was optimized using BayesTraits and MacClade. RESULTS: Our phylogenetic hypothesis for Tabanidae is largely concordant with previous hypotheses, but also indicates several newly recognized groupings that will form the basis for focused taxonomic revisions. Divergence time estimates based on nucleotide data and fossil evidence reveal rapid radiations in the early history of horse fly diversification. CONCLUSIONS: The current taxonomy of Tabanidae at the subfamily and tribal level needs to be revised based on new molecular evidence. Additionally, long-proboscides appear to be ancestral, suggesting that tabanids are ancient pollinators.