i 4 2 MIND IN EVOLUTION CHAP, vn 







which he examined. Thus he was dealing on each occasion 

 with a different selection of individuals from the culture, 

 and it is possible that those that took the harmful 

 substance at first died of the effects, while only those 

 that avoided it or took it in smaller quantities survived. 1 

 In fact, Schaeffer, repeating these experiments, obtained 

 contrary results, and found no evidence that either 

 stentor or paramecium can learn by experience to improve 

 its selection of food. 2 



Whether the confirmation or inhibition of an action 

 through experience of results is in reality a higher or 

 more difficult form of learning by experience than 

 selective modification is a question which I must leave 

 undetermined. I note only that for the lowest class 

 of organisms the former appears to be at present more 

 definitely established than the latter. But at any rate in 

 its most elementary forms we must apparently carry the 

 power of learning by experience down to the lowest types 

 of animal life. Just as we find no type of organism 

 whose actions are purely mechanical, so it now seems 

 probable we shall find none whose behaviour is entirely 

 fixed by heredity in relation to present stimulus. We 

 can at anjr'rate say with certainty that the powers of 

 assimilating the direct results of conation are traceable 

 from a very low stage. 



1 Comptes Rendus de la Societe de Biologic, Vol. LXXIV. pp. 702, 705. 



2 Selection of Food in Stentor Caeruleus^ summarised in J. A. B., 1911, 

 p. 400. 



