THE MOST PRIMITIVE KNOWN REPTILE. 299 



have this peculiar structure. I have ah^eady brigaded Livinoscelis 

 with the Captorhinids. Seymouria is so much more primitive 

 that it is remarkable to find in it a character of this kind 

 connecting it clearly with a definite series of more advanced 

 reptiles. 



It remains to be seen whether the other reptiles have lost this 

 type of fenestra ovalis, or whether, as seems more probable, their 

 ancestors never possessed it. 



It is a remarkably interesting fact that the primitive Pelyco- 

 saurs such as Varanosaurus retain primitive features found in 

 Seymouria which have been lost by the Captorhinids. Such 

 are : — 



The retention of a supratemporal lost in Labidosaurus. 



The retention of a quadrate which is U-shaped in horizontal 



section. 

 The retention of a suture between the pterygoid and squamosal 



behind the quadrate. 

 The non-ossification of a posterior coracoidal element in 



Varoops. 

 The retention of a considerably expanded lower end of the 



clavicle and a largely expanded interclavicle. 

 The retention of a posterior process on the ilium. 



The characters in the above list represent features in which 

 the Captorhinids are more advanced than the early Pelycosaurs ; 

 they are, however, all acquired by later Anomodonts and represent 

 parallel evolutionary changes which have gone on in the Capto- 

 rhinids and Pelycosaurs independently since their separation. 

 The most striking of these characters is the peculiar flattening 

 and latei-al compression of the quadrate. In principle the 

 quadrates of Caj)to7-hmus and Dimetrodon are identical, difiering 

 only in the wide articular condyle of the Cotylosaur. As this 

 type of quadrate is restricted to the members of the Superfamily 

 Sphenacodontidse, we have a very remarkable case of parallel 

 evolution, which in the original stock does not follow the main 

 line of the evolution of the derived group, but duplicates a 

 structure restricted to a single aberrant side line. 



The study of the Seymouria skeleton which fills this paper 

 shows that that reptile is far more primitive than any other, 

 presenting a strange mosaic of characters derived without change 

 from the Embolomerous Labyrinthodonts Avith common reptilian 

 characters marking distinct advance over an amphibian structure, 

 and one or two which are restricted to the mammal-like reptiles 

 and such Cotylosaurs as can be brought into connection with 

 them. 



These facts have an important bearing on reptilian classifica- 

 tion, suggesting as they do that all reptiles may be divided into 

 two groups, one composed only of the Anomodontia and the 

 Captorhinomorpha, the other of all other reptiles. 



