38 CORA D. REEVES 



Two years after the tests just mentioned, one of the 

 sunfish' which had meantime been trained in the discrim- 

 ination of the colored plates, was tested by training meth- 

 ods for brightness discrimination of white light. A 1 to 

 2 intensity difference was not found effective even after 

 long training. At first, slits of 4 mm. and 0.6 mm. were 

 used, and food was given in front of the duller plate. The 

 first fourteen responses were positive, i. e., to the duller 

 side. The width of the wider slit was then reduced. When 

 the slits were 1.8 mm. and 0.6 mm. the fish showed 70 per 

 cent of correct choice. When the slit- widths were 1.2 mm. 

 and 0.6 mm., a long training failed to secure more than 50 

 or 60 per cent of responses to the duller plate. It is there- 

 fore certain that under training methods this sunfish was 

 able to make use of a difference of intensity of 1 to 3 corre- 

 sponding to slit-widths 0.6-1.8 (see fig. 7, graph for Point 

 0) but not of a smaller difference under the conditions of 

 illumination already described. It is possible, as indicated 

 in the preceding experiment, that its unlearned response 

 is the more accurate. 



5. Wave-Length Discrimination. Blue-Red by Sunfish, with Red Decreas- 

 ing IN Intensity 



a. General Account. — Three sunfish were used for this 

 test, and each was separately tested. 



A 10 cm. female sunfish (Eupomotis gibbosus) named 

 Large Sunfish was put into the experiment aquarium at 

 the same time and subjected to the same training test 

 under the same conditions as the dace Md and H (p. 

 25). The red and blue stimulus patches were presented, 

 but for a time the fish lurked in the dark corners at the 

 farther end of the discrimination compartment. Then the 

 fish when tested approached to the blue illuminated area 

 in a series of rapid mouse-like movements, but it refused 

 to eat before the lighted plates. It required more than a 

 month to so build up the food association with the blue 

 area that shifts could be made in the intensity of the red 

 plate without inhibiting the response; and even then sev- 



^ This sunfish was " Small Sunfish " and was about 8 cm. long when used for 

 these tests. 



