HUMAN REACTIONS IN A MAZE 107 



III. RELATIVE RATE OF ELIMINATION DURING LEARNING PROCESS 



The rapidity of elimination of the gross values is much greater 

 during some parts of the learning process than during others. 

 The extremely rapid initial descent of the curves representing 

 animal learning is a well known phenomenon. The purport of 

 this section is a comparison of the three groups in this respect. . 

 The comparison is facilitated by a translation of the gross values 

 into percentage terms. By dividing the value of each trial by 

 the value of the first trial, a series of percentage results is secured 

 decreasing from ioo to zero. The value for any trial indicates 

 the percentage of the initial value yet to be eliminated. These 

 percentage values for the several groups are given in Table IV. 

 The values for time, error, and distance for each trial are aver- 

 aged, giving a set of results for each group of animals which 

 represents a combination of the three criteria of time, error, and 

 distance. Since all initial values are 100%, this method allows 

 a superposition of curves, a device which renders comparison 

 easy. In constructing the curves from these tables, we have 

 averaged all results except for the first trial by twos, in order 

 to eliminate extremes of variability. These curves are repre- 

 sented in Figures 2, 3, and 4. Several facts are obvious. 



1. All curves exhibit the rapid initial descent. In a general 

 way, 85% of the values are eliminated during the first half of 

 the learning process. Obviously, this extremely rapid initial 

 elimination cannot be interpreted as due to the functional 

 efficiency of rational processes. 



2. In the character of this initial elimination, certain differ- 

 ences obtain between the three groups for all the curves, and 

 the rats represent a mean between the two extremes represented 

 by the two human groups. The children make the most rapid 

 initial elimination; the rate of the adults is the slowest, while 

 both sets of rats occupy a mean position. But an inverse rela- 

 tion obtains between the rapidity of elimination and the number 

 of trials over which it is sustained. The rate for the children 

 is the most rapid, but it persists for the shortest time (3-4 

 trials). The rate for the adults is the slowest, but it persists 

 for the longest time (up to the 9-10 trial). The initial spurt 

 of the rats terminates around the 5-7 trials. 



Since these differences are relatively pronounced, characterize 

 all three sets of data, and apply to two sets of rats run by dif- 



