ANIMAL AND RATIONAL INTELLIGENCE 207 



the latter must be used as our auxiliary, if the test 

 of training is to be applied. The man who would 

 govern an animal, must first govern himself. 



The results of continued friendly companionship 

 with man supply the most reliable tests of animal 

 intelligence. As the dog has been most favoured in 

 this respect, evidence is most abundant from its life. 

 This evidence is better gathered from its ordinary 

 services to man, in carrying, retrieving, pointing, or 

 driving sheep, than from any attempts at general 

 education. These serve more to prove the relative 

 inferiority of his intelligence, than his natural capa 

 bilities. Evidence, at once familiar and reliable, is 

 supplied by interpretation of signs for direction of 

 conduct, such as fetch, come to heel, go wide, and 

 go home ; and many more, familiar to the collie. 

 It is better to give less heed to restraints upon 

 action, and more to full activity of natural powers. 

 We must, therefore, first find the basis of our observa 

 tions in the ordinary life of the animal, and after 

 wards seek expansion of this by reference to domesti 

 cation. By power of scent, the dog traces his master 

 in the crowd; intelligence performs no part here. 

 The same dog hears his name called by his master, 

 and is at once arrested. Susceptibility performs 

 here also a necessary part, but there is an association 

 of the sound with past experience to such an extent 

 as to involve a higher phase of discrimination. 

 When the animal is at first arrested, this may be a 

 purely sensible result, the impression being more 

 acute on account of development of sensibility by 

 exercise; but when the call or command of the 

 master is translated into action, it is impossible to 



