LIGHT OF DIFFERENT WAVE-LENGTHS BY FISH 89 



D. Final Statement of Results 



This paper embodies an attempt to learn whether fish 

 are capable of responding differentially to lights of longer 

 and shorter wave-lengths. The experimental evidence 

 presented is of two sorts — that in which the fish went for 

 food to a patch of light of shorter wave-length when pre- 

 sented at the same time with a patch of longer wave-length, 

 and that in which the unlearned responses of the fish to 

 light of different wave-length were observed. 



By the first method all the seven fish used learned to 

 go to the patch of blue light of constant intensity before 

 which they were fed, when this was presented at the same 

 time with a red patch. There was then established a food 

 association or sensory-motor habit. When the intensity 

 of the red patch was gradually reduced the -fish continued 

 to go to the blue until the brightness of the red was ap- 

 proximately that of the blue for the human dark-adapted 

 eye. At this value of the red most of the fish at first failed 

 to discriminate the two patches and went about as often 

 to the one as to the other. But as the trials were continued 

 at this matched brightness the fish again discriminated and 

 continued to do so with further reduction in the intensity 

 of the red. The value of the red variable at which tempor- 

 ary failure to discriminate and subsequent recovery take 

 place is believed to be that at which the red and blue are 

 of matched brightness for the fish. The behavior of the 

 fish at this value thus becomes an indicator for matched 

 brightness. The fish then learned to discriminate the red 

 from the blue at all values of the red variable used includ- 

 ing one of matched brightness for them. (See graphs figs. 

 13 to 16 and discussion of them elsewhere). Check tests of 

 the apparatus used and of the method of manipulating 

 it seem to have excluded the possibility that the fish were 

 guided in their discrimination by anything else than the 

 stimulus patches, and at matched brightness these patches 

 differed for the fish and for man only in the wave-length 

 of light coming from them. 



This evidence from training experiments is reinforced 

 by that obtained from a study of the unlearned response 

 of the fish. From this evidence it appears that in the 



