THE ALTERNATION PROBLEM 373 



facts: — 1. The repetitions and the alternations are practically 

 equal in number for the first 50 trials. Evidently no animal 

 came to the problem with a preference for either mode of choice. 

 2. Three rats maintained this neutral attitude for 150 trials, 

 and then rapidly developed a pronounced preference for the 

 alternating mode of attack. One animal immediately developed 

 a slight preference for alternation and maintained this attitude 

 for 400 trials, relapsed into a neutral attitude, and then rapidly 

 developed the habit of alternation. Three animals rapidly de- 

 veloped a repeating preference for 300 to 400 trials, and then 

 shifted quite rapidly to the opposite mode of attack. The re- 

 maining animal first developed a slight preference for alternation, 

 shifted to the repeating tendency for 100 trials, and then per- 

 fected the habit of alternation in 300 trials. 3. The correctness 

 of the choices due to repetition is a matter of chance. Each 

 rat closely approximated a score of 50% of correct choices for 

 every stage of learning. 4. The correctness of the choices due 

 to alternation is at first a matter of chance. All rats approx- 

 imated a score of 50% for the first 50 trials. Finally the rats 

 learn to adapt their alternate choices to the objective series and 

 approximate a score of 100% for this mode of attack. 5. Four 

 rats rapidly learned the trick of adapting their alternate choices 

 to the objective sequence. A score of 90% or better was at- 

 tained in 150 to 250 trials. One of these individuals lost the 

 trick for quite a long period and then remastered it. The other 

 four animals at first increased their percentage of wrong alter- 

 nations for 160 to 280 trials, and then quickly learned to adapt 

 their choices to the objective series. 6. There is no correlation 

 between initial ability to alternate and success in adapting this 

 to the objective sequence. Of the four rats that immediately 

 developed a preference for alternate choices, two succeeded in 

 adapting these to the objective sequence and two did not. Of 

 the four animals that decreased the initial number of alternate 

 choices, two succeeded in adapting them to the given order of 

 presentation and two did not. 



Our problem thus presents four distinct difficulties which must 

 be mastered: — 1. The rat must learn to choose correctly the 

 initial entrance for each day's trials. 2. The animal must learn 

 to keep its choices equally distributed between the two exits, 

 or, in other words, it must inhibit all tendency toward the 



