152 HARVEY CARR 



of trials; in other words the number of errors in the initial blind 

 alleys is decreased more rapidly than for the final cul de sacs. 

 2. The sensory character of the maze influences the distribu- 

 tion of errors in other ways than by minimizing the number of 

 returns. The tendency to follow the edge or an olfactory trail 

 evident in mazes I-d, I-e, and Il-b operated to reduce the total 

 number of errors. This fact is well demonstrated in Miss Vin- 

 cent's paper. The operation of the tendency was relatively 

 more effective in the initial than in the final stages of learning, 

 and Miss Vincent interpreted this fact as due to the shift in 

 control from olfactory and cutaneous cues over to kinaesthetic 

 factors; when the maze is run in terms of kinaesthetic stimuli, 

 the trail factor is no longer present and the rat is subject to the 

 enticements of curiosity, distractive odors, etc. I wish to sug- 

 gest a similar explanation of the greater frequency of entrance 

 into the final cul de sacs in these three mazes. Starting from the 

 entrance box, the rats are at once dominated by the stimulus 

 characteristic of the trail. This sensory trail is followed and 

 possible exploratory excursions due to curiosity, fear, and other 

 motives are eliminated. As a consequence relatively few errors 

 are made in the first sections of the maze. As the final sections 

 are reached, however, the dominance of the trail motive is weak- 

 ened and other enticements begin to operate. The weakening 

 of the trail motive may be conceived under such terms as habit- 

 uation or adaptation. The strength of other motives such as 

 fear, curiosity, and the returning tendency may progressively 

 increase with the distance traversed; the concept of summation 

 of stimuli may be applicable here. The strength of the olfactory 

 stimulus from the food box must necessarily increase as the 

 final sections of the maze are reached. One can hardly suppose 

 that the actions of an organism so complex, alert, variable, and 

 thoroughly alive as is the rat can long be continuously domi- 

 nated by a single motive. Some shift of motives during a run 

 must be presupposed irrespective of the explanatory difficulties 

 involved. Our hypothesis then assumes that the animals in 

 these three mazes start out on each run under the dominance 

 of the trail motive which tends to prevent errors. As the final 

 section of the maze is reached, there occurs a shift of motives, 

 and these new motives, such as food odor, fear, curiosity, etc., 

 tend to produce errors. As a consequence of this shift of motives 



