826 



NATURE 



[August 25, 192 1 



New Facts of Colour Vision. 



Bv Dr. F. W. Edridge-Green. 



n^HE White Equation. — The fact that when two or 

 -^ three simple spectral colours are combined a 

 white is produced which matches that from which 

 the spectrum has been formed is the basis of many 

 theories of colour vision. It is therefore of funda- 

 mental importance to any theory of colour vision. 



In a recent paper (Proceedings of the Royal Society, 

 B, vol. xcii., 192 1, p. 232) it was pointed out that 

 when an exact match of a red of A 6670-6770 A., a 

 green of A5144-5156 A., and a violet of A 4250-4267 A. 

 with white was made, after fatigue with r^d light in 

 the region of A b'jofxji, there was no longer a match 

 between the simple and mixed whiter, the mixed 

 white appearing bright green, an^ in (order that a 

 match could be made the green l^d to oe reduced to 

 about one-half of the amount 'require(r> by an un- 

 fatigued eye. It is obvious, therefore, that the under- 

 lying physiological processes ar^ -no| the \ same with 

 the mixed and simple whites. It shtjuld be noted, 

 however, that no change rln the equation is seen 

 when the eye is fatigued/with red light in the region 

 of A 780/x/x. ^ — ^ / 



Another factrof ^b^»t6ur fatigue bears on this point; 

 red of A 6yofifj. can be matched with red of the end 

 of the spectrum by varying the intensity, and so it 

 has been stated that red A SyofxiJi, as well as the 

 terminal red, affects only the hypothetical red sensa- 

 tion. If, however, the eye be fatigued with red of 

 the region of A 760/i/x, and red of the region A Oyo/xfi 

 be afterwards viewed, this appears yellow, or even 

 greenish-yellow, which could not be the case if the red 

 sensation only had been affected. 



The Change of Hue produced by the Addition of 

 White Light to Spectral Colours. — White light is a 

 purely relative term. The white light of the sun is 

 not the same as that from an artificial source ; the 

 term is therefore employed as meaning the combined 

 light of the source which is used. In making the 

 experiments described, the light was that of a 

 looo-candle-power tantalum arc, which, compared with 

 sunlight, is 3'ellow. The apparatus used in these ex- 

 periment^ was that described in the Proceedings of 

 the Royal Society, B, vol. xcii., 1921, p. 232. 



Various spectral colours were isolated on a screen 

 coated with magnesium oxide, and definite proportions 

 of white light taken from the source added. The scale 

 of white light is arbitrary, the maximum amount of 

 light it is possible to add being 100 divisions, A 

 comparison white light taken from the source was 

 used. Each colour became less saturated on adding 

 white light. Red first became orange, then yellow. 

 Orange became yellower. A 585/x/i, pure yellow, did 

 not change in hue. Orange-yellow and yellow-green 

 became yellow. Green became yellow-green. Blue, 

 A 480/ji/x, became white, the comparison white appear- 

 ing yellow. The violet end of the spectrum from 

 A 480JU/X, making a blue on the screen, changed to 

 violet on adding 33 divisions of white light ; light 

 purple, on adding 100 divisions. Wave-length 585/x/x, 

 the point where the addition of white light produces 

 no change of hue, is also the centre point of pure 

 yel'ow and the apex of the luminosity curve. 



The result of these experiments shows that the com- 

 ponent part of white light which has the greatest 

 luminosity effect is the hue to which all colours tend 

 on the addition of white light. 



The Anomalous White Equation without Colour- 

 blindness. — Just as a man mav make an anomalous 

 Rayleigh equation without any evidence of colour- 



NO. 2704, VOL. 107] 



i blindness (Proceedings of the Royal Society, B, 

 ! vol. Ixxxvi., 1913, p. 164), so may a man make an 

 I anomalous white equation without being colour-blind. 

 As an example of this, a man was examined who pre- 

 sented no sign of colour weakness. He passed my 

 card test, lantern test, and spectrometer with the ease 

 and accuracy of an absolutely normal-sighted person. 

 His luminosity curve was taken by the flicker method 

 and corresponded with the normal. The wave-length 

 of the apex of the luminosity curve was at 585/i/i, 

 which is the normal point. When, however, his white 

 equation was taken, he put onlv eight scale-divisions 

 of green instead of thirteen and a half or fourteen, 

 which is normal, and the mixed light appeared red to 

 the normal-sighted. An important fact was noted, 

 namely, that after fatigue with red in the region of 

 A SyofifjL the equation changed to him in the same way 

 as the normal-sighted, and he required only four scale- 

 divisions of green instead of eight. It is quite obvious 

 that this was not a case of partial red-blindness. 



The White Equation and Colour-hlitidness. — The 

 colour-blind have been classified by some as red- or 

 green-blind, in accordance with their white equations, 

 those who put too much red in the equation being 

 classed as red-blind and those who put too. much 

 green in the equation being classed as green-blind. 

 There are, however, many who, whilst agreeing with 

 the normal equation, are quite satisfied when a con- 

 siderable additional amount of green or red is added 

 to the equation. This explains why in certain cases 

 some have been described as red-blind by one ob- 

 server and green-blind by another. 



A remarkable fact which does not seem to have 

 been previously observed is that many colour-blind per- 

 sons who strongly object to the normal match, but are 

 satisfied with an anomalous equation, will completely 

 agree with the normal equation when the comparison 

 white light is increased in intensity so that it is much 

 too bright to a normal-sighted person. This clearly 

 shows that the normal mixed white produces the 

 same effect so far as colour is concerned, but has a 

 more powerful effect as to luminosity. This is in 

 complete accordance with other observations, and is 

 found in those cases in which there is abrupt and 

 slight shortening of the red end of the spectrum. If 

 there be shortening of the red end of the spectrum 

 which does not affect A 6-jonn, and A 670^/x has its 

 normal light value, the mixed light will be more 

 luminous than the simple white in the exact propor- 

 tion of the shortening. This portion of red light not 

 producing any effect has to be subtracted from the 

 white light. 



These facts are quite inconsistent with a hypo- 

 thetical red sensation which is affected by light of all 

 wave-lengths. Another illustration may make this 

 point clear. A man with shortening of the red end 

 of the spectrum and normal colour discrimination 

 will put together as exactly alike a pink and a blue 

 or violet much darker. If, however, the pink and 

 blue be viewed by a normal-sighted person through 

 a blue-green glass which cuts off the red end of the 

 spectrum, both will appear Identical In hue and 

 colour. This proves conclusively that the defect is 

 not due to a diminution of a hypothetical red sensa- 

 tion, because all the rays coming through the blue- 

 green glass are supposed to affect the red sensation, 

 and yet we have been able lo correct the erroneous 

 match by the subtraction of red light. On the other 

 hand, there are colour-blind persons who, whilst dis- 



