48 HARRY MILES JOHNSON 



fallibly localized by the human subject. There was a notice- 

 able difference in intensity, that of buzzer i, placed in this 

 experiment over door X, being considerably the louder. The 

 quasi-tones were also of different pitch, that of buzzer 2, over 

 door X', being near the third of buzzer i. This selection was 

 made deliberately. The timbre of the two respective sounds 

 also differed. The sound of buzzer i was decidedly nasal, while 

 that of buzzer 2 was quite brilliant. Since these conditions un- 

 fortunately are not reproducible elsewhere this description should 

 suffice. In a word, the two stimuli differed in frequency, am- 

 plitude, form and direction of the sound waves. 



Contrary to the method employed at the beginning of the 

 experiments on tone-discrimination, the animals in this experi- 

 ment were not "put through" at the beginning, but left to form 

 the associations between a particular sound and a particular 

 food-box for themselves. Punishment was not given in case 

 of incorrect choice. Care was taken to make the duration of 

 each stimulus one-half second, or as near one-half second as 

 possible. 



The problem assigned the dogs was the association of the 

 sound of a given buzzer with choice of the food-compartment 

 over which the buzzer was placed. For the first two days after 

 the animals were introduced to the problem each dog tended 

 to react negatively to the stimulus. This was followed, except 

 in case of Dog 4, by a tendency to choose food-compartment 

 F' regardless of the stimulus presented. This was broken up 

 on the third day of the experiment by sounding buzzer i a 

 second time after the animal had wrongly gone into alley D'. 

 After only two or three repetitions this produced a returning 

 into alley D. It was not continued after this day, however, 

 as the experimenter feared that the animal might make it, 

 rather than the actual sound of the respective buzzers, the cue 

 for reaction. The learning tables, which follow, reveal a situa- 

 tion quite different from that of the tone-discrimination problem. 



This experiment shows plainly that the dog can learn very 

 quickly and without help to discriminate between two auditory 

 stimuli. The question remains whether the discrimination in 

 this case was on the basis of pitch, intensity, timbre or localiza- 

 tion. During the experiment the animals often pricked their 

 ears and turned their heads toward the sounding buzzer, so it 



