264 HARVEY CARE 



fact would indicate that the rat does not rely upon stimuli from 

 the extraneous environment during the later stages of the 

 learning process. 



Uncovering Maze. The animals first mastered the maze while 

 it was covered with the canvas top. After mastery this top was 

 removed and the animals tested. Two slightly different exp e 

 ments were performed. 1. The maze was mastered when the 

 top was open on one side allowing poor daylight illumination 

 of the interior. The top was now removed. Seven rats were 

 tested and none were disturbed by the changes. 2. The maze 

 was mastered while entirely closed and the interior illumined 

 by an electric light. The top was now removed. There re- 

 sulted the substitution of a heterogeneous for a uniform optical 

 environment, and the introduction of daylight for artificial 

 illumination. Ten rats were tested, and five were disturbed. 

 The effects persisted from 1 to 6 trials. The errors were dis- 

 tributed irregularly, and perfect records were secured in 70 per 

 cent of the tests. The average error record for those affected 

 was 1.07 as compared with a previous normal of .20. The 

 total number of errors per animal varied from 3 to 11. 



Increase of Illumination. The maze was learned while entirely 

 covered with the canvas top but without interior illumination. 

 The interior was now illumined by the electric light. A well 

 lighted uniform environment was thus substituted for a subdued 

 one. Ten rats were tested, and seven were affected. The 

 disturbance lasted from 1 to 6 trials. Errors were present in 

 but 40 per cent of the tests. The average error record was 

 1.35 as compared with the previous record of .51. The total 

 number of errors per individual varied from 4 to 47. 



Decrease of Illumination. An open maze was mastered. It 

 was situated in front of an open window giving a good illumina- 

 tion. After the maze was learned this window was covered so 

 that practically all light from this source was excluded. This 

 procedure decreased the illumination in the maze and altered 

 its direction, without changing the character of the environing 

 objects. Ten rats were tested and seven were disturbed. The 

 effects lasted for 1 to 8 trials. The maximum effect occurred 

 on the second test. Many trials were without error. The 

 average error record was 3.18 as compared with a previous 

 normal of .21. One animal made 40 errors in eight trials. After 



