1224 MIXING OF COLOURS. [BOOK m. 



on the other hand absorbs the red and yellow but reflects the 

 blue and a good deal of the green. Hence when we look at a 

 yellow gamboge patch our retina is excited not by those rays 

 alone which form the yellow of the spectrum, but by many other 

 rays as well; the colour is not a 'pure' colour, does not cor- 

 respond to one of the colours of the spectrum, but is a mixture of 

 more than one. And this is the case with most pigments ; hence 

 when they are employed in experiments on the mixture of 

 sensations, difficulties and even errors arise which are avoided 

 by the use of the colours of the prism. We may here incidentally 

 remark that mixing the sensations excited by looking at pigments 

 gives very different results from mixing the pigments them- 

 selves. Thus when gamboge and indigo are mixed the mixture 

 is green because the gamboge absorbs the blue and the indigo 

 absorbs the red and yellow, while both reflect the green. We 

 shall see presently that when the sensation excited by gamboge 

 is mixed with the sensation excited by indigo the result is a 

 sensation not of green but of white ; and we shall see why this 

 is. What we have just said with regard to surfaces coloured 

 with pigments applies also to glasses stained with pigment, it 

 being understood that the colour of stained glass, seen as a 

 transparent object, corresponds to the rays which it does not 

 absorb. When pure pigments, i.e. pigments corresponding as 

 closely as possible to the prismatic colours, are used, satisfactory 

 results may be gained, either by using the reflected image of one 

 pigment, and arranging so that it falls on the retina at the same 

 spot as the direct image of the other pigment, or by allowing the 

 image of one pigment to fall on the retina before the sensation 

 produced by the other has passed away. The first result is easily 

 reached by the simple method of placing two pieces of coloured 

 paper a little distance apart on a table, one on each side of a glass 

 plate inclined at an angle. By looking with one eye down on the 

 glass plate the reflected image of the one paper may be made to 

 coincide with the direct image of the other, the angle which the 

 glass plate makes with the table being adjusted to the distance 

 between the pieces of paper. In the second method, the ' colour 

 top ' is used ; sectors of the colours to be investigated are placed 

 on a disc made to rotate very rapidly, and the image of one colour 

 is thus brought to bear on the retina so soon after the image of 

 another that the two sensations are fused into one. 



757. When by any of the above methods sensations corre- 

 sponding to the red and yellow of the spectrum are mixed together 

 in certain proportions the result is a sensation of orange, quite 

 indistinguishable from the orange of the spectrum itself. Now 

 the latter is produced by rays of certain wave-lengths, whereas 

 the rays of red and of yellow are respectively of quite different 

 wave-lengths. The orange of the spectrum cannot be made up 

 by any mixture of the red and the yellow of the spectrum in the 



