AMERICAN RUMINANT-LIKE MAMMALS. 263 



Uinta type. This elongation of the teeth accompanies the 

 extension of the muzzle and, as it were, prevents the formation 

 of diastemata, though these appear in the later species, P. 

 labiatum, in which the growth of the rostrum outstrips that of 

 the premolars. In the John Day the tendency is changed, and 

 the premolars of Gonip/iotherium revert almost to the Uinta 

 type in form, while Procaviehts and the subsequent genera of 

 the phylum are remarkable for the reduction of their premolars 

 both in size and number. 



Wortman has called attention to a third example of this fluc- 

 tuation in the tylopodan phylum, affecting the form of the 

 tympanic bulla. In Protylopiis the bulla is small and hollow, 

 but in Poebrotheriiim it has become greatly inflated and filled 

 with cancellous bone, the inflation especially affecting the 

 medial portion of the bulla. In GompJiotherhmi, once more, 

 the direction of development is changed and the outer portion 

 of the bulla begins to enlarge at the expense of the inner, a 

 change which reaches its culmination in the existing genera. 



So general are these minor fluctuations, that it would be 

 difficult to point out a single genus which in every minute 

 detail is exactly fitted to be the ancestor of a later genus, assum- 

 ing that these fluctuations do not occur. Just how great they 

 may be in degree we have at present no means of determin- 

 ing. It seems, a priori, improbable that, after a structure has 

 been lost or reduced to a rudimentary condition, it can ever be 

 regained, or become functional once more, and yet certain 

 cases do suggest that even such regeneration may occasionally 

 take place. At all events, it would be premature to deny the 

 possibility of changes of this character. 



The features of alternating, up and down, or zigzag develop- 

 ment, to which attention has been called, are, after all, of a very 

 trifling nature. When we survey the successive and closely 

 connected genera of a long and crowded phylum, we cannot 

 fail to be impressed with the steady, orderly, unswerving 

 advance in all important structural features. This advance is 

 not always, perhaps not even usually, uniform in all parts of 

 the structure. One part may be accelerated and another 

 retarded, and what was retarded in one genus may be accel- 



