the two sets of shades. It is a very common experience that the sky of a landscape in a drawing turns out too pale after the rest of the drawing is completed. Contrasted with the white paper of the unfinished sketch, it may look quite right; but, after the deeper tones of the distance and foreground are added, may become quite insignificant. Again, a few decidedly black touches in a drawing will often by contrast lighten up portions that had previously seemed considerably too dark; or a few touches of pure white will apparently darken spaces that had seemed pale and weak. In each case the observer is furnished by external means with a standard for measuring the depth of the shade, and induced to use it rather than his memory. By the skillful employment of contrast, drawings in light and shade can be made to appear luminous and brilliant, or rich and deep; neglect of this element produces tameness and feebleness. The contrast of light with dark shades is not inferior in power to that of warm with cool tints; and, in point of fact, the contrast of white with black is the strongest case of contrast possible. We have on the one hand the presence of all the colours, on the other their total absence. Hence, as has been noticed before, the contrast that takes place between light and shade will sometimes mask or even reverse that which occurs with different colours. We can perhaps better tolerate a shortcoming in the matter of colour-contrast than in that of light and shade; if the latter is right and powerful, we forgive a limited amount of inferiority in the former, merely remarking that the work is rather slight or pale in colour, but not on that account pronouncing a verdict of total condemnation. On the other hand, if the colour as such is right, but the depth of the different tints mostly defective, then the whole is spoiled, and we contemplate the tints, lovely enough in isolation, with no satisfaction. We forgive, then, a partial denial of the truths of colour more easily than those of light and shade, which probably is a result of the nature of the opti

cal education of the race. For the human race, thus far, light and shade has been the all-important element in the recognition of external objects; colour has played only a subordinate part, and has been rather a source of pleasure than of positive utility.

All that has been said with regard to the contrasts of white, black, and grey, with slight modification, applies to any single colour taken by itself; for instance, to drawings executed in one colour only, such as blue or brown. From this it results that every colour is capable of exhibiting two kinds of contrast, viz., that involved by competition with other colours and that of mere light and shade.

The contrast of white, black, and grey with the series of positive colours remains to be noticed. Taking up these in order, we find that red when placed on a white ground appears darker and rather more intense in hue; on a black ground it becomes tinted somewhat orange-red, and looks of course more luminous. Both these effects are probably

due ultimately to mere contrast of light and shade; the white ground makes the red by contrast look darker. But we are accustomed to see red when it is darkened recede from orange and approach pure red, or even perhaps to become somewhat purplish; hence it appears so in this case; it is an instance of expectant attention. When red is placed on a black ground, it is made by contrast to look more luminous; but we are accustomed to see luminous red become tinted with an orange hue; hence the result. Red on grey grounds of various depths undergoes modifications corresponding to those just mentioned. Pale red, i. e., red mixed with much white, on a white ground, gains in intensity of colour; on a black or dark-grey ground it loses intensity, and approximates to pure white in appearance. Here the contrast of light and shade is so strong as to cause the colour to pass almost unperceived; or we may say pale red really does approach much nearer to pure white than black, and hence is at last accepted for it. Dark, dull red

on a white ground may be mistaken for brown; on a black ground it appears more luminous and more red. Orange on a white ground looks darker and more reddish, on a black ground more luminous and yellow. The other effects correspond with those described in the case of red. Yellow on a white ground appears darker and more greenish than on a black ground; in the latter case it is particularly brilliant, and the black also looks well, taking on a bluish tint. Dark yellow on a white ground looks brown or greenish-brown; on a black ground its colour is displayed to more advantage. Pale yellow on a white ground is apt to look greenish, on a black ground to appear whitish. Yellow and grey or black constitute a pleasant combination, of which extensive use has been made in nature and art. Green on a white ground looks deeper and richer, on a black ground somewhat paler; by contrast the black is made to look somewhat reddish or rusty. Green causes grey to appear reddish; the effect is particularly marked when the grey has about the same luminosity with the green, also when both are in the shade. Cyan-blue on white appears darker and perhaps more greenish than on black. Blue on white appears dark and rich, but shows no tendency to green; on black, by contrast, it becomes more luminous. The same is true with blue on grey; the latter acquires a somewhat yellowish or rusty hue. The action with violet is similar to that of blue.

From the foregoing it is evident that the contrast of black, white, and grey with the colours depends mainly on an apparent increase or diminution of their luminosity, whereby in most cases their apparent hue is affected owing to association. In the case of the colder colours, the tint of the grey or black ground is affected, and shows a tendency toward a hue complementary to the colour employed. We associate grey with blueness, and where the effect is such as contradicts this habitual association it is disagreeable; on the other hand, grey with yellow forms

an agreeable contrast, as the yellow tends to make the grey look more bluish, and thus corrects any yellowish or rusty appearance connected with it. It is claimed in some works on colour that the complementary tints furnished by the pure grey react on and strengthen the colours which call them forth. An eye which is tired by gazing at green is indeed rested by looking at its complement, i. e., at a mixture of red and violet, and afterward will see the green with more vividness; but it is difficult to understand how the presentation of red, violet, and green, or, what is the same thing, grey light, can materially refresh the eye, or restore its temporarily exhausted power. In the case of pale tints, an effect of this character does indeed seem to take place, but we must attribute it rather to an act of judgment than to a physiological cause.

CHAPTER XVI.

ON THE SMALL INTERVAL AND ON GRADATION.

In the preceding chapter we have seen that, when two colours which are nearly identical are contrasted, each is made to appear less intense or saturated: red with orangered, yellow with orange-yellow, cyan-blue with blue, are examples of such combinations. From this it might be supposed that, in chromatic compositions, it would not be allowable to place colours thus nearly related in close juxtaposition. It is, however, found in practice that colours which are distant from each other in the chromatic circle by a small interval can be associated without detriment under certain conditions. If the two colours express a variation in the luminosity of one and the same coloured surface, they do not come into hurtful competition, and we receive the impression of a single coloured surface, more highly illuminated in certain portions. The scarlet coat of a soldier when shaded appears red; the sunlit portion is orangered. Grass in the sunshine acquires a yellowish-green hue; in the shade its colour is more bluish. But neither of these cases produces on us a disagreeable effect, for we regard them as the natural consequences of the kind of illumination to which these objects are exposed. The effect is not disagreeable even in mere ornamental painting, if it is seen that the two tints are intended to express different degrees of luminosity of the same constituent of the design, even though this be only arabesque tracery. From this explanation it follows that the two contiguous tints should