Mat 26, 1922] 



SCIENCE 



559 



appearing color pairs — they produce a more 

 primitive white, or yellow (see above). If these 

 facts are held distinctly in mind, the appro- 

 priate chemical conception almost forms itself. 

 I represent, purely diagrammatically, of course 

 (Burdon-Sanderson^ especially noted this point 

 when my theory first came out) that portion of 

 a molecule which is capable of being dissociated 

 out of light in the way indicated in Figure 4. 



Stage I 



Stage III 



Fig. 5. The cleavage products in tlie three stages 

 of the color sense. This diagram does not rep- 

 resent the entire light-sensitive molecule, but 

 only the specific cleavage products which, ac- 

 cording to the Ladd-Frankhn theory, constitute 

 the several nerve excitants for the color sensa- 

 tions (From Woodworth's Psychology). For 

 other diagrams, see Psychological Jteview, 23: 

 247, 1910 ; Zeitschrift f. Psychologie, Bd. 6, etc. 



The development required is that of a greater 

 and greater specificity to the electro-magnetic 

 vibrations of the visible spectrum. A portion 

 of a molecule which at first is broken off indif- 

 ferently by the whole spectrum becomes in a 

 second stage more specific, — by a fresh aggre- 

 gation of atoms a portion T responds to the 

 yellow end of the spectrum, a portion B to 

 the "blue end of the spectrum. But what 

 happens when yellow and blue light fall at 

 once on this chemical substance? The T and 

 the B (since they are the chemical constituents 

 of W, because the assumption is that they were 

 segregated out of it) will chemically unite and 

 will produce W, the nerve-excitant of the sensa- 

 tion white. In the same way in the third, and 

 latest, stage, the newly segregated B and G, 

 when torn off from the molecule by light of 

 low and of high middle frequency, will revert 

 to the mother substance Y ; and if light of high 

 frequency, "blue," is now added, we shall again 

 have the nerve-excitant of white. That is to 

 say, just as when we have in a test-tube the 



6 Presidential Address, British Association, 

 Nature, Vol. 48, p. 469. 



chemical constituents of HCl (namely, H and 

 CI) they chemically unite, under proper con- 

 ditions, and produce HCl, so in a cone we have 

 B -]-G = Y 



Y + B(=B-l^GJ^B) — W 



Observe that we have now, quite incidentally, 

 explained how it happens that lights of only 

 three specific, homogeneous wave-lengths 

 ("red," "green" and "blue") are sufficient to 

 reproduce the whole gamut of color sensations, 

 including yellow and white, which the psys- 

 icists have never noticed the existence of.'' Yel- 

 low is a secondary product, and so is white, but 

 they are both perfectly good unitary sensations. 

 The theory explains at the same time, of course, 

 how it is that the primitive white mediated by 

 the rods is the same sensation as the white 

 made out of (in the highly developed cones) 

 yellow and blue, or red and green and blue. 

 There is no reason, of course, as Professor Carr 

 has pointed out to me, why there should not be 

 also some of the more primitive chemical sub- 

 stances in the cones of the central retina. 



Since the interesting work of Hecht, it seems 

 to be quite certain that the first effect of light 

 on the retina is photo-chemical (which is, of 

 course, the same thing as electrical). It is 

 here, without question, that is found that 

 "transformer mechanism," as I have called it, 

 by which what should look to us like 165 

 different colors in the spectrum is replaced by 

 a paltry four — the best that Nature could ac- 

 complish with only one small cone to work in. 

 Five unitary colors (including white) and mix- 

 tures of them — the color blends — are all that 

 we can see in the 30,000 discrimindble sensa- 

 tions that are given us by light. It is nerve 

 impulses produced by retinal chemical stimuli 

 of the character which I have described (or of 

 some other character) that mediate the pro- 

 cesses which take part in the final "neuro- 



' Crowther, a prominent English physicist, actu- 

 ally says that the white produced by mixing 

 homogeneous yellow and blue light-rays is not a 

 "real" white— that, he thinks, rmist be a whole- 

 spectrum white. Professor Titchener, on the 

 other hand, has called my attention to the fact 

 that on my view white is always (aside from the 

 most primitive stage) due to a union of yellow 

 and blue constituents. 



