Yesterday we looked at the evolution of the temporal fenestrae in lepidosauromorph reptiles. Today we’ll look at the same sequence in archosauromorph reptiles. Note that both develop their temporal fenestrae by convergence from “anapsid”-type precursors.
Starting with the basal “anapsid” reptile Protorothyris, the lateral temporal fenestra developed large in Elliotsmithia and small in Arachaeothyris, both considered basal synapsids. Heleosaurus is also a basal synapsid, but phylogenetically it nests among the protodiapsids alongside Millerosaurus (not related to Milleretta). Early diapsids include Eudibamus, Spinoaequalis and Petrolacosaurus.
Contra tradition, Petrolacosaurus is not the basalmost diapsid.
Thereafter this lineage of diapsids splits into a major marine clade, the Enaliosauria and a more terrestrial clade, the Thadeosauria, which gave rise to crocs and dinos along with several other terrestrial reptiles.
Preceding and within the Enaliosauria the lateral temporal fenestra fills in with Araeoscelis and both upper and lateral fenestrae fill in with Mesosaurus. The lower temporal bar is lost or modified in several marine taxa. The upper temporal fenestra is squeezed shut in thalattosaurs, but for the most part the temple and cheek architecture remain conservative in the archosauriformes.
At the first appearance of the lateral temporal fenestra, the postorbital, squamosal, jugal and quadratojugal all border the rim in Aerosaurus. The quadratojugal is excluded in most other synapsids due to a long jugal process. Heleosaurus follows the Aerosaurus pattern. Millerosaurus has an odd and autapomophic squamosal process that divides the lateral temporal fenestra into anterior and posterior holes. The contribution of the quadtratojugal in two of the diapsids is less clear, but Petrolacosaurus demonstrates contact with the ltf.
The upper temporal fenestra appears in Eudibamus bordered by the postorbital, parietal and squamosal, a pattern retained by the other diapsids listed.
Reconstruction of the in situ skull bones demonstrates the autapomorphic lateral placement of the upper temporal fenestra in Petrolacosaurus, different from the dorsal placement in prior reconstructions.
A reordering of the reptile family tree is required
Due to these various appearances and disappearances of temporal fenestrae, reptiles must be nested and classified according to their overall morphologies, not strictly according to their cheek and temple architecture.
Moreover, the diapsid skull architecture did not appear all at once, but evolved from a simpler architecture;
Archosauromorph cheek and temple architectural patterns are distinct from those in the new Lepidosauromorpha, as we saw yesterday. These two clades share no common ancestor with a diapsid skull architecture, which is a heretical departure from the traditional paradigm.