148 HARVEY CARR 



TABLE I 



Correlation Between Quickness of Elimination and Nearness 



to the Food Box 



Varying degrees of positive correlation between the two factors 

 were obtained for six of the nine mazes. None of our values 

 are as high as that obtained for Miss Hubbert's maze. There 

 are, however, three exceptions to a uniform positive correlation; 

 in these three mazes the errors were mastered in proportion to 

 their nearness to the maze entrance. This lack of a uniform 

 positive correlation can be interpreted in two ways. 1. If food 

 satisfaction is an effective agency of elimination, its influence is 

 overcome by some other selective factors which are peculiar to 

 three mazes. 2. On the other hand we may suppose that food 

 satisfaction is non-effective and that all of the above correla- 

 tions (both positive and negative) are to be explained in terms 

 of a single principle. The latter hypothesis is the preferable one. 



Miss Vincent suggested in her paper the possibility that the 

 ease of elimination is a function of the strength of the tendency 

 to enter an alley, and that in a general way the relative attrac- 

 tiveness of the various alleys can be measured by the frequency 

 with which they are entered. Cul de sacs with the greatest 

 error score offer the most inducement to entrance, and the 

 stronger the attraction, the larger will be the number of trials 

 necessary for elimination. On this basis a negative correlation 

 will obtain between the number of errors for each alley and the 

 order of mastery. To test the hypothesis, the various cul de 

 sacs were now ranked in the order of number of entrances made 

 by the group for the successive stages in the mastery of the 

 maze. Correlation data were computed by the ranking method 

 with much detail but since these results were uniformly consis- 

 tent for all stages of learning, we give in table II the values 



