18 INDIVIDUAL DIFFERENCES AND FAMILY EESEMBLANCES 



evident throughout the subsequent trials, yet it did not appear to 

 slow up the formation of a new habit to as great a degree as might 

 be expected from the nature of the behavior that was tested. The 

 quick recovery from the interference effect suggests that the be- 

 havior of the animals of this investigation shows a considerable 

 amount of flexibility. As was previously stated, the animals took 

 only half as much time to go through the maze on the day when the 

 interference was given as they did on the first day they were tested, 

 and this shows that an accelerating transfer effect from the previous 

 training was operating to counteract the slowing up of the inter- 

 ference. 



In order to more carefully analyze the behavior of the mice the 

 median record was calculated for each day's performance, and a 

 curve based upon the same is represented by the dotted lines in Fig. 

 4. It may be noted that for each test the curve based on the median 

 falls considerably below the corresponding curve for the average. 

 The essential character of the curves, found by these different meth- 

 ods, is the same; except that the curve based upon the median is 

 more regular than the one found from the average. In the inter- 

 ference test the curve for the median records is much more regular 

 than that of the average curve for the same test, while no disturb- 

 ance at all is to be noted at the sixth day of that test. The character 

 of the curve based on the median supports what has already been 

 said concerning the permanence of the interference effects over a 

 number of trials. 



The retention test shown in Fig. 4 was given immediately after 

 the mice had been tested in the multiple choice test, and may repre- 

 sent a certain amount of training acquired there. The average time 

 for the last ten trials of the interference test is 60.26 P.E. 4.7, 

 while the average time for the ten trials in the retention test is 52.81 

 P.E. 4.7. The superiority of the average retention test in time is 

 nearly twice its P.E., indicating a fair reliability. The superiority 

 of the retention test may be due to the presence of one or more of 

 the following conditions: (1) A mere carrying over of capacity at- 

 tained in the interference test. (2) The dying out of bonds devel- 

 oped in the negative test (which in this case was the initial learning 

 test), and the strengthening of bonds developed in the interference 

 test. (3) The transfer of capacity developed in the multiple choice 

 test. Now 1 is not likely because the average record made in the 

 retention test was much better than that in the interference test 

 even after a considerable interval of time. It is possible that the 

 condition in 2 may account for the facts, but there is no direct con- 



