LIGHT OF DIFFERENT WAVE-LENGTHS BY FISH 49' 



the blue, as well as the number of correct choices of the 

 three (35 blue to 7 red) made it plain that they discrimin- 

 ated without the aid of the slit on the red side when bright- 

 nesses match for human dark-adapted eyes. For the 

 dace the record shows 41 red and 41 blue choices. These 

 fish had long since ceased to avoid red. If their high per- 

 centage of blue choices in the preceding tests had been 

 merely choices of the no-slit side, they should, in the series 

 now described, have made a high percentage of no-slit 

 choices, and in that case would have chosen red. Taken 

 alone this set of data certainly indicates lack of ability 

 on the part of dace to discriminate patches of light which 

 for man match in brightness. The results, however, show 

 that the slit was not the means for discrimination by these 

 trained fish. 



In some of these check tests the slit was not present on 

 either side, in others it was transferred to the blue side. 

 In the later procedure the fish did not show by high red 

 choice that they were avoiding the slit and so using it as 

 a guide. Further evidence that they were not guided by 

 the slit appears in the section of this paper dealing with 

 equated energy values (p. 53). 



The changes in the sharpness of outline of the red stimu- 

 lus patches may account for the tendency of the dace to 

 go toward it to investigate the unusual, Table I. In other 

 words, changes in the appearance of the stimulation patches 

 were sufficient to inhibit temporarily the association in 

 those individuals in which it was weakest. The series was 

 not long enough to re-establish the habitual response. 



E. Tests with changed quality of light. 



A further attempt was made to determine whether the 

 quality of the light of the stimulus patches was the means for 

 the discrimination shown by the fish in the red-blue series. 

 The red variable had been presented at all intensities 

 from very bright to very dull as described in section C. 

 In the last tests (described in section C) the changes in 

 intensity were rapid. It was thought that substituting 

 a patch of light of moderate intensity but of different quality 

 for this red variable might show the significance of wave- 

 length for fish. 



