CEREBRAL FUNCTION IN LEARNING 81 



SO far as can be determined from so few cases, the relation 

 between extent of lesion and amnesia seems to be continuous. 

 When the group is divided into lower, middle, and upper 

 thirds with respect to extent of lesion, we obtain the follow- 

 ing averages : 



The small number of cases in group b and the scattering 

 of the records make it uncertain, however, whether the 

 series represents an all-or-nothing or a continuous function 

 of the extent of lesion. 



For group a, with total destruction of the critical area for 

 pattern vision, the results are unequivocal. Total destruc- 

 tion of the area by a small lesion produces only a partial loss 

 of the brightness habit, as is evident in cases 56, 60, and 76. 

 The destruction of the area plus other parts of the visual 

 cortex results in more severe loss and the severity of the 

 amnesia is rather closely proportional to the extent of lesion. 

 The correlation for the group (p trials ^0.72; p errors = 

 0.71) is as high as that found for the entire series of fifty 

 cases, in spite of the marked reduction in range of variation 

 introduced by this selection. "When the cases of group a are 

 divided by fives from least to greatest with respect to extent 

 of lesion, the following result is obtained : 



There is only one reversal of order in the table of fifteen 

 places, so it seems certain that the correlation is continuous 

 throughout the range of the data. Thus the results with 

 this group show that the correlation between extent of lesion 

 and the degree of postoperative amnesia is not merely an 

 artifact arising from invasion of the area necessary for pat- 



