380 HARVEY CARR 



expiration of the given time interval, they were again placed 

 in the apparatus for the succeeding trial. These conditions are 

 radically different from those under which the problem was 

 mastered. With the previous conditions the animals devoted 

 themselves during the period of delay to the act of eating and 

 they usually maintained a relatively constant position. On the 

 table the rats were free to run around and react to whatever 

 stimuli that may attract their attention. The purpose of the 

 experiment was twofold. 1. We wished to determine the de- 

 pendence of the choices upon the activities obtaining during the 

 period of delay. To this end, we repeated the tests for the 

 intervals of 50, 75 and 95 seconds. 2. We wished to continue 

 the experiment with larger time intervals than the previous 

 conditions permitted. With the new conditions the usual num- 

 ber of trials per day can be given even though very large time 

 intervals are employed. 



The results secured for these conditions are represented by 

 curve 2 of fig. 8. The percentage value for the interval of 

 50 sec. is based upon a total of 1070 trials. The remaining 

 values are each based upon a total of 100 trials. The following 

 conclusions have been derived from these data. 1. The intro- 

 duction of the novel conditions during the delay has decreased 

 the percentage of correct choices by about 27%. The validity 

 of this conclusion is readily apparent from a comparison of the 

 two curves of fig. 8. 2. All of the animals were able to approx- 

 imate a record of 70% of correct choices for the interval of 50 

 sec. 3. No improvement was manifested for the 50 sec. interval 

 although the rats were tested daily for a period of 15 days. 

 4. The co-ordination was again disrupted for intervals greater 

 than 50 seconds. The similarity of the two curves for the 

 intervals of 50 to 95 seconds is striking. This fact indicates 

 that the 60 sec. interval is a critical point. 5. Further increases 

 beyond 75 sec. seem to be without effect. 6. The larger time 

 intervals did not wholly destroy the functional efficiency of the 

 co-ordinations for six of the eight animals. The group averages 

 for these larger intervals are all at least 60%. Two rats made 

 records of but 51 and 52% for the four large intervals. The 

 percentage records of the remaining animals are at least 60%. 

 The highest record was 70% and this score was made by two 

 rats. Since these values are based upon a total of 52 trials 



