250 MENTAL EVOLUTION IN ANIMALS. 



of barking, although perhaps acquired as a result of domesti- 

 cation, is so innate and general among most of the breeds, 

 that it deserves to be regarded as an instinct. Yet UUoa 

 noticed that in Juan Fernandez the dogs did not attempt to 

 bark till taught to do so by the importation of some dogs 

 from Europe — their first attempts being strange and un- 

 natural. Linnaeus records that the dogs of South America 

 did not bark at strangers. Hancock says that European 

 dogs when conveyed to Guinea " in three or four generations 

 cease to bark, and only howl like the dogs natives of that 

 coast." Lastly, it is now well known that the dogs of 

 Labrador are silent as to barking. So that the habit of bark- 

 ing, which is so general among domestic dogs as to be of the 

 nature of an instinct, is nevertheless seen to vary with 

 geographical position. 



Specific Variations of Instinct. 



To the above instances of the local variations of instinct, 

 I shall now add a few cases of what we may call specific 

 variations of instinct — that is to say, instincts which occur in 

 a species of a character strikingly different from the instincts 

 which occur in the rest of the genus. After what has been 

 said on the local variations of instinct, the attesting value of 

 the cases which we are about to consider must be evident. 

 For we should expect that if the conditions which determine 

 a local variation of instinct are constant over a sufficient 

 length of time, the variation should become fixed by here- 

 dity, and so give rise to a change of instinct in the species 

 affected — which change ought to become observable in the 

 contrast exhibited by the instincts of this species and those 

 of the rest of its allies. This head of evidence becomes of 

 special value when we remember that it is the nearest 

 approach we can hope to obtain of anything resembling a 

 palaeontology of instincts. Instincts, unlike structures, do 

 not occur in a fossil state, and therefore in the course of their 

 modification they do not leave behind them any permanent 

 record, or tangible evidence, of their transformations. But 

 we obtain evidence of transformation almost as conclusive in 

 the cases to which I now allude; for if a living species 



wallings" of cats; for, according to Eoulin (quoted by Dr. Carpenter in 

 Contemp. Rev., voL xxi, p. 311), the domestic cats in South America do not 

 make these sounds. 



