474 SCIENCE PROGRESS 



3. The physiological difference would probably correspond 

 roughly to the physical difference; that is to say, the largest 

 and smallest waves would be the first to be differentiated. 



4. Yellow should be a simple but secondary sensation. 



5. A pure spectrum should be divisible into a series of mono- 

 chromatic areas, the size of these areas depending upon the 

 development of colour discrimination. 



6. There should be defects of light perception distinct from 

 defects of colour perception ; shortening of the red or violet 

 end of the spectrum should be distinct defects, and not neces- 

 sarily associated with defective colour discrimination. 



7. There should be innumerable varieties of dichromic vision. 



8. There should be trichromic cases of defective colour 

 perception, three colours being seen in the bright spectrum, 

 yellow being seen as red-green, and blue as green-violet. 



9. All colours when reduced sufficiently in luminosity and 

 area should appear white, the colour disappearing first in the 

 least developed portions of the retina. 



10. Simultaneous and successive contrast should be increased 

 in those with defective colour discrimination. 



11. Those with defective colour discrimination should see 

 like those with better discrimination in conditions of more 

 difficulty. 



Now all these predictions have been fulfilled, and are found 

 to be actual facts, so that whilst the facts support the theory, 

 they present difficulties to be solved by any other theor}'. 



We will now review a number of facts of colour vision in 

 order to show the requirements of any colour-vision theory. 

 Further details will be found in the papers to which references 



are given. 



I. The Facts of Colour Mixing 



The fact that any colour may be matched by a combination 

 of three selected spectral colours is the foundation of the trichro- 

 matic theory which was propounded in order to explain it. The 

 trichromatic theory, which may be represented by ax + /3y + yz = 

 any colour, is only one possible explanation of the facts. Let 

 us take, for instance, the fact that when spectral red and spectral 

 green are mixed in appropriate proportions they match spectral 

 yellow. The other explanations of this fact are that red and 

 green each contain a yellow element, when the two are mixed 



