404 Dr. R. A. Houstoun on a 



the separation is greater still, the resultant curve has two 

 maxima separated by a hollow. This is the cnse of blue 

 superposed on red to give purple. Many educated people 

 who have not studied light regard purple as composite, as 

 something analogous to a chord in music ; this view finds its 

 natural expression in the two maxima of the curve. 



The general rule for the mixture of spectral colours is, 

 that if they are not far apart in the spectrum, the mixture 

 has the same hue as an intermediate spectral colour, but is 

 paler than the latter. If the separation is increased, the 

 two colours become complementarj', and the mixture is 

 white. This is what we would expect if the effect of mixing 

 colours is obtained by superimposing curves of the type 

 shown in fig. 1. 



§ 2. It is, however, not sufficient to explain the facts of 

 colour mixing qualitatively ; a quantitative explanation 

 must also be given. In order to do this, it is first of all 

 necessary to find out what the facts are ; this is not so 

 easy, because the different investigators are not in good 

 agreement. 



The principal papers are by Maxwell *, Konig and Die- 

 tericif, Abney % (1900), Abney § (1906), and Rayleigh||. 

 Any colour whatever can be expressed in terms of three 

 fundamental colours. It is the object of the first four 

 papers cited to do this for the spectral colours. The results 

 are given in different units and it is impossible to compare 

 them. Konig and Dieterici, indeed, say that their results 

 are in essential agreement with Maxwell's, but no comparison 

 has as yet been made between Abney's results and those of 

 Konig and Dieterici. Maxwell gives two sets of results, 

 one for an observer J (himself) and another for an observer- 

 K, Konig and Dieterici each give a set of figures, and Abney 

 gives two, one in each of his papers. As a preliminary to 

 studying the results I found it necessary to reduce all six 

 sets to the same units and exhibit them in parallel columns. 



* " On the Theory of Compound Colours, and the Relations of the 

 Colours of the Spectrum," Phil. Trans, cl. p. 57 (1861). 



t " Die Gvundempfindung-en und ihre Intensitatsvertheilung ini 

 Spektruni," Bed. Ber. p. 805 (1886). 



X "The Colour Sensations in Terms of Luminosity," Phil. Trans, 

 cxciii. A, p. 259 (1900). 



§ " Modified Apparatus for the Measurement of Colour and its 

 Application to the Determination of the Colour Sensations," Phil. Trans, 

 ccv. A, p. 833 (1906). 



|| " On Colour Vision at the ends of the Spectrum," Scientific Papers, 

 vol. v. p. 569. 



