402 Bulletin American Museum of Natural History. [Vol. XXVII, 



(e. g., Leptochccrus), the lower molars were not far from the tuberculo sec- 

 torial type. Even the earliest known selenodont Artiodactyl skulls (from 

 the Upper Eocene) differ from the Lower Eocene Perissodactyl type in many 

 details (apart from the dentition) which sufjcrest a wide separation of the 

 two groups, while the primitive Artiodactyl Achoenodon presents only a very 

 superficial resemblance in its skull to the Perissodactyl Palwosyops. 



A comparison of recent material naturally reveals still wider differences 

 between the two orders both in the skeleton and in the soft parts (Weber, 

 1904, pp. 597-690). 



In the structural details of the fundus of the eye, the Perissodactyls are 

 very sharply separated from the Artiodactyls (Johnson, 1901), while as 

 regards the azygous veins Beddard states (1907, p. 219) that "....the 

 Perissodactyles have their own plan of azygosus tructure, which happens to 

 agree with that of some Cervidfe, but distinguishes the Suborder from the 

 Piss and hollow-horned Ruminants." 



The Perissodactyla, it is suggested (p. 396), were probably derived 

 from some unknown family of Condylarths in which the premolars and 

 molars resembled those in Euprotogonia, the manus and pes were becoming 

 mesaxonic, the head of the astragalus had begun to flatten and the susten- 

 tacular facet to become J shaped. There is little or no evidence that the 

 Artiodactyla were derived from such a source. From what source then did 

 they originate? 



Certain conditions of the hypothetical prototype of the Artiodactyla 

 seem to be realized in the dentition of the Basal Eocene Mioclfenidse (c/. 

 Osborn, 1907, figs. 147, 148, 152, 157). These have been thought to be 

 Condylarths and may also be allied to the supposed Insectivore Hyopsodus 

 (Matthew, 1909, p. 512). Earle (1893, pp. 377-379) has suggested that 

 Protogonodon may hold the same relationship to the Artiodactyla that 

 Phenacodus was supposed to hold to the Perissodactyla, but in the absence 

 of skeletal remains such comparisons are very indecisive and Matthew (1897, 

 p. 302) has adduced evidence to show that Protogonodon is related to Eu- 

 protogonia rather than to the Artiodactyla. 



IlemitMoeus howalevskianus, a Basal Eocene Periptychid figured by 

 Osborn (1907, p. 165), in its tritubercular molars and bicuspid premolars 

 also fulfills nearly all the desired conditions for the Artiodactyl prototype 

 (except the loss of the paraconid) and further resembles the Artiodactyla 

 in the early development of a broad contact between the astragalus and the 

 cuboid {cf. Matthew, 1897, p. 297, fig. 11); but in other features the as- 

 tragalus is very different from that of the Artiodactyls. 



There are also some suggestions of relationship between the x\rtiodactyla 

 and early Tertiary Insectivores and Primates. In Poebrotherium a primi- 



