M. H. Helmholtz on the Theory of Compound Colours. 531 



fixed at right angles to the table at about tlie distance of distinct 

 vision. The observer looks obliquely through the j^latc down- 

 wards towards the table, and places the wafers in a position 

 which is suitable for the combination of their colours. The 

 nearer both are brought to the imaginary intersection of the 

 plane of the table with the glass plate, the more obliquely will 

 the rays fall u])on the ])late, the fewer will pass through, and 

 the greater will be the number reflected ; so that, in this case, 

 the colour of the reflected light will be jjredominant. Conversely, 

 the colour of the transmitted light will be predominant when the 

 wafers are removed to a distance from the line of intersection ; 

 and in this way it is possible to alter the relative intensities of 

 the combining colours in any required degree. In this experi- 

 ment both wafers are placed upon a black ground* ; or if whitish 

 combinations of colour be required, which it is necessary to com- 

 pare with pure white, one of the wafers (the brightest is best) 

 is placed upon a white, the other upon a dark ground. Observed 

 through the glass plate, the wafer appears in the compounded 

 colour upon a white ground. It is manifest that in this way 

 the colours of all coloured surfaces whatever, as also those of 

 coloured glasses, may be combined. 



Colours thus composed are distinguished by their brightness 

 and clearness from those obtained l)y the mixture of the colour- 

 ing matters ; they do not always agree with the latter, but, on 

 the conti-ary, yield the same results as those obtained when the 

 prismatic colours are united. Blue and yellow, in particular, do 

 not give green, but white. As the representative of the yellow, 

 I made use of paper disks which I had v/ashed with bright 

 chrome-yellow or gamboge. Of blue colouring matters laid on 

 disks in the same manner, a beautiful sky-blue cobalt gave, with 

 both kinds of yellow, pure white ; artificial ultramarine, reddish 

 white ; and bright Prussian blue, a weak greenish white. Cin- 

 nabar combined with blue gives rose-colour ; the same red colour 

 combined with green gives yellow, &c. In short, these experi- 

 ments prove, that not only the simjile coloured rays of the spec- 

 trum have other laws of action than those hitherto generally 

 assumed, but that quite similar laws apply to the combination of 

 the colours of pigments. It does not appear to me doubtful that 

 these new laws will su})ersede the old ones which were based upon 

 the mixture of the colouring substances. 



It is best, however, to commence with the simple colours of 



* I have repeated the experiment with a single yellow wafer on a blue 

 ground, the image of the wafer j)rojected upon the hitter gives u white 

 spot. — J. T. 



3 M :;i 



