44 JOHN LINCK ULRICH 



rats went without hesitation to the latch, held their snouts 

 under it, but for a short time made no decisive movement to 

 raise it. A similar loss in retention was recognized in the solu- 

 tion of the inclined plane box when some of the rats went to the 

 plane but did not push it down. In the maze there was an 

 increase in the time of the runs and there was a reappearance 

 of errors. With some a slowness of action alone was evident 

 when rats were first given the problem, but, when an initial 

 start was made, speed was regained. In all these cases a loss 

 in retention was particularly evident. In all such cases it was 

 apparent that there had been an actual interference with or 

 loss of the integrations. But interference was not always present, 

 because for some individual rats retention was perfect. The 

 integration of movements was as perfect as that displayed in 

 the last trial in the original training series, and the time was 

 the same. This statement must be restricted to the latch box 

 and the inclined plane box, the retention for the maze cannot 

 be said to have been ever perfect, for errors were always present 

 after a period of no practice. 



As in the case of learning, only average results are given 

 and these are presented on Tables I, II, III, IV and V for reten- 

 tion. The averages of the last trial, the first retention test, and 

 the mean averages of both are offered. Upon examining these 

 tables critically when one problem was given with one or more 

 trials the retention test averages are higher than the averages 

 of the last trial. The same thing is true when three problems 

 are given concurrently, the tables here presenting great differ- 

 ence in this respect. There is no exception to this statement. 

 In regard to the effect of varying the number of trials, the 

 results on retention vary. When the latch box was used alone 

 retention was best with three trials. On the other hand when 

 two additional problems were given with the latch box, the 

 retention average was best wdth one trial. With the maze alone 

 the average was lowest with five trials. When the maze was 

 used with two additional problems the three trials method was 

 best. With the inclined plane box the retention was best when 

 one trial was given. 



Little further can be said about these results for they do not 

 point to anything very definite. Before anything conclusive 

 can be said, better methods for the preparation of the tests 



