COLOR VISION IN FISHES 473 



Alburnus alburnus, later with the marine Spinacbia spinachia and 

 Syngnathus acus, Graber found a preference for darkness as opposed to 

 white light, and a decided preference for red light over blue, produced 

 by glass filters. Equating the colored lights in brightness for his own eyes 

 (and of course assuming that they were then of equal brightness to the 

 fishes) , he determined that the animals preferred red to green, and went 

 to green or blue when either was paired with a blue + ultra-violet (by 

 which is meant, here and elsewhere in this discussion, human ultra-violet 

 — that is, wavelengths beyond the short-wave end of the human visible 

 spectrum. What is truly 'ultra-violet' for an animal may commence at a 

 longer or shorter wavelength than the one which is just visible to man) . 

 By making a red light twenty times as intense as a blue one, Graber 

 could force his fishes to show a preference for the blue. 



The earliest use of the method of training seems to have been that of 

 Zolotnitzky who, in 1901, fed fishes for a time on red midge larvae and 

 then attempted to deceive them with bits of colored yam glued to a card 

 which was held against the glass side of the aquarium. The fishes tried 

 to get at the red pieces, ignored those of other colors. A first attempt to 

 eliminate the possibility of discrimination on a basis of brightness-differ- 

 ence was made five years later by Washburn and Bentley, who induced a 

 *red=food' association in Semotilus atromaculatus by feeding from a 

 red-marked forceps presented simultaneously with an empty green one. 

 The dace continued to go to the red forceps even when it contained no 

 food, and even when the shade of red was changed considerably in either 

 direction of brightness. 



In 1908, Reighard concluded that Lutianus griseus discriminates hues 

 as such, for he failed to find what has since come to be called the step- 

 wise phenomenon. Offering both red and blue baits to wild gray snap- 

 pers in the open sea, he found that they avoided the red ones. They also 

 preferred white to blue. Back in the laboratory, Reighard found that the 

 brightnesses of the stimuli he had used were in the order white->red-> 

 blue. Since the fishes preferred the brighter member of one combination 

 (white-blue) and the duller member of another (blue-red) , he concluded 

 that they were guided by color rather than by brightness. He was further 

 convinced of this when he found that fishes negative to red baits refused 

 all shades of red, at the same time accepting other colors which must have 

 been matched in brightness for them by one shade or another of red. 



From 1909 to 1915, a flood of papers appeared in which the capacity 

 of fishes for hue-discrimination was debated pro and con, with Hess tak- 



