40 CORA D. REEVES 



eral minutes might elapse after the sliding door had been 

 opened before the fish would leave the discrimination com- 

 partment. The inconvenience of this slow approach was 

 in part compensated for by the rapid rush back to the 

 discrimination compartment when the test was over. A 

 series of tests was finally made with this fish, with gradual 

 reductions of the width of slit. 



Two small sunfish, which had had the same treatment 

 as Large Sunfish, were brought into the experiment aquar- 

 ium about the time that dace Y P and Yl were introduced. 

 Records were made of their first approach and behavior 

 toward the red and blue stimulus patches in contrast with 

 no-illumination. The eye movements of the fish, their 

 orientation, and the straighter paths followed in approach- 

 ing the illuminated plates, showed that both colored patches 

 were effective in modifying behavior. Tests for the two 

 small fish, Small Sunfish, of undetermined sex, and 552,, a 

 male, began in November, 1914. 



The graphs of these three fish, made by plotting the 

 percentage of correct choices for each twenty successive 

 trials and distributing these according to the width of slit 

 used with the red variable, are shown in figure 15. The 

 slit widths appear at the top of the graphs and the corre- 

 sponding intensity of illumination of the red filter may be 

 read from the curve, figure 7. Like the graphs of the dace 

 these may be divided into two parts, one in which the 

 red was held at maximum intensity (35 mm. slit) and one 

 in which the red was of decreasing intensity. They are 

 also divisible into the three parts shown at the top of the 

 figure, in the first of which the red is brighter than the 

 blue for the human dark-adapted eye, while in the second 

 the colors are matched in brightness, and in the third the 

 red is the duller. 



In the ' red maximum ' part of the graphs it is to be 

 noted that that of Large Sunfish is essentially like the 

 graphs for dace (fig. 13). It shows (1) an initial avoid- 

 ance of red. indicated by a high per-cent of blue choices 

 (trials 1-20), (2) a subsequent getting used to the red 

 (adaptation) (trials 21-40), shown by a drop in the curve 

 to the 55 per cent level, and (3) a final establishment of 



