LOCALIZATION OF LEARNING PROCESS 203 



This hypothesis does not imply that the thalamic 

 connections drawn in Figure 50, diagram III, are not 

 employed in the learned reaction, but evidently under 

 the conditions postulated they are incapable of carry- 

 ing the reaction through unaided. Cortical reinforce- 

 ment continues to be indispensable. As has already 

 been brought out, there is no evidence that in the 

 process of automatization of a habit the cortical loop, 

 if once developed, is eliminated by overtraining and 

 the reaction short circuited through the thalamus. Ac- 

 cordingly, if the visual cortex is removed from such an 

 animal, the brightness-discrimination habit is lost. 

 Effective cortical reinforcement of this habit is a func- 

 tion of a particular part of the cortex only, namely, 

 the occipital cortex, probably because here alone are 

 the direct anatomical connections between the visual 

 centers of the thalamus and the cortex. 



Of the actual nature of this process of cortical fa- 

 cilitation or of the mechanism employed, we have no 

 conclusive experimental evidence. But the line of ap- 

 proach to the question of cortical participation in 

 trial-and-error learning which has just been sketched 

 seems to be congruous with the whole trend of physio- 

 logical work with the nervous system. In particular, 

 the conception of physiological gradients and physio- 

 logical dominance as developed experimentally by 

 Child (1924) and as applied theoretically to the nerv- 

 ous system by the writer (1924) presents interesting 

 possibilities. Transmission of excitations in nervous 



