138 MIND IN EVOLUTION CHAP. 



pounced on the eggs. The result was at once to reawaken 

 the father stickleback's protective instincts. He attacked 

 the females, and then undertook the defence of the nest as 

 before. The evidence when put together suggests that the 

 stickleback does not in our sense of the term recognise the 

 eggs as his own. The truer view is that at breeding time 

 his parental instincts are awakened by the perception of his 

 eggs. If the action of the instinct is broken off, the sight 

 of his own eggs may fail to renew it, while a chance 

 excitement may set the whole machinery in motion again. 

 The attachment of hereditary modes of action to certain 

 objects or individuals is certainly an effect of experience. 

 But we cannot infer from it knowledge of those objects or 

 individuals as objects or individuals. It is rather a form 

 of the defining or particularising of instinct in which we 

 found the most elementary operations of experience. 



According to the general account of the mental qualities 

 of Fish given in Brehm's Thierleben? Fish have a certain 

 c< understanding," but very little. They can distinguish 

 enemies from those who are not hostile. They become 

 wary of traps (Nachstellungen) and note places of safety. 

 They accustom themselves to a keeper, to feeding time, 

 and a signal like the sound of a bell that food is coming. 

 They choose suitable places for hunting, where they lie in 

 wait for their prey, learn to overcome obstacles, and with- 

 draw themselves from danger, form a more or less intimate 

 connection with their fellows, and hunt in co-operation. 

 In some cases they show a measure of care for their young. 

 Broadly speaking, this account seems to tally well with the 

 elementary form of intelligence which we have described. 

 Both among Reptiles and Amphibia a certain capacity for 

 training is verified by many examples. They " come to 

 know" their keepers, for example, in the sense of coming 

 at call or at sight of them, 2 but in this case the knowledge 

 often has very marked limitations. It would seem, for 



1 Brehm, VIII. p. 12. The account of reptile intelligence in the same 

 work (VII. pp. 24, 25) points to a similar type. 



2 Brehm, VII. p. 547, for tortoises. The most intelligent and most 

 tameable kind, the fresh-water tortoises, however, do not distinguish 

 individuals (pp. 562, 563). For snakes, ib. p. 220 ; for lizards, pp. 34, 35 ; 

 crocodiles, p. 525 ; toads, p. 700. 



