Brazilosaurus holotype – Shikama (1966)
Brazilosaurus (Fig. 1) was described by Shikama (1966) as a sort of mesosaur with a longer neck, shorter teeth and other minor differences. From the available online pdf (created from photocopies), I can only read the line drawing, but photos were published originally (if anyone has them, I’d like to see them, they’re too dark on the pdf).
The illustrated cheek region is much too large compared to all sister taxa, so it may represent the occiput, which it is the right size to be. The jugal looks similar to those in Stereosternum and Mesosaurus, so likewise probably framed a lateral temporal fenestra. Guesswork at this point.
Brazilosaurus referred specimen – Rossman (2002)
published this image (Fig. 1) of a specimen he attributed to Brazilosaurus. While not too far off from each other phylogenetically, no one would ever confuse the two. They’re just too different. Note the scale bars. Those tells us the Rossman specimen is not just a short-rostrum juvenile. Overall it’s the same size but has a shorter/taller rostrum and larger orbit.
My re-interpretation of the Rossman bones is in color. The reinterpretation is closer to the interpretation of sister taxa. The parietal goes on top of the head, not posterior to the orbit, for instance. And an unidentified bone sticks up from the cranium and that bone has what appears to be half of a pineal / parietal opening. When I see hirez photos I’ll revisit this if necessary.
The referred specimen is likewise known to me only as a drawing with flaws (Fig. 2). Again, I wish I had access to the original specimen, or to a good photo. Or to the post-crania. All attempts at contacting Rossman have failed. I understand he is in poor health.
Xinpusaurus suni (Fig. 3) is a basal thalattosaur (close to ichthyosaurs like Utatsusaurus) sharing many traits with the Rossman specimen. It has no upper temporal fenestra, but it has a large lateral temporal fenestra. So these fenestra appear to come and go, as they do in basal diapsids like Araeoscelis.
Adding Brazilosaurus to the large reptile tree nests it with Mesosaurus. Adding the Rossman specimen nests it basal to all thalattosaurs and therefore close to Xinpusaurus. Hupehsuchus + Utatsusaurus are outgroups.
So, one more mesosaur – thalattosaur connection.
Rossmann T 2002. Studien an Mesosauriern (Amniota inc. sed., Mesosauridae): 3. Neue Aspekte zur Anatmie, Erhaltung und Paläoökologie aufgrund der Exemplare im Paläontologischen Institut der Universität Zurich. Neues Jahrbuch fur Geologie und Paläontologie, Abhandlungen 224, 197-221.
Shikama T 1966. On a reptilian skeleton from the Palaeozoic formation of San Paulo, Brazil. Transactions and Proceedings of the Palaeontological Society of Japan, New Series 64:351-358.