IMPERFECTION OF IXSTIXCT. 169 



Among Mammals it must be deemed a mistaken instinct 

 which leads the Norwegian lemming to swim out to sea in its 

 migrations, and perish by millions in consequence. Under 

 existing circumstances it is an imperfect instinct which leads 

 the quadrupeds in South Africa, mentioned by Mr. Darwin in 

 the Appendix, to migrate, seeing that by so doing they expose 

 themselves to persecution. The shrewmouse, also mentioned 

 by Mr. Darwin in the Appendix, which " continually betrays 

 itself by screaming out when approached," is another, and 

 perhaps a better instance. The instincts of rabbits with 

 regard to the attacks of weasels appear to me to be imperfect, 

 or not completely formed. For, as I observe in " Animal 

 Intelligence " (p. 359), I have witnessed the mode of capture 

 practised by weasels in the open field, and it consists merely 

 in the rabbit " toddling along, with the weasel toddling 

 behind, until tamely allowing itself to be overtaken . . . 

 There seems to have been here a remarkable failure of natural 

 selection in doing duty to the instincts of these swift-footed 

 animals " — a failure, however, which time would doubtless 

 remedy, if weasels were sufficiently numerous in relation to 

 the breeding power of the rabbit to give natural selection the 

 opportunity of perfecting the instinct of escape from this 

 particular enemy. 



Many other instances of the imperfection of instinct 

 might be quoted, but enough have now been given to render 

 unquestionable the only point with which we are concerned, 

 viz., that although well established instincts are, as a rule, 

 adjusted with astonishing nicety to certain definite and 

 frequently recurring circumstances, the adjustment is made 

 only with reference to these, so that a very small variation in 

 them is sufficient to lead the instinct astray. It is also of 

 interest here to note what seems to be a complementary truth, 

 viz., that small variations taking place in the organism itself 

 when not in normal converse with its environment, are suffi- 

 cient to throw the delicate mechanism of instinct out of gear 

 when this is afterwards brought into such converse. This fact, 

 for instance, is familiar enouo-h in the case of tamed animals 

 (which when again " turned down " in their native haunts 

 are not at first at home in them), but is brought out in a much 

 more striking manner by the experiments of Mr. Spalding. 

 Thus he says : — 



"Before passing to the theory of instinct, it may be 



