Washburn and Bentley, Color-Discriminatioji. 115 



of wood 5x5 mm. across and 70 mm. long. In all the earlier 

 experiments the strips of wood attached to one pair of forceps 

 were painted red, while those attached to the other pair were 

 green of a shade — to the experimenter's eye — somewhat brighter 

 than the red. The strips were fastened to the forceps by small 

 rubber bands and projected in both cases about 5 to 10 mm. 

 beyond the metallic points. The first attempt to apply our 

 general method may be described as follows: 



/. The Method of Inhibition. — When the fish reached a cer- 

 tain position at the bottom of the tank, a young live grasshopper, 

 held in one of the two pairs of forceps, was quickly thrust under 

 the surface of the water. The fish was allowed to take the food 

 from the red forceps, but when it snapped at the green pair the food 

 ivas quickly withdrawn. For one day, the red forceps only were 

 used and for the succeeding four days (six feedings) the red and 

 green were used in irregular sequence. One of us applied the 

 stimulus while the other recorded, by means of a stop-watch, the 

 time of reaction — from the instant the food touched the water till 

 it was snapped at by the fish. At the second red-green feeding, 

 we noticed that the reactions seemed, at times, to be prematurely 

 released by the sight of the approaching hand that held the for- 

 ceps. In order to eliminate the possibility of reaction to a warning 

 signal, the side of the tank next the experimenter was enveloped 

 with a black cloth screen and half the top was covered with heavy 

 gray cardboard. In the subsequent tests, the baited forceps were 

 slipped over the edge of the cardboard top and directly into the 

 water at the center of the tank. In this way, the fish saw no move- 

 ment until the object appeared at the surface of the water. 



The recorded times were, of course, too inaccurate to be con- 

 sidered as "reaction times"; but it was not with reaction times 

 that we were primarily concerned, but rather with the discrimina- 

 tion of color-tones. It seerfied probable, however, that a discrimi- 

 nation of red and green — if it occurred — might be expected to 

 lead, in time, to an inhibition, or at least to a retardation, of the 

 green (unsuccessful) reaction. 



In the last four days (six feedings) one hundred and thirty-one 

 trials were made, sixty-two with red and sixty-nine with green. 

 The longest reaction was five seconds, and one hundred and thir- 

 teen of the whole number fell within one to three seconds. The 

 average time for red and for green was the same, 1.4 seconds. 



