1910.] The Artiodactyla: Achcenodon. 403 



tive Tvlo])0(I of the Upper Eocene (figured by Wortinan, 1898, p. 11], 

 fig. 7) the premaxillary is elongate, the small incisors are arranged in a 

 longitudinal series, the lower incisors are procumbent and the canines are 

 small and incisiform. The skull also has weak zygomata, an elongate 

 muzzle, a post-orbital ring, a delicate mandible, etc., all features which are 

 seen also among the Insectivora. But here the analogy ends for both the 

 upper molars and the limbs of Poebrotherium are fully Artiodactylous. Even 

 stronger general resemblances to the Lemuroid-Insectivore type are exliib- 

 ited in the skull of the small Dichobime leporina as figured by Stehlin (1900, 

 fig. Ixxi). I'here are also marked analogies in the dentition and skull be- 

 tween the earliest Artiodactyls and the Upper Eocene Lemuroids. But again 

 the objection against deriving the Artiodactyls from the Primates is the radi- 

 cal difference in the character of the tarsus. In the Lower Eocene Artio- 

 dactyl genus Trigonokstes the astragalus is of the typical Artiodactyl type, 

 while on the other hand in the Middle Eocene Primate Notharctus (now 

 known from excellent material in the American Museum) all the limb bones 

 including the astragalus are of characteristically Primate type and shoAv 

 no significant resemblances to the Artiodactyla. 



These comparisons of Artiodactyls with Insectivores and Primates are 

 also weakened by the entirely different facies of the skull of Achoenodon 

 robustus a Middle Eocene Suilline described by Osborn (1883, pp. 23-35, pi. 

 vi). 



This important skull is full of Creodont resemblances. The stout 

 incisors are arranged in a transverse series, the canines are very stout and of 

 the bluntly caniniform long- rooted type; the lower premolars are large 

 and conical, the upper fourth premolar bicuspid; the zygomata are very 

 heavy and there is a high sagittal crest, the glenoid facets for the mandibles 

 are set well back as in the Mesonychids and are equally prominent and 

 transversely extended; the nasals spread proximally as in Mesonychidse 

 (Matthew) the lachrymal is large, the postorbital process of the frontal is 

 stout, the postorbital constriction marked. It is not stated whether the 

 alisphenoid canal is lacking as in recent Artiodactyla, but if so this fact 

 would not constitute an important objection to the derivation of the Artio- 

 dactyla from the Creodonta, since in certain Creodonts (p. 306) the alisphen- 

 oid canal is known to be variable (p. 430). The mandible also is very heavy 

 but differs from the Mesonychid type in the high position of the condyle and 

 in the broad angle. The quadritubercular upper molars are a bunodont 

 modification of the tritubercular type; the lower molars show no characters 

 that are inconsistent with derivation from a tuberculosectorial type much 

 like that of Triisodon. These observations are in line with Matthew's 

 suggestion (1909, p. 485) that the Artiodactyla have been derived from 

 unknown Creodonts allied to the Mesonychidse. 



