Phylogenetic Relationships of Extinct Cetartiodactyls: Results of Simultaneous Analyses of Molecular, Morphological, and Stratigraphic Data

Abstract

Although some recent morphological and molecular studies agree that Cetacea is closely related to Hippopotamidae, there is little consensus on the phylogeny within Cetartiodactyla. We addressed this problem by conducting two analyses: (1) a simultaneous cladistic analysis of intrinsic data (morphology and molecules) and (2) a stratocladistic analysis, which included morphological, molecular, and stratigraphic data. Unlike previous simultaneous analyses, we had the opportunity to include data from the recently described hindlimbs of protocetid and pakicetid cetaceans. Our intrinsic dataset includes 73 taxa scored for 8,229 informative characters, of which 208 are morphological and 8,021 molecular. Both analyses supported the exclusion of Mesonychia from Cetartiodactyla and a close phylogenetic relationship between Hippopotamidae and Cetacea. Many polytomies in the strict consensus of the most parsimonious trees for the intrinsic dataset can be attributed to differing positions for Raoellidae, which in some trees is the sister-group to Cetacea. Pruning Raoellidae and 18 other taxa from all most parsimonious produced a fully resolved agreement subtree, which indicates that the Old World taxa Cebochoerus and Mixtotherium are successive stem taxa to Whippomorpha (i.e., Cetacea + Hippopotamidae). The main result of adding stratigraphic information to the intrinsic dataset was that we found fewer most parsimonious trees, which in most respects were congruent with a subset of the shortest trees for the intrinsic dataset. Our stratocladistic analysis supports species of Diacodexis as the most basal cetartiodactyls, a clade of suiform cetartiodactyls, a monophyletic Tylopoda that includes Protoceratidae, and a monophyletic Carnivora. We were unable to identify any pre-Miocene stem taxa to Hippopotamidae, thus its ghost lineage is still 39 million years long. The relatively low Bremer support for many nodes in our trees indicates that our phylogenetic hypotheses should be subjected to further testing.