374 HARVEY CARR 



development of a position preference. 3. The animal must learn 

 to alternate its choices, and 4, it must further master the trick 

 of adapting these to the temporal order of presentation. 



The progressive mastery of the above aspects of the problem 

 accounts for the peculiarities of the various curves of learning. 

 An analysis of three typical learning curves into their four com- 

 ponents will be given as illustrations. 



The dotted line curve of fig. 2 exhibits the most pronounced 

 initial rise and this rat was the first to master the problem with 

 any degree of perfection. This animal also made the most rapid 

 progress in mastering the initial choice, developed no serious 

 position preference, belonged to the group which made the 

 greatest progress in learning the habit of alternation, and was 

 the first to learn the trick of adapting its alternate choices to 

 the objective sequence. 



Curve 1 of fig. 4 exhibits a rapid descent for 220 trials and 

 this is followed by a normal rate of ascent until the problem 

 was mastered. Likewise we find that the percentage of correct 

 initial choices rapidly decreases for 250 trials and then increases 

 at a normal rate. The animal also developed a position prefer- 

 ence which reached its maximum strength at the 330th trial, 

 and which was then quickly eliminated. The rat also developed 

 a repeating preference up to the 390th trial, and then shifted 

 very quickly over to the system of alternate choices. The per- 

 centage of correctness of the alternate choices decreased for 

 220 trials, and the animal then began to learn to adapt these 

 to the objective sequence. 



Curve 1 of fig 3 exhibits four aspects, an initial rise at the 

 100th trial, a pronounced fall at the 150th trial, a rapid rise to 

 the 330th trial, and a subsequent plateau period. The cor- 

 responding percentage record of the initial choices is represented 

 by curve 2 of the same figure. The animal first succeeded in 

 choosing correctly, then failed dismally, and again succeeded. 

 This rat also exhibited for 150 trials a position preference which 

 was then quickly eliminated. The rat made no progress in 

 increasing the number of alternations for 150 trials, and then 

 practically perfected the habit in 150 trials. The curve repre- 

 senting the percentage of successful alternations is practically 

 a replica of the learning curve of fig. 3. 



The most important aspect of the problem is the ability to 



