298 COLOUR VISION 



explanation of the ps?/chological evolution of colour perceptions. It 

 must, however, be regarded as a pure hypothesis and must stand or 

 fall bv the accumulation of evidence for or against it. So far as the 

 physiological processes underlying colour sensations are concerned it 

 fails to afford any satisfactory explanation. In particular it does not 

 account for the trichromatism of normal colour vision as revealed by 

 the mixture of pure-colour stimuli. Edridge-Green says : "We must 

 therefore explain in another way the apparent trichromatism of normal 

 colour vision, which is so well known to every photographer, especially 

 those who are concerned with colour photography. If my theory of 

 the evolution of the colour sense be the correct one, and we have cases 

 of colour blindness corresponding to every degree of the evolutionary 

 process, we have an explanation of the facts. In past ages all saw the 

 rainbow made up of only three colours — red, green, and violet. When 

 a new colour (yellow) appeared between the red and green, it is obvious 

 that a mixture of red and green would give rise, not to red-green, but 

 to the colour which had replaced it— namely, yellow^." This is obviously 

 at most a very partial explanation of the trichromatism of normal 

 colour vision, which is a fact and not a theory. 



It must be admitted that the evidence which has been collected from 

 various sources on the evolution of the colour sense is of an uncertain 

 character, but such as it is it affords no support to the theory. The dis- 

 cordant results and conclusions arrived at by different observers on the 

 colour perceptions of lower animals are recorded in Part I, Section VII, 

 Chap. II. If we consider only those arrived at by v. Hess, as being the 

 most recent, most exhaustive, and in many respects the most accurate, 

 we find that mammals have the same spectral limits as men ; in birds 

 and reptiles the spectrum is shortened at the violet end only ; amphibia 

 resemble mammals ; and fishes are totally colour-blind. With regard 

 to primitive races I agree with Myers that we have not sufficient evidence 

 to show that the colour sense materially differs in different peoples, 

 save that, as shown by Rivers's careful observations, their colour 

 sense is defective for light of the violet end of the spectrum. On the 

 recapitulation theory that the developing child passes rapidly through 

 the stages of evolution of the race one might expect to obtain useful 

 information from the development of colour perceptions in the infant. 

 I do not think that the experiments recorded afford any evidence on 

 this point, but only on the affective values of different colours for the 



^ Hunferian Lectures, [>. 23 1911. 



