KHO.] CarjniH of Perissodactyl.s. 451 



(r) Th(> .small size of the nia<i:nuiii in Kiiprofogonia, Pantolmnbda, 

 Creodonts and many lower Unguienlate orders. 



(4) In the ancestral Perissodaetyl the lunar in back view rest(Ml on 

 both magnum and unciform {of. Fig. 25). This condition is jjreserved in all 

 Perissodactyls and is a direct inheritance from the unguiculate orders. 



(5) In the ancestral Perissodaetyl tiie proximal row of carpals were 

 vertically dee])er than the distal row, for: 



(a) This is so in the Lower Eocene forms. 



(h) It seems to be a primitive character in Creodonts and other Ungui- 

 culates. 



(6) Even the oldest known Perissodactyls had already attained the 

 ordinal characters in the astragalus and in the reduced number of digits, 

 and since the manus had already lost digit I, it is probable that the carpus 

 also had advanced considerably in the Perissodaetyl direction. 



(7) If the carpus had advanced considerably in the Perissodaetyl 

 direction, then in the light of the foregoing considerations, in its earlier state 

 it must have approached the unguiculates in the following characters, in 

 comparison with those of Heptodon (Fig. 23, p. 393): 



(a) ]\Iagnum small in front view (about as in Heptodon). 



(b) Smaller lunar-unciform contact. 



(c) Smaller scaj^ho-centrale-magnum contact. 



(d) Larger trapezium and digit I. 



(e) Relatively smaller digit III, with narrower proximal abutment on 

 the unciform. 



(/) Unciform more horizontal. 



(g) Carpal facets less oblique. 



In other words, the carpus of the ancestral Perissodaetyl was less com- 

 pletely interlocking, and somewhat more "serial," than that of Heptodon 

 which appears to the writer to be on the whole the most primitive Perisso- 

 daetyl manus so far discovered.^ 



(8) From this point of view Cope and Osborn were justified in regard- 

 ing the carpus of later Perissodactyls as "displaced" only in so far as the 

 scapho-centrale-magnum and lunar-unciform contacts became secondarily 

 widened, but it is very probably incorrect to state that in the ancestral 

 Perissodaetyl the "scaphoid rested solely on the trapezoid and the lunar on 

 the magnum." 



Ancylopoda. The carpus of Moropus and ChaJicotherium is charac- 

 terized by: (a) the extreme convexity of the proximal facets, (b) the back- 

 ward extension of the superior and inferior facets on the magnum, so that 



1 In Eohippus the scapho-centrale-magnum contact is very broad, the magnum is broad 

 and digit III is very broad. 



