408 M. H. Helmholtz on Sir David Brewster's 



of the glass appeared a dull yellow; through a still greater 

 thickness it appeared to be a greenish-white ; and on introdu- 

 cing other colouring matters, particvilarly solutions of copper 

 and red ink, it finally changed into white. This white he further 

 asserts is not to be decomposed by the prism ; bxit if I rightly 

 comprehend the meaning of his expressions, he has never tried 

 this by a second prism — which indeed could not be effected mth- 

 out a considerable modification of the method of experiment — 

 but infers it merely from the fact that this white light has passed 

 undecomposcd through the first prism. 



The blue glass which I had at my disposal showed the phsE- 

 nomena described by Brewster in the following manner. Seen 

 through one plate, the yellow stripe in a spectrum produced by 

 the light of the firmament was very feebly luminous, and a 

 greenish-yellow ; but in the spectrum obtained from the portion 

 of the heavens which lay near the sun it was a pure and shining 

 yellow. Observed through two plates, the stripe obtained fi'om 

 dayhght disappeared totally; with direct sunlight it appeared 

 almost white; with greater intensity of light it verged into 

 greenish-yellow, and with diminished intensity into blue. Ad- 

 jacent to this moderately illuminated band the blue and violet 

 appeared of course splendidly bright, and the extreme red was 

 also strongly luminous. Seen through three plates in direct 

 sunlight, the yellow band appeared a bluish-white. The altera- 

 tion of the colours was somewhat less when the plates, instead 

 of being introduced between the prism and the eye, were placed 

 before the aperture, that is, between the source of light and the 

 aperture. When we consider that the sun is upwards of 50,000 

 times brighter than the brightest white surface which he illumi- 

 nates, and that the yellow in the original spectrum possesses the 

 intolerable brightness of the sun, but seen through two blue 

 glass plates appears as a moderately illuminated surface of paper, 

 in the absence of more exact measm'ements we shall not be far 

 from the truth in assuming that the hundredth part of the yel- 

 low passes through one glass, and the ten-thousandth through 

 two. Now supposing that only the ten-thousandth part of the 

 coloured light which passes unweakened through the plates is 

 caused, by the little irregularities before spoken of, to fall upon 

 the same portion of the retina as that which receives the yellow, 

 we must certainly obtain colours very different from the latter. 

 By the mixture of indigo-blue light with yellow, we obtain, as I 

 have shown in my investigation on compound colours, first a 

 whitish yellow, then white, which finally passes into a bluish- 

 white. The colours in the smalt-glass spectrum which lie next 

 to the yellow, namely red and green, can, by mixing in various 

 proportions, cause the white to approximate to red or green, as 



