354 K. S. LASHLEY 



restriction of the data to a few subjects leaves the experimental 

 results inconclusive. A simpler technique is therefore desirable; 

 one by which the subjects may be trained rapidly and data 

 gathered upon a large number within a reasonable time. The 

 training of larger numbers might, perhaps, be accomplished by the 

 use of automatic training and recording apparatus but the appara- 

 tus of this character that has been devised is bulky, expensive, 

 and not altogether dependable. Further, many animals seem to 

 be much more disturbed by mechanical contrivances in the train- 

 ing box than by manipulation in the hands of the experimenter 

 and the preliminary training necessary to accustom them to 

 moving doors, and the like, nearly counterbalances the advan- 

 tages of automatic training. The alternative method for gather- 

 ing larger amounts of data involves the training of the animals in 

 some easily acquired habit for which no great expenditure of 

 tune will be required by any one animal. 



A widely accepted objection to the use of simpler habits arises, 

 however, from the possibility that these may fail to demonstrate 

 individual differences which would appear in more complex forms 

 of learning. This is perhaps true where the primary interest of 

 the study is in the comparative intelligence of the animals but 

 where the problem is primarily that of the mechanism of learning, 

 the nervous changes involved in the reintegration of conduction 

 paths, the phenomena resulting from the complexities of the habit 

 only serve to confuse the data and make it impossible to distin- 

 guish between the effects of simple reintegration and those result- 

 ing from the simultaneous formation of a number of habits. This 

 confusion is illustrated by the results of studies of the effects of 

 the distribution of practice upon the rate of learning. Such 

 results, as I have shown in previous articles (Lashley, '15 and' 17) 

 are due at least in part to the interference of simultaneously 

 formed habits and, as the present study indicates, are probably 

 always due to the complexity of the habit and not to any funda- 

 mental character of the learning process. The failure of a simple 

 habit to reveal a difference which appeared when more complex 

 habits were studied should indicate, therefore, that the difference 

 found was due to some factor introduced by the formation of 



