Abstract

Sauria is the crown-group of Diapsida and is subdivided into Lepidosauromorpha and Archosauromorpha, comprising a high percentage of the diversity of living and fossil tetrapods. The split between lepidosauromorphs and archosauromorphs (the crocodile-lizard, or bird-lizard, divergence) is considered one of the key calibration points for molecular analyses of tetrapod phylogeny. Saurians have a very rich Mesozoic and Cenozoic fossil record, but their late Paleozoic (Permian) record is problematic. Several Permian specimens have been referred to Sauria, but the phylogenetic affinity of some of these records remains questionable. We reexamine and review all of these specimens here, providing new data on early saurian evolution including osteohistology, and present a new morphological phylogenetic dataset. We support previous studies that find that no valid Permian record for Lepidosauromorpha, and we also reject some of the previous referrals of Permian specimens to Archosauromorpha. The most informative Permian archosauromorph is Protorosaurus speneri from the middle Late Permian of Western Europe. A historically problematic specimen from the Late Permian of Tanzania is redescribed and reidentified as a new genus and species of basal archosauromorph: Aenigmastropheus parringtoni. The supposed protorosaur Eorasaurus olsoni from the Late Permian of Russia is recovered among Archosauriformes and may be the oldest known member of the group but the phylogenetic support for this position is low. The assignment of Archosaurus rossicus from the latest Permian of Russia to the archosauromorph clade Proterosuchidae is supported. Our revision suggests a minimum fossil calibration date for the crocodile-lizard split of 254.7 Ma. The occurrences of basal archosauromorphs in the northern (30