We can test the truth of this argument in a very simple way. If we add to the colour with which the match has to be made a small quantity of white light from the reflected beam, cutting off more or less by the rotating sectors, we can get the exact hue of the impure blue-green made by the mixture of the colours coming through the two slits; and further we shall find that the amount of white added corresponds with the amount of red which would be required when the components of the white light in the terms of the three colours are taken into account. For spectrum colours between the violet and the green it may therefore safely be said that no match can be effected by the mixture of violet and green light; but that it always gives the intermediate colour diluted with white light. For colours between the green and the red of the spectrum, a very close, if indeed not an exact match, can be made with the red and green slits, without the addition of white. If we take from the second apparatus light from above the position of the violet slit in the first apparatus, that is, nearer the limit of visibility, it will be found that a match is made, for at all events a very considerable way with the violet slit alone, by merely reducing the aperture, thus showing that the colour is the same, only less intense. In the same way it will be seen that the rays coming from any point between the lower limit of the spectrum to a little below the C line are identical in colour. As we have arrived at the fact that in colour mixtures of violet and green, white light is to be found in the colour produced, it follows that either the violet or the green, or both, must themselves contain some small proportion of white. It might perhaps be said that violet is really a mixture of red and blue, and hence the white in the mixture with the green; but if in the first apparatus we place one slit in the purest blue we can find, and the other in the red, and throw a violet patch on the screen from the second apparatus, we shall be unable to form the same hue of violet by any means; it will always be diluted with white. Now as the very blue we are using, if matched as above by green and violet, requires white light to be added to it, and as to match the violet with the same blue and red, white light has also to be added to it, it follows that the violet must be freer from white light at all events than the blue. There is one other experiment that must be mentioned before leaving for a time this part of our subject, viz. the formation of white by a mixture of yellow and blue. If one of the slits be placed in the yellow of the spectrum, a position will be found in the blue where, if a second slit be placed, and the apertures are adjusted, an absolute match with the reflected white of the apparatus can be secured. This experiment will be referred to later on, when considering the question of primary colours.