94 HERMANN VON HELMHOLTZ 



all other colours must be derived by combination; Newton 

 assumed (without proving it by any prolonged experiments) 

 that the same results must obtain for the composition of 

 coloured light also : and on this supposition of three primitive 

 colours, Thomas Young built up his hypothesis that the 

 particles which lie upon the surface of the retina are capable 

 of specific vibrations ; that, at any spot, particles of three 

 different vibration-rates, corresponding with the oscillation- 

 frequencies of the three primary colours, are in immediate 

 juxtaposition ; and that, lastly, mixed sensations are produced 

 by excitation of the ends of the specifically reacting nerve- 

 fibres. Helmholtz now discovered that the mixture of colouring 

 substances gave quite different results from the blending of 

 spectral colours. It is only when the two colours happen to 

 lie near each other in the spectrum, that the fusion of coloured 

 light yields almost the same results as the mixture of pigments, 

 because the resulting colour then resembles the intermediate 

 colour-tones of the spectrum ; the disparity is most marked 

 in the combination of blue and yellow, which yield green 

 when pigments are mixed and white with the corresponding 

 mixture of spectral colours. Now green is among the colours 

 most imperfectly produced by the fusion of spectral colours ; 

 it was therefore necessary for Helmholtz, if the hypothesis 

 of the composition of all colours out of three elementary 

 colours was to be maintained, to select for these three 

 primaries, not red, yellow, arid blue, but red, green, and 

 violet; by the mixture of these three all the weaker com- 

 pound colours can be obtained, while if the saturated colours 

 of the spectrum are to be imitated, at least five primaries, 

 red, yellow, green, blue, and violet, are required. In order 

 to settle, in the first place, whether there are three elementary 

 colours, from which all possible colours are, or at any rate 

 may be, built up, this rrypothesis had to be tested upon 

 the prismatic colours, as the purest and most saturated. By 

 observing through a prism set vertically the spectra of the 

 two limbs of a \/-shaped slit (letting the bands of colour 

 run in the one from left above to right below, and in the 

 other from right above to left below, so that each coloured 

 band from the one spectrum intersected all the bands of the 

 other in the common field), he obtained all the combinations 



