SKULL OF A PARIASAURIAN REPTILE. 179 
occiput is plate-like with a wide supraoccipital separating the 
posttemporal fosse. Any bones which could be postparietals or 
tabulares are on the posterior surface overlapping the supra- 
occipital, the brain-cavity is very high, and the ear, as shown 
extremely well by the ‘ brain-cast” figured by Case, is very low 
in the skull. 
The angular is flat and notched; there are two coracoidal 
elements, the interclavicle is flat and wide and not T-shaped, and 
the carpus and tarsus are of thoroughly Therapsid type. In fact, 
every character that is found in all South-African Therapsids is 
also present in Dimetrodon, which must hence be included in the 
same group. [It is, however, much more primitive than any 
South-African type in many features. | 
With Dimetrodon, Hdaphosaurus must go to the Therapsids, for 
it also has the characteristic occiput and angular. 
Varanosaurus, as shown by the magnificent type-specimen of 
V. acutirostris in Munich, has an occiput of the same type, 
although of course it is very incompletely known. Unfortunately 
the angular of this type is completely unknown; the form has, 
however, the two coracoidal elements, shown extraordinarily 
clearly in the Munich specimen, a flat and not T-shaped inter-~ 
clavicle, and, as shown by Williston’s excellent description, the 
feet only differ by lack of ossification of centralia. It can, I 
think, also be regarded as a Therapsid. 
When one comes to Ophiacodon the problem becomes much 
more difficult. The whole of the post-cranial skeleton of that 
type, as described by Williston and Case, seems to be essentially 
identical with Varanosaurus, but the extraordinary skull with two 
temporal vacuities is apparently very different. It is exceedingly 
unfortunate that no description of the occiput or lower jaw is 
possible, and, in my opinion, in the absence of that knowledge, 
we are not justified in discussing the position of Ophiacodon 
amongst primitive Reptilia. 
The foregoing discussion will, I hope, have made clear what, in 
my opinion, are the really important characters of the Therapsids ; 
it remains only to examine Pariasaurus in the light of them. 
It shares with the Therapsids the possession of two coracoidal 
elements and a single squamosal bone. The occiput is not in the 
least plate-like, the postparietals and tabulares are quite on the 
upper surface of the skull, the brain-cavity is long and low, 
the opening from the brain-cavity to the inner ear is high on the 
side-wall of the brain-cavity, the angular is boat-shaped, and the 
internal mandibular vacuity, which, by its excessive enlargement, 
gives rise to the flat Therapsid angular, is extremely small; the 
interclavicle is T-shaped and narrow; there are no centralia on 
the carpus or tarsus, and the proximal tarsals are fused. _ 
It is thus certain that so far from being at all closely related 
to the Therapsids, Pariasawrus represents an extremely different 
12* 
