M, H. Helmboltz on the Theory of Compound Colours. 533 



choose the three simple ones which admit of the least perfect 

 imitation, namel_y red, green, and violet : we should then obtain 

 a j'ellow and blue, which, when compared with the colours of 

 our pigments, would appear saturated, but which would not bear 

 comparison with the yellow and blue of the spectrum. These 

 are the three which Thomas Young proposed as the three primi- 

 tive colours. Red, green, and blue would not answer so well ; 

 for were these three chosen, the mixed violet would appear worse 

 than the mixed blue of the former three. The three primitive 

 colours commonly chosen are altogether insufficient, because from 

 them green can never be obtained. 



According to the above we must also abandon the theory of 

 three jjrimitive colours, which, according to Thomas Young, are 

 three fundamental qualities of sensation. If the sensation of 

 yellow by the yellow rays of the spectrum were due to the fact 

 that by them the sensations of red and green were simultaneously 

 excited, and both working together produced yellow, exactly the 

 same sensation must be excited by the simultaneous action of 

 the red and green rays ; nevertheless by the latter we can never 

 obtain so bright and vivid a yellow as that produced by the yellow 

 rays. The same remarks apply to blue, which would be formed 

 from the mixture of green and violet ; and to violet, which would 

 be formed from the mixture of blue and red. To retain in this 

 sense the theory of primitive colours, five such, at least, must be 

 assumed. On the contrary, to represent and classify the dull 

 and comparatively impure colours of natural bodies, in the sense 

 of Lambert and Forbes, three primitive colours would be quite 

 sufficient. But, for a sure and a scientific classification, it would 

 be necessary to apply a method of combining colours diffiirent 

 from the mixing of pigments. 



By the union of every two simple colours we are met by two 

 new impressions, namely white and purple-red, with their de- 

 grees of transition into the simple colours before named. The 

 purple-red belongs to the saturated colours which cannot be 

 otherwise obtained than from the extreme red and violet, with- 

 out a loss of brightness. AVhite, on the contrary, can be obtained 

 in an infinity of ways, without the eye being able to distinguish 

 one white from the other. Wc obtain it for example from simple 

 yellow and blue, from simple red, green, and violet, or from 

 these five simple colours taken together ; and besides these, from 

 several more complicated combinations. In contrast with coloxu's 

 it is therefore regarded as indifferent light. The remaining 

 combinations of evei-y two simple colours a])pcar to the eye as 

 transitions of the simple colours and purple into white; but in 

 further combinations, as above remarked, they behave in a 



