Some Krtr Primitive Thcriodonts. 155 



order ; and in a paper recently read before the South African Associa- 

 tion for the Advancement of Science (4) I proposed to form a new 

 Order— Therocephaha — ^for their reception. 



Besides the three new genera described above, which may be taken 

 as typical of the new order, there are a number of others which are 

 apparently more or less closely allied and which evidently belong to 

 the same order. 



^lurosaurus, Oivcn, is fairly closely allied to Scylacosaurus. 

 Seeley had the palate of the type specimen partly cleared of matrix, 

 but he does not seem to have realised that the type of palate is, as I 

 pointed out a few years ago (5), quite unlike that of the Theriodonts 

 proper. So far as can be made out, the palate is essentially similar 

 to that of Scylacosaurus, though there is a very much greater develop- 

 ment of small palatal teeth. In the restoration which Seeley gives 

 of ^lurosaurus (6), owing to his having mistaken an incrustation of 

 lime for bone, he places the articulation for the lower jaw below the 

 back of the orbit. Doubtless the posterior part of the skull in 

 -3^]lurosaurus is very similar to that in Ictidosuchus and Lyco- 

 suchus. 



Ictidosuchus (2), in which the skull and a number of other bones 

 of the skeleton are known, is, like iElurosaurus, a near ally of 

 Scylacosaurus. The palate in the type is so imperfect that when 

 describing the specimen I did not think it safe to say more than 

 that " there is no distinct secondary palate." When the crushed 

 fragments are viewed in the light of the fairly complete palate of 

 Scylacosaurus, it is at once apparent that the palate of Ictidosuchus 

 has been essentially similar. 



Lycosuchus (1), which is known by an almost perfect skull, is 

 most probably also a Therocephalian. Unfortunately the palate is 

 unknown, but the structure of the lower jaw, arches, and dentition 

 render it fairly certain that Lycosuchus belongs to this new order. 



Titanosuchus is doubtless closely allied to Scymnosaurus. It has, 

 as I have shown (7), a pair of large prevomers, and the position of 

 the internal nares renders it evident that this genus of Owen's 

 belongs to the Therocephalia. 



Gorgonops (8) differs from all the other Therocephalian genera in 

 having the temporal fossae roofed, but in most points the resem- 

 blances to the typical Therocephalians are so marked as to render it 

 not improbable that it should be placed in the same group. The 

 presence of the minute canine in front of the large one is almost 

 exactly similar to what we find in Ictidosaurus and Scylacosaurus. 

 The internal nares are situated near the front of the palate, and are 



12 



