LIGHT OF DIFFERENT WAVE-LENGTHS BY FISH 81 



matched for the fish, the method under discussion affords 

 no proof that this is so. If it is not so, the fish may dis- 

 tinguish the two stimulus objects because of brightness 

 difference, although to the human eye they appear equally 

 bright. 



The third training procedure hitherto employed attempts 

 to meet the difficulty of the first two by varying the in- 

 tensity of one of the two stimuli while the other is held at 

 constant intensity. It assumes that some intensity of the 

 variable will be used at which its brightness equals, for 

 the fish, that of the constant. If at this intensity the fish 

 continue to discriminate, they must do so by means of 

 quality differences, and they, therefore, have color vision; 

 if they fail to discriminate, they are color-blind. Frisch 

 used this method when he offered his fish food concealed 

 in a colored tube, which was one of a series of gray tubes. 

 On the assumption that the colored tube is of the same 

 brightness as one or more of the gray tubes, a fish that 

 learns to go to it and continues to ^o to it when it is empty, 

 must distinguish it by the quality of light coming from it, 

 and has, therefore, color vision. A fish that is capable 

 of learning, but cannot distinguish the colored tube amongst 

 the grays is color-blind. The danger in the employment 

 of this method lies in the assumption that the colored tube 

 is matched in brightness for the fish by one or more of the 

 gray tubes. If a long series of closely approximated gray 

 tubes is used, it is probable that one of them matches the 

 colored tube, but it always remains possible that none of 

 them matches for the fish. There may remain a bright- 

 ness difference between the colored tube and any of the 

 grays great enough to enable the fish to discriminate them. 

 To obviate this, the brightness limen must first be found. 

 It cannot be too often pointed out that brightness is not 

 measurable by physical means, but is a psychical or physio- 

 logical quality, so that colors matched in brightness for 

 one animal may or may not be matched for another. What 

 is brighter or duller each of us can know directly only for 

 himself. It is, therefore, important to use a training method 

 in which the behavior of the fish affords, a means of judging 

 when the two color stimuli are matched in brightness. 



