MAZE STUDIES WITH THE WHITE RAT 273 



involved in the cleansing of the maze but susceptible to the 

 first experiment. Many similar illustrations can be given. 4. 

 A rat may make a very large number of errors in some tests and 

 very few errors in others. Ten rats were ranked as to the number 

 of errors made in each of five experiments. The rankings given 

 to one rat for the five experiments were 1, 2, 9, 2, and 2. Similar 

 rankings for another animal were 8, 1, 1, 7, and 9. This lack 

 of consistency may be shown by dividing the animals into two 

 groups on the basis of number of errors. Only one of the ten 

 rats belonged to the better half in all five experiments. In 

 another group of eleven rats but four manifested any high 

 degree of consistency; two were found in the better half for all 

 experiments, while two invariably belonged to the poorer half. 

 The rankings for one experiment were correlated with those 

 for the other four experiments, and positive values of .369, 

 .690, .414, and .068 were obtained. 5. Affected animals make 

 a relatively high percentage of perfect runs in one experiment 

 and a low percentage in another. 6. One would naturally expect 

 a high degree of correlation between the total number of errors 

 made in an experiment and the number of trials in which a 

 disturbance was present. Two groups of animals were ranked 

 in both respects for five experiments and the correlation values 

 were computed. Small negative values were obtained in every 

 case. These results mean that those animals which make an 

 extremely large number of errors in one trial are likely to become 

 adapted to the alteration and run the subsequent trials without 

 error. 7. Animals that do well for one position in the experiment 

 on maze rotation do not necessarily make good records for 

 other positions. The correlation value between two positions 

 for a group of nine rats was but .434. Animals that do well 

 for one position do not necessarily make good records when the 

 test for this position is repeated. Such a correlation by the 

 ranking method for the above group of rats gave a value of 

 but .024. 



The above emphasis upon the irregular and accidental character 

 of the disturbances must not blind one to the fact that some 

 rats manifest a relatively high degree of consistency in the 

 various experiments. Some animals are quite susceptible and 

 make a large number of errors in every experiment. Other 

 rats are prone to immunity; they either fail to be disturbed or 



