82 K. S. LASHLEY 



tern vision by large lesions and non-invasion by smaller le- 

 sions, but represents a genuine function of the extent of 

 lesion. 



The nature of this function is still obscure. I have pre- 

 sented evidence that the mass relation cannot be ascribed to 

 the extent of scotoma, in that there is no retardation in the 

 initial learning of animals operated before training. A sec- 

 ond possibility, which I have not before considered, is the 

 following. If we assume, as now seems probable, that habits 

 based on brightness discrimination are mediated by subcor- 

 tical centers, and that pattern vision is a function of the 

 visual cortex, it is possible that the normal animal learns the 

 discrimination of brightness in the Yerkes box, not as a sim- 

 ple reaction to brightness, but as a reaction to luminous 

 patterns in which the figural aspect of the stimuli dominates 

 the reaction. In such a case, destruction of the occipital cor- 

 tex would abolish the capacity to react to the figural aspect 

 of the stimulus and force the animal to relearn in terms of 

 brightness only. A dim residual pattern vision in the periph- 

 eral retina might facilitate this transfer from one aspect of 

 the stimulus to another. The present analysis indicates that, 

 if the central area for pattern vision is spared by a lesion, 

 there is no postoperative amnesia for habits based on bright- 

 ness discrimination, and when the central area is destroyed, 

 the amnesia is proportional to the extent of destruction 

 within the surrounding region, which is perhaps functional 

 in peripheral vision. Thus the present results are in har- 

 mony with the above hypothesis. 



But there are two difficulties for the hypothesis which 

 seem to exclude it entirely. In the first place, it has been 

 impossible to demonstrate any reaction to the figural aspects 

 of the stimulus in normal animals under the conditions of 

 training employed in the Yerkes box. In one unpublished 

 experiment twenty-five animals were trained to a differential 

 reaction to stimuli which differed both in brightness and in 

 pattern. When the reaction had been established, the bright- 

 ness of the patterns was gradually equated until only the 



