176 



BINNIE D. PEARCE 



structed as follows: The total number of trials given a rat 

 prior to the 40 made at the standard is divided into 10 parts. 

 The percentage of correct reactions in each one-tenth is then 

 computed and an average for all rats taken. The resulting 

 curve shows the progress of error elimination independently of 

 the absolute number of trials and is thus representative through- 

 out its length. N indicates the records during the 40 trials 

 made at the standard percentage. 



% 



1 



/o 



7i 



Figure 2. — Curves of learning. A — record for all rats on habit No. 1; B — record 

 for all rats but No. 15 on habit No. 2; C — record for rat No. 15 on habit No. 2; 

 D — record for Hunter's rats on learning a first habit in audition. 



Curve A shows the succeeding average percentages of correct 

 reactions for all the rats during the formation of habit 1. It is 

 essentially the same as curve D. Curve D is taken from Dr. 

 Hunter's ' Auditory Interference ' study, table 7. It is a 

 normal curve of four rats on sound. The practical identity of 



