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



the solar spectrum, because these are the purest and most per- 

 fect, and even with a moderate intensity of light make an almost 

 dazzling impression ; beside which, all pigmentary colours appear 

 diUl and gray. Newton has already given the rule, that each 

 simple colour can be obtained from the miion of the two next it. 

 My owTi investigations corroborate this. I must, however, at the 

 same time remark, that the distance of the combined colours 

 must not be too great, when it is sought to obtain from their 

 union a colour similar to that which lies between them. This is 

 particularly the case in the central portion of the spectrum. Red 

 and yellow give an orange, the appearance of which appears to 

 be quite the same as that of the simple orange; and, in like 

 manner, the indigo which results from the combination of blue 

 and violet is scarcely to be distinguished from the simple indigo. 

 On the contrary, yellow-green and blue- green give a green the 

 tone of which indeed corresponds to the intervening tone of the 

 spectrum, but which is decidedly didler and more whitish, so that 

 the simple green can be obtamed from such colours only as 

 scarcely differ from it in appearance. Yellow and blue appear 

 in this respect less sensitive than green. The former may be 

 pretty well obtained from orange and yellow-green, but very pale 

 from red and green ; the latter, again, may be well obtained 

 from the combination of blue-green and indigo, but is very dull 

 when formed of green and violet. With regard to the end colom-s 

 of the spectrum, red and violet, Newton in his coloured disk 

 places them in contact with each other, and subjects them to the 

 rule which refers to the union of neighbouring colours mentioned 

 above. From indigo-blue and very little red it is indeed possible 

 to generate a kind of violet, which however always approximates 

 more to white or rose-colour than the simple violet. Much more 

 incomplete appears to my eyes the imitation of red by orange 

 and violet ; their combinations always pass into tones of carmine- 

 red or of white, and I have not succeeded in obtaining a tolerable 

 imitation of the pure red of the spectrum. 



Hence if we pro])ose to ourselves the problem of imitating the 

 colours of the spectrum by the union of the smallest possible 

 number of simple colom-s, we tind at least live of the latter 

 necessary for this purpose, namely red, yellow, green, blue, 

 violet. I must, however, leave the question undecided, whether 

 these are completely sufficient, and whether with better appa- 

 ratus, which would permit of the illumination of larger surfaces 

 by the simple colours, and by the corresponding compound ones 

 placed adjacent, a practised eye might not detect diiferences 

 which with my ajjparatus could not be recognized. If, however, 

 we wish to limit ourselves to three colours, it would be best to 



