42 CORA D. REEVES 



widths, as did three untrained sunfish which were tested. 

 Two years after the experiments represented by the graph 

 of this fish, Large Sunfish was again tested with blue against 

 red with a 1.8 mm. slit. The first responses, made after 

 16 months absence from the experiment aquarium, showed 

 great differences in behavior when compared with the 

 earlier tests of this fish. It at once went toward the stimulus 

 patches and either waited at a little distance or jumped 

 toward the end of the food-bar. There was apparently 

 retention and recall of kinesthetic behavior. I could dis- 

 cover no memory of the blue-food association. The avoid- 

 ance of red was less marked than in the first responses of 

 two years previous. A " position habit " soon developed. 

 More than ten days passed before I discovered any differ- 

 ential response toward either plate but then noted that 

 there was a difference in the length of time for the response 

 to blue as compared with that to red. Then some days 

 later the discrimination between the blue and red with 1.8 

 mm. slit became definite and exact. The ability of this 

 fish to discriminate at slit widths 1.8-1.5 mm. does not 

 then differ from that of the other sunfish. Its earlier failure 

 at these widths may have been due to some internal factor 

 or some undiscovered effect of manipulation. No further 

 trials were made with it at slit width 0.9 mm. The curve 

 of this fish indicates that it formed the blue-food associa- 

 tion more slowly than the smaller sunfish. Eighty trials 

 are required to bring the curve to the 85 per cent line (trials 

 40-120), while in the other fish this level is reached in twenty 

 trials. This slowness is perhaps due to the greater age 

 of Large Sunfish and the fear behavior at first so dominant. 



c. Small Sunfish. — Small Sunfish learned the food asso- 

 ciation very quickly and gave high percentages of correct 

 choice with the red at maximum intensity, but did not 

 show the initial avoidance of the bright red plate shown by 

 all the other fish. For this exceptional behavior I have no 

 explanation other than that the fish lacked fear of the 

 bright red plate as shown by the fact that it fed immedi- 

 ately from the bar. As the red light was reduced in intensity, 

 there was no lack of discrimination. The percentage of 

 correct choice remained about 80 per cent until a 0.9 mm. 



