TRANSFER OF RESPONSE IN THE WHITE RAT 3 



is needed here, however, to make sure the statement. 

 Richardson's(4) experiments on white rats also furnish some 

 data. Two sets of rats were obliged to learn to respond 

 correctly to box III. One set had previously learned to 

 respond correctly to boxes I and II. The curves for the 

 two sets show a decided advantage for the trained rats. 

 Yoakum's(5) experiments with squirrels show that training 

 is sometimes a disadvantage. Thus animals trained on 

 box I and box II took a longer time to learn to open box III 

 than animals which had had no training with boxes I and 

 II. The responses required in the first two boxes, such 

 as scraping, butting up the latch with the nose, were car- 

 ried over to box III where a different response was required. 

 Hunter(6) offers two sets of experiments bearing on this 

 question, only one of which will be mentioned here, the 

 other being discussed later. In his work on pigeons he 

 found that there was no difference as to number of trials 

 required for trained and untrained animals in learning a 

 given maze where training had been on another maze. 

 The excess time and errors during early trials were greater 

 for the trained, yet the excess effort was eliminated in a 

 smaller number of trials by the trained. 



Bogardus and Henke(7) also made some tests upon the 

 question of transfer in the case of rats learning a maze. 

 The animals were taught a certain maze. Then by use of 

 doors for blocking the true pathway at certain points and 

 for opening up the pathway at new points, the correct 

 pathway was altered in certain respects. Then the ani- 

 mals were taught to run the new pathway. An effort was 

 made to make the maze different, each time any alteration 

 was made, in but one respect. These were arranged ac- 

 cording to what was judged to be their relative difficulty 

 in order I, II, IV, III, V, number I being the original maze. 

 The actual order of difficulty in learning these five proved 

 to be, as shown by the table of results, V, IV, III, I, II. 

 This leads to the conclusion that the disadvantages of 

 the old habits rather overshadow their advantages, that 

 is, that responses to the previously learned maze were 

 carried over to the new maze and actually proved a hin- 

 drance to the formation of the new habit. In Watson's 



