FOSSILS, TOPOLOGIES, MISSING DATA, AND THE HIGHER LEVEL PHYLOGENY OF EUTHERIAN MAMMALS

The claim that fossils strongly influence the outcome of phylogenetic analysis for extant taxa has been recently demonstrated in a study of higher amniote relationships (Gauthier, Kluge, and Rowe, 1988, Cladistics 4:105-209). Similar, albeit more limited, effects are also apparent in the case of higher (eutherian) mammal phylogeny. For example, the incorporation of selected fossil taxa shifts the balance of evidence concerning hyracoid relationships in favor of a close affinity between this order and sirenians and proboscideans (Tethytheria). The use of introduces character conflict to a grouping of hyracoids and perissodactyls, indicating that some traits shared by extant hyracoids and extant perissodactyls were acquired independently. Although analysis of the combined (Recent and fossil) data set favors the hyracoid-tethythere clade over the hyracoid-perissodactyl clade in the most parsimonious tree, the former grouping is highly vulnerable to collapse when minor homoplasy is introduced. This instability is expected because the fossils function as transitional taxa, with the effect of reducing the number of characters for particular nodes. At the same time, such transitional taxa can more precisely represent the actual transformation events and accordingly provide important information on phylogenetic history. The overall effects of fossils in analysis of higher eutherian mammals is much more subtle than in the amniote case. A primary reason for this discrepancy appears to be the large proportion of characters for extant amniote taxa that show marked transformation relevant to early splitting events. This level of information loss though marked transformation is not as evident in the case of the mammalian orders. Both extant and fossil taxa can introduce major shifts to the original topology of a given tree. Nonetheless, the tendency for fossils to preserve characters lost in living taxa suggests a pivotal role for fossils in many current and pending studies.