FACTORS IN LEARNING BY WHITE RATS 361 



and my own 19 results show in general the same progressive elim- 

 ination of errors. 



Hubbert and Lashley classified the errors in the circular maze 

 into those of wrongly passing a door (type I) and those of turn- 

 ing in the wrong direction (type II). The errors of type I they 

 found to be eliminated in less than two-thirds the trials neces- 

 sary for the elimination of those of type II. The serial back- 

 ward elimination of errors of type I was found to agree in the 

 main with results of the other studies cited in the preceding 

 paragraph, but no such serial elimination of the type II errors 

 took place. They find, in accord with our own results, that 

 the animals seem to orient themselves to the maze as a whole, 

 favoring in the several blind alleys the inward direction which 

 in the circular maze is always toward the food box. If the 

 individual reactions of each animal had been studied more in 

 detail these experimenters would likely have found the expla- 

 nation of the differences in the method of elimination of the 

 two types of errors. Our results, in the monograph already 

 referred to, show that the animal soon learns to keep its general 

 forward orientation in the maze, and also that the final stages 

 of the elimination of blind alleys are frequently accompanied 

 by a confusion to the animal which results in entrances to cut 

 de sacs nearer the food box, already eliminated. A study of the 

 situation in the circular maze seems to suggest that the early 

 development of the forward orientation tendency would tend 

 to throw the animal into the blind alleys entrance to which 

 constitutes errors of the second type. In every case a rat keep- 

 ing its general forward direction and avoiding the error of type I 

 would be thrown into an error of type II. This condition cer- 

 tainly would seem to invalidate the authors' general conclusion 

 as to the relative frequencies of elimination of the two types 

 of error. Moreover, since the rate of elimination is studied 

 in terms of the number of trials required to avoid successfully 

 the entrance to a blind alley, it must be recalled that the final 

 trials for elimination of the cut de sacs first encountered will 

 bring about confusions resulting in entrances to some of those 

 already eliminated, further along the trail. From these con- 

 fusions errors of type II would most probably result, for the 

 reason already indicated. The matter seems to need further 

 investigati on. These difficulties make plain how necessary it 

 19 Cited in note 5. 



