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



compound light formed by them must in the prismatic analysis 

 behave as simple light. To this Brewster replies, that such rays 

 may be separated from each other by taking advantage of their 

 difference of absorption in coloured media ; and he has attempted 

 by means of this method to prove, that in all portions of the 

 spectrum, rays of all three descriptions, and consequently the 

 white light due to their union, is to be found. The facts which 

 he calls in to his support prove, he considers, that homogeneous 

 light, in the sanse of Newton, that is, light composed of rays 

 of equal refrangibility (wave-length) only, sometimes suffers 

 a change of colour in its passage through coloured media, while 

 the universally accepted theory of Newton asserts that the colour 

 of homogeneous light depends solely upon its refrangibility 

 (wave-length) ; that such light may be weakened, nay, com- 

 pletely extinguished, in its passage through coloured media, but 

 can never exhibit a change of colour. We must certainly grant 

 that, if a single case be established in which the colour of homo- 

 geneous light is changed by absorption in a coloured medium, 

 Newton's theory must be abandoned, and that of Brewster, or 

 one similar, must be assumed in its stead. 



I remark in the first place here, that the number and nature 

 of the three primitive colours assumed by Brewster are based 

 upon indirect inferences. In this respect he has retained the 

 pretty generally received theory of the mixture of colours, accord- 

 ing to which red, yellow, and blue are the components of all 

 others; yellow and blue, for example, producing green. I have 

 shown in another place* that this theory is based upon the 

 results of the mixture of the coloured substances merely, but that 

 the mixture of such substances is by no means equivalent to the 

 mixing of lights of the same colours, — to cite a particular example, 

 yellow and blue united do not produce green, but white. The three 

 colours, red, yellow and blue, can therefore compose no green, 

 and must, if we are to retain the idea of three primitive colours, 

 be superseded by others, say red, green and violet, as already 

 assumed by Thomas Young. By this alteration Brewster's theory 

 would undergo no essential change; single conclusions only 

 would require modification. I will therefore not enter further 

 upon this subject here, but limit myself to the investigation of 

 the question, " Is the colour of homogeneous light altered by 

 coloured media or not ?" 



Hitherto Airyf, Draper J, and Melloni§ have sought to refute 



* " Uebcr die Theorie der zusaiiimengesetzten Farben," Miiller's Archiv 

 fur Anatomie und Phyniolor/ie, 1852. 



t Phil. Mag. vol. xxx. p. 73 ; Pogg. Ann. vol. Ixxi. p. 393. 



J Silliman's Journal, 184/, vol. iv. p. 388 ; Phil. Mag. vol. xxx. p. 345. 



§ Bibl. Univ. de Gen. Aoftt, 1847 ; Phil. Mag. vol. xxxii. p. 262. 



