Mat 26, 1922] 



SCIENCE 



557 



700 



600 600 400 



Fig. 1. B, G, and B, resonance curves. These are the curves of Kbnig and 

 Dieterici corrected to new determinations of the points of section, a, 6, c, d. 

 Abscissae, wave-lengths of the interference spectrum of the arc light; ordi- 

 nates, arbitrary scale. (F. Esner.) 



two types of yellow-and-blue vision)- that the 

 number of "primary" colors is three and not 

 four. (Fig. 1). This is fact. (2) The Hering 

 school has only to ask for the most cursory ex- 

 amination of the gamut of color-sensations to 

 show that the nmnber of its different chromatic 

 constituents is four, not three. This is fact. 

 But the group of facts subsumed under (1) — 

 the facts of "matching by mixture" (Fig. 2)- — 

 is absolutely incompatible with the theory of 

 Hering,^ and the group of facts subsumed 

 under (2) is absolutely incompatible with the 

 theory of Helmholtz. It is little to the credit of 

 any association of scientists (for instance, the 

 Optical Society of America) that they still 

 solemnly discuss the theories of Helmholtz and 

 of Hering. The situation is simple : each of 

 these theories is absolutely refuted by the facts 

 which are the groundwork of the other. 



1 have devised a simple diagram by means 

 of which one can keep in mind the impossibility 

 at once of the Helmholtz and of the Hering 



2 Dictionary of Philosophy and Psychology, Art 

 Vision, II, 788. 



3 The attempt of v. Kries to supplement the 

 Helmholtz theory by supposing that the three 

 colors resolve themselves into four at a higher 

 level of the visual nerve system is a purely ad hoe 

 hypotheses, and without significance. See my arti- 

 cles on ' ' The Theory of Color Theories, ' ' Comptes 

 rendiis du V/e Congres intern, de Psychologic, 

 Geneve, 1909, and Psychological Review, May, 

 1922. 



theory. Color diagrams are immensely more 

 illuminating if they are done up in color.* 

 But lacking that one can make shift with ap- 

 propriately striated surfaces. I call this dia- 

 gram my Quadrigeminal Color Body (a term 

 suggested by the corpora quadrigemina) , but it 

 is at the same time triaxial. (The triangle 

 should always be drawn with the YB line a 

 horizontal (fundamental) line, as indicative of 

 the fact that yellow and blue were developed 

 first — that red and green were a later addi- 

 tion. (The actual spectral line approaches 

 nearer to the point W in the green region on 



Fig. 2. The color triangle, which exhibits ' ' match- 

 ing by mixture. ' ' 



* Stoelting is putting on the market for me my 

 complete set of colored color diagrams. 



