THE ORIGIN OF VARIATION 55 



In a highly adaptable mammal like the Grey Squirel (Sciurus 

 carolinensis) (see Middleton, 1931) almost endless variations in 

 habit are recorded, e.g. in food, use of burrows instead of trees, 

 etc. 



These data suggest that the fundamental genetic basis of 

 behaviour is very easily modified by the environment. It also 

 appears to be subject to spontaneous change, though the 

 origin of this change is obscure. It is similarly difficult to 

 distinguish the various roles of heredity and tradition. Some 

 authors have suggested that ' traditions ' ultimately become 

 hereditarily fixed. 



(e) Summary of Data on the Inheritance of Induced 

 Modifications. — Much of the experimental evidence is un- 

 satisfactory, but it is difficult to avoid the impression that some 

 types of impressed modifications are in certain circumstances 

 inherited. 



The indirect evidence appears to require one of three 

 possible hypotheses : 



(a) That the modifications are all mere fluctuations. This 

 is scarcely tenable. 



(b) That where the modifications are inheritable, it is due 

 to the selection of adapted variants. 



(c) That acquired modifications, long impressed, have 

 become inherited. 



A serious objection is brought forward by those who hold 

 that in any particular case the correlation between the varia- 

 tion and the environment may be due merely to the selection 

 of variants best suited to that environment. This objection is, 

 quite literally, unanswerable, but it assumes what can never 

 be proved, at any rate with our present knowledge. It is a very 

 large assumption to maintain that a graded series of variations 

 in a species corresponds to a parallel gradient of adaptation 

 to the altering environment, if only because of the extra- 

 ordinarily discriminative selection required. It appears to us 

 that neither of these rival theories can be dismissed by a priori 

 argument. Both are possible, both are at present incapable of 

 final proof and must in each case be judged by the balance of 

 the evidence. The extent to which the discriminative power 

 of Natural Selection is developed is discussed in more detail 

 elsewhere (Chapter VI I). We shall merely record our opinion 

 that an adaptive explanation of much of the data on pp. 44-50 is 



