# Popular Science Monthly/Volume 14/November 1878/The Contrast of Colors

THE

POPULAR SCIENCE

MONTHLY.

NOVEMBER, 1878.

 THE CONTRAST OF COLORS.[1]
By Professor O. N. ROOD.

IN a previous chapter we have studied the changes which colored surfaces experience when viewed under various kinds of illumination, or when modified in appearance by the admixture of more or less white or colored light. The appearance which a colored surface presents to us can, however, be altered very materially, by a method which is quite different from any of those that have thus far been mentioned: we can actually change color to a considerable extent without at all meddling with it directly, it being for this purpose only necessary to alter the color which lies adjacent to it. The truth of this can be seen by a very simple experiment: If we cut out of a sheet of red paper two square pieces, about an inch in size, and then place one of them on

Fig. 1.—Sheets of Red and Green Paper with Red Squares

a sheet of red, the other on a sheet of green paper, as indicated in Fig. 1, it will be found that the red square on the red paper will not appear nearly so brilliant and saturated in color as that placed on the green ground; hence the observer will be disposed to doubt whether the two squares are really identical in hue. By a somewhat analogous proceeding we can cause a surface which properly has no color of its own, which is really gray, to appear tinted red, blue, green, etc. These changes, and others of a like character, are produced by what is called contrast, and are partly due to actual effects generated in the eye itself, and partly to fluctuations in the judgment of the observer. The subject of contrast is so important that it will be worth while to make a somewhat careful examination of the laws which govern it, and it will be well for the reader to repeat some of the simple experiments described below. If we place a small piece of bright-green paper on a sheet of gray drawing-paper, in the manner indicated in Fig. 2, and then for several seconds attentively look at the small cross in the centre of the green slip, we shall find, on suddenly removing it, that in its place a faint image of a rose-red color makes its appearance (see Fig. 3). This

 Fig. 2.—Gray Paper with Green Slip. Fig. 3.—Gray Paper with Rose-colored Image.

red image presently vanishes, and the gray paper resumes its natural appearance. The rose-red ghost which is thus developed has a color which is complementary to that which called it into existence, and this will also be the case if we employ little squares of other colors: red will give rise to a greenish-blue image, blue to a yellow, violet to a greenish-yellow, etc., the color of the image being always complementary to that which gave rise to it. Upon this account these images are called negative, since as far as the color goes they are just the reverse of the images which are first presented to the eye of the observer. They are also often spoken of, in older treatises on optics, as "the accidental colors." It is quite easy to explain their production with the aid of the theory of Young and Helmholtz. Let us take as an example the experiment just described: According to our theory the green light from the little squares of paper, acting on the eye, fatigues to some extent the green nerves of the retina, the red and violet nerves meanwhile not being much affected. When the green paper is suddenly jerked away by the string, gray light is presented to the fatigued retina, and this gray light may be considered to consist, as far as we are concerned, of red, green, and violet light. The red and violet nerves, not being fatigued, respond powerfully to this stimulus; the green nerves, however, answer this new call on them more feebly, and in consequence we have presented to us, mainly, a mixture of the sensations red and violet, giving as a final result rose-red or purplish-red. The green nerves, of course, are not so fatigued that they do not act at all when the gray light is presented to them, but the only effect that their partial action has, is to render the rose-colored image somewhat pale or whitish in appearance. The fatigue of the optic nerve mentioned here does not differ essentially from that which it undergoes constantly even under the conditions of ordinary use, where the waste is constantly made good by the blood circulating in the retina, and by the little intervals of rest constantly occurring. In our experiment we have merely confined the fatigue to one set of nerves, instead of distributing it equally among the three sets.

