1907.] OP MAMMAL-LIKE REPTILES. 1049 



The early orders of the Diapsida are less fully known than 

 those of the Synapsida. The Pelycosauria are, thanks to the 

 labours of Cope, Baiir, and Case, now fairly well known, though 

 there are still a number of serious blanks in our knowledge. The 

 digital formula is not certainly known and more definite know- 

 ledge is required of the structure of the tarsus. Even in the 

 skull there is still a little doubt about the structure of the 

 posterior temporal region. Most authorities, however, seem 

 agreed in placing the order in the Diapsida. 



The Procolophonia are much better known, nearly every detail 

 in the osteology of Procolophon being as fully known as in recent 

 animals. In most of its characters the latter comes nearer to the 

 ■early Rhynchocephalians than to the mammal-like forms. The 

 digital formula is that of the lizards — 2, 3, 4, 5, 3 ; thei-e are well- 

 marked abdominal ribs, the vertebrae are notochordal, the pre- 

 vomers carry teeth, there is a quadrato-jugal bone, and the very 

 small coronoid process is formed by a distinct coronoid bone and not 

 by the dentary. It shows affinities, however, with the mammal- 

 like groups in having a well- developed precoracoid, but this is a 

 character which must have been possessed by early Diaptosaurians, 

 as it is met with in the Pelycosaurians, in Mesosaurus and in 

 Heleoscmrus. Most of the other important characters are common 

 to the early Diapsidan and early mammal-like forms — e. g., the 

 plate-like pelvis, the intercentra, the pro-atlas, and the columella 

 cranii. From the consideration of these points I have expressed 

 the opinion that Procoloijhon should be placed among the early 

 members of the Diapsida rather than among the Synapsida. 



"While the Pelycosauria and the Procolophonia seem to be 

 Diapsidan orders, it must nevertheless be admitted that both 

 show certain resemblances to the mammal-like groups. As 

 already mentioned. Cope believed the Pelycosaurs to be closely 

 allied to the South African " Theriodonts," and Procoloj)hon has 

 been placed among the mammal-like forms by Seeley (8), 

 Boulenger (6), and others. The most striking i-esemblance is in 

 the shoulder -girdle with its well-developed precoracoid. As, 

 however, an ossified precoracoid is found in the " Cotylosauria " 

 and even occasionally among the Stegocephalia, we should 

 naturally expect it to be met with in the early forms of both 

 Synapsidan and Diapsidan reptiles. The digital formula of the 

 Therocephalians and other Therapsida, viz. 2, 3, 3, 3, 3, marks 

 them off faii-ly distinctly from the Diapsidans with a typical 

 formula of 2, 3, 4, 5, 3. Still, when dealing with Permian reptiles, 

 we find the Diapsidan and Synapsidan types approach each other 

 so markedly that we are constantly in doubt about the position 

 of individual forms. ISTo distinction can be found in the 

 shoulder-girdle, the palates are similar and both have plate-like 

 pelves ; and it becomes manifest that the two groups have had a 

 common ancestor, or that one of the groups has sprung from 

 a member of the other. 



I have been inclined to find the common ancestor in the some- 



