1910.] Principlei^ of Mammalian Phylogeny. 100 



(p. 400), while the feet of AnopJotherhmi show that it is an aberrant Artio- 

 dactyl. 



Foot-struciure also subject to titc deceptive effects of parallel evolution. — 

 Reduction in the number of digits, pursued independently in different 

 genetic lines, brings about some curiously close parallelisms. 



(a) The hind feet of Theosodon, a Patagonian Litoptern, are tridactyl, 

 and analogous in certain respects to those of the Perissodactyl genus Colodon 

 of the North American Oligocene; while the monodactyl feet of Thoatherium, 

 another Litoptern, are remarkably horse-like. These resemblances have 

 caused Ameghino {e. cj., 1904, pp. 518-521) to place the Litopterns in the 

 order Perissodactyla. But the detailed relations of the carpals and of the 

 tarsals in the Litopterns suggest that they have no close relationship with the 

 Perissodactyls ; and this conclusion is supported by much other evidence 

 (p. 379). 



(6) Both the manus and pes of Nesodon, a Santa Cruz Toxodont, are 

 similar to those of a Rhinoceros (Gaudry, 1908, p. 9, fig. 34) and this general 

 similarity extends also to the hind feet (Gaudry, /. c, p. 11, figs. 39-41); 

 while the general characters of the skvdl and cheek teeth also recall those 

 of the Rhinocerotoid Metamynodon. These resemblances seem to have 

 been regarded by Lydekker (1896, pp. 83, 84) as indicating a genetic relation- 

 ship between Rhinoceroses and Toxodonts. But notwithstanding these 

 resemblances the two groups differ in so many seemingly "non-adaptive," 

 or palseotelic, characters that the relationship between them is probably 

 similar to that obtaining between Litopterns and Hippoids ; namely, in both 

 cases we seem to be dealing merely with the terminal members of adaptively 

 parallel and partly convergent series, which are probably related only very 

 indirectly through descent from different families of the Condylarthra. 



(c) The manus of the Perissodactyl genus Titanotherium, with its four 

 somewhat spreading digits and broad carpus is superficially similar to that 

 of the Artiodactyl Hippopotamus; but here again the wide ordinal separa- 

 tion of the two forms is faithfully revealed in the detailed characters of the 

 carpus, which in Titanotherium conforms to the Perissodactyl, in Hippo- 

 potamus to the Artiodactyl, type. 



Effect of close kinship upon parallel evolution. — Where there is a close 

 relationship between ancestors the results of parallel evolution in descen- 

 dants become still more striking. 



(a) The similarity between the manus of Titanotherium and that of 

 Metamynodon, an Oligocene Rhinocerotoid, is even greater than the similar- 

 ity between the manus of Titanotherium and that of Palwosyops, its own 

 collateral ancestor. In this case even the detailed relations of the carpals 

 (which served to separate Titanotherium from the Artiodactyl THppopotamus) 



