DISTRIBUTION AND ELIMINATION OF ERRORS IN MAZE 155 



that a choice between two such possibilities might be determined 

 to a large extent by tendencies aroused in traversing the previous 

 zig zag course. On this hypothesis the attractiveness of any 

 cul de sac in a maze is in part a function of the motor tendencies 

 developed in the prior sections of the maze. The habits aroused 

 in a group of individuals in traversing the initial section of a 

 maze must necessarily have much in common, and yet any part 

 of a maze presents possibilities of wide divergence in the char- 

 acter of the pathway actually traversed. A group will thus 

 approach a cul de sac with some degree of uniformity of dispo- 

 sition toward it, but radical exceptions are possible. 



b. Final distribution. — The final distribution of errors repre- 

 sents the order of elimination, for eliminated cul de sacs are 

 those which are not entered. As previously developed there is 

 a correlation between the initial distribution of errors and the 

 order of mastery of the blind alleys, but this correlation is far 

 from perfect. It is thus evident that the order of elimination, 

 or the final distribution of errors, is also dependent upon other 

 factors than initial attractiveness. Certain peculiarities of the 

 cul de sacs constitute a determining condition. Some blind 

 alleys are relatively difficult and others are relatively easy to 

 master, and this ease or difficulty of a cul de sac is to some 

 extent independent of its initial attractiveness. 



The above principle is well illustrated by mazes III and IV. 

 The eleven cul de sacs of maze III fall into four rather well 

 defined groups as to rate of elimination. The progress in mastery 

 of four cul de sacs typical of these groups is represented by the 

 curves of fig. 1. The alleys chosen as types are 1, 2, 5, and 11. 

 The values represented are the total number of errors made by 

 the group for each successive five trials. Curves 1, 2, and 11 

 illustrate the first principle that the order of mastery is inversely 

 proportionate to the number of initial errors. No. 11 elicited 

 but few initial errors and was mastered first; No. 1 was the most 

 attractive but the most difficult, while No. 2 occupies a median 

 position between the two. As to progress of mastery the three 

 alleys are to be ranked in order, 1, 2, and 11. No. 5 is the 

 exception which illustrates the influence of the second factor. 

 This cul de sac was the hardest of the eleven to master, and 

 yet its initial attractiveness was no greater than its position 

 would justify. The number of entrances into this alley rapidly 



