n.i THE NATURE OF INSTINCT 173 



always perfect. It has recently been shown that the 

 Ammophila sometimes kills the caterpillar instead of 

 paralyzing it, that sometimes also it paralyzes it incom- 

 pletely. 1 But, because instinct is, like intelligence, fallible, 

 because it also shows individual deviations, it does not at 

 all follow that the instinct of the Ammophila has been 

 acquired, as has been claimed, by tentative intelligent 

 experiments. Even supposing that the Ammophila has 

 come in couisse of time to recognize, one after another, 

 by tentative experiment, the points of its victim which 

 must be stung to render it motionless, and also the special 

 treatment that must be inflicted on the head to bring about 

 paralysis without death, how can we imagine that elements 

 so special of a knowledge so precise have been regularly 

 transmitted, one by one, by heredity? If, in all our present 

 experience, there were a single indisputable example of 

 a transmission of this kind, the inheritance of acquired 

 characters would be questioned by no one. As a matter 

 of fact, the hereditary transmission of a contracted habit 

 is effected in an irregular and far from precise manner, 

 supposing it is ever really effected at all. 



But the whole difficulty comes from our desire to ex- 

 press the knowledge of the hymenoptera in terms of in- 

 telligence. It is this that compels us to compare the 

 Ammophila with the entomologist, who knows the cater- 

 pillar as he knows everything else — from the outside, and 

 without having on his part a special or vital interest. 

 The Ammophila, we imagine, must learn, one by one, 

 like the entomologist, the positions of the nerve-centres 

 of the caterpillar — must acquire at least the practical 

 knowledge of these positions by trying the effects of its 

 sting. But there is no need for such a view if we suppose 

 a sympathy (in the etymological sense of the word) between 



1 Peckham, Wasps, Solitary and Social, Westminster, 1905, pp. 28 ff. 



