522 M. H. Helmboltz on the Theory of compound Colours. 



and green sensations would be simultaneously aroused, and tbus 

 the mixed sensation of yellow generated. 



I have been equally unsuccessful with Forbes in my efforts to 

 find among Newton^s followers, up to the latest period, experi- 

 ments on the mixture of the single prismatic colours. It appears 

 as if the question was regarded as completely exhausted by the 

 experiments with the mixed powders. Even the divergent results 

 given by the rotating disk were insufficient to convince experi- 

 mentei's that difficulties lay concealed here. 



The referring of all colours to the three primitive ones has, in 

 the case of the different observers, three different senses : — 



1. That the primitive colours were s\ich as permitted of the 

 formation of all others from their combinations. 



2. Or, as supposed by Mayer and Brewster, that the primitive 

 colours correspond to three different kinds of objective light. 



3. Or, as supposed by Thomas Young, that they correspond 

 to three primitive modes of sensation experienced by the visual 

 nei'ves, and from which the remaining sensations of colour are 

 composed. 



To the second of these views and the reasons by which Brewster 

 has endeavoured to support it, I will return in another place, and 

 believe that I am in a position to refute it. The two others 

 must, at all events, be tested by the prismatic colours, these 

 being the pm'est and most saturated that we possess. This shall 

 be the object of the present paper. 



The means which I have made use of to obtain the combi- 

 nations of the colours of the spectrum, two by two, is as follows : 

 I cut in a black screen two sufficiently narrow slits {\ of a line 

 wide) which together form a V. Both are inclined at an angle of 

 45° to the horizon, the angle which they enclose being thus a right 

 angle which points downwards. This slit is observed from a 

 sufficient distance (12 feet) through a telescope and prism. The 

 prism is placed close before the object-glass of the telescope, in 

 the position of minimum deflection, and the edge of its refracting 

 angle stands vertical. It is known that, looking through a ver- 

 tical prism at a vei'tical slit, a rectangular spectrum is observed 

 in which the coloured bauds and the lines of Fraunhofer occupy 

 a vertical position. If through a vertical prism an oblique slit 

 be observed, the spectrum assumes the form of an oblique-angled 

 parallelogram, with two opposite sides horizontal and two others 

 parallel to the inclined slit. The bands of colour and the lines 

 of Fraunhofer are here, of course, parallel to the slit. When our 

 compound angular slit is thus observed, the spectra of its two 

 legs partially cover each other, and, as in the one the bands of 

 colour are directed from the left above to the right below, in the 

 other from the right above to the left below, they mutually in- 



