Giant stem tetrapod was apex predator in Gondwanan late Palaeozoic ice age.

Current hypotheses of early tetrapod evolution posit close ecological and biogeographic ties to the extensive coal-producing wetlands of the Carboniferous palaeoequator with rapid replacement of archaic tetrapod groups by relatives of modern amniotes and lissamphibians in the late Carboniferous (about 307 million years ago). These hypotheses draw on a tetrapod fossil record that is almost entirely restricted to palaeoequatorial Pangea (Laurussia)1,2. Here we describe a new giant stem tetrapod, Gaiasia jennyae, from high-palaeolatitude (about 55° S) early Permian-aged (about 280 million years ago) deposits in Namibia that challenges this scenario. Gaiasia is represented by several large, semi-articulated skeletons characterized by a weakly ossified skull with a loosely articulated palate dominated by a broad diamond-shaped parasphenoid, a posteriorly projecting occiput, and enlarged, interlocking dentary and coronoid fangs. Phylogenetic analysis resolves Gaiasia within the tetrapod stem group as the sister taxon of the Carboniferous Colosteidae from Euramerica. Gaiasia is larger than all previously described digited stem tetrapods and provides evidence that continental tetrapods were well established in the cold-temperate latitudes of Gondwana during the final phases of the Carboniferous-Permian deglaciation. This points to a more global distribution of continental tetrapods during the Carboniferous-Permian transition and indicates that previous hypotheses of global tetrapod faunal turnover and dispersal at this time2,3 must be reconsidered.

Highly specialized mammalian skulls from the Late Cretaceous of South America.

Dryolestoids are an extinct mammalian group belonging to the lineage leading to modern marsupials and placentals. Dryolestoids are known by teeth and jaws from the Jurassic period of North America and Europe, but they thrived in South America up to the end of the Mesozoic era and survived to the beginnings of the Cenozoic. Isolated teeth and jaws from the latest Cretaceous of South America provide mounting evidence that, at least in western Gondwana, dryolestoids developed into strongly endemic groups by the Late Cretaceous. However, the lack of pre-Late Cretaceous dryolestoid remains made study of their origin and early diversification intractable. Here we describe the first mammalian remains from the early Late Cretaceous of South America, including two partial skulls and jaws of a derived dryolestoid showing dental and cranial features unknown among any other group of Mesozoic mammals, such as single-rooted molars preceded by double-rooted premolars, combined with a very long muzzle, exceedingly long canines and evidence of highly specialized masticatory musculature. On one hand, the new mammal shares derived features of dryolestoids with forms from the Jurassic of Laurasia, whereas on the other hand, it is very specialized and highlights the endemic, diverse dryolestoid fauna from the Cretaceous of South America. Our specimens include only the second mammalian skull known for the Cretaceous of Gondwana, bridging a previous 60-million-year gap in the fossil record, and document the whole cranial morphology of a dryolestoid, revealing an unsuspected morphological and ecological diversity for non-tribosphenic mammals.

The Stapes of Gomphodont Cynodonts: Insights into the Middle Ear Structure of Non-Mammaliaform Cynodonts.

The stapes is known in several non-mammaliaform cynodonts, although it has only been cursorily studied. Here we thoroughly analyze the stapedial anatomy of several basal cynodonts in a phylogenetic framework. Our study shows that the stapedial anatomy is more variable than previously thought. The morphological variation of the stapes led to the recognition of 11 phylogenetic characters that were included in a total evidence data matrix centered in the analysis of gomphodont cynodonts. Stapes morphology does not provide evidence to suggest a direct connection between the stapes and a postquadrate tympanic membrane (if present) and the hypothesis of a dorsal process as the site of attachment of a small ligament or the stapedial muscle is supported. The re-evaluation of the theories concerning the position of the tympanic membrane in non-mammaliaform cynodonts allowed us to conclude that the hypothetical postquadrate tympanic membrane associated with the squamosal sulcus is at best relictual and most likely non-functional (not connected with the stapes). The sound waves were most likely transmitted to the stapes from a postdentary tympanic membrane through the quadrate. Our analysis results in a better understanding of the auditory system in basal cynodonts and its evolution, highlighting the variability of the stapedial anatomy.