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



Conclusion. 



The ancestral Perissodactyl as thus conceived resembled the more 

 primitive Condylarths such as Euprotogonia in many features. It had the 

 same dental formula, the same kind and arrangement of incisors, canines 

 and premolars, its sexitubercular upper molars were essentially similar in 

 the relative positions of the cusps to those of Euprotogonia, and differed 

 only in their somewhat better adaptations to a herbivorous diet. The 

 architecture of the skull, the arrangement of the cranial foramina and the 

 dorso-lumbar vertebral formula were about the same as in PJicnacodus, 

 while the peculiarities of its scapula and pelvis (p. 391) are not of a radical 

 nature. The manus and pes, in spite of the differences noted below, re- 

 tained clear traces of derivation from a pentadactvl incipiently mesaxonic 

 type, and in so far are similar to those of Condylarths. 



The known Lower Eocene Perissodactyls had, as stated above, a larger 

 brain than Phenacodus but reasons have been given (p. 389) for regarding 

 this as a not insuperable objection to the supposed Condylarth affinities of 

 the stem Perissodactyl. 



Most of the differences between the stem Perissodactyl and the known 

 Phenacodonts may be regarded as more advanced adaptations to cursorial 

 habits. Under this heading may be cited the changes in the scapula and 

 ilium, the strengthening of the sacrum and, in the humerus, the total loss 

 of the entepicondylar foramen, the reduction of the entocondyle, the flat- 

 tening of the capitellum and widening of the radio-ulnar trochlea, and the 

 early reduction of the digital formula to 4-3. The differences between 

 the serial manus of Phenacodus and the interlocking or displaced manus 

 of Perissodactyls are probably the result of divergent adaptations from a 

 more central type (p. 451). The com])lex form and function of the mag- 

 num in the earliest Perissodactyls demand a much simpler form in their 

 remote predecessors, and this condition is retained in Euprotogonia. The 

 difference in the astragalus in the two types, while very radical, is doubtless 

 also an expression of improved cursorial powers in the Perissodactyls. In 

 the Condylarths and earliest Amblypods the diverging distal end of the 

 astragalus (doubtless a Creodont and remotely arboreal inheritance) exposes 

 the head of the astragalus in the rear view and makes a relatively movable 

 but unstable joint between the astragalus and calcaneum. By a develop- 

 ment of the external (calcaneal) side of the neck or by a twisting of the 

 trochlea, the head of the astragalus was finally brought downward and out- 

 ward into contact with the distal end of the calcaneum. The "distal facet" 

 thus established, together with the cuboid facet thus made possible, greatly 



