ASSOCIATION AND COLOR DISCRIMINATION 469 



periment and the one following are tabulated so as to show ex- 

 actly what occurred at each feedmg and with each intensity of 

 hght (fig. 3, D, and tables 1 and 2). 



Beginning with January 12th, this experiment was repeated 

 with the same mudminnow and continued for sixty-nine days. 

 An especially interesting fact in connection with this experiment 

 was that no failures were noted during the first three days, 

 although the time which had elapsed since the completion of the 

 firsf experiment was forty-two days (table 2). This occurrence 

 may have been an accident, but if so, it was the only time a fish 

 made no failures at the beginning of an experiment. It seems 

 more logical to suppose that the fish had preserved the acquired 

 association of green light with food and red with paper through- 

 out this interval. The experiment seems to show a clear case of 

 color discrimination. The number of errors was in all only 

 seven. Three of these occurred during the time when the lights 

 were being varied, but as with fish no. 27 and fish no. 40, there 

 seemed to be no correlation between the errors and the relative 

 intensities of the lights. On one occasion the fish snapped at 

 the bait in the red light when the light which shone through the 

 red filter was 1.4 candlemeters without the color screen and that 

 which shone through the green filter 4.9 candlemeters (February 

 25). At another time, the fish was confused when the light 

 passing through the red filter was 4.9 candlemeters and that 

 illuminating the green 2.5 candlemeters (March 4). On an- 

 other day the red was 2.5 candlemeters and the green 4.9 candle- 

 meters (March 8). On two occasions when errors were made, 

 the light passing through the green filter was the more intense, 

 but in the other instance, the red was stronger. (Fig. 3, E, and 

 tables 1 and 2.) 



Experiments with green and yellow filters 



Mudminnow no. 60 was given the problem of distinguishing 

 green no. 74 from yellow no. 73. The yellow, which, it may be 

 noted, had two absorption bands, appeared to the human eye 

 considerably brighter than the green. On this account a weaker 



