DISTRIBUTION AND ELIMINATION OF ERRORS IN MAZE 147 



and 5, and skip 2,4, and 6. The matter is further complicated 

 by the phenomenon of returns. The tendency to return toward 

 the entrance box is very persistent during the early stages of 

 learning, and in any trial we may have a temporal order of 

 1-3-6-1-2-3-4-5-4. According to the hypothesis this trial will 

 tend to eliminate alleys 3, 4, and 5 prior to 6. 



Hubbert's results, however, neither prove nor disprove the 

 efficacy of food satisfaction as an eliminative agency. Although 

 a high degree of correlation obtains between rate of elimination 

 and nearness to the food box, one should still refrain from gen- 

 eralizing on the basis of one maze. Different maze patterns 

 may give other results. A correlation between two factors does 

 not always indicate a causal relation between them; both may 

 be the result of some more fundamental condition. 



This paper presents data on two mazes, and in addition Miss 

 Vincent has kindly furnished me records for six mazes. A cul 

 de sac was considered eliminated when but one entrance was 

 made in ten successive runs. The number of trials necessary 

 to eliminate each alley was determined for each animal and 

 the average number of trials for the group was thus computed 

 for the various cul de sacs. These values constitute the order 

 of elimination for the group. This temporal order of mastery 

 was correlated by the rank method with the spatial arrangement 

 of the alleys in relation to the food box. Table I designates 

 the various mazes, gives the number of cul de sacs in each, the 

 number of rats employed, and the correlation data for each maze 

 pattern. Mazes I-a to I-e have the same pattern, a slight 

 modification of the Hampton Court arrangement; they differed 

 only in the arrangement of sensory factors. I-a presented a 

 uniform sensory interior; I-b had the true path painted white 

 and the blinds black; I-c had the true path painted black and 

 the blinds white; in I-d an olfactory trail was laid in the true 

 path, while the trail was inserted in the cul de sacs for I-e. Il-a 

 and Il-b were alike except that the alleys of Il-b were without 

 sides. The records for the above mazes were obtained by Miss 

 Vincent. Mazes III and IV were somewhat similar in pattern 

 but radically different from patterns I and II. Maze V refers 

 to the circular maze used by Miss Hubbert whose data are given 

 for comparison. 



