54 



CORA D. REEVES 



what the duller at equated energy, but with the slit on the 

 red side, and water cell and slit on the blue side it was 



TABLE III 



Table showing percentages of red and blue choices for four 

 dace and three sunfish when the radiant energy of the two 

 stimulus patches was equated. 



possible to so adjust them that the red was brighter or 

 duller than the blue or matched it at equated energy values. 



When the energies of the two plates had been equated, 

 the fish were tested by feeding before the blue plate as in 

 the red-blue series. The results are shown in Table III. 



The record of Large Sunfish is difficult to interpret. 

 For the first six tests this fish went to the blue. Then 

 while keeping the energies equated, I increased the bright- 

 ness of the red; the six following trials were to the red, then 

 one to the blue, one to the red, and again five to the blue. 

 The conditions were duplicated some days later, and when 

 this fish was retested, discrimination was accurate. In 

 the record of Y P of nine trials, six choices were red and 

 three were blue. The curve for this fish (fig. 13) shows 

 that there were abrupt changes in the slit widths during 

 its series of red-blue trials, and that in the region of matched 

 brightness no trials were made. The fish apparently had 

 not retained the blue-food association. The remaining 

 fish show percentages of correct choices ranging from 72 

 to 100. 



It appears that when the energy is equated, the fish can 

 still discriminate, and that consequently relative energies 

 do not condition the discrimination. (See postscript.) 



