658 LAND MAMMALS IN THE WESTERN HEMISPHERE 



assumed the size and complexity of molars ; beginning at the 

 hinder end of the series, these teeth one by one become molari- 

 form, not in irregular and haphazard fashion, but by perfectly 

 graded stages. The same gradual and direct process was main- 

 tained in the oft-recurring reduction of digits among the hoofed 

 animals, differing for each group according to the symmetry 

 of the foot. In the horses, for example, the first digit became 

 vestigial and disappeared, and then the fifth followed, leaving a 

 three-toed foot, in which the median digit was notably the 

 largest and bore most of the weight. Throughout the Oligo- 

 cene and Miocene epochs the horses were all tridactyl, but 

 there was a gradual enlargement of the median digit and dwin- 

 dling of the laterals, until these became mere dew-claws, not 

 touching the ground, and the weight was carried entirely upon 

 the median one. Finally, the laterals lost their phalanges and 

 were farther reduced to splints, which is the modern condition. 

 In the same gradual and unswerving manner the higher artio- 

 dactyls went through a process of digital reduction from five to 

 two, and numberless other instances of similar sort might be 

 adduced. 



On the other hand, the direction of change long followed 

 may be departed from, the deviation being due to the introduc- 

 tion of a new factor. In the earliest deer the males were horn- 

 less, but they developed effective weapons of defence by the 

 enlargement of the upper canine teeth into long and sharp, 

 sabre-like tusks. When antlers appeared, the work of defence 

 was transferred to them, and the tusks began to dwindle, being 

 eventually suppressed in those deer which had large and 

 complex antlers, though persisting to the present time in the 

 hornless Musk Deer and in the small-antlered Muntjaks, 

 which can defend themselves with their sharp tusks. 



It would be inaccurate to say that fluctuations in the size 

 and effectiveness of parts never occurred ; on the contrary, 

 there is evidence that such fluctuations in details were not in- 

 frequent, and may have been even more common than we sup- 



