208 BULLETIN OP THE BUREAU OF FISHERIES. 



from the background below, in accord with Keeble and Gamble's hypothesis, then an 

 increase in the amount received from below, without any change or with a decrease in 

 the amount received from above, ought to cause the animals to become abnormally 

 pale, no matter what the shade of the background may be. On a gray background the 

 animals ought to become maximum white. On the other hand, if this hypothesis holds, 

 an increase in the amount of light received from above, without any change or with a 

 decrease in the amount received from below, ought to cause the animals to become 

 abnormally dark. Under these conditions they ought to become gray on a white 

 background. 



In the following experiments these conditions of illumination were produced, but 

 not all of the results obtained support the hypothesis in question. It was found that 

 abnormally increasing the reflected light induces the fish to become abnormally pale, 

 but that abnormally decreasing it has no effect. It was also found under the former 

 conditions that black spots in the background affect but little the pattern in the skin 

 and that color in the background is simulated only to a slight degree. 



Reference to text figures 2 and 3 (p. 207, 213) will make clear the essential features 

 of the apparatus used in these experiments. We shall refer to the two sets of experi- 

 ments performed by means of the apparatus represented in these figures as (A) and 

 (B), respectively. In (A) the background was illuminated from below, thus making the 



"ratio — 5 light" abnormally small. In (B) the eyes were more highly illuminated 



from above than was the background, making the reflected light received by them 



relatively low. Thus the "ratio — ^ light" became abnormally large. 



Experiments (A), Light from Below Relatively Stronger Than Normal. 



I. White background. — On August 10, 9 a. m., a specimen of P. albiguttus 18 cm. 

 long was taken from a yellowish white aquarium and put into the glass crystalUzing 

 dish, over an ordinarj' sheet of white writing paper 25 cm. square in the apparatus 

 shown in text figure 2. When the specimen was put into the apparatus it was yellowish 

 white. All the air was removed from under the dish, so as to avoid any interference 

 with the light reflected from the mirror below. The black cylinder was then put in 

 place and the upper end closed with a piece of opaque cloth. At 12 m. the specimen 

 was very light gray with a brownish cast. The cloth was then removed, admitting light 

 from above. No appreciable change occurred either in the shade or the color of the 

 fish. 



August 1 1 the cylinder was removed and the light from the mirror intercepted by 

 inserting a piece of bristol board under the dish. The brownish color disappeared, and 

 the fish consequently became sUghtly paler. These tests were repeated several times 

 under other conditions with different individuals and similar results were obtained. 

 The brownish color was, no doubt, due to the collection of brown sediment on the mirror. 

 The water contained so much soUd material, ooze, diatoms, and the like that it required 

 only a very short time to form a perceptible layer on everything submerged. On a 

 white background, then, increase in the relative amount of light from below does not 

 appreciably alter the reactions of the chromatophores. 