The above experiments and explanation will enable us easily to comprehend the more complicated case, where, instead of placing our little green square on gray, we lay it on a sheet of colored paper. Instead, then, of gray, let us take yellow paper, placing the green square on it as before (see Fig. 4). On suddenly withdrawing the green square, we find it replaced by an orange-colored ghost (Fig. 5), which

 Fig. 4.—Yellow Ground with Green Slip. Fig. 5.—Yellow Ground with Orange-colored Image.

we account for thus: As before, the green nerves are fatigued, the red and violet nerves remaining fresh; when the square is removed, yellow light is presented to the retina, and this yellow light, as explained in Chapter IX., tends to act on the red and green nerves equally, but the green nerves in the present case do not respond with full activity, hence the action is more confined to the red nerves, and, as explained in Chapter X., the resultant tint is necessarily orange, that is to say, we have a strong red sensation mingled with a weak green sensation, and the result is the sensation called orange. In this experiment the violet nerves do not come into play to any great extent. If the green square is placed on a blue ground the image becomes violet, for the reason that the blue light which is presented to the fatigued retina acts, as explained in Chapter IX., on the green and violet nerves; but the green nerves being already fatigued, the action is mostly confined to the violet nerves, and hence the corresponding sensation. In this case the red nerves hardly come into play at all.

It follows, from the above examples and reasoning that the final effect is, that we obtain as an after-image what amounts to a mixture of the complementary color of the small square with the color of the ground; and, by recollecting this, we can easily retain this class of facts in the memory.

There is another similar experiment, which is simpler than those just described, but which nevertheless is instructive: A small square of black paper is to be placed on a sheet of red paper, and the attention in this case is to be directed to a mark on the edge of the former (see Fig. 6). When the black square is suddenly removed, the observer sees in place of it a more luminous spot, which in the case before

 Fig. 6.—Red Ground with Black Paper. Fig. 7.—Red Ground with Intense Red Image.

us will, of course, be red; but what is remarkable is the circumstance that this red image will be more intense or saturated in color than the rest of the ground. The rest of the sheet of red paper will look as though gray had been mixed with its color (Fig. 7). This experiment will, of course, succeed with paper of any bright color, and Helmholtz has found that the same effect can be obtained with the pure colors of the prismatic spectrum. The explanation, according to our theory, runs about thus: While we are in the act of looking at the edge of the black square, red light is passing into the eye, and is fatiguing all those portions of the retina that are not protected by the presence of the black square; it thus happens that the ability of the larger portion of the retina to receive the sensation of red is considerably diminished; the ability of the protected portion, of course, suffers meanwhile no such change. When the black square is suddenly removed, the unfatigued portion of the retina receives a powerful impulse from the red surface, but the effect produced upon the rest of the retina is inferior in degree. This accounts for the fact that the image of the square is brighter or more luminous, and we can easily understand why it is at the same time more intense or saturated in color, if we remember, as explained in Chapter IX., that red light excites into action not only the red nerves, but to a lesser extent the green and violet nerves. Now, as the red nerves begin to be fatigued, the action of the other two sets will be relatively more powerful than at first, so that gradually the sensations of green and violet begin to add themselves to that of red, or, what is the same thing, the sensation of white mingles itself with that of red and makes the red color of the paper look a little grayish. The success of the experiment with the pure colors of the spectrum, which contain no white, is easily accounted for by the explanation just given.

All these phenomena are cases of what is called successive contrast, because we look in succession from one surface to another. When colored surfaces are placed near each other and compared in a natural manner, successive contrast plays an important part, and the appearance of the colors is more or less modified according to its laws. If we attempt to confine our attention to only one of the colored surfaces, this still holds good, for the eye involuntarily wanders to the other, and to prevent this requires a good deal of careful practice, for fixed vision is quite opposed to our natural habit. It follows from this that, in the natural use of the eye, the negative images, although present to some extent, are not sharp and distinct, and hence usually remain unobserved by persons not trained to observations of this character. Nevertheless these images modify to a considerable extent the appearances of colored surfaces placed near each other, and the changes of hue are visible enough to the most uneducated eye.

