506 STUDIES, SCIP:NTIF1C AND SOCIAL chap. 



Narratives of the most marvellous character have been 

 published, but, unfortunately, no systematic experiments 

 appear to have been made, except a few by the late Mr. 

 Romanes. These showed that a dog could exactly track 

 its master's footsteps by scent, and it was proved that 

 the scent was derived from his shoes, which must of 

 course be full of perspiration and other emanations from 

 the skin, because when he wore new shoes or those of 

 another person his track could not be followed. In Nature, 

 vol. vii., there is a considerable discussion of this subject, 

 and many remarkable cases are narrated by various corre- 

 spondents, but in none are all the data given for arriving 

 at a rational conclusion on the question. I then wrote 

 summarizing the discussion {Nature, vol. viii., p. Qb), and 

 suggested a series of experiments, which would give us 

 the real information required. My suggestion was that a 

 dog whose antecedents were known should be taken by 

 a circuitous route by rail and road to some spot where he 

 had never been before, and should be there handed to 

 some person he did not know, who should carry him a 

 short distance, and on releasing him should keep him in 

 sight, either on foot or horseback, till the animal returned 

 home, noting carefully every movement and action. A 

 moderate number of such experiments would settle the 

 question of instinct .or sense-observation, and it is to be 

 hoped, now that a more intelligent interest is taken in 

 the subject, such experiments will be made. 



Theoretically, any instinct of direction is almost incon- 

 ceivable, because quadrupeds in a state of nature do not 

 require such an instinct. They learn, step by step, the 

 surroundings of their birth-place, extending their range, 

 perhaps, year by year, but never requiring to go back over 

 a country they have not previously traversed. In their 

 case ordinary memory, assisted by very acute senses, would 

 be all they would need. In the case of domestic animals 

 returning home, we find that the recorded time elapsed 

 varies from a few hours to several days or weeks, even 

 when the distance can be easily traversed fn a day. This 

 shows that an instinctive sense of direction cannot always 

 be the agency employed ; and, as in no one case is the 



