368 



THE AMERICAN NATURALIST [Vol. LIII 



Biologists of the future may recognize the importance 

 of determining experimentally whether the germinal 

 variations of a species ever respond to changed life con- 

 ditions in such a way as to shift the mode of any char- 

 acter in the direction of greater adaptation. If such a 

 general tendency as this were revealed, and if, at the 

 same time, the transmission of somatic modifications 

 were rigidly excluded, we should be brought to a crisis 

 in the history of our science. The question at issue 

 would not be merely the adequacy or this or that hypoth- 

 esis. It would be the adequacy of our recognized scientific 

 methods to deal with such problems. Despite the lengthy 

 arguments with which I have sought to defend a purely 

 naturalistic position, I should not, in advance, be su- 

 premely confident as to the outcome of such experiments. 

 It might, after all, turn out that there was just such an 

 ''immanent teleology" in living things as the vitalists 

 claim. If this should prove to be true, science would 

 have to re-survey its territory and set itself new bound- 

 aries well within the old ones. 



Such an undertaking, like that of settling once for all 

 the "acquired characters" question, would doubtless be 

 beset by great technical difficulties. But these difficul- 

 ties should not be insuperable. So long, however, as 

 "genetics" is held to be nearly or quite synonymous with 

 Mendelism, evolution along dynamic lines is likely to lan- 

 guish. We must grant the enormous strides which have 

 been made in our knowledge of the inheritance of certain 

 types of variations, but the much more fundamental ques- 

 tion of the causes of these variations is almost as far 

 from solution as in the days of Darwin. 



In conclusion, I would say a few further words in re- 

 gard to my use of the expressions "contingency" and 

 "chance" throughout these pages. It is needless to say 

 that I have not used these w^ords as synonymous with un- 

 caused. I have spoken of an event as contiiiiieiit, merely 

 in the sense of its being causally mirclatcd to Mniicthing 

 else: for example, a variation in relation to a jiccd to be 

 fulfilled. Whether or not, in the last analysis, all things 



