334 



DASHIELL 



day. Thus the CA 1 maze of figure 4, identical with the PA 

 maze of figure 2, was run the first day, CA 2 the second day, 

 CA 3 the third, and so on for the twenty-five daily trials. The 

 changes made in the maze patterns between each two successive 

 trials were not great; and certain constant features were preserved 

 throughout: the outside dimensions remained identical, the 

 entrance door was always on the same side, and the exit door on 

 the same side (opposite to the side on which the entrance door 

 was placed). 



The results of the twenty-five trials are presented graphically 

 in figure 5. A slight, very general, downward slope seems to 

 indicate improvement, but interpretation is difficult because of 

 the extreme irregularity of the curve, and especially because 

 what is indicated is not only the improvement in the subjects' 

 reactions but also (perhaps mainly) the differences in difficulty 

 to be found between the various mazes used. A somewhat more 

 definite way of measuring whatever improvement there was in 

 the subjects' reactions is given in table 2. Here are set down all 

 the cases of repetition of the same error in the same trial. In 

 the running of maze CA 3, for example, the blind alley A was 



