248 5. W. WILLISTON 



small size of the limbs and the attachment of the anterior pair 

 almost at the angle of the mandible; in the shape of the humerus, 

 very unlike that of the terrestrial types, with the planes of its 

 extremities but slightly divergent and with strong muscular rugosi- 

 ties, as in the plesiosaurs and marine turtles, situated low down; 

 in the absence of the adductor ridge on the femur; in the relatively 

 very short epipodials, a certain indication in crawling reptiles of 

 swimming habits; in the conspicuous lack of ossification at the 

 ends of the long bones; in the almost certain chondrification of 

 the mesopodials and pubes; in the short scapulae, utterly unlike 

 those of terrestrial amphibians; in the structure of the clavicular 

 girdle, so like that of Diplocaulus as to be almost indistinguishable 

 at first sight; and in the certainly bare skin. Further evidence is 

 scarcely needed. 



It may be granted that the first amphibians were aquatic ani- 

 mals, perhaps as much so as is our modern Necturus, although of 

 that I am not convinced. If then, Trimerorhachis is a generalized 

 amphibian, it naturally follows that it has retained its primitive 

 structure and habits from Devonian times; that its peculiar 

 clavicular girdle is primitive, as also the unossified mesopodials 

 and the relatively feeble ossification of the vertebrae, the very 

 small size of the pleurocentra, etc. It necessarily follows that if 

 the clavicular girdle is primitive, that of the stereospondyl amphib- 

 ians of the Upper Trias is also, for the buckler-like structure in 

 these, the last of the stegocephs, is much more pronounced. On the 

 other hand, I believe that the clavicular bones in the direct line of 

 the ancestry of the reptiles were never large and rugose, that they 

 began as small dermal elements and continued so in those amphib- 

 ians which gave origin to the reptiles. 



The occipital condyle of Trimerorhachis is remarkably different 

 from that of all other known American Permian amphibians, as 

 will be seen by reference to the figures (Fig. 5, 1, J, K), in that it is 

 single and deeply cupped, fishlike. Granted that this a primitive 

 character, and it may be, it does not necessarily disprove the 

 argument I make that Trimerorhachis was descended from terres- 

 trial forbears, that its aquatic characters are acquired, not heredi- 

 tary. Watson urges that the primitive amphibians had a closed 



