150 VINNIE C. HICKS 



association discussed in a preceding paragraph and need no 

 further comment. The curiosity factor, however, is present 

 long after the association has been established, and after errors 

 are largely eliminated. It often leads the animal into the cul 

 de sacs which were eliminated in an early stage of the experiment, 

 and consequently this feature finds expression in all three curves. 

 The tendency also causes the animal to stop and examine parts 

 of the true path even after its behavior has approximately reached 

 the automatic stage. The time of the run may often be doubled 

 or tripled in this manner, a result that is represented only by 

 the time curve. This feature of their behavior probably is partly 

 responsible for the later rise of the time curve relative to error 

 and distance. It may be urged that the elimination of curiosity 

 and timidity constitutes no part of the maze problem per se and 

 hence should not be represented. This contention is erroneous 

 for two reasons : First, the whole problematical situation cannot 

 legitimately be conceived as confined to the maze alone. The 

 problem is one of adaptation of an animal with a given nature 

 to the maze. If the nature of the animal is entirely eliminated, 

 comparative psychology is without point. The purpose of any 

 experiment is to investigate the ability of a certain animal to 

 learn a given problem, and all native peculiarities, either advan- 

 tageous or disadvantageous to the process, must be considered. 

 Second, the tendencies cannot be eliminated even if desired. 

 Their results, as already shown, are partially reflected in both 

 the error and distance curves. If these features are to be repre- 

 sented at all, they should be represented as adequately as possible. 

 The assumptions that these three factors account for the gross 

 differences between the time and distance curves and that they 

 are essential elements of the learning process are supported by 

 curve II, figure 4. This represents the learning by twelve rats 

 of a maze with no cul de sacs. With the exception of the returns, 

 the three hypothetical elements mentioned are the only factors 

 of the problem. The error and time data are given in table I. 

 That a real problem is presented to the animals is evident at a 

 glance. More than a third of any maze problem is constituted 

 by the factors under consideration. It is noteworthy that this 

 maze may be regarded as learned in the region of the twelfth 

 trial, which approximates the position at which time ceases to 

 decrease relative to distance in the original experiment. Accord- 



