154 HARVEY CARR 



ably attractive for it was entered much more frequently than 

 one would expect if the spatial order of the cul de sacs were the 

 only factor determining the distribution of errors in the initial 

 trials. Likewise No. 3 was relatively difficult in two cases, and 

 No. 4 three times. Comparing mazes Il-a and Il-b of like 

 pattern but with a different distribution of errors, we find that 

 No. 7 is easy in each case while No. 4 tends to be difficult. 

 Chance will account for the deviations in part, but chance can 

 not invariably favor certain cul de sacs. The tendency toward 

 uniformity in the character of the deviations for the same maze 

 pattern must be due to peculiarities of the cul de sacs other 

 than their spatial position in the maze. Certain ones are rela- 

 tively easy to avoid, while others are prone to be entered. 



The susceptibility of the animals to these favored cul de sacs 

 is, however, an individual matter. In maze III, No. 8 received 

 more than its due share of entrances, yet this alley was not 

 entered at all in the first five trials by four of the sixteen rats; 

 in fact one animal did not enter this alley once in fifty trials, 

 while another rat entered but three times in fifty runs. This 

 avoidance of the difficult errors by certain rats can in part be 

 explained in terms of the habits acquired by those animals in 

 traversing the first section of the maze. Let us suppose that 

 the first section of a maze contains no blind alleys but consists 

 of a series of runways so arranged as to necessitate alternating 

 right and left turns of 90°. At the end of this section there is 

 presented a choice between two paths, — a ' straight ahead ' 

 error, and a turn of 90° leading to the food box. It is conceivable 



TABLE IV 



Deviating Cul de Sacs 



Number of 

 Maze cul de sacs Plus Minus 



A. Mazes exhibiting a positive correlation 



I-a 7 3, 6 1 



I-b 7 4, 6 1 



I-c 7 4, 6 1, 2 



Il-a 7 4 1, 2, 3 



III 11 8 7, 10 



IV 9 3, 4, 8 6 



B. Mazes exhibiting a negative correlation 



I-d 7 3, 6 1 



I-e 7 3, 6 1, 5, 7 



Il-b 7 4, 6 3, 7 



