REWARD AND PUNISHMENT IN HABIT FORMATION 237 



PRESENT PROBLEM 



The problem of the relative values of reward and punishment 

 is of importance both for practical experimentation and for its 

 bearing upon the question of the role of ' incentives' in determining 

 the fixation of arcs in habit. Its solution demands the discovery 

 of the most effective methods of employing both incentives 

 and the comparing of these two. In order to avoid the error of 

 comparing a degree of hunger to a strength of electric shock which 

 might in no way be comparable, we have attempted to work out a 

 curve of relative values of different degrees of hunger and a curve 

 of the relative values of different strengths of electrical shock. 

 We found the relative values for each for four different condi- 

 tions using in all some 80 subjects. While we do not contend that 

 we found absolutely the most favorable condition of hunger and 

 the most favorable strength of electric shock for the learning 

 process we do contend that the most favorable condition in each 

 case lies somewhere within the limits used and that we approached 

 fairly close to that condition. 



The writer at this time desires to acknowledge his indebtedness 

 to Major Robert M. Yerkes for his supervision and many val- 

 uable suggestions in carrying out this experimental study. 



Subject used. As the white rat possesses the desirable qual- 

 ities of lending itself readily to laboratory methods and treat- 

 ment, and of breeding rapidly we chose it as our subject. All 

 individuals used in this experiment were bred in the Harvard 

 Laboratory for Animal Psychology from a pure albino "pet 

 rat stock" secured from Miss A. E. C. Lathrop, Granby, Massa- 

 chusetts. They were fed, when not being used as subjects, at 

 8.30 in the forenoon and at 5.00 in the afternoon each day. 

 The morning meal consisted of mixed grains and the afternoon 

 meal of bread soaked in milk. A constant supply of water was 

 provided by means of automatically feeding bottles which also 

 served as weights to keep the doors closed. 



Apparatus. The control box, figure 1, was 55 cm. long by 

 39 cm. wide by 20 cm. deep. It was divided into a nest box, A, 

 15 cm. long by 21 cm. wide, an entrance chamber, B, 20 cm. long 



