A reappraisal of the morphology and systematic position of the theropod dinosaur Sigilmassasaurus from the "middle" Cretaceous of Morocco.

Abstract

Sigilmassasaurus brevicollis is an enigmatic theropod dinosaur from the early Late Cretaceous (Cenomanian) of Morocco, originally based on a few isolated cervical vertebrae. Ever since its original description, both its taxonomic validity and systematic affinities were contentious. Originally considered to represent its own family, Sigilmassasauridae, the genus has variously been suggested to represent a carcharodontosaurid, an ornithischian, and, more recently, a spinosaurid. Here we describe new remains referrable to this taxon and re-evaluate its taxonomic status and systematic affinities. Based on the new remains, a re-evaluation of the original materials, and comparisons with other spinosaurids, the holotype of Sigilmassasaurus brevicollis is identified as an anterior dorsal, rather than a cervical vertebra, and differences between elements referred to this taxon can be explained by different positions of the elements in question within the vertebral column. Many characters used previously to diagnose the genus and species are found to be more widespread among basal tetanurans, and specifically spinosaurids. However, the taxon shows several autapomorphies that support its validity, including the presence of a strongly rugose, ventrally offset triangular platform that is confluent with a ventral keel anteriorly in the mid-cervical vertebral centra and a strongly reduced lateral neural arch lamination, with no or an incomplete distinction between anterior and posterior centrodiapophyseal laminae in the posterior cervical and anterior dorsal vertebrae. We argue furthermore that Spinosaurus maroccanus, also described on the basis of isolated cervical vertebrae from the same stratigraphic unit and in the same paper as Sigilmassasaurus brevicollis, is a subjective synonym of the latter. Both a detailed comparison of this taxon with other theropods and a formal phylogenetic analysis support spinosaurid affintities for Sigilmassasaurus. However, we reject the recently proposed synonymy of both Spinosaurus maroccanus and Sigilmassasurus brevicollis with Spinosaurus aegyptiacus from the Cenomanian of Egypt, as there are clear differences between the vertebrae of these taxa, and they do not share any derived character that is not found in other spinosaurids. Together with a comparison with other spinosaurid vertebral material from the Kem Kem, this suggests that more than one taxon of spinosaurid was present in the Kem Kem assemblage of Morocco, so the referral of non-overlapping material from this unit to a single taxon should be regarded with caution.