468 



GERTRUDE MAREAN WHITE 



the association, the intensity of the Ught was varied, but no in- 

 crease was noted in the number of mistakes, nor did the dis- 

 crepancies seem to be correlated with the relative intensities of 

 the light (fig. 3, C, and table 1). 



Mudminnow no. 55 was tested in two sets of experiments. 

 The first set was carried on in the fall of 1916 between October 



-— s 



Fig. 5 Representing the conditions of experiments with gelatin filters il- 

 luminated by an electric lamp. The electric lamp in box, B, was connected 

 with the rheostat, R, by which the intensity of the light could be varied. During 

 the experiments the jar containing the fish was surrounded by an opaque screen, 

 S, and the colored light was flashed upon the fish at the same time that food 

 was offered from forceps. 



25 and November 29 (thirty-six days), with green light indicating 

 the presence of food and red indicating paper. No errors were 

 recorded after November 1, the eighth day of the experiment, as 

 the fish consistently refrained from snapping at paper in red 

 light, but took food in green light. Beginning with November 

 15th, the lights were varied in intensity. The results of this ex- 