One of the most common cases belonging here is represented in Fig. 8. We have a gray pattern traced on a green ground; the

Fig. 8.—Gray Figure on a Green Ground.

tracery, however, will not appear pure gray, but tinged with a color complementary to that of the ground; that is, reddish. We can, of course, substitute for the green any other bright color, and it will always be found that the gray pattern is more or less tinged with the complementary hue. As black is really a dark gray, we should expect to find it also assuming, to some extent, a color complementary to that of the ground, and this is indeed the case, though the effect is not quite so marked as with a gray of medium depth. Chevreul, in his great work on the simultaneous contrast of colors, relates an anecdote which illustrates the matter now under consideration: Plain red, violet-blue, and blue woven stuffs were given by certain dealers to manufacturers, with the request that they should ornament them with black patterns. When the goods were returned the dealers complained that the patterns were not black, maintaining that those traced on the red stuffs were green; on the violet, dark greenish-yellow; and on the blue, copper-colored. Chevreul covered the grounds with white papers in such a manner as to expose only the patterns, when it was found that their color was truly black, the effects which had been observed being entirely due to contrast. The remedy in such cases is not to employ pure black, but to give it a tint like that of the colored ground, taking care to make it just strong enough to balance the hue generated by contrast. If we substitute a white pattern for the black, something of this same effect can often be observed, but it is less marked than with gray or black. In cases like those now under consideration, the contrast is stronger when the colored surface is bright and intense or saturated in hue. The effect is also increased by entirely surrounding the second color with the first; the circumscribing color ought also to be considerably larger than its companion. When these conditions are observed, the effect of contrast is generally noticeable only on the smaller surface,

Fig. 9.—Arrangement to show the Effects of Simultaneous Contrast (One-Half Size).

the larger one being scarcely affected. When, on the other hand, the two colored surfaces are about equal in extent, then both suffer change. If it is desired to produce a strong effect of contrast, the colored surfaces must be placed as near each other as possible. This is beautifully illustrated in one of the methods employed by Chevreul in studying the laws of contrast. Two colored strips were placed side by side in contact, as shown in Fig. 9, duplicate strips being arranged in the field of view at some distance from each other. The tints of the two central strips were both altered; those placed at a greater distance apart suffered no change. In the experiment represented in Fig. 9, the central ultramarine by contrast is made to appear more violet, the central cyan-blue more greenish; the color of the outlying strips is scarcely affected.

As it requires a little consideration to predict the changes which colors undergo through contrast, we give below a table containing the most important cases:

 Pairs of colors Change due to contrast Pairs of colors Change due to contrast ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Red. becomes more purplish. ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Orange. becomes more yellowish. Orange. " yellowish. Violet. " bluish. ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Red. " purplish. ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Yellow. " orange-yellow. Yellow. " greenish. Green. " bluish-green. ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Red. " brilliant. ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Yellow. " orange-yellow. Blue-green. " "⁠ Cyan-blue. " blue. ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Red. " orange-red. ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Yellow. " brilliant. Blue. " greenish. Ultramarine-blue. " "⁠ ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Red. " orange-red. ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Green. " yellowish-green. Violet. " bluish. Blue. " purplish. ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Orange. " red-orange. ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Green. " yellowish-green. Yellow. " greenish-yellow. Violet. " purplish. ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Orange. " red-orange. ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Greenish-yellow. " brilliant. Green. " bluish-green. Violet. " "⁠ ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Orange. " brilliant. ${\displaystyle \scriptstyle {\left\{{\begin{matrix}\ \end{matrix}}\right.}}$ Blue. " greenish. Cyan-blue. " "⁠ Violet. " purplish.

