THE MAMMALIA OF THE DEEP EIVER BEDS. 



121 



accompanied by an ebb and flow in certain minor characters, illustrating what Galton 

 has called the " regression to mediocrity." This does not imply an all-ronnd, inde- 

 terminate variation ; the changes are alternately towards and away from a certain 

 definite standard, and are sometimes repeated in one succeeding form after another, 

 while in other cases a new set of characters are affected. Indeed, if we admit the 

 possibility of parallel developments, and that, at least, is demonstrated by the fossils 

 beyond peradventnre, the possibility of alternations follows of itself. 



While there is very little in favor of the view of indeterminate variation to be 

 derived from a series of fossil mammals, yet Anchitherium does display some varia- 

 tions which appear to be of this character. Thns Filhol has noticed the instability 

 of the tubercle which sometimes appears in the entrance to the median valley of the 

 upper teeth. "II semble qu'il n'y ait aucnne tendance a, des modifications de la 

 structure des dents. Le seul fait que j'ai pu constater, et qui a une bien petite im- 

 portance, consiste dans la presence ou l'absence d'un denticule d'email qu'on trouve 

 aux dents superieures, entre les denticules internes. Sur la piece que j'ai fait repre- 

 senter on observe ce denticule sur les trois dernieres premolaires et sur la dernierne 

 molaire; sur un autre echantillon il existe sur toutes les dents, alors que sur deux 

 autres il fait absolument defaut sur les molaires vraies" (No. 13, pp. 193, 194). In 

 the specimen of A. equinum which I have described, a similar tubercle occurs in the 

 posterior valley of the first lower molar, which is doubtless of the same variable 

 character. In none of the equine genera do these tubercles attain any importance, 

 and they have, therefore, the appearance of being indeterminate variations. 



Another principle may be deduced from the facts of equine descent, viz., that a 

 slight degree of specialization in a direction away from that taken by the main line 

 is not incompatible with a place in that line. Thus the elbow joint of Mesohippus is 

 curiously specialized in a fashion that does not occur in any of the later horses ; the 

 outer portion of the humeral trochlea projects laterally and is flared in a peculiar 

 manner, forming with the corresponding surface on the radius a joint which allows 

 an extraordinary degree of flexion. But for the obliquity of the trochlea, which 

 throws the radius outward when the arm is flexed, the two bones could be brought 

 into contact for almost their whole length without dislocation. It may be objected 

 that no known species of Mesohijjpus is really in the direct series, and that the 

 ancestral species did not have this peculiarity, but this seems improbable from many 

 points of view, especially when it is remembered that a trace of the same structure 

 may be observed in Miohippus. If true at all, this principle is of wide application, 

 but it must not be pushed too far, for nothing seems better established than the 

 belief that premature specialization in any conspicuous degree is fatal to the perma- 



