With 149 currently recognized species, Hypostomus is one of the most species-rich catfish genera in the world, widely distributed over most of the Neotropical region. To clarify the evolutionary history of this genus, we reconstructed a comprehensive phylogeny of Hypostomus based on four nuclear and two mitochondrial markers. A total of 206 specimens collected from the main Neotropical rivers were included in the present study. Combining morphology and a Bayesian multispecies coalescent (MSC) approach, we recovered 85 previously recognized species plus 23 putative new species, organized into 118 'clusters'. We presented the Cluster Credibility (CC) index that provides numerical support for every hypothesis of cluster delimitation, facilitating delimitation decisions. We then examined the correspondence between the morphologically identified species and their inter-specific COI barcode pairwise divergence. The mean COI barcode divergence between morphological sisters species was 1.3 ± 1.2%, and only in 11% of the comparisons the divergence was ≥2%. This indicates that the COI barcode threshold of 2% classically used to delimit fish species would seriously underestimate the number of species in Hypostomus, advocating for a taxon-specific COI-based inter-specific divergence threshold to be used only when approximations of species richness are needed. The phylogeny of the 108 Hypostomus species, together with 35 additional outgroup species, confirms the monophyly of the genus. Four well-supported main lineages were retrieved, hereinafter called super-groups: Hypostomus cochliodon, H. hemiurus, H. auroguttatus, and H. plecostomus super-groups. We present a compilation of diagnostic characters for each super-group. Our phylogeny lays the foundation for future studies on biogeography and on macroevolution to better understand the successful radiation of this Neotropical fish genus.
Keywords: Cluster credibility index; DNA barcode; Hidden diversity; Hypostominae; Multispecies coalescent approach; Species delimitation.
Copyright © 2019 Elsevier Inc. All rights reserved.