270 HARVEY CARR 



while entirely covered by the canvas top and illuminated by 

 an electric light. After learning, both maze and top were 

 rotated as a unit. Tests were given for the three positions 

 on successive days. On the fourth day a normal record was 

 secured for the original position. The above procedure was 

 then repeated several times. Ten rats were utilized in the 

 experiment. In the first shift eight rats were disturbed; these 

 gave an average error record of 1.29 for 48 trials, although 

 errors were present in but 31 per cent of the trials. The shifts 

 were now repeated three times and no tendency toward adapta- 

 tion was in evidence. The percentages of animals affected in 

 the four successive shifts were 80, 80, 90 and 70. The percentages 

 of trials in which error was present were 31, 35, 55, and 36. 

 The error records were 1.29, 1.81, 1.18, and 1.32. The largest 

 disturbance occurred for the 180 degree position. This result 

 is a function of the position and not of the temporal order of 

 the shifts, inasmuch as a different temporal order of the three 

 positions w r as given in the successive series. 



This experiment is comparable with the first test of the 

 previous section in all respects except the environmental condi- 

 tions. In the previous test the maze was rotated in reference 

 to the environment, while here both maze and environment 

 were rotated. Rotation in reference to a stationary environment 

 produced much the greater effect at first, and allowed a pro- 

 nounced degree of adaptation when the experiment was repeated. 

 No adaptation was present when both maze and environment 

 were rotated, and the records secured were practically identical 

 with those in the former experiment after the rats had become 

 adapted. 



C. Alteration of conditions while learning the maze 



Rotation of Maze. Animals were required to master the 

 standard maze when its cardinal orientation was changed for 

 each day's test. The daily shifts in position were 90 degrees, 

 each position being repeated every fourth day. These records 

 are compared with those representing the mastery of a stationary 

 maze, and we are able to estimate the relative effect upon learn- 

 ing of a stable vs. a variable relation to the objective environ- 

 ment. The following records for a rotated maze were obtained 

 from ten rats without previous laboratory experience: — the 



