TRANSFER OF RESPONSE IN THE WHITE RAT 21 



the responses would be correct. From a study of the daily 

 records it was found that there is no evidence in any case for 

 believing that the animals at any time used the double 

 alternation position habit. A study of the records does 

 show, however, that at first the choice of response was 

 practically a matter of chance. In a very few instances the 

 position habit of always going to one side was picked up 

 after a series or two but was not maintained for longer than 

 one day at a time. Then in most cases there seems to be a 

 gradual shift to the simple alternation position habit, but in 

 none of the cases in this list of eleven animals was such a 

 habit actually learned and consistently maintained while 

 learning the light. A careful study of the daily records did not 

 reveal a single series where a perfect simple alternation was car- 

 ried clear through the series while learning the light. This was 

 not true, however, after the animals had learned the light and 

 had been transferred to the sound. There the animals 

 seemed to be working largely on a basis of simple alternation 

 but not with any degree of consistency. Out of 215 series 

 where sound was given as a stimulus after light had been 

 learned only 20 showed a perfect simple alternation. From 

 a study of the records, however, one must conclude that there 

 is a strong tendency toward using simple alternation in a 

 series of the type used, but that there is little ability to 

 maintain it consistently. 



In order to make it as certain as possible that the animals 

 were actually responding to the light stimulus when the 

 required response was considered learned, first, a high 

 standard of excellence was chosen and second, control series 

 were introduced. The standard adopted as indicating that 

 the problem had been mastered was 95 correct choices out of 

 100 chances or five series in which there was not more than 

 one mistake to a series on the average. The actual fact was 

 that most of the animals met even a higher standard than 

 that, as a glance at the sample individual curves will show. It 

 was not considered probable that the animals used would be 

 able to develop a position habit of sufficient complexity to en- 

 able it to respond correctly, 95 times out of 100 even with a 

 series apparently as simple as the one adopted. As indicated 

 above the results justified this assumption. But to be cer- 



