28 HARRY MILES JOHNSON 



room were two food-boxes, each above a chair which the animal 

 was to mount with her fore -feet. The two chairs were fastened 

 with sides together, and with their fronts secured to the wall, 

 so that the animal had to face west in mounting the chair to 

 be associated with the tone of the c-fork ; and east in mounting 

 the one to be associated with the tone of the e-fork. When proper 

 choice was made food was dropped on to the chair from a chute 

 above, which was opened by the pulling of a string in the hand 

 of the experimenter. In case of incorrect choice the animal was 

 recalled without being fed. The reactions were observed by the 

 experimenter from without the room, who watched the work 

 through a hole in the wall three inches square. 



In the beginning of this experiment the animals were first 

 given each tone separately, and "put through" by the operator. 

 After eight or ten stimuli had been given the animal showed a 

 tendency to go to the box of her own accord. After each animal 

 had performed fifty consecutive reactions to the c-fork correctly 

 without being put through, she was trained to go to the other 

 box at the striking of the e-fork and allowed a like number of 

 reactions. Then both stimuli were given irregularly in a given 

 series, and the animals' unaided reactions were recorded. Each 

 early developed a "position habit," at first usually choosing 

 the box where : he was last fed or avoiding the one from which 

 she has last been recalled without being fed. Then followed 

 in each dog a preference for one box or the other, causing its 

 selection from 70% to 90% of the chances. Then each dog 

 acquired a rhythmical habit, going to one box, now to the other, 

 regardless of the stimulus presented. To break this up the 

 animal was worked against her preference — the order of pre- 

 sentation being so arranged that the animal's method would 

 seldom secure food for her. Later both animals reacted irreg- 

 ularly, but as appears in table 5, with little regard to the 

 stimulus given. 



After 37 days — 505 trials each, the experiment had to be 

 interrupted. There was little evidence of discrimination in the 

 dogs at the end of the test. 



There were some indications, however, that they might be 

 failing to react correctly, not because of inability to discrimi- 

 nate, but because of lack of "attention." If on leaving the cage 

 the animal immediately began a wide detour to the right or 



