Bever et al. (2014) report that Eunotosaurus (Fig. 1) is a turtle ancestor. This has been falsified in phylogenetic analysis (and see below). We learned earlier that Eunotosaurus is a turtle-mimic that actually nests with Acleisotorhinus, leaving no known descendants. Phylogenetic turtle ancestors include Stephanospondylus (Fig. 2), which does not have temporal fenestration. However, bolosaurids (Bolosaurus and Belebey), taxa known only from skulls, are also close to the base of this lineage and do have lateral temporal fenestra.
From the Bever et al. abstract: “The reptile skull is an increasingly utilized model for understanding the evolution and development of vertebrate adaptation. Turtles are an important yet enigmatic piece of this puzzle. The earliest uncontroversial stem turtles exhibit a fully anapsid skull with an adductor chamber concealed by bone. If this lack of fenestration reflects conservation of the ancestral condition, then turtles are an extant remnant of an early reptile radiation that excludes the other living forms. If turtles are nested within crown Diapsida, then their anapsid skull is a secondary configuration built on a diapsid structural plan. No direct paleontological evidence yet exists for this reversal, a situation that epitomizes a general lack of consilience between the fossil record and the molecular signature of living taxa and one that obfuscates attempts to synthesize broad evolutionary patterns across Reptilia.
“Eunotosaurus africanus is a 260 Ma fossil reptile whose status as an early stem turtle continues to be strengthened by new cranial and postcranial synapomorphies. Here we use computed tomography (CT) to study the temporal region of Eunotosaurus and to formulate a model for the origin of the anapsid and diapsid skulls of modern amniotes. Expression of a lower temporal fenestra (LTF) supports the hypothesis that the closed cheek of modern turtles is secondary (1). The ventrally unbounded nature of the LTF places Eunotosaurus at odds with parareptiles (2), but also with pandiapsids where an unbounded LTF is known only in conjunction with the more conservative upper temporal fenestra (UTF) (3). The region housing the diapsid UTF is overlain by an elongate supratemporal in Eunotosaurus. In contrast to the plesiomorphic condition, digitally removing the supratemporal reveals a moderate-sized opening circumscribed by the same elements that define the UTF (4). Additional evidence that this covering is secondary is drawn from the observation that in Eunotosaurus the supratemporal overlaps the postorbital, whereas plesiomorphically these two elements are abutting or the postorbital overlaps the supratemporal. We propose (5) Eunotosaurus captures an early step in the evolution of the anapsid turtle skull in which the UTF was secondarily covered by the supratemporal before being obliterated through expansion of neighboring dermal elements (6). The recognition of such a critical transitional form facilitates the articulation of meaningful transformational and functional models that can be tested with future paleontological discoveries and rapidly emerging developmental data.”
1). Pure speculation when done without phylogenetic analysis.
2). The post-crania of Eunotosaurus is clearly derived, so the skull is also, reduced from the ‘synapsid’ grade skull found in the following (non-synapsid) millerettids: Acleistorhinus, Feeserpeton, Australothyris, Oedaleops, Eothyris, Ennatosaurus, Casea and Cotylorhynchus, taxa closer to Eunotosaurus than Eunotosaurus is to turtles.
4). The same could be said of any of the taxa in (2).
5). “Propose” is another word for “speculate without evidence.” For this idea to have weight, they should “show” or “demonstrate,” but they cannot do this phylogenetically if they include bolosaurids and Stephanospondylus.
6). Except in basal turtles the supratemporal is a long bone rimming the posterior cranium, as in related pareiasaurs, bolosaurids and Stephanospondylus.
Turtles are so firmly nested in their present tree topology that you can remove (as I did) Stephanospondylus, all four pareiasaurs, and both bolosaurids and the tree topology does not change. I even deleted the Macroleter clade. Turtles still don’t nest with Eunotosaurus. This has been known online for the last three years.
Bever GS, Lyson T and Bhullar B-A 2014. Fossil evidence for a diapsid origin of the anapsid turtle skull. SVP 2014 Abstracts pg. 91