82 THOMAS WILLIAM BROCKBANK 



comes the question of economy in retention must remain 

 more or less suspended. The results of the present experi- 

 ment show conclusively that of the two distributions here 

 considered, namely, three trials per day and one trial per 

 day, the greater distribution in trials produces the greater 

 economy both in learning and after retention. The totals 

 given in the tables and the curves in the plates only in- 

 dicate that economy is present; but, as may be expected 

 from any statistical method, they do not give any physio- 

 logical explanation for the economy itself. 



(3) Redintegration After a 30-Day Retention Period 

 (a) Retention Period for the Maze 



The effects of the respective retention periods of 45 and 

 70 days, which have been considered in the preceding 

 pages, aid only negatively in the determination of the 

 approximate period during which the maze habit may be 

 retained with a precision approaching that of the norm 

 in learning. No rat, after either of the two retention periods 

 already considered, completed redintegration tests in the 

 learning norm of 15 trials. This would have required a 

 practically perfect retention of the habit and also its perfect 

 exercise in the first six trials of redintegration. 



In order to ascertain the approximate period after which 

 integration appears perfect in the redintegration tests, it 

 was determined to experiment with a group of rats by 

 the one trial per day method and try redintegration after 

 a 30-day retention period. With this immediate end in 

 view, two litters, the first of three rats, namely, 1, 2, and 3 

 in Tables V-A and V-B, and the second of four rats, namely, 

 4, 5, 6, and 7 in the same tables, were set to learn the 

 maze. All the rats of both litters completed learning and 

 redintegration. The same norm and method was used with 

 this litter as with the groups already considered. 



The " error " results of these seven rats are presented 

 in Table V-A, and the trial and time records in Table V-B. 

 In two rats of the seven, redintegration was as good or 

 even better than the learning; while a third rat required 

 but one more trial, or 16. This may be understood more 



