in ? for five lainutes, but perraitted no more food until, the 

 completion of the next dny ' a experlinent . Each i^ft.t was used 

 daily until it had learned the course perfectly, the criterion 

 of perfection beint; five perfect trials for each of three suc- 

 cessive days. A perfect trial consisted in runnin^j the course 

 within six seconds, a period so short that it was practical- 

 ly impossihle for the rat to make p detectable error p.nd reach 

 the centre v;ithin that tine. Those rats failin^; to learn v-'ith- 

 in one hundred days (500 trials) were no longer used for ex- 

 perimentation. Those rats learning the mase were, at the con- 

 clusion of the experiment, fed for sixty days in a rvinv/ay tven- 

 ty-five feet long with a feeding box at the far end. At the 

 end of this period they were tested for retention and relearn- 

 ing. 



The results of the experiment for the inbred rats 

 are given in tables la, lb, and Ic; for the normal control 

 series in tables Ila, lib, and lie. These tables give only 

 the averages of the five daily trials of individual rats. The 

 shortest period of learning for an inbred rat was twelve days; 

 for a normal control, ten days. Two inbred rats and one nor- 

 raal failed to learn the maze at all. This paper does not pre- 

 tend to take up individual differences, but certain of the 

 normal control series showed peculiarities of behavior simi- 

 lar to those of the inbred series. These peculiarities, for 

 the most part, consisted in disorientation and persistent er- 

 rors. Strain B of the control series exhibited these peculiar- 

 itif^a to such an extent and were so slow in learning (the con- 

 trol rat failing to learn at all was from this strain) that 

 the investigator suspected this strain to be of lesser rela- 

 tive brain weight; and, when the returns were received from 

 the V/istar Institute, this was indeed found to be the case. 



