114 THE MAMMALIA OF THE DEEP RIVER BEDS. - 



upper teeth, which are so characteristic of Anchitherium and which have, perhaps, as 

 much as any other feature, led to the view of its derivation from Palceotherium. 

 (3) Anchitherium has now been found in America in strata which are probably 

 as old as, if not older than, those of Sansan and Steinheim, and there is thus no 

 geographical or geological objection to assuming that the two species of this genus 

 have both been derived from some species of Miohippus as yet not identified. We 

 may hope to learn much upon this subject when the various species of the John Day 

 genus have been more fully described and their variations in tooth and foot structure 

 correlated. 



So far as the relations of Anchitherium to later genera of the equine series are 

 concerned, I think the evidence now available strongly confirms Sehlosser's view, 

 already quoted, that this genus is an abortive side branch of the main phylum, which 

 died out, leaving no successors behind it. (1) The teeth of Anchitherium are in 

 some important respects less characteristically horse-like than those of the more 

 ancient genera, as may be seen from the following facts, (a) In the lower molars 

 and premolars of the Pquidai no feature is more characteristic than the two pairs of 

 internal cusps (a a 1 , b b 1 ), which originated at a very early period and steadily 

 increase in size and importance until they reach their maximum development in the 

 modern forms. Now, in Anchitherium, these elements are reduced in prominence ; 

 on some teeth they are missing, and, as in the case of disappearing structures gen- 

 erally, they are very variable. Thus, in the case of the large A. aurelianense from 

 Steinheim, figured by Fraas, the anterior pillar is absent on in. 3, present on the 

 others ; the posterior pillar is much reduced on m. i, absent on m. 2. Kowalevsky's 

 specimens are too advanced in wear to show these features, but in a lower jaw from 

 Sansan containing m. 2 and m. 3, which the Princeton museum owes to the kindness 

 of Prof. Gaudry, the anterior pillar is not found in the last or the penultimate molar, 

 and iu. 2 has no posterior pillar. The same description will apply to A. equinum. In 

 this connection it is important to note that in the milk molars of A. aurelianense 

 these pillars are much more conspicuously developed than in the permanent teeth 

 (see Kowalevsky, PI. Ill, Fig. 58). (b) In the upper cheek-teeth the posterior con- 

 ule retains its importance throughout the equine series, and yet in AncJiitherium it is 

 so much reduced as to be hardly recognizable, (c) The external cusps of the supe- 

 rior molars and premolars in MesoJiippus, Miohippus, Desmatippus, Protohippus and 

 PJquus are but slightly concave and do not project inward in any marked degree, 

 while in Anchitherium these cusps are more decidedly concave than in any of the 

 earlier or later genera, and their apices project inward in a way that recalls the 

 molars of the titanotheres. All of these features tend to indicate that the dentition 



