120 THE MAMMALIA OF THE DEEP RIVER BEDS. 



River up into the John Day. Hence the exceptional value of those two formations 

 for the study of phylogenetic problems. 



Admitting the full weight of the difficulties above mentioned, some general 

 principles stand out clearty from the confusion. The clearness may be deceptive, but 

 that must be determined by wider investigation. In comparing the series of horse- 

 like genera, we are struck at once by two facts : first, the steady advance of differen- 

 tiation in the main, and, secondly, the continually alternating progress and regres- 

 sion in certain minor details. Every genus is in some respect or other, often very 

 trivial, less modernized than its predecessor. For example, we have seen that, in 

 certain details of the carpus and tarsus, Mesoluppus is more advanced than Mioliip- 

 pus, Mioliippus more than Desmatippus, and the latter again than Protoliippus. It 

 is worthy of note that not always the same structures are affected by the retrogres- 

 sion in the various genera; we do not find continual advance in some respects bal- 

 anced by continual retreat in others. On the contrary, each genus would appear to 

 recover part, at least, of the ground lost by its predecessor, but to lose in some other 

 direction itself. Part of this appearance of alternation is no doubt due to individual 

 variation, for a character is often long subject to great variation before becoming 

 finally established, and, as already stated, there is in each successive genus a tend- 

 ency to run through similar cycles of variation. 'No doubt, also, allowance must be 

 made for the difficulty of constructing a phyletic series of species, so that, if these 

 appearances were confined to the horses, no great weight could be attached to them, 

 but every plrylum which I have been able to carefully examine displays the same 

 phenomena. In view of the very close connection between the John Day and White 

 River beds, there can be very little doubt that Miohippits has descended from one or 

 more species of Mesoliippus ; but if so, and unless the ancestral species of the White 

 River genus has not yet been discovered, the principle must be admitted. The sulci 

 which invade the articular surfaces of the tarsal bones, and are so conspicuous in the 

 recent horses, have already commenced in Mesoliippus. In that genus they are 

 variable, but, so far as I have been able to observe, they are more constantly present 

 and larger than in Mioliippus, its successor. !N~one of the known species of the latter 

 genus fulfills all the conditions which are required of a form ancestral to Ancliitherium, 

 Some do so in one respect, others in another, none in all. If this alternation in 

 minute details is to be rigidly excluded, I know of no species among fossil mammals 

 which can claim to be ancestral to another, and unless, therefore, we are prepared to 

 admit that no two species which have been found in successive formations stand, in a 

 direct relation of descent to one another, there would seem to be no escape from the 

 conclusion that, in some cases at least, the general differentiation of a line may be 



