466 GERTRUDE MAREAN WHITE 



but seldom came to the top. She also was confused by the 

 dimming down of the lights, as is shown by the sudden rise in 

 curve B, figure 3 (a-a), but with the renewing of the batteries 

 the errors ceased. The experiment was continued until ten 

 consecutive errorless days were recorded. 



Using the same mudminnow, the colors were then reversed. 

 (Fig. 3, B, c.) Although the forceps under the green light now 

 contained paper, the mudminnow persisted in leaping out at 

 them, but refused to do so when red light with food was sub- 

 stituted. For five days this fish could not be induced to take 

 food in red light, so that it was necessary to, drop bits of food 

 into the water without colored illumination in order to keep her 

 in proper condition for experiment. On the sixth day she ate 

 once in red light after snapping at paper in green light, but 

 would not take food again in light of either color. On the 

 seventh and eighth days she would not eat except when food 

 was dropped into the water without colored lights. She sprang 

 out after food in light of either color on the ninth day. Then 

 followed eleven days of perfect record when she consistently 

 took food under red light, but refused to do so under green 

 light, (Fig. 3, B, and table 1.) 



Experiments with constant and varied intensity of illumination 



Since a flashlight does not yield a uniform illumination for 

 any considerable length of time, an apparatus was contrived in 

 which an electric lamp could be used for passing light through 

 the gelatin filters (Fig. 4) . An oblong tin box was constructed, 

 6cm.x6cm.xl2.5 cm. The electric light bulb was fitted into 

 one end of the box. At the other was an opening 4 cm. in 

 diameter before which was a groove to hold the gelatin filters in 

 position. Two plates of glass (P, P',) were inserted between the 

 light and the gelatin plates to prevent the plates being over- 

 heated. A sliding cover was fitted tightly over the top of 

 the box. 



The e'ectric lamp was connected with a rheostat (fig. 5, R) 

 by which the intensity of the light could be regulated and varied. 



