A gene-tree test of the traditional taxonomy of American deer: the importance of voucher specimens, geographic data, and dense sampling

The taxonomy of American deer has been established almost entirely on the basis of morphological data and without the use of explicit phylogenetic methods; hence, phylogenetic analyses including data for all of the currently recognized species, even if based on a single gene, might improve current understanding of their taxonomy. We tested the monophyly of the morphology-defined genera and species of New World deer (Odocoileini) with phylogenetic analyses of mitochondrial DNA sequences. This is the first such test conducted using extensive geographic and taxonomic sampling. Our results do not support the monophyly of <i>Mazama</i>, <i>Odocoileus</i>, <i>Pudu</i>, <i>M. americana</i>, <i>M. nemorivaga</i>, <i>Od. hemionus</i>, and <i>Od. virginianus</i>. <i>Mazama</i> contains species that belong to other genera. We found a novel sister-taxon relationship between "<i>Mazama</i>" <i>pandora</i> and a clade formed by <i>Od. hemionus columbianus</i> and <i>Od. h. sitkensis</i>, and transfer <i>pandora</i> to <i>Odocoileus</i>. The clade formed by <i>Od. h. columbianus</i> and <i>Od. h. sitkensis</i> may represent a valid species, whereas the remaining subspecies of <i>Od. hemionus</i> appear closer to <i>Od. virginianus</i>. <i>Pudu</i> (<i>Pudu</i>) <i>puda</i> was not found sister to Pudu (Pudella) mephistophiles. If confirmed, this result will prompt the recognition of the monotypic <i>Pudella</i> as a distinct genus. We provide evidence for the existence of an undescribed species now confused with <i>Mazama americana</i>, and identify other instances of cryptic, taxonomically unrecognized species-level diversity among populations here regarded as <i>Mazama temama</i>, "<i>Mazama</i>" <i>nemorivaga</i>, and <i>Hippocamelus antisensis</i>. Noteworthy records that substantially extend the known distributions of <i>M. temama</i> and "<i>M.</i>" <i>gouazoubira</i> are provided, and we unveil a surprising ambiguity regarding the distribution of "<i>M.</i>" <i>nemorivaga</i>, as it is described in the literature. The study of deer of the tribe Odocoileini has been hampered by the paucity of information regarding voucher specimens and the provenance of sequences deposited in GenBank. We pinpoint priorities for future systematic research on the tribe Odocoileini.