228 HAROLD E. BURTT 



bers are the results for the trials indicated, — -the figures denoting 

 the numbers of the boxes entered in the order of entrance. The 

 last five columns give the number of correct and incorrect first 

 choices in a given series of ten, the total number of right and 

 wrong first choices on a given day and the daily ratio of right 

 to wrong. At the bottom of the tables for problem 1 appear 

 the results in a check series of ten trials with ten new settings. 

 For example, referring to table I, we see that rat A entered com- 

 partment No. 3 on the first trial and after punishment by con- 

 finement entered No. 1 which was correct; in the 56th trial she 

 entered No. 8 three times in succession, being punished for each 

 entrance and then entered No. 6 the correct one. The time of 

 the trials, although recorded as a matter of routine is not included 

 in the results. 



The tables enable one to obtain quickly information concern- 

 ing the forms of response and the changes therein during the 

 course of the experiment. Further comment will be made in 

 the following section. 



DISCUSSION OF RESULTS 



Problem 1, which consisted in the selection of the reaction 

 mechanism on the subject's right, proved soluble for the four 

 rats, i. e., they were able to make correct first choices in ten 

 consecutive trials on the ten settings. 



Outbred female A was successful after 170 trials and her 

 brother D after 200. Inbred male J, of about the same age, 

 solved the problem in 170 trials, while the younger inbred female 

 C required 350. 



In this problem the total number of doors open in the ten 

 settings is 35. Of these 10 were of course correct. Hence the 

 probability of a correct first choice in any trial is 10-35 and the 

 probable ratio of correct to incorrect first choices in any series 

 apart from experience is 1 to 2.5. The tables show that D and 

 J started with a ratio for the first day of 1 to 9 which is much 

 less efficient than was to be expected, while A and C were 

 more efficient than expectation with ratios of 1 to .66. These 

 differences were perhaps due to a tendency to go toward the right 

 on the part of A and C, while with D the opposite tendency was 

 manifest throughout. 



After solution of the problem each animal was given 10 trials 



