13^ Bulletin American Museum of Natural History. [Vol. X, 



In Procamelus, on the other hand, the distal ends of the 

 proximal phalanges, as well as those of the second row, have the 

 articular surfaces reaching well back upon the dorsal aspect, just 

 as in the modern Camel and Llama, and this we know to be 

 accompanied by the digitigrade gait and the peculiar cushioned 

 foot. There is likewise a marked change in the character of the 

 ungual phalanges, which are much reduced and flattened vertically 

 especially in Camclus. 



The conclusion is obvious, therefore, that this peculiar tylopod 

 character is secondary, and was developed from the unguligrade 

 condition. Just what led to it is impossible to say. Zittell's 

 explanation that it resulted as a consequence of the coossification 

 of the metapodials can hardly be the correct one, since the same 

 thing occurred in all of the Pecora in which the unguligrade gait 

 was retained. 



The coossification of the metapodials and the consequent 

 formation of a cannon bone took place comparatively late in the 

 history of the phylum, much later in fact than the union of the 

 ulna and radius. In Protylopus and Poeb rather ium the metapo- 

 dials were not only free but capable of considerable independent 

 movement, as indicated by the facets. The opposed surfaces of 

 these bones, moreover, are relatively smooth and less flattened 

 than in the succeeding Gomphotherium. In this latter genus the 

 bones are more closely applied to each other, the surfaces rough- 

 ened, and the articular facets, by which they join one another, 

 much reduced. The feet of Protolabis are wholly unknown. In 

 one species of Procamelus {P. gracilis), there is evidence that the 

 metapodials of the fore feet were not united until late in life, if 

 at all, while those of the hind foot were fully coossified into a 

 cannon bone. Procamelus occidentalis had the posterior metapo- 

 dials fully united and those of the fore foot only partially coossi- 

 fied, being free at their proximal ends for some distance. 



In the Pleistocene species the metapodials were fully united 

 into a cannon bone early in life, and in the modern Camelid^ bony 

 union of these elements takes place before birth, almost as soon 

 in fact as bony tissue is thrown down in the shafts. 



It will thus be seen that the coossification of the metapodials, 

 as well as that of the ulna and radius, was a gradual process, and 



