Reassessment of Cenozoic Longirostrine crocodiles of East Africa

Phylogenetic analyses made using molecular and morphological datasets yield conflicting results for placement of the modern Indian gharial (Gavialis gangeticus) and false gharial (Tomistoma schlegelii) among other crocodylians. Morphological analyses indicate they are distantly related, last shared a common ancestor more than 80 million years ago and are similar because of evolutionary convergence. Molecular analyses instead support a close relationship and a divergence within the past 40 million years, with similarities resulting from common ancestry. The continued disagreement between methods limits our ability to use Crocodylia as a model clade for the integration of the earth and life sciences. Numerous undescribed Neogene slender snouted crocodylians from Kenya and Uganda reveal a substantial diversity of gharials in the region. They preserve characters unique to both modern lineages, suggesting the closer relationship produced by molecular methods may be correct. Present phylogenetic analyses draw some of these close to Gavialis and others close to Tomistoma, but this may reflect incompleteness in some forms and insufficient character sampling. Ontogeny may also be a confounding factor in assessing modern relationships. Preliminary results indicate similar morphology between juvenile Tomistoma and adult Gavialis, suggesting paedomorphosis may be inaccurately drawing gavials towards the more basal thoracosaurs. Inclusion of the novel African taxa when thoracosaurs are excluded from morphological phylogenetic analysis supports the close gharial relationship found by molecular methods. The persistence of the morphological topology when all taxa are included indicates additional sampling of taxa and morphology is needed to resolve the obscured relationships of the modern gharials.