than those that produce the sensation of red, but longer than those that give us that of green; referring to Fig. 1, we see that no especial set of nerves has been provided for this case, but a moment's reflection will suggest that these intermediate waves, according to our theory, ought to set into moderate action both the red and green nerves, that the stimulation of the former should predominate as the length of our intermediate waves is made longer; the green set, on the other hand, coming more into play as it is shortened. This accounts, then, for the mode in which waves of a certain length, or light of a certain kind, gives us the sensation of yellow or orange. The light may be simple, and of only one kind, but it produces a compound sensation, made up of the two simple sensations, red and green. From all this it follows that, on the other hand, if we actually present to the same eye mixtures of red and green light, the sensations of yellow or orange ought, according to our theory, to be produced. This is a matter that we can easily test by experiment. With the same apparatus used a moment ago for combining blue and yellow light, I throw upon the screen a large square of red light, and superimpose on it one of green, and, as you see, the result is a fair yellow; on reducing the brightness of the green component, the yellow passes into orange (Fig. 2). I call your attention, in passing, to the circumstance
Fig. 2.
that, according to the old theory, the result ought not to have been yellow, but rather an approach to white, all the colors, according to its doctrines, being present. Restoring the green squares to their original brightness, and reducing the intensity of the red light, we easily obtain a greenish yellow, completing thus this series of tints.
And now to account for the blue: pure blue light has a wave-length intermediate between that of green and violet light, and hence sets both the green and violet nerves into action, and, though the light itself may be simple, it produces a compound sensation which we call blue. Corresponding to this, I ought to be able to reproduce on the screen blue light by mixing together green and violet light. The experiment is now arranged, and, as you see, we actually do obtain a quite good blue in this way, and can cause it to run through all
