62 CORA D. REEVES 



noted when the fish used to blue-red patches were shown 

 patches matched in brightness but with the gray substi- 

 tuted for the red. The evidence that red differs from gray 

 is emphasized by the differential behavior of untrained 

 fish. It is equally emphasized by similar behavior of the 

 fish which were modified by months of experience, as 

 recorded in the blue-red series. They showed toward 

 gray the same differential response which wild fish showed 

 toward red. 



C. Dark-adaptation and sensitivity to light. — To test this 

 question I made tests from time to time to determine 

 whether there was a change in the relative brightness of 

 colors when the fish were in strong and weak light. These 

 tests were made during the red-blue training experiments. 

 They were never long continued but merely interpolated 

 in the red-blue series, when a point had been reached in 

 the adjustment of the red at which its brightness for the 

 fish was judged to be equal to that 'of the blue. The fish 

 were then tested at night after two hours dark-adapta- 

 tion. This dark-adaptation did not greatly change the 

 accuracy of the response. There were no marked changes 

 in behavior even when the two stimulus patches were of 

 the intensities to which the fish had responded by similar 

 behavior before the food association was established. This 

 seemed to show either that there was no shift in relative 

 brightness, or that the response was to color rather than 

 to light intensity. It was too dark for the fish to secure 

 the food as it fell. Only when it fell through the water 

 between the fish and the colored plate was it ever seized. 

 As there was danger under these conditions of breaking 

 up the reaction of the fish to the positive stimulus plate, 

 the experiments were not carried further with the trained 

 fish. The work was continued with red and white stimulus 

 plates and with untrained dace, sunfish, and blunt nosed 

 minnows (Pimephales). While these were light-adapted, 

 they were presented with the stimulus patches until red 

 avoidance disappeared, and the intensity of the plates 

 was then adjusted until the fish went to the two with equal 

 frequency and in the same manner. The fish were then 

 allowed to become dark-adapted. No marked change 



