in North America of rare Extinct Vei'tebrates. 217 



belong to the same genus as that founded on those teetli in 

 1846. Other parts of the skeleton included cervical, dorsal, 

 and caudal vertebrae, and, what is still more important and 

 suggestive, the bones of both fore and hind limbs, permitting a 

 restoration of the feet, as in the figures copied in PI. XI. 

 figs. 1 & 2, from Marsh's memoir *. 



In the Section vii. of the undercited workf, containing an 

 attempt to develop Cuvier's idea of the classification of Pachy- 

 derms by the number of their toes|, I referred the genus 

 Coryphodon to the Perissodactyle series. The first confirmation 

 from the limb-bones supplied by the North- American fossils 

 is derived from the femur. I had noted that " the trochan- 

 ters of the femur are two in the Artiodactyles, but three in the 

 Perissodactyles " § ; but at that date I could not apply this pro- 

 position to the genus in question. Prof. Marsh writes (1877), 

 " the femur of Coryphodon is of tlie perissodactyl type, and has 

 a distinct third trochanter " ||. 



In my work above cited, when treating of digital characters, 

 I referred to Coryphodon, together with Lophiodon, Palceothe- 

 rium, Acerotherium, and Ilippotherium, " as links filling up 

 the now broken series of perissodactyle or odd-toed Ungulates 

 represented by the existing genera Rhinoceros, Hyrax, Tapirus, 

 Equus " II. 



But the importance of the link supplied by Coryphodon could 

 not have been divined before Prof. Marsh's discovery. This 

 genus, older in time, earlier in date, than Palceotherium or 

 Lophiodon, retained the digits which they had lost. They are 

 present in what may be termed the mammalian typical num- 

 ber, 5, on both fore (PI. XI. fig. 1) and hind (ib. fig. 2) feet. 



A form of hoofed limb may yet be discovered (and I should 

 expect it in the predecessors of the Hyracotherioids) of a more 

 generalized type than that in Coryphodon — one, viz., in which 

 the perisso- or the artio-dactyle characters will be less distinc- 

 tively marked. 



It is not that the greater robustness of the third raetapodial 

 indicates the tendency to perissodactylism ; for such is the 

 character of that bone in the artiodactyle Hyopotamus. The 

 third metatarsal (second through loss of the first) differs both 

 by breadth and length from the fourth metatarsal, to which it 



: * Tom, cit. pi. iv. 



t * Contribution8 to the History of the British Fossil Mammals,' 4to, 

 1848, p. 30. 



X See ' Ossemens Fossiles,' torn. iii. ed. 1822. 4to, p. 72. 



§ Contributions &c. p. 59. 



II Loc. cit. p. 83. 



^ Loc. cit. p. 55. 



Ann. ct Mag. N. Hist. Ser. 5. Vol. ii. 15 