It is easy and instructive to study the changes produced by contrast with the aid of a chromatic circle (Fig. 10), and it will Fig. 10.—Chromatic Circle. be found that alterations in color produced by contrast obey a very simple law: when any two colors of the chromatic circle are brought into competition or contrasted, the effect produced is, apparently, to move them both farther apart. In the case, for example, of orange and yellow, the orange is moved toward the red, and assumes the appearance of reddish-orange; the yellow moves toward the green, and appears for the time to be greenish-yellow. Colors which are complementary are already as far apart in the chromatic circle as possible, hence they are not changed in hue, but merely appear more brilliant and saturated. This is indeed the effect which a strict application of our rule leads to: The two colors are to be moved farther apart; they are already situated on the opposite extremities of a diameter of the circle, and, if they are to recede still farther from each other, they can accomplish this in no other way than by moving outside of the circumference of the circle; but this corresponds, as explained in the previous chapter, to an increase of saturation. If the experiments indicated in the previous table are carefully repeated, it will be found that all the pairs of colors there enumerated are not equally affected by contrast. The changes of tint are greatest with the colors which are situated nearest to each other in the chromatic circle, and much less with those at a distance. Thus both red and yellow are much changed by contrast, the red becoming purplish, the yellow greenish, while red with cyan-blue, or blue, is much less affected in the matter of displacement or change of hue. On the other hand, the colors which are distant from each other in the chromatic circle, while suffering but slight changes. in hue, are made to appear more brilliant and saturated—that is, they are virtually moved somewhat outside of the circle, the maximum effect taking place with colors which are complementary. Colors which are identical are affected by contrast in exactly the opposite way from those which are complementary—that is, they are made to appear duller and less saturated. The author finds that these and other effects of contrast can be studied with great advantage by the aid of two identical chromatic circles laid down on paper. One set of these lines should be traced on a sheet of transparent paper, which is afterward to be placed over the companion-circle. The use of these circles will best be made evident with the aid of an example: Let us suppose that we wish to ascertain with their aid the effect produced by red, as far as contrast goes, on all the other colors, and also on red itself. We place the transparent circle on its companion so that the two drawings may coincide in position, and we then move the upper circle along the diameter joining the red and green-blue some little distance, so that the two circles no longer have the same common centre. We then transfer the points marked red, orange, yellow, etc., on the upper circle, by pricking with a pin through to the lower circle; these pin-marks on the lower circle will indicate the changes produced on all the colors by competition with red. Fig. 11 gives the result. The dotted circle with the crosses represents the new positions of the different colors when contrasted with red. If we examine it we find that red, when contrasted with greenish-blue, causes this last color to move away from the centre of the circle in a straight line; hence, as the new point is on the same diameter, but farther from the centre, we know that the greenish-blue is not made more or less blue or green, but is simply caused to appear more saturated or brilliant. The new point for the red lies also on the same diameter, but is nearer to the centre of the circle; that is, the color remains red, but appears duller or less saturated. Experience confirms this: if a considerable number of pieces of red cloth are examined in succession, the last one will appear duller and inferior in brilliancy to the others, but it will still appear red.

Proceeding with the examination of the effects produced on the other colors, we find that orange has been moved toward yellow, and also toward the centre of the circle; hence our diagram tells us that red, when put into competition with orange, causes the latter to appear

Fig. 11.—Chromatic Circle displaced by Contrast (shows the effects produced by red upon the other colors).

more yellowish, and at the same time less intense. Advancing along the circumference of the circle, our diagram informs us that yellow is not much affected in the matter of saturation or intensity, but is simply made to appear more greenish. The two circles during superposition cut each other near the position of yellow; from this point onward the effect changes as far as intensity or saturation is concerned, the greenish-yellow being moved decidedly outside of the original circle as well as toward the green; it is made, therefore, by contrast with red, to appear more brilliant as well as more greenish. Green is made to appear somewhat bluish and more brilliant. Greenish-blue has been considered. Cyan-blue is made to appear slightly more greenish as well as much more brilliant; the same is true of blue, though its increase in brilliancy by contrast with red is rather less than is the case with cyan-blue. Violet has its hue considerably altered toward blue; its saturation is diminished. Purple is made to look more violet and is much diminished in saturation.

