THE ALTERNATION PROBLEM 371 



will give a percentage of 50 of correct choices and no improve- 

 ment will be possible until the habit is broken. An equal dis- 

 tribution of choices will give a score of 50% with no improve- 

 ment so long as the choices are a matter of chance. When the 

 alternation system is mastered, the choices will still be equally- 

 distributed and a score of 100% will be attained. 2. The rats 

 may develop the tendency either to repeat or alternate from 

 the previous choice. An invariable repetition of the previous 

 choice irrespective of whether it was correct or incorrect is 

 equivalent to a position habit and it will give a score of 50% 

 with no improvement. Alternation from the previous choice 

 will give a score of zero if each day's initial choice was incor- 

 rect, while a perfect score of 100% will be attained if each day's 

 sequence gets started properly. 3. The rats may also develop 

 the tendency either to repeat or to alternate from the previous 

 exit that gave food. The repeating tendency will necessitate a 

 wrong alternation with a score of zero. The alternating ten- 

 dency will solve the problem and give a score of 100%. 



All possibilities thus reduce to two, the development of a 

 position preference, or the acquisition of a habit of alternation 

 and this alternating sequence of choices may or may not conform 

 to the objective sequence. Our results were now analyzed and 

 tabulated with the purpose of studying the development of 

 these two tendencies. 



The relative number of R and L choices, irrespective of their 

 correctness, was determined for the successive stages of learning. 

 The group exhibited a slight preference for the R exit for the 

 first 100 trials. A pronounced L preference was now developed 

 and this persisted with some degree of strength until the 500th 

 trial, after which point the choices were equally distributed 

 between the two exits. The L exit was consistently chosen in 

 two-thirds of the trials for a period of 200 trials. The develop- 

 ment of the L preference was confined to five of the eight animals, 

 while the other three rats maintained a practically neutral atti- 

 tude towards the two exits throughout the entire period of 

 learning. The L preference began to develop somewhere in the 

 period from the 50th to the 170th trial and it persisted for a 

 period of 300 to 600 trials. Four of the five animals at times 

 chose the L exit in 80% of the trials. The development of this 

 preference may be both advantageous and detrimental to the 



