2 CORA D. REEVES 



months three small fish, of undetermined species, had been 

 living in a crystallizing dish on my desk. One morning 

 the dish was placed so that one-half of it was in the direct 

 sunlight and one-half in the shade. The fish came to rest 

 most frequently in the shade. Then a prismatic solar 

 spectrum was thrown on the bottom of the shaded half 

 of the dish and the fish were observed to start or leap as 

 they entered the red-orange area. They often swam 

 around this area, to rest with the head in the blue or green, 

 facing in the direction of the incident rays. Some shiners 

 {Notropis cornutiis) were then tested by throwing the 

 intense light from a projection lantern, one meter distant 

 into the glass-walled aquarium in which they were kept. 

 At first some of the fish gave a start as they entered the 

 lighted area, while others swam along the side of the lighted 

 space but did not enter it. 



It was evident in both of these experiments that light 

 called out differential responses, presumably unlearned or 

 instinctive; but whether intensity or wave-length differ- 

 ences induced the response in the test with the spectral 

 band was in no way indicated. 



These reactions to the intensity of white light, as well 

 as to different parts of the spectral band, were evanescent. 

 They were obtained in the first trials, but were not evident 

 in later trials, when the fish, having become used to the 

 stimulus, no longer responded to it. These first responses 

 showed the behavior of an untrained animal when stimu- 

 lated but cannot be assumed to have shown the animal's 

 inherent power of discrimination. To secure sustained 

 responses and to show by them the full capacity of the 

 animal it seemed essential to resort to habit-forming ex- 

 periments. For these experiments two large intense 

 patches of light of restricted and definitely known wave- 

 length were used as stimuli and were presented simul- 

 taneously. In order to form a food association and thereby 

 to secure continued response the fish were always fed before 

 the same patch and the position of this positive stimulus 

 was, of course, irregularly shifted, while it was kept at 

 constant intensity. When an association had been formed 

 with the positive patch of light it became necessary to 



