CANADA PORCUPINE 23 



At first No. 3 was fed loo morsels of cabbage and then loo 

 morsels of carrot, but as this length of series -was great enough 

 to reestablish the habit each time the number was dropped to 

 50 before a change, then to 25 and to 10 in a series, and lastly 

 with no definite number, ranging from three to eight before a 

 change of food was made. All of this is clearly indicated in 

 the table. At first, before starting again after a change, he was 

 given a piece of the food of the kind to be used and was not 

 allowed to use his hands in taking it. Later this was discon- 

 tinued. The result of the entire practice was that after about 

 350 such changes of food and hands involving nearly 5000 tests 

 and 160 errors No. 3 reached for cabbage with his right hand 

 whenever, wherever, and by whomsoever it was oftered. He 

 was considered perfect when he made the irregular changes 100 

 times in succession without an error. Thirty days later with 

 no practice with his hands in the meantime, he showed little 

 or no loss of ability to make the discrimination. The 100-day 

 memory test yielded results indicative of almost perfect power 

 of retention. His dislike for cabbage at that time, however, 

 made results rather indefinite. 



Nos. 4 to 10 inclusive except No. 6 were used as direct checks 

 upon this and showed equal skill. Little would be added by a 

 detailed account other than what is given in the modified checks 

 'given below. One of the most convincing arguments that the 

 experimenter did not present the food in a different manner and 

 thus suggest the change of hands each time is seen in the con- 

 duct of the animal when mistakes were made as to how some 

 particular animal had been trained. Experimenters occasionally 

 recorded a dozen or more errors in the beginning of a day's 

 work and then stopped to learn the cause of the trouble, only 

 to discover that the animal was reacting according to his habit 

 at the time they were expecting the opposite and recording 

 errors. This could not be voluntarily tested as that would have 

 destroyed the condition necessary for its operation as a check. 

 Three times, however, without arousing suspicion, the writer 

 intentionally gave wrong directions to an assistant. Without 

 exception, the complaint was soon made that something must 

 be wrong as the porcupine was reacting in just the opposite 

 way. As this little deception was not divulged until all experi- 

 ments were completed, the writer is convinced that sufficient 



