0. C. Marsh — Recent Polydactyle Horses. 351 



observed in the modern horse has its counterpart in some 

 extinct species now known, and it is not at all improbable that 

 future discoveries will bring to light examples corresponding 

 to the present apparent exceptions. 



All the examples of polydactylism in the horse which the 

 writer has had opportunity to examine critically are best 

 explained by atavism, and many of them admit no other 

 solution. Taken together with the fact now known of their 

 great frequency, they clearly indicate the descent of the horse 

 from comparatively recent polydactyle ancestry. 



The writer has traced back the genealogy of the horse 

 through various stages, to a form, Eohippus, in which all five 

 toes were represented, and has likewise given an explanation 

 of the change which in succeeding forms has reduced the num- 

 ber to one functional digit, as in the existing horse. It is in 

 these extinct species that the true solution of the problems 

 relating to the extra digits of the modern horse is to be sought. 

 In the various papers cited at the beginning of the present 

 article will be found the main facts relating to the ancestry of 

 the horse as made out by the writer from American forms. 



The oldest ancestor of the horse, as yet undiscovered, 

 undoubtedly had five toes on each foot, and probably was 

 not larger than a rabbit, perhaps much smaller. This hypo- 

 thetical predecessor of the horse can now be predicated with 

 certainty from what is known of the early hoofed mammals. 

 It may be called Hippqps, and its remains will be found at the 

 base of the Tertiary, or more likely in the latest Cretaceous. 

 A still more primitive ancestral form, and next older in the 

 series of Ungulates, will show the more generalized characters 

 of the group called by the writer Holodactyla (Dinocerata, p. 

 172), from which both the Perissodactyles and the Artiodac- 

 tyles branched off before the equine line became distinct. 



It is impossible to say from what generalized form the horse 

 line first separated, but at present the probabilities point to a 

 genus allied to the Eocene ifyracothermm, Owen (1839), as the 

 stem. In the latter genus, the molar teeth are of the bunodont 

 type, the tubercles being conical and distinct. Similar teeth 

 are characteristic of suilline mammals, and from the teeth 

 alone, the two groups could not be distinguished, but it is now 

 probable that the latter appeared as a distinct group later than 

 the equine mammals, and that their oldest representatives were 

 very diminutive in size. 



The American representative of Hyracotherium was first 

 described by the writer under the name Iielohyus* and 

 another form, nearly allied, as Thinotherium, all three being 

 *This Journal, vol. iv, pp. 207, 208, September, 1872. 



