eyes from the comparison. It is easy to mix a paint which will be called black by one and gray by another, and with a little less illumination the most sensitive eye might detect no gray whatever in the mixture; and even among a number of pigments, all of which would be classed as undoubtedly black, one may by comparison see differences and be able to select some which are "blacker" than the rest. Crumple a piece of white paper, and it exhibits lights and shades of greatly different degrees, some of the shades perhaps being deep enough to be designated black, and all intermediate shades may exist, but any two persons would not be likely to agree upon exactly at what particular shade should be drawn the line between gray and black. The retinal impression, therefore, under ordinary circumstances is not a reliable guide to the classification of the cause which produces it.
Consider, then, how we get impressions of color from objects. The sun emits waves of light varying in length by infinitesimal gradations between the extreme red and the extreme violet of the solar spectrum. As far as our purpose is concerned, we may disregard the ultra-violet and ultra-red rays, which are without perceptible effect upon the retina. These luminous or visible rays, acting together, produce in the eye the impression of white; separately, the longest waves produce red; those a little shorter, orange, and so on to the shortest, which produce violet. Aubert calculated that there were at least one thousand distinguishable primary color-impressions to be obtained from the solar spectrum. These rays of various lengths falling upon the things about us are partly absorbed, partly reflected, the latter portion producing in the eye sensations of color. Nearly all of our color-sensations are produced by this "selective reflection," and it will be unnecessary here to consider the other causes of color-production. Reference will be made, however, to subjective color-impressions further on. Now it is very rare indeed, or never, that but one kind or length of waves is reflected by a pigment or surface; usually several kinds are present, and even surfaces having apparently a pure color not uncommonly reflect rays differing considerably in wavelength from those of the predominant kind. Or, to put it another way, the rays from a surface having a definite hue may find their representatives in the solar spectrum not only in the portion corresponding to that particular hue, but also in one or more remote parts of the spectrum. For instance, the light from green leaves contains not only, in predominance, green rays, but some red rays and some violet rays, which find their representatives in the middle and each end of the spectrum respectively. It is true, further, that almost every hue in nature or art is made up not only of several kinds of rays, but of all kinds found in the spectrum; that is, some white light is almost always present in that which we receive from illuminated surfaces.
Now, if blackness were the complete absence of light, the question
