Updated May 10, 2016 with a new therapsid cladogram
Skulls are the most plentiful elements
preserved (or excavated) from the therapsid beds of South Africa and Russia. What were the phylogenetic patterns within the base of the Therapsida. How did all those clades diverge?
The Traditional Tree
The paradigm of therapsid evolution (Rubidge and Sidor 2001, Fig. 1) holds that therapsids were derived from sphenacodonts. Tetraceratops and Biarmosuchus are the traditional basal therapsids. Thereafter dinocephalians (estemmenosuchids, tapinocephalids and anteosaurids) branch off. Another branch includes anomodonts (venjukovids, dromasaurs and dicynodonts) and finally theriodonts (gorgonopsians, therocephalians and cynodonts, which give rise to mammals).
This is the tree I will be testing.
Obviously the vast majority of the sister taxa within each clade are undoubtedly sister taxa. What I want to know is how do the clades themselves relate to one another? Lumping anomodonts and theriodonts together doesn’t seem tenable at first glance. Tetraceratops doesn’t belong here. Stenocybus doesn’t appear to be a good match for its putative sister, Stryacocephalus (Fig. 1). Other perceived problems are minor by comparison.
Setting Up a New Study
As mentioned earlier, the large phylogenetic analysis bumped sphenacodonts off the inclusion set of therapsid outgroup taxa, replacing them with Archaeothyris and Ophiacodon. The same character traits used to recover a family tree of the Reptilia are not going to work with just the Therapsida and just their skulls. Several new traits are used here (matrix available on request). There are several new taxa not used in the Rubidge and Sidor (2001) study (compare Figures 1 and 2).
The Heretical Tree as Told in Skulls
PAUP came up with this tree (Fig. 2). I was most interested in how basal taxa separated and there were a few surprises when rarely used taxa were included. Here again, Ophiacodon and Archaeothyris are the outgroup taxa, as reported earlier.
The basalmost therapsid appears to be Cutleria, which nests at the base of the a largely fangless Anomodontia (dicynodonts, dromasaurs and kin) on one branch and fanged Hipposaurus on the other. That means fangs were secondarily derived in dicynodonts. And widely flaring cheekbones were convergent with dinocephalians and therocephalians.
Stenocybus nests next at the base of the Anomodontia based on its overall similarity to sphenacodonts. Kammerer (2011) considered it a juvenile Sinophoneus.
The Burnettidae/Ictidorhinosauridae is a clade that includes Herpetoskylax and the “horned” taxa, Lemurosaurus, Proburnetia and Burnetia. They progressively reduced their teeth and have wider, lower, knobbier skulls.
Biarmosuchus nested at the base of the remainder of the Therapsida, all fanged taxa until certain herbivores appear.
Eotitanosuchus and Scymnognathus (representing the Theriodontia) nested next. This clade gave rise to gorgonopsids and mammals.
The skull of Sinophoneus was reported as dorsoventrally crushed, but as is it is very close in appearance to Eotitanosuchus. Kammerer (2011) reinflated its appearance. Scoring did not change either way. Sinophoneus is basal to Deuterosaurus and Estemmenosuchus on one branch and several round-snouted herbivores on the other.
Phthinosuchus nests at the base of the Anteosauria. The first anteosaur branch includes Syodon and two other syodonids (both with a gracile lateral temporal arch). The second includes a putative anomodont, Biseridens (Liu et al. 2009), and two long-faced tapinocephalids, Jonkeria and Struthiocephalus. The second branch also includes Titanophoneus and two species of Anteosaurus, A. magnifius and A. abeli.
Struthionops and the short-snouted tapinocephalids, Moschops and Ulemosaurus arise from Biseridens.
Changes From the Traditional Tree
Here I used several taxa not typically used. Here the carnivorous Dinocephalia were not separate from the herbivores. The tapinocephalids were split into two distinct yet convergent clades. The traditional tree (Fig 1) nested tapinocephalids with Tapinocaninus and titanosuchids (includes Jonkeria) and Estemmenosuchus.
Dr. Chris Kammerer’s Convergence
Kammerer (2011) reported, “Several characters appear to be homoplastic between anteosaurs and tapinocephalians, with the highest degree of convergence occurring between the giant anteosaurs Titanophoneus and Anteosaurus and the tapinocephalids.” Here that “convergence” was found to be homoplastic. I’m always eager to see if change results when more data comes in. I understand a comprehensive report on these taxa is due soon.
The Value of the Reconstruction
Reconstructing taxa, I think, is key to understanding traits and correctly scoring a matrix. Placing reconstructions of sister taxa next to one another in phylogenetic order to see if they do indeed greatly resemble one another is the first test of a recovered tree. In that way mismatches are identified and corrective measures can be initiated.
The Struthiocephalus Problem
Struthiocephalus nested with Jonkeria in the above tree, but not in the tree below, which includes several more tapinocephaline taxa. Just goes to show what happens when you add enough taxa. Things can change, especially when the best data comes from freehand line drawings.
Kammerer C 2011. Systematics of the Anteosauria (Therapsida: Dinocephalia). Journal of Systematic Palaeontology 9(2):261-304.
Liu J, Rubidge B and Li J 2009. A new specimen of Biseridens qilianicus indicates its phylogenetic position as the most basal anomodont. Proceedings of the Royal Society B 277 (1679): 285–292. doi:10.1098/rspb.2009.0883. PMC 2842672. PMID 19640887.
Rubidge BS and Sidor CA 2001. Evolutionary Patterns Among Permo-Triassic Therapsids. Annual Review of Ecological Systems 32:449–80. online pdf