■202 Transactions of the South African Philosophical Society, 



twice as broad as the largest of the incisors. It is very raarkedly 

 flattened, with a rounded anterior and a serrated posterior border. 

 The only other Theriodont skull, so far as I am aware, in which two 

 €anines are known is the Theriodont snout in the Albany Museum, 

 which Seeley has made the type of ?Cynognathus leptorhinus. In 

 it there are two canines side by side, and as in the case of the 

 Lycosuchus skull the anterior canine has both edges serrated. It is 

 impossible to say from the evidence whether one canine corresponds 

 to the mammalian permanent one and the other to the deciduous 

 canine, or whether both canines belong to the set. 



Close behind the 2nd canine, and partly resting on it, there is 

 seen on the right side the remains of the root of a medium-sized 

 tooth. This root is not in its natural position, and may be the 

 •displaced remains of the 3rd incisor which has apparently been just 

 shed. 



Some little distance behind the large canine there is seen on both 

 sides a small simple-pointed molar. It is less than half the size of 

 •one of the incisors, and shows no indication of serration. I cannot 

 find any evidence of more than the one molar on each side, and 

 there does not appear to be room on the maxillary for more than 

 three at the most. 



Affinities of Lycosuchus. 



In endeavouring to trace the affinities of Lycosuchus the chief 

 ■difficulty lies in the fact that the majority of the hitherto discovered 

 Theriodonts are known only by fragments of the snout. If, however, 

 we neglect the very imperfectly known forms w^e find that the 

 remaining more perfectly preserved Theriodonts arrange themselves 

 into four fairly well marked-groups. Of these, two are made up of 

 primitive Theriodonts, and two of highly specialised forms. Of the 

 primitive forms w^e have (1) the Lycosaurus type with simple molars, 

 powerful dentary, well-developed quadrate, and large parietal 

 foramen, and (2) the Ictidosuchus type, somewhat similar, but with 

 very slender lower jaw^ and temporal arch. The highly specialised 

 Theriodonts belong — (1) to the Cynognathus type, characterised by 

 having cusped molars, rudimentary quadrate, two occipital condyles, 

 and the parietal foramen small or absent ; or (2) to the Gom- 

 phognathas type, allied to the forrner, but with flattened grinding 

 molars. 



The affinities of Lycosuchus are undoubtedly with the primitive 

 types Lycosaurus and Ictidosuchus ; and though Lycosuchus has a 

 powerful jaw, in some respects it seems to come nearer to Ictido- 



