56 



JOHN B. WATSON 



rately the complete individual records of all twelve animals, the 

 averages by trials, and the average of each succeeding three 

 trials. The curve is plotted from the daily average. 



After the above records had been obtained by the one trial a 

 day method, it occurred to me that a disturbance might result 

 if the animals were given two trials per day. I then gave the 

 test for three days, each animal receiving two trials in immediate 

 succession. For convenience in comparing the records, table 2 is 

 plotted as a continuation of the graph of table 1. The graph of 

 table 2 begins at A., figure 2. 



TABLE 2 

 Showing effect of giving two trials daily 



It will be seen that there was a steady continuation of the 

 learning process and that no evil results followed from this short 

 test. Attention is called to the fact that animals 1, 2, and 3 of 

 the immediate feeding group were slow on the first trial on the 

 first day. This was undoubtedly due to the fact that some wild 

 rats ran over the problem box during the night. These three 

 records were not included in the averages. 



It occurred to me finally that it would be desirable to see 

 what would happen if the feeding methods were reversed for a 

 few trials. Accordingly three of the animals from the delayed 

 feeding group (the three sixty day animals) were forced to run 

 for eight trials receiving their food immediately at the end of 

 each trial. On the other hand three of the animals from the 

 immediate feeding group (the three sixty day animals) were 

 forced to work by the delayed feeding method. 



