450 Bulletin A77icricau Museim2 of Natural History. [Xol. XXVII, 



Peri.f.sodacti/Ia. According to Cope's original view, the characteristic 

 arrangement of the carpals in Perissotlactyls, in which the scaphoid in front 

 view rests largely on the magnum, the lunar on the unciform, was derived 

 from the serial condition of Phenacodus and was produced by the shifting 

 or displacement of the proximal row upon the distal row. Osborn, in his 

 study on the ' Evolution of the Ungulate Foot' (1889, pp. 5(30-569) took the 

 more correct view that this "displacement" was brought about by meta- 

 trophic growth in response to certain mechanical requirements, the theory 

 of which he developed at length. But both Cope aufl Osborn took the serial 

 carpvis of Phenacoduft as the starting j)oint. From what has already been 

 said, however, it will be clear that this assum})tion is not in accordance with 

 the present evidence. A reexamination of the carpus of the known Eocene 

 Perissodactyla and of the trend of evolution in the various families leads to 

 the following conclusions : 



(1) So far as ])rescnt evidence indicates, in the ancestral Perissodactyl 

 the manus M-as narrow, more like that in Ileptodon (Fig. 23), and not broad 

 like that in Pala'osijop.s. Reasons: 



(a) Small light-bodied primitive forms often have a narrow manus, 

 especially cursorial forms (e. g., cursorial Polyprotodonts, Creodonts, Fissi- 

 pedes and Rodents (Agutis)). 



(h) The manus of known Lower Eocene Perissodactyls, representing 

 the principal families, are all relatively narrow. 



(c) Several families exhibit an apparent progression from narrow footed 

 cursorial to broad footed slow moving forms. 



Compare Lamhdotherium and Eotitanops with Palceosyops and Titano- 

 therium. 



Compare Heptodon with Lophiodon. 



" PaloploiJien'iim " Pala'otherium. 

 " Hi/rarhi/iis " Metamynodon and TeU'oceras. 



(2) In the ancestral Perissodactyl the middle digit was already in the 

 ascendant but perha]:)s to a less degree even than in Heptodon; for 



(a) The manus is already mesaxonic in Meniscotherinm and Panto- 

 lambda, and even in the Artiodactyls digit III was primitively longer than 

 digit IV (cf. p. 405). 



(h) Digit III is somewhat longer than its fellows in the majority of the 

 Unguiculates. 



(3) In the ancestral Perissodactyl the magnum in front view was small. 

 This is supported by: 



(a) The progressive widening of the magnum in Titanotheres, Horses, 

 Amynodonts, and Lophiodonts. 



(h) The small size of the magnum in the Lower Eocene Heptodon, 

 Lamhdotherium, and Eotitanops. 



