REWARD AND PUNISHMENT IN HABIT FORMATION 239 



doors, H, H, which a rat could easily push open after which it 

 would enter the nest box but could not pass back into the alley. 

 Small food receptacles, /, /, were inserted into the side at the 

 rear of the control box. These receptacles could be removed 

 and the openings closed by small metal doors. On the floors 

 of the electric boxes were series of copper plates connected in a 

 circuit with a no. 6 Columbia dry cell and a calibrated Porter 

 inductorium. The. current was kept constant at 1.5 amperes 

 by means of a resistance coil. Fastened to the rear of the control 

 box was a lamp stand, /. 



Room and light. The experiment was done in a dark room 

 lighted by a " Champion" 16 watt ground glass tubular lamp 

 K, figure 1, the lighting capacity of which was reduced by a 

 resistance coil to 0.61 candle power. The lamp was attached to 

 the stand at a distance of 80 cm. from the bottom of the control 

 box and so placed that the front of the electric boxes cast a per- 

 pendicular shadow. One of the electric boxes was made dark 

 by placing a cardboard just the size of the top of the box over the 

 box. This cardboard was shifted from one box to the other in 

 irregular order, not remaining on one side for more than two 

 trials at one time. Thus the problem for the rat was to learn 

 to discriminate between a box the lighting of which was 0.95 

 candle-meter and a box the lighting of which was such as was re- 

 flected from the dark surface of the control box through an open- 

 ing 10 cm. by 15 cm. at the front of the electric box, and to choose 

 the light box. 



Time and number of trials. Each subject was given ten trials 

 every third day. This unusually long time between training 

 periods was necessary to make it possible to use a hunger period 

 of forty-eight hours. In order to eliminate differences due to 

 tendencies of varied activity of the rat at different hours of the 

 day, all experiments were done between three and five o'clock 

 in the afternoon. The animals trained with electrical shock were 

 given their trials from three to four o'clock and those with hunger 

 from four to five. The habit was considered perfected when the 

 subject made ten correct choices on the same day. While 

 this is a smaller number of correct choices than is usually required 



