COLOR VISION IN MAMMALS 507 



the 89 samples of the entire Bradley color set, the cat would pick out not 

 only it, but several others as well. Pavlov, it may be noted, was never able 

 to establish in the cat a conditioning of reflexes to hue, and was partially 

 successful with a dog only after 3000 trials. 



Gregg et al, in 1929, attempted to train a cat positive and negative to 

 different combinations of filtered lights arranged like Ardois signals; but 

 when gray stimuli of equivalent brightnesses (for the human) were sub- 

 stituted for the various colors, the animal responded just as though the 

 colors were still there. The investigators concluded that the cat is totally 

 color-blind, or that at any rate colors have absolutely no significance 

 for her. 



Only Kalischer has claimed that cats easily discriminate hues. His 1929 

 report on the subject is very sketchy. He claims to have varied the intens- 

 ities of his colored lights sufficiently to exclude a brightness-discrimi- 

 nation, but he does not give enough details to enable one to be at all 

 sure — especially when it is borne in mind that cats certainly see short- 

 wave lights much brighter (as indicated by their pupil responses) and 

 probably see long-wave ones much dimmer (because of the great predom- 

 inance of red-blind rods) than we do. It is particularly reprehensible, in 

 the case of nocturnal mammals, to assume that the relative brightnesses 

 of colors are the same as they are for humans. It can very reasonably be 

 assumed, always, that they are not. Another method of Kalischer's — 

 'training' the cat positive to undyed, negative to dyed, meat — is open to 

 the serious criticism that he made no attempt to rule out olfaction. So 

 finicky a feeder as the cat would assuredly need no training to avoid 

 food which did not smell quite right to her. We can be quite sure that 

 the cat has no hue-discriminatory capacity at all; and we might para- 

 phrase the old saw to read : "Day and night, all cats see gray." 



None of the various researches on the raccoon is very complete. Cole's 

 first work, in 1907, was not properly controlled. With Long, in 1909, he 

 succeeded in getting raccoons to select a colored paper, or the gray, from 

 a series of five colors and one gray all of which had the same albedo (in 

 flicker photometry) for the human eye. These investigators also tested 

 the animal's ability to discriminate brightnesses, and found it excellent. 

 But their conclusion — that the animal has some color vision — was unjus- 

 tified inasmuch as they made no effort to match a color with a gray in 

 brightness for the raccoon. Davis, in 1907, was not even able to train 

 raccoons to colored stimuli which were of equal brightnesses for man. 

 Gregg et al, with the same procedure which had yielded only negative 



