LIGHT OF DIFFERENT WAVE-LENGTHS BY FISH 47 



The difference in the appearance of the red plate might 

 easily account for the behavior of 17 and YP, while the 

 discrimination of H and the sunfish showed that they were 

 not depending upon the presence of the slit in order to 

 reach their food. 



Further evidence was furnished when a " new " blue 

 filter was put into use. This let through the same range 

 of wave-lengths as the old, or standard blue filter, but 

 transmitted more light. In order to reduce the intensity 

 of its color patch to match the standard blue, it was neces- 

 sary to use a slit. The seven trained fish were tested with 

 this " new " blue with the slit, and the red at maximum 

 value, i. c, without the slit. The result of this arrange- 

 ment was to transfer the slit to the blue side and leave 

 the red side without a slit. Of 34 trials 30 were to the blue. 

 It was clear that when the colors were not matched in 

 brightness the fish could discriminate with the slit on the 

 blue side quite as well as they had hitherto discriminated 

 when it was on the red side. The slit did not afford them 

 a clue. 



A short time after this the same fish were tested again, 

 by the use of the slit with the new blue filter as in the last 

 tests. To reduce the intensity of the red without the aid 

 of a slit three separate means were used; the carbon suspen- 

 sion already mentioned, a solution of saffranin, or a solu- 

 tion of potassium sulfocyanide with ferric chloride. By 

 use of the flicker photometer the new blue with the slit, 

 and red reduced by one of the solutions were matched by 

 varying the amount of the solution above the red plate. 

 By this arrangement the slit was on the blue side while 

 the red was without a slit. The experiment differs from 

 the other experiment only in that the colored plates were 

 matched in brightness. The distribution of the light over 

 the two stimulus plates was again different because of 

 reflections from the surface of the solutions employed on 

 the red side. The results obtained are presented in 

 Table I. 



The dace YP and 17 failed to discriminate. The record 

 of H may also show lack of discrimination. In the case 

 of the sunfish the definite straight lines of their paths to 



