474 GERTKUDE MAREAN WHITE 



light (2.5 cm.) was frequently used with the yellow filter (4.9 

 cm. being used for the green). The experiment lasted forty- 

 one days (February 25 to April 6). On each of nine days one 

 mistake was made, but errors were never recorded on two con- 

 secutive days. The fish never made a pei'fect record for so long 

 a period as did the fishes in the experiments previously described, 

 but in two instances a perfect record was shown for five succes- 

 sive days, and in two instances for four successive days. It 

 seems safe to conclude that mudminnows are able to distinguish 

 green from yellow, but that these colors are not so readily dis- 

 criminated as are green and red, and blue and red (fig. 1, E). 



Experiments with gray filters 



In order to check up Hess' theory that fishes perceive colors 

 as shades of gray, as a totally color-blind human being perceives 

 them, further experiments were performed. Two photographic 

 plates were exposed to the light one second and four seconds, 

 respectively, producing on one a light and on the other a dark 

 gray surface. These plates were cut to the size of the gelatin 

 filters and the gelatin surface protected by glass so that they 

 could be inserted without scratching in the box shown in figures 

 4 and 5, and the experiment performed under the same condi- 

 tions as the color tests. If the fishes were reacting to shades of 

 gray, they ought to be able to distinguish between the two pho- 

 tographic plates just as they had done with the colored filters. 

 Mudminnow no. 60, which had distinguished green and yellow, 

 did not show the slightest sign that it could perceive any differ- 

 ence between the two photographic plates during fifty-one dasy 

 (April 21 to June 10). It took food a little irregularly, as some- 

 times happens in the spring, but the results were consistent 

 throughout, since the fish attempted in the same manner to take 

 whatever was offered with both plates (fig. 6, B, and table 1). 



Mudminnow no, 55, which had learned the red-green combi- 

 nation, was also used with the two grays. The fish also gave no 

 evidence of a perception of difference between the light and dark 

 plates (fig. 6, A, and table 1). 



