E. F. shorn — Mammalia m Worth America. 383 



vative preservation of its ancestral three toes. For these 

 reasons the teeth and feet, owing to the frequent parallels of 

 adaptation, may wholly mislead us if taken alone ; while, if 

 considered together, they give us a sure key ; for no case of 

 exact parallelism in both teeth and feet between two unrelated 

 types has yet been found, or is likely to be. This, I believe, 

 is the one lesson of later work which reverts to older methods ; 

 we should not base either classification or descent upon the 

 teeth or feet alone. Every additional character diminishes the 

 chances of error. 



The evolution of foot structure has now become a science 

 and advances have been made in the principles of progression 

 from the plantigrade, pentadactyl serial types to the unguli- 

 grade, monodactyl alternating types which are of the greatest 

 importance in classification and phylogeny. It is surprising 

 how little attention was given to ungulate foot structure 

 between the time of Cuvier and Kowalevsky. Owen's gener- 

 alization as to the Artiodactyl and Perissodactyl pes formed 

 the one bright exception. Kowalevsky first directed attention 

 to the importance of the more median metacarpals displacing 

 or spreading to gain a stronger foothold upon the carpals as 

 the lateral toes disappeared. Ryder also worked out the laws 

 of reduction. The discovery of Phenacodus led Cope to the 

 final generalization that the primitive ungulates were not only 

 plantigrade but had some of their carpals and tarsals in vertical 

 rows like bricks clumsily set with unstruck joints — and that 

 one great law of evolution towards digitigradism was to pro- 

 duce diplarthry or alternating joints. As he found this alter- 

 nation differed both in degree and kind in different groups, 

 he revived the comprehensive " Ungulata," of Linnaeus and 

 divided all hoofed mammals exclusively upon their foot struc- 

 ture into five great orders. 



Rutimeyer and myself have shown that however successful 

 and convenient this system appears, Cope's lines of division 

 ignore the fundamentally different modes of evolution of the 

 fore and hind feet ; an animal may be a taxeopod in front and 

 a diplarth behind or vice versa. Numerous exceptions to 

 Cope's definitions are also found. The discovery of the aber- 

 rant ungulate foot types of South America further invalidates 

 Cope's system and sustains the principle that to be permanent 

 classification must be based upon at least two entirely diverse 

 sets of characters. This does not diminish the importance of 

 the primitive taxeopod plantigrade type as one great key to 

 the still unsolved problems of the primary relationships of the 

 Condylarthra, Hyracoidea, x\mblypoda, Proboscidia, Toxo- 

 dontia, Litopterna, Artiodactyla and Perissodactyla. All these 

 orders still stand apart in the dim past like so many mile-posts. 



