44 HARRY MILES JOHNSON 



Dog 3 developed a variety of i^osition habits. After perhaps 

 three weeks both became accustomed to conditions so that 

 they did not hesitate in choosing the alley to a food-box; but 

 neither could be made to turn back immediately a shock was 

 received. Both would proceed to door X or X' as the case 

 might be, or cross the grill at least, before turning and going to 

 the proper place. Dog 2 was little affected by punishment. 



The learning records, showing percentage of daily accuracy, 

 for ninety-two days of this work are shown in table 6. Not 

 only was the problem not learned in that time, but no animal 

 showed promise of improvement. 



In the belief that the two tones may have been so nearly 

 alike that the dogs could not discriminate, the experiment was 

 abandoned for the time, and the animals were given the prob- 

 lem of associating release from the home -box at the stimulus- 

 tone c-256 d.v. with food in food-compartment F, and release 

 from the home-box without a stimulus-tone with food in food- 

 compartment F'. Ten days (150 trials) was allotted for this 

 work. There was no change in the dogs' behavior after the 

 problem had been changed, beyond a return for one or two 

 days to old position-habits which had been abandoned. At 

 the end of ten days, as appears in the learning table below, 

 discrimination was not established, nor was any improvement 

 shown by any animal. This was taken as indication that the 

 animals had been disregarding the stimulus-tones entirely. The 

 results of this experiment are shown in table 7. 



The explanation of these unexpected results is not easy. It 

 is of course possible that in ordinary noises the dog is stimulated 



