VALUE AND ORIGIN OF COLOR VISION 465 



all we can say at present. But with transmutation (pp. 163-8) in mind, 

 we know that we must be less ready than ever before to assume this, and 

 we must realize that even if a series of color-vision systems is ever found 

 in vertebrates, only a far more accurate knowledge of the course of phy- 

 logeny than we now have can ever make it possible to say which simpler 

 systems are stages in the evolution of human color-vision, and which of 

 them are independent inventions. 



For, just as it is certain that rhodopsin has been invented many times, 

 it is almost certain that color vision has been repeatedly evolved in differ- 

 ent vertebrate groups. When the geckoes' rods were secondarily made 

 from ancestral lizard cones, was color vision lost? If so, then has it 

 returned in the tertiarily diurnal gecko Phelsuma? Do snakes have 

 color vision — and if so, could it conceivably have been inherited directly 

 from the lizards through the snakes' underground, degenerate-eyed begin- 

 nings? Did those first placental mammals keep any of the cones (and 

 color vision) of their ancestors? Or is it because the cones (and color 

 vision) of the placental mammals and primitive snakes were reinvented 

 by those groups, that there are no oil-droplets or double cones among 

 them as there are in their respective marsupial and lacertilian forebears? 

 We cannot answer positively any of these questions or others like them; 

 but every one of them is absorbing to anyone who is both interested in 

 color vision and convinced of the past occurrence of cone-to-rod and rod- 

 to-cone conversions. 



Withal, the very existence of any capacity for hue-discrimination has 

 been proven for so very few groups — none of them anywhere near the 

 direct road of primate evolution — that much pioneering work remains to 

 be done before anyone need concern himself with 'systems' and their phy- 

 logenetic significance or lack of it. It would be very fine if a moratorium 

 on new genetic theories of color vision could be enforced, until a great 

 many more cold facts have been garnered. 



Evidence for Color Vision — The techniques, some of them very in- 

 genious and some of them very stupid, which have been used to ascertain 

 whether particular animals discriminate hues, defy classification. All of 

 them however involve an all-important potential pitfall which at first 

 went unrecognized, was later disposed of most inadequately by one 

 means or another, and nowadays constitutes the careful investigator's 

 most time-consuming concern. This pitfall is the danger of concluding 

 that the animal has discriminated between two color-stimuli on the quali- 

 tative basis of hue, when he has actually discriminated them on the quan- 



