MAMMAL-LIKE REPTILES, 775 
Finally, so much of the palate of Vitanosuchus as is known is 
quite unlike that of any Gorgonopsid, whilst it strikingly resembles 
in detail that of Lamiasaurus. We are, I think, therefore 
justified in regarding the two as divergent branches of the great 
order Deinocephalia. 
That the Deinocephalia are Therapsids has never been questioned 
of recent years. The whole character of the limb-bones, par- 
ticularly the humerus, is quite lke that of the corresponding 
parts of an Anomodont, and is unknown in any other group. 
The shoulder-girdle with the precoracoid excluded from the 
glenoid cavity is extremely like that of a Gorgonopsid, and, except 
for its lack of an acromion, that of Dicynodon. No similar 
arrangement is known in any other group of reptiles. In the 
more fundamental features of its skull-structure, the wide 
occipital plate with the small and greatly separated post-temporal 
fosse, the high brain-cavity, and the low position of the vestibule 
of the ear, it agrees exactly with all South African Therapsids. 
As is shown clearly by Dr. Broom’s figures of the lower jaw 
and by that of 7itanosuchus, it has the characteristic flat angular 
and notch of the Therapsids. 
The group has, however, a very special importance, because alone 
amongst the South African forms it retains a large quadrate. 
The Tapinocephaloid quadrate is relatively as large as in any 
other reptile—as large, for instance, as that of a Tortoise. The 
importance of this is that it shows that the reduction of the 
quadrate, which is so noticeable a feature of the Anomodonts 
and “‘Carnivorous” Therapsids, is not an essential feature of 
Therapsid structure, but has developed in comparatively ‘late 
times in at least two branches independently (I intend to adduce 
evidence in support of this statement in a later paper). Some 
years ago, in describing the skull of Diademodon, I pointed out 
that this reduction of the quadrate was part and parcel of a whole 
series of changes which led to the reduction of all regions of the 
skull (the basisphenoidal, basioccipital, and exoccipital regions 
in particular) which lay below the base of the brain. This 
generalisation is justified by all subsequent work on Therapsids, 
and is well illustrated by the great depth of the basicranium in 
Deinocephalia, an arrangement which may to some extent be 
paralleled in Deinosaurs. 
The exact relation of the Deinocephalia to other South African 
groups of Therapsids is not easily determined. 
The skull of Zitanosuchus bears a very considerable resem- 
blance to that of a Gorgonopsid, modified, of course, by its great 
thickness and the large bony bosses over the orbits. It is dis- 
tinctly more primitive in many features—for instance, in the less 
widely spread zygomatic arches. The skulls of the more primi- 
tive Gorgonopsids, however, show no trace of the vertical area 
of basioccipital below the condyle; and there are no features of 
the auditory region of the brain-case which in any degree suggest 
that they have been derived from a type with a Deinocephalian 
