Comparative Assessment of tyrannosaurid interrelationships

Abstract

Synopsis We employ a new comparative method to four cladistic analyses of tyrannosaurid dinosaurs to identify root causes for differences between phylogenetic results. The comparative method is a three‐step procedure that (1) adjusts competing hypotheses so they share equivalent taxonomic scope, (2) isolates the character data relevant to the common problem, and (3) divides relevant character data into shared and novel partitions. It is then possible to quantify the degree of similarity between character data using three indices (ancestor similarity index, character similarity index and character state similarity index). The most parsimonious cladograms generated by the four analyses of tyrannosaurids appear fairly congruent, with two subclades present in all four analyses (Albertosaurus and Gorgosaurus versus Daspletosaurus, Tarbosaurus and Tyrannosaurus). A comparative examination of the underlying character data, however, highlights striking differences in character selection and significant differences in character state scores. Character selection and differences in scoring are root causes for phylogenetic incongruence. Comparative analysis reveals the existence of many data‐level differences that remain largely obscured when comparison is limited to the most parsimonious cladograms.