76 Mr. D. M. S. Watson on the 



in size of the temporal fossa the parietal and jugal may enter 

 into its border, but the postfrontal never does so. This last 

 point is of interest in connection with the fact that in several 

 Nothosaurians the postfrontal, although present, takes no 

 part in the formation of the temporal fossa. 



This ne%Y view of the homology of the Therapsid temporal 

 fossa raises no difficulties and allows of the occasional pro- 

 duction of incipient upper temporal fossae in Dimetrodon 

 and lower temporal fossae as in the type skull of Cynognatlms 

 crateronotus. 



The Gorgonopsidffi in the characters of their parietal region 

 agree very closely indeed with the Anomodontia ; in many 

 other features of the skull they also present resemblances to 

 them, and it becomes interesting to speculate on the possible 

 relation between the two types. Typical Anomodonts of 

 small size are known from the Pariasaurus-zone, and Gor- 

 gonopsids first appear, so far as is at present known, in the 

 succeeding Endothiodon-zoue, so that direct derivation is not 

 possible. It seems to me most probable that they agree so 

 closely owing partly to actual descent of primitive features to 

 them from ancestral Tiierapsids and partly to the direction 

 imposed on their evolutionary change by latent homoplasty *. 

 They both agree closely in the constitution of their cranial 

 roof with the Deinocephalia, although the preparietal is not 

 known to occur in that group. The remarkable general 

 resemblance between the apparent method of development of 

 the secoridary palate in Crocodiles, where in Metriorhynchus 

 there is a large median vomer, and that of Cynodonts as 

 outlined above suggests that tiie development of the secondary 

 palate of Tiierapsids may have taken place independently on 

 several lines of descent. 



It seems to me very probable that Bauria is the terminal 

 type of a completely different but parallel line of descent to 

 Cynognathus. Bauria with a secondary palate retains large 

 suborbital vacuities, as shown in a specimen collected by 

 myself, which is exceedingly like Bauria cynops, but perhaps 

 belongs to an allied species. The family, as shown by 

 Microgomphodon oligocynus, has also a large interpterygoid 

 vacuity f. The Permian type Arctognatlms curvimola {cf. 



* I owe this convenient phrase to Baron Fr. Nopsca, Jun. 



t Coiuparisou of the anterior part of the dentition of Microgomphodon, 

 Seelev (Phil. Trans, vol. 186 B. pi. i. tig. 4), with that oi Bauria, Broom 

 (Proc. Zool. Soc. 1911, p. 896, fig. 168), will render the close relation of 

 the types certain. 



