Color Mixing
Color-tone depends on the wave-length, as has been said, but this is far from the whole truth; the whole truth, indeed, is one of the most curious and significant facts about color vision. We have said that each color-tone is the response to a particular wave-length. But any color-tone can be got without its particular wave-length being present at all; all that is necessary is that wave-lengths centering about this particular one shall be present. A mixed light, consisting of two wave-lengths, the one longer and the other shorter than the particular wave which when acting alone gives a certain color-tone, will give that same color-tone. For example, the orange color resulting from the isolated action of a wave-length of 650 is given also by the combined action of wave-lengths of 600 and 700, in amounts suitably proportioned to each other.
A point of experimental technique: in mixing colored lights for the purpose of studying the resulting sensations, we do not mix painter's pigments, since the physical [{215}] conditions then would be far from simple, but we mix the lights themselves by throwing them together either into the eye, or upon a white screen. We can also, on account of a certain lag or hang-over in the response of the retina, mix lights by rapidly alternating them, and get the same effect as if we had made them strike the retina simultaneously.
By mixing a red light with a yellow, in varying proportions, all the color-tones between red and yellow can be got--reddish orange, orange and yellowish orange. By mixing yellow and green lights, we get all the greenish yellow and yellowish green color-tones; and by mixing green and blue lights we get the bluish greens and greenish blues. Finally, by mixing blue and red lights, in varying proportions, we get violet, purple and purplish red. Purple has no place in the spectrum, since it is a sensation which cannot be aroused by the action of any single wave-length, but only by the mixture of long and short waves.
To get all the color-tones, then, we need not employ all the wave-lengths, but can get along with only four. In fact, we can get along with three. Red, green and blue will do the trick. Red and green lights, combined, would give the yellows; green and blue would give the greenish blues; and red and blue would give purple and violet.
The sensation of white results--to go back to Newton--from the combined action of all the wave-lengths. But the stimulus need not contain all the wave-lengths. Four are enough; the three just mentioned would be enough. More surprising still, two are enough, if chosen just right. Mix a pure yellow light with a pure blue, and you will find that you get the sensation of white--or gray, if the lights used are not strong.
[Footnote: When you mix blue and yellow pigments, each absorbs part of the wave-lengths of white light, and what is left after this double absorption may be predominantly green. This is absolutely different from the addition of blue to yellow light; addition gives white, not green.]
Lights, or wave-lengths, which when acting together on the retina give the sensation of white or gray, are said to be complementary. Speaking somewhat loosely, we sometimes say that two colors are complementary when they mix to produce white. Strictly, the colors--or at least the color sensations--are not mixed; for when yellow and blue lights are mixed, the resulting sensation is by no means a mixture of blue and yellow sensations, but the sensation of white in which there is no trace of either blue or yellow. Mixing the stimuli which, acting separately, give two complementary colors, arouses the colorless sensation of white.
Blue and yellow, then, are complementary. Suppose we set out to find the complementary of red. Mixing red and yellow lights gives the color-tones intermediate between these two; mixing red and green still gives the intermediate color-tones, but the orange and yellow and yellowish green so got lack saturation, being whitish or grayish. Now mix red with bluish green, and this grayishness is accentuated, and if just the right wave-length of bluish green is used, no trace of orange or yellow or grass green is obtained, but white or gray. Red and bluish green are thus complementary. The complement of orange light is a greenish blue, and that of greenish yellow is violet. The typical green (grass green) has no single wave-length complementary to it, but it does give white when mixed with a compound of long and short waves, which compound by itself gives the sensation of purple; so that we may speak of green and purple as complementary.