1010.] Carpus of Artiodactyla. 449 



with the theory that these orders have been derived horn different famihes 

 of the Condylarthra. The strongly interloeking- ty})es represented in 

 Homalodotherium and the Typotheres have doubtless been evolved by a 

 progressive increase in the obliquity of the facets and by the broadening 

 of the lunar-unciform and scapho-centrale-magnum contacts. In the 

 larger Toxodontia, as in the Perissodactyls, this is associated with a marked 

 tendency toward mesaxonic tridactyly. In the Litopterns the carpals have 

 flattened down and the lunarimciforni contact has been disrupted (p. 379). 

 Artiodactyla. The theory that the Artiodactyla (p. 403) have been de- 

 rived from relatives of the Mesonychid Creodonts, and the Perissodactyla 

 (p. 396) from relatives of the stem of the Condylarthra is supported in large 

 measure by the nature of the resemblances and differences between the 

 manus of Artiodactyls and that of Perissodactyls. The manus of Artio- 

 dactyls has been derived from a progressively paraxonic, functionally 

 tetradactyl type with interlocking carpus, more or less resembling the manus 

 of Dromocyon (p. 404) ; the manus of Perissodactyls has been derived from a 

 mesaxonic progressively tridactyl type more or less like that of Ewproto- 

 gonia; but in this type the carpus, while interlocking, was narrower and 

 higher than was the case in the Artiodactyla. In both orders digit III was 

 originally the largest but in the ancestral Artiodactyl its superiority over digit 

 IV was less than was the case in the ancestral Perissodactyl. The general 

 resemblance of the carpus of Ancodns, a primitive number of the Anthraco- 

 theriidffi to that of Dromocyon has already been noted above. Nor can 

 there be any doubt that the carpus of Ancodus is approximately prototypal 

 to that of all higher Artiodactyla; for it compares with those of the hippo- 

 potamus and pig on the one hand, while on the other by further emphasis 

 of its didactylism it would lead into the carpus of the larger Anthracotheres, 

 and of the Oreodontidse, Hypertragulidse, Canielidie, Cervidae, Bovidte, etc. 

 The unity of derivation of all the more modified types of carpus in the Artio- 

 dactyla and the contrast in important details with all other ordinal types of 

 carpi, becomes plain from a study of the Tertiary Artiodactyls exhibited 

 in the American Museum of Natural History and from the figures given by 

 Kowalevsky (1873), Scott (1895) and others. From all this it is evident 

 that it is misleading to say as Cope did that the "carpus of Artiodactyls is 

 of the highly displaced type." In the oldest and indeed in all Artiodactyla 

 the lunar-unciform contact is broad and the scapho-centrale process or hook 

 is in contact with the magnum. Very probably the lunar-unciform contact 

 and the scapho-centrale-magnum contact were narrower in the ancestral 

 Artiodactyls than they were in later Artiodactyls, but the interlocking 

 arrangement is no more a true "displacement" in the Artiodactyla than it is 

 in the Basal Eocene Creodonts or in Euprotogonia or Pantolambda. 



