An Almost Perfect Skull of a New Primitive TJieriodont, 203 



suchus. Unfortunately, the temporal region of Lycosaurus is not 

 well known, but it seems probable that Lycosuchus is somewhat 

 more primitive than Lycosaurus. The squamosal in Lycosuchus 

 closely resembles that in Ictidosuchus, but the two genera differ 

 markedly in the structure of the postorbital arch. In Ictidosuchus 

 the arch is apparently entirely formed by the postfrontal, while in 

 Lycosuchus the ascending process of the jugal forms at least half of 

 the arch. In both Ictidosuchus and apparently Lycosaurus the 

 dentary forms a larger proportion of the lower jaw than in Lyco- 

 suchus. It is very interesting to observe that in those points in 

 which Lycosuchus differs from Ictidosuchus it makes a nearer 

 approach to the Anomodonts. 



In Udenodon the postorbital arch is formed almost exactly as in 

 Lycosuchus. The Anomodont squamosal is more developed and 

 more highly specialised, but it agrees with that in Lycosuchus in 

 having an anterior zygomatic process, an internal portion articulating 

 with the parietal and a descending portion which supports the quad- 

 rate. The lower jaw in Lycosuchus also agrees closely with that in 

 Udenodon, the only essential difference being that in the former there 

 is a large coronoid process, while in the toothless Anomodont the 

 process is rudimentary. In both the angular forms the greater part 

 pi the posterior half of the jaw. 



The whole structure of the skull in Lycosuchus is so essentially 

 similar to that in the Anomodonts that there can be but little 

 doubt the Anomodonts are descended from Theriodont ancestors. 

 Lycosuchus, though much nearer the Anomodont ancestor than 

 Ictidosuchus, has the posterior part of the squamosal too specialised 

 to have itself been the ancestor ; but it is probably allied to the 

 ancestral form. Pristerodon and another small Endothiodont [Pro- 

 dicynodon], which I am elsewhere describing, almost completely 

 bridge the gap between the Theriodont and Dicynodon. 



While Lycosuchus is probably allied to the Theriodont, from 

 which the Anomodonts are sprung, it is probably also allied to the 

 Theriodont, which was the ancestor of the Monotremes and higher 

 mammals. Cynognathus and Gomphognathus in many respects 

 make very near approaches to the mammals — as in the specialisation 

 of the teeth, in the great reduction of the quadrate, in having two 

 occipital condyles, and in having a well-developed secondary palate. 

 In some respects, however, both these genera are too specialised to 

 have been the ancestor of the Monotremes. x\nd it is probable that 

 the first mammals sprang from a Theriodont ancestor somewhat in- 

 termediate between Lycosuchus and Cynognathus, having the simple 

 type of skull of the former, but with a full set of simple molars. 



