776 MR. D. M. 8. WATSON ON 
structure. It will be remembered that Rhophalodon, which is in 
many ways extremely like a Gorgonopsid, has this characteristic 
Deinocephalian structure. So far as the lower jaw goes, the 
Gorgonopsids could be derived directly from a lightly built 
Titanosuchid. The post-cranial skeleton, so far as is known, is 
also very similar, allowing for the differences due to the great 
weight of all well-known Deinocephalians or, using as a term of 
comparison, the relatively slender bones from the Ural copper- 
mines. 
In fact, it is legitimate to assume that Deinocephalia and the 
Gorgonopsids arose from a not very distant common ancestor, but 
that they subsequently pursued quite different paths. 
Perhaps the most interesting comparison is with the Anomo- 
donts. In many features (the upturned parasphenoid and the 
structure of the ethmoid, for example) there is a distinct resem- 
blance between the two groups, which also resemble one another 
to some extent in the shortness of the brain-cavity—in some types, 
at any rate. Further, many Anomodonts have a considerable 
vertically placed area of bone below the basioccipital condyle, 
which may be compared with that in Tapinocephaloids. As this 
area does not exist in Yndothiodon, whilst it is most pronounced 
in Lystrosaurus, it is probably a secondary specialisation inde- 
pendently acquired within the group. Connected with this 
feature, however, is the development of that curious process of 
the vestibule which carries the fenestra ovalis down to the lower 
border of the skull. This curious detail is found even in 
Endothiodon, a rather primitive type, and is apparently quite a 
fundamental character of the group. 
Although the cavity of the inner ear has not been cleared in 
any Deinocephalian, there can be no doubt that a similar structure 
did occur in that group—a fact which suggests a nearer relation- 
ship between the two groups than either of them hold with the 
Gorgonopsid line. 
At the same time the Deinocephalia differ from all other 
South African Therapsids in the mode of articulation of the ribs 
in the dorsal region. 
The most interesting comparison is between the Deinocephalia 
and the Pelycosauria, using that term to cover all the “Texas ” 
Therapsids. 
If we compare a Deinocephalian with the very primitive 
Varanosaurus, we find many resemblances. We have, to begin 
with, the characteristic Therapsid characters of the occiput and 
lower jaw. In addition, we find that there are striking resem- 
blances between the quadrates of the two types in their size, 
relation to the squamosal and quadratojugal, and in the power- 
ful step for the distal end of the stapes. Another resemblance is in 
the form of the squamosal, which in both types is a simple sheet 
folded round the back of the quadrate, and not produced outwards 
as in the Anomodonts and Therocephalia. The occurrence of a 
