1894.] 249 [Scott. 



was followed by Cope and by myself. Osborn and Wortniau referred 

 their genus Artionyx to the Anoylopoda. with ChaUcotherium, construct- 

 ing a new suborder, the Artionychia, for its reception. While much re- 

 mains to be learned regarding Agriochainis, some inferences from the 

 facts of structure described in the preceding pages are reasonably clear. 

 In the first place, I cannot agree with Osborn and Wortman in removing 

 this genus altogether from the Artiodactylaand assigning it to the Ancylo- ' 

 poda. Such a removal implies that all of the artiodactyl features of 

 structure have been independently acquired, and this is highly improbable. 

 I have, it is true, repeatedly insisted upon the reality and frequency of 

 parallelism in development, but it is very easy to push this doctrine to 

 unwarranted extremes. Among mammals, at least, no such extreme case 

 of this mode of evolution is known as would include the skull, dentition, 

 limbs, carpus and tarsus, and in fact everything but the phalanges. The 

 only evidence which could justify such a conclusion would be the finding 

 of a succession of species by which the independent origin of the two 

 groups could be traced out step by step. The agreement of Agnoc7i(Brus 

 with ChaUcotherium is of the slightest and most superficial character, con- 

 sisting only in the fact that both genera have claws. But the ungual 

 phalanges are of a very different pattern in the two genera, and it is surely 

 the less dangerous horn of the dilemma to conclude that this single corre- 

 spondence is due to parallelism rather than that the numerous and im- 

 portant characters in which AgriocJicerus agrees with the artiodactyls are 

 the result of such a process. 



But a difficulty arises here ; is not the distinction between hoof and claw 

 a fundamental one, established long before the artiodactyls had arisen ? 

 To answer this question with any certainty would require a much more 

 exact knowledge of the genesis of both kinds of phalanx than we at 

 present possess, but there is much reason to believe that while the ungu- 

 late and unguiculate types of mammals are radically distinct, yet the dis- 

 tinction does not rest, as the names imply, upon the character of the un- 

 guals. At all events both hoof and claw are found in closely allied 

 genera, as for example, among the rodents, creodonts and edentates. 

 These examples arc, it is true, all found among the unguiculates, but there 

 is no a priori reason for assuming that a similar diversity among the ungu- 

 lates may not occur. Tliat it is rare in the latter group is doubtless due to 

 the fact that ungulates almost uniformly employ the feet only fur purposes 

 of locomotion, and indeed it is difficult to conjecture what the function of 

 such feet as those of Agriochairus and ChaUcotherium may have been. In 

 view of all the characteristic artiodactyl structures which Agriochcerus 

 displays, one can hardly escape the conclusion that in this case the transi- 

 tion from hoof to claw has actually taken place and that this genus is the 

 culmination of a series of aberrant artiodactyls. In the second place it is 

 obvious that Leidy's separation of Agriochcerus from the Oreodontidm is 

 entirely proper. A more obscure problem is to determine the relationship 

 between the two families, and to this end it will be necessary to briefly 