If we wish to study the effects produced on the colors of the chromatic circle by contrasting them with yellow, we have of course merely to displace the upper circle along the line joining yellow and its complement, ultramarine-blue, and then proceed as before.

It is quite evident that this contrast-diagram will furnish correct results only on condition that the colors in it are properly arranged; if the angular positions of the colors are laid down falsely, the results in the matter of increase or diminution of brilliancy will also be false. The author has made many experiments to settle this question, and in Fig. 12 gives his result in the form of a diagram.

From the foregoing, then, it is evident that the effect of contrast may be helpful or harmful; by it colors may be made to look more beautiful and precious, or they may damage each other and appear dull, pale, or even dirty. When the apparent saturation is increased, we have the first effect; the second, when it is diminished. Our diagram (Fig. 11) shows that the saturation is diminished when the contrasting colors are situated near each other in the chromatic circle, and increased when the reverse is true. It might be supposed that we could easily overcome

Fig. 12.—Contrast-Diagram, according To O. N. Rood.

the damaging effects of harmful contrast by simply making the colors themselves from the start somewhat more brilliant; this, however, is far from being true. The pleasure due to helpful contrast is not merely owing to the fact that the colors appear brilliant or saturated, but that they have been so disposed and provided with such companions that they are made to glow with more than their natural brilliancy. Then they strike us as precious and delicious, and this is true even when the actual tints are such as we would call poor or dull in isolation. From this it follows that paintings made up almost entirely of tints that by themselves seem modest and far from brilliant, often strike us as being rich and gorgeous in color; while on the other hand the most gaudy colors can easily be arranged so as to produce a depressing effect on the beholder. We shall see hereafter that, in making chromatic compositions for decorative purposes or for paintings, artists of all times have necessarily been controlled to a considerable extent by the laws of contrast, which they have instinctively obeyed, just as children in walking and leaping obey the law of gravitation, though hardly conscious of its existence.

The phenomena of contrast as exhibited by colors which are intense, pure, and brilliant, are to be explained to a considerable extent by the fatigue of the nerves, as set forth in the early part of the present chapter. The changes in color and saturation become particularly conspicuous after somewhat prolonged observation, and are often attended with a peculiar soft glimmering, which seems to float over the surfaces, and, in the case of colors that are far apart in the chromatic circle, to lend them a lustrous appearance. Still, upon the whole, the effects of contrast with brilliant colors are often not strongly marked at first glance, from the circumstance that the colors, by virtue of their actual intensity and strength, are able to resist these changes, and it often requires a practiced eye to detect them with certainty. The case is quite otherwise with colors which are more or less pale or dark—that is, which are deficient in saturation, or luminosity, or both. Here the original sensation produced upon the eye is comparatively feeble, and it is hence more readily modified by contrast. In these cases the fatigue of the nerves of the retina plays but a very subordinate part, as we recognize the effects of contrast at the first glance. We have to

Fig. 13.—Shadow of Rod in a Darkened Room.

deal here with what is known as simultaneous contrast, the effects taking place when the two surfaces are, as far as possible, regarded simultaneously. In the case of simultaneous contrast, the changes are due mainly to fluctuations of the judgment of the observer, but little to the fatigue of the retinal nerves. We carry in ourselves no standard by which we can measure the saturation of color or its exact place in the chromatic circle; hence, if we have no undoubted external standard at hand with which to compare our colors, we are easily deceived. A slip of paper, of a pale but very decided blue-green hue, was placed on a sheet of paper of the same general tint, but somewhat darker and more intense or saturated in hue. The small slip now appeared pure gray, and by no effort of the reason or imagination could it be made to look otherwise. In this experiment no undoubted pure gray was present in the field of view for comparison, and, in point of fact, the small slip did actually approach a pure gray in hue more nearly than the large sheet; hence the eye instantly accepted it for pure