488 ADAPTATIONS TO PHOTIC QUALITY 



recently brought the conditioned-reflex method to bear upon the question 

 of hue-discrimination in fishes, his revelation of the ease and speed with 

 which simple discrimination can thus be proven, as compared with the 

 training method, has come too late to save any investigator's time or 

 trouble. Bull's methods cannot rival the training technique for the labor- 

 ious working-out of a color-vision system, and it is this that remains to 

 be done. No reasonable student of the problem any longer doubts that 

 fishes — all duplex teleosts at least — can experience hue as a sensation- 

 quality apart from brightness. 



The 1924-1926 work discussed above has revealed some preliminary 

 data on the nature of the hue system of Phoxinus, which has such a head 

 start that probably no American species will be worked out sooner. 

 Hungry, partly trained individuals have been found to snap for food in 

 bands of wavelengths neighboring the ones used for training. Thus, red- 

 trained animals will snap also at yellow while still incompletely trained, 

 yellow-trained ones snap also at red and green, 'blue' animals at green 

 and violet. Some of these confusions persist after thorough training, 

 enabling the determination of a curve of hue-discrimination which is 

 interestingly different from the human one, and has seeming maxima at 

 MSOm^l, A,485m[X, A,590m[l, and probably at X655m[l. The human is 

 nowadays believed to have but two genuine maxima, at 7,490m[A and 

 7,580m|l, though from one to three other secondary maxima were 

 described by the older investigators. Probably some of the 'maxima' of 

 Phoxinus will disappear in future investigations. 



A valuable finding upon imperfectly trained Ellritze was that those 

 undergoing training to red made most of their erroneous snaps in the 

 violet and ultraviolet, and vice versa. This demonstrates a recurrence of 

 redness in the short-wave sensations, closing the color circle through the 

 red-blue and red-violet mixtures which the human sees as the extra- 

 spectral purples. It would be most interesting to know, though sadly 

 unknowable, whether ultraviolet is 'red' to fishes that can see it, or would 

 look red to us if our optic media did not fluoresce it so completely into 

 pallid short-wave visible light. For many fishes, the penetrability of the 

 ocular media ceases in the (human) violet; but in Gasterosteus aculeatus, 

 for example, the wavelength 313m[X can reach the retina. Merker, the 

 leading investigator of the biology of the ultraviolet, notes that these 

 sticklebacks can be trained to snap for food in the wavelength band 

 313-253m[l, but thinks that in such very low wavelengths the food is 

 seen as a shadow cast by visible light into which the fluorescence of the 



