17 

 results. It may have affected the retention tests but 



not materially. If anything of interest resulted it 

 was the daily variations or fluctuations after learning 

 was assumed to he complete. These were, at times, higher 

 than the norm. Again the retention for some rats was 

 as good as the last trial. but the averages for all the 

 rats are higher than those of the last trials. It can- 

 not be said that the retention is better for one than 

 for three or five trials. Three trials seem to be the 

 best of all. These results cannot be considered a true 

 test of retention, for better methods of preparing the 

 rats for the tests are necessary than those adopted, before 

 something more definite can be determined. 



The number of trials, or total time and days which 

 each rat required before it learned the problem is indi- 

 cated on the distribution curve, Plate I. Though the rats 

 were run in groups and the norm was an average of one 

 second in three successive days for all the rats in the 



