MAMMAL-LIKE REPLILES. ata 
vestigial ‘ supratemporal” in Varanosaurus is a primitive feature 
of little importance. 
Comparison of a Tapinocephaloid with Dimetrodow leads to 
some interesting results, as Broom has already pointed out. 
There is a resemblance in shape, owing to the very short teim- 
poral regions and high and compressed form of the skull in 
both groups. The basicranial region of Dimetrodon, although 
more specialised than that of Varanosaurus, shows no suggestions 
of the characteristic Deinocephalian structure, and is, in fact, 
extremely like that of the more primitive of the Gorgonopsids. 
The very peculiar parasphenoid is identical in the Deinocephalia 
and Dimetrodon, and it is probable that the ethmoids are also 
similar. The quadrate of Dimetrodon is modified from its 
primitive condition by that development backwards of the 
posterior angles of the skull which is shown by the curious 
backwardly directed processes of the outer ends of the paroccipital 
processes. Making allowance for this modification, its relations to 
the surrounding bones are essentially those held by the corre- 
sponding bone of Mormosaurus, so far, at any rate, as can be 
judged from published descriptions. In the palate also there are 
distinct resemblances in the development of the high vertical 
flanges from the pterygoids and prevomers. The two types are, 
however, specialised in directly opposite directions, one with a 
short face for a herbivorous diet, the other with the elongated 
gape which is necessary to a carnivorous animal with the charac- 
teristic reptilian habit of grasping its prey after a single snap. 
The occiput of Hdaphosaurus, as figured by Williston and Case, 
is extremely like that of Pnigalion, and, if we may judge by 
Varanosaurus, was essentially similar in structure. 
Finally, as Broom has pointed out, the face of Dimetrodon is 
structurally very similar to that of such a Deinocephalian as 
Mormosaurus—it is perhaps even more similar to that of Deutero- 
SaUrUs. 
On the whole, the Deinocephalian skull resembles that of the 
Pelycosaur more closely than it does any other South African 
Therapsid. | 
As Broom has already shown, the lower jaw of a Tapino- 
cephaloid, except for the differences due to the oppositely 
specialised dentitions, is structurally very similar to that of 
Dimetrodon. It differs, however, in the more greatly developed 
Therapsid notch. 
Broom has already shown the very striking resemblance between 
the Pelycosaur vertebrae (particularly those of Dimetrodon) and 
those of Tapinocephaloids. 
In the rest of the post-cranial skeleton, however, there is very 
much more similarity between the Tapinocephaloids and the 
other South African Therapsids than between that group and 
the earlier Pelycosaurs. 
In the cartilaginous shoulder-girdle the precoracoid is entirely 
