ASSESSING THE PHYLOGENETIC PERFORMANCE OF MORPHOLOGICAL PARTITIONS USING META-ANALYSIS AND MOLECULES

Morphology is fundamental for phylogenetic reconstruction and testing evolutionary hypotheses, especially for fossil taxa. It is, however, widely acknowledged that morphology is problematic; higher levels of homoplasy, subjectivity and linkage are observed in phenotypic characters relative to genotypic characters (i.e. molecular sequence data). To investigate the performance of morphological partitions we undertook meta-analyses of phenotypic datasets of extant clades. Analysis of 28 mammal datasets finds osteological data and dental data to be significantly different, and of the two, osteology is significantly more consistent with molecular trees. Analysis of bird datasets also finds osteological characters to be significantly more consistent with molecular trees relative to non-osteological morphology (plumage, myology etc.). As such, results indicate that osteological data may more accurately reflect evolutionary history than the other aspects of morphology. The congruence of osteology and molecules is generally encouraging from a palaeontological perspective, but is problematic for mammals given the taphonomic bias toward preservation of teeth and the dominance of dental data. Investigating morphology in this way will allow a more molecularly informed approach to fossil phylogenies and potentially identification of the misleading aspects of morphology for clades across the tree of life.