fi44 PHYSIOLOGY 



away than would correspond to the complementary colours we obtain 

 again a coloured sensation, which, however, is unsaturated, being 

 mixed with a certain amount of white light. By taking three colours, 

 such as red, green, and violet, or four colours, such as red, yellow, 

 green, and violet, it is possible by mixing them in various proportions 

 to form either white light or any colour of the spectrum, besides 

 the various purples which do not occur in the spectrum. 



These experiments of mixing colours can be carried out in various ways : 

 In every case we aim at stimulating the retina simultaneously with rays 

 of different wave-lengths, or successively at such short intervals of time that 

 there is complete fusion of the sensations resulting from the individual excita- 

 tions. In order to determine fine differences in shade a coloured surface is 

 always provided with which a colour, produced by the fusion of the different 

 rays under experiment, may be compared : 



(1) The most exact method of mixing colours is to employ a couple of spectra 

 and by means of prisms to bring different parts of the spectra on one and the 

 same white surface on which the result of the mixture can be compared with 

 sample colours. 



(2) In Maxwell's ' colour top' discs with a radial slit are placed one over 

 the other on a disc which can be made to revolve with considerable rapidity. 

 By means of the slit two or three discs of different colours can be slid one into 

 the other, so that the disc is composed of sectors of variable extent of the two 

 or three colours which we wish to mix. These discs are generally mounted 

 on a background of a larger disc, which can be used as a sample colour to com- 

 pare with the result of the mixture of the colours in the centre. If we are 

 determining the relative amount of different colours necessary to produce white, 

 the outside disc would be partly white and partly black, so that on rotation 

 the effect of grey is produced, i.e. a weak white. Thus in one experiment the 

 small discs in the centre were red, green, and blue. It was found that when 

 the sectors were chosen in the following proportion : 165 red, 122 green, and 

 73 blue, a grey colour was produced equal to the grey obtained in the outer 

 disc by mixing 100 white with 260 black. 



These methods and all others, in which coloured surfaces are used 

 suffer from the defect that no pigments give perfectly pure colour- 

 sensations. On looking at a red painted surface, for instance, in 

 a bright light with a spectroscope, it will be found that the spectrum 

 contains yellow and green rays as well as red. On this account no 

 information as to the effect of mixing colour-sensations can be 

 obtained by mixing the pigments themselves. Thus a familiar way 

 of producing a green pigment is to mix a blue and a yellow pigment. 

 A mixture of blue and yellow rays will give a sensation of white. 

 The fact that blue and yellow pigment mixed together give a green 

 pigment is due to the fact that the blue pigment cuts oft the red 

 and yellow rays, and reflects, or allows to pass, the blue and green 

 rays, while the yellow pigment retains the blue and violet rays, but 

 reflects the red, yellow, and green rays. When we mix the two 

 we get, not an addition, but a subtraction ; the blue pigment absorbs 

 the red rays which the yellow pigment allows to pass, and the yellow 



