22 Transactions of the Boyal Society of South Africa. 
with a fragment of the maxilla, and it seems fairly certain that the skull is 
that of Titanosuchus ferox or a closely allied form. Though very imperfect, 
the fragment is sufficient to give a fair idea of the skull. The parietal 
region is very broad and the temporal fossae small. The pineal foramen 
large, and surrounded by a marked thickening of the bone. The occiput 
slopes downwards and backwards. The orbits are very far apart, and 
over each is a huge rounded boss of bone, which completely overhangs it. 
The snout is relatively very long, and on the upper nasal region is a large 
median thickening as if covered by a horny plate. An attempted restora- 
tion is made in the figure. The temporal arches are evidently weaker 
than in Tapinocephalus, and the quadrates are not carried much forward, 
being probably situated more as in Therocephalians. Few of the sutures 
are distinct, but it is manifest that the large supraorbital bosses are post- 
frontal, and that the inner walls of the temporal fossae are postorbital. 
The low median boss is mainly frontal but partly nasal. 
The fragment of the skull as preserved measures 440 mm., and 
probably the length of the perfect skull would be about 600 mm. The 
width across the supraorbital bosses is 310 mm. 
Theriognathus. 
In the British Museum " Catalogue of the South African Fossil Eep- 
tiles," published by Owen in 1876, he describes and figures a fairly large 
but very imperfect skull under the name Theriognathus microps. Owen 
evidently believed it to be an Anomodont, and speaks of it as " seem- 
ingly edentulous," but as evidently distinct from Oudenodon. Lydekker 
came to the conclusion that the skull really belonged to Endothiodon 
uniseries, and in his Catalogue of 1890 says : " The teeth are not shown, 
but from a comparison with the next specimen [Endothiodon uniseries'] 
the generic position of the specimen is quite evident. This is especially 
shown by the great width of the interorbital region and the natural cast 
of the right orbit, which evidently had a bony roof identical with that of 
the next specimen." 
The specimen is in such a very bad state of preservation that it is 
impossible to determine its affinities with certainty, but there is, in my 
opinion, little doubt that it is not in any way nearly related to either 
Oudenodon or Endothiodon. No doubt the interorbital region was broad, 
as pointed out by Lydekker, but so it is in all Therocephalians. The 
lower jaw of Endothiodon uniseries is known ; in fact, the mandible of the 
type individual is in the South African Museum, and it in no way 
resembles the jaw of Thei-iognathus. This latter jaw so far as preserved 
is typically Therocephalian. There is a large coronoid process, a develop- 
ment unknown in any Anomodont, while the angular and articular bones 
