No. 480] EXTINCTION OF MAMMALIA 



859 



natural selection, in each generation the inthvichials which ten. I 

 to varv in the same direction will survive, while at tlio same liiiic. 

 as already pointed out, their capacity for variation l)e(oiiies more 

 and more restricted. The consequence of this will be that the 

 more highly specialized any stock becomes, the more the individ- 

 uals composing it will come to resemble one another, until at 

 length the same results as arise from close inter-breeding, viz., 

 weakening of the stock, and, finally, extinction, may follow." 



This is purely a question of evidence and all the evidence we 

 can muster is negative. Invertebrate paleontologists cite cases of 

 extinction being preceded by an efflorescence of new structures. 



Among the Vertebrata no evidence has been adduced of extinc- 

 tion being preceded by an arrest of variation, i. c, of evolution. 

 On the contrary extinction often occurs at the high tide of change 

 and not after a prolonged period of stability. 



C0XCLUSI0>f 



The chief induction which can be made from this extensive 

 survey of the causes of extinction seems to this: folio in' >i(f the 

 diminidion in number which may arise froiii a vlitvj or on^/inal 



From weakening its hold upon hfe at one {)()int an aiumal is 

 endangered at many other points. 



