Taxonomic congruence versus total evidence, and amniote phylogeny inferred from fossils, molecules, and morphology.

Abstract

Taxonomic congruence and total evidence are competing paradigms in phylogenetic inference. Taxonomic congruence focuses on deriving a consensus from the results obtained from separately analyzed data sets, whereas total evidence uses character congruence in the search for the best-fitting hypothesis for all of the available character evidence. Explicit or implicit use of taxonomic congruence is usually employed when an investigator either has both molecular and morphological data sets or has different gene-, rRNA-, or protein-sequence data sets available. Indeed, a taxonomic congruence rationale is frequently used as the basis for exploring classes of data, thus allowing comparison between the phylogenetic signal emerging from a particular data set and those of other such classes. Problematic aspects of employing the taxonomic congruence approach include the potentially misleading and arbitrary choices of both a consensus method and the division of characters into subsets. If the goal of an analysis is to provide the best estimate of genealogy afforded by the available character evidence, then taxonomic congruence is substantially more arbitrary than a total evidence approach. The theoretical advantages of phylogenetic estimates based on total evidence are argued in the present study and are illustrated with an example of amniote relationships. We report conflicting results from total evidence and taxonomic congruence approaches, with analyses of previously reported data from both fossil and living amniotes and from both morphology and molecules, the latter including available 18S rRNA, 28S rRNA, and protein sequences. We conclude that a more highly resolved and robust phylogenetic hypothesis of amniotes, the traditional one, emerges when a total evidence approach is employed.