Passing through a prism is not, however, the only means by which white light can be split up. When the sun shines upon a blue flower, for instance, the blue petals perform a partial separation; they reflect the blue part of the sunlight, and absorb all the rest. A red flower likewise reflects the red part of the sunlight and absorbs the rest. It is because things can thus discriminate, reflecting some kinds of light and absorbing the remainder, that we perceive things in different colours.
It follows, therefore, that when we look upon a landscape, or a field of flowers, we receive into our eyes an enormous variety of coloured lights. The white sunlight furnishes each thing we see with a flood of white light, and each thing according to its nature, reflects more or less. A white flower reflects the whole, a pure black object reflects none, but the great majority of things reflect some part or other of that infinite variety of which white light really consists.
So a view at all varied sends to our eyes a variety of colours, almost as manifold as the colours of the spectrum, which, as has been said, are infinite. And the task of reproducing them, or even of producing a similar general effect, upon a piece of paper seems at first sight beyond the bounds of possibility.
But fortunately there is a way by which we can produce, approximately at all events, the intermediate colours by mixtures of the others. The second colour of the spectrum, for example, orange, can be obtained by mixing its neighbours on either hand—namely, red and yellow. We can, indeed, imitate very closely the imperceptible change from red to yellow through orange, by skilful mixture of red and yellow pigments. First there is the pure red, then just a suggestion of yellow is added; more and more yellow brings us to orange; after which by gradually diminishing the amount of red we reach the pure yellow. Next, by introducing blue pigment, we can gradually change the yellow into green, and further manipulation of the same two colours will lead us on to pure blue. Indeed by mixtures of red, yellow and blue we can obtain almost all the perceptible varieties of colour.
And it must be remembered that when, by mixing blue and yellow pigments, we get the effect of green, that is only the result of an optical illusion. The particles of which the yellow pigment is made remain yellow, and the particles of blue remain blue. The one sort reflect yellow light to our eyes, the other sort reflect blue light, and owing to what in one sense may be called a defect in our vision, these two mingling together look as if the whole were green. In the spectrum we see real green light; from green paint made by mixing yellow and blue, we only see an imitation or artificial green. If the particles were large enough, we should see the yellow and the blue ones quite separate, but since they are too small for us to see at all, except in the mass, our eyes blend the whole together into the intermediate colour.
Thus we see that, although the variety of colours is infinite, we can for practical purposes reproduce as much difference as our eyes can perceive by the judicious blending of three—namely, red, yellow and blue.
And there is a further fortunate fact—we can filter light. The red glass with which the photographer covers his dark-room lamp looks red, and throws a red light into the room, because it is acting as a filter to the light proceeding from the lamp behind it. The lamp is sending out light of many colours, but the glass is only transparent to the red. It holds up all the others but lets the red pass freely. So if we were to take a photograph through a red screen, we should get on the plate only those parts which were more or less red in colour. For example, if we thus photographed a group of three flowers, one red, one orange and one yellow, the red one would come out prominently, the orange one would come out faintly, and the yellow one not at all.
Then suppose we took the same picture again through a yellow screen. In that case the yellow flower would be prominent, the orange would again be faint, but the red would be absent.
Having got, in imagination, two such negatives, let us make two carbon prints, one off each. And let the print off the first negative be red, while that off the second is yellow. Let each be, in fact, of the same colour as the screen through which the picture was taken. Finally, let the two films be placed in contact one upon the other. On holding the two up to the light, what should we see?
We should see a red flower, for there would be a red flower clearly defined upon one film coinciding with a blank transparent space upon the other film. We should see, too, a yellow flower, for a clearly defined yellow flower on the second film would coincide with a clear space upon the first. We should see also an orange-coloured flower, for there would be a faint red image of it, and a faint yellow image of it, one on each film, lying one over the other, producing the same effect as a mixture of yellow and red pigments. Thus by taking two negatives through two coloured screens, and then colouring the prints to correspond, we can obtain three colours in the finished picture.