Phylogenetic and morphometric assessment of the evolution of the longirostrine crocodylomorphs

Study of the evolution of crocodile-line archosaurs (Pseudosuchia) has intensified recently. Along with this increased attention have come dramatic changes in our understanding of the evolutionary relationships of the group. Modern crocodylians are generally similar-looking with most of their morphological disparity expressed in their skulls. However, phylogenetic studies have demonstrated that skull shape is highly convergent, with similar skull forms arising in distantly related lineages. Yet, crocodylians and their relatives (crocodyliforms) have traditionally been divided into taxonomic and ecomorphological groups based on skull morphology. Many of the crocodyliform groups with the most controversial affinities also share the same skull shape (long, slender snouts). This issue was assessed from multiple angles including geometric morphometrics, detailed assessment of anatomy, and expanded phylogenetic analyses. Previously proposed skull shape categories were based largely on assessments of snout length and width but ignored important features in the rest of the skull. I applied two-dimensional geometric morphometrics including landmarks and sliding semilandmarks to characterize the skull shape of 131 extant and extinct crocodyliforms. Newly developed shape categories based on normal mixture analysis show a weak correlation to phylogeny corroborating hypotheses of convergent evolution. Different skull shapes are not evenly distributed through time and the results presented here indicate that crocodyliform disparity peaked in the Late Cretaceous followed by a dramatic decline in the Maastrichtian (preceding the end-Cretaceous extinction). The emptied region of morphospace, exemplified by a short and narrow snout, was never explored by crown-group crocodylians. Modern crocodylian disparity is much lower than in the Cretaceous, but is similar to Jurassic levels. The recent increases in knowledge of pseudosuchian relationships have important implications for phylogenetic studies of crocodylomorphs. Previous studies of crocodylomorph and crocodyliform relationships have rooted trees on outgroups that in many cases are either quite distantly related to the group under study (e.g. Gracilisuchus stipanicicorum, a basal suchian), or could actually belong within the ingroup. Thalattosuchia, one of the earliest occurring groups of crocodyliforms (Early Jurassic, Sinemurian), has a controversial phylogenetic position relative to other crocodyliforms — they are recovered in either a basal position, nested high up in the tree, or sister to Crocodyliformes. Thalattosuchians lack several crocodyliform apomorphies, but share a longirostrine skull shape with highly derived neosuchian groups. These groups share a similar ecological habit, suggesting the derived position of thalattosuchians may be the result of convergent evolution. Several of the "shared" characters uniting these groups are due to ambiguously worded character state definitions - structures that are superficially similar (though anatomically different in detail) are scored the same. A new analysis of crocodylomorphs recovers the thalattosuchians as the sister-group to Crocodyliformes outside of the "protosuchians" and all other crocodyliforms. It is also demonstrated that the addition of the rauisuchid Postosuchus kirkpatricki to the outgroup of previously published analyses draws thalattosuchians down the tree as the sister-group of Crocodyliformes. These results demonstrate the importance of careful outgroup sampling, and highlight issues with current descriptions of morphological characters.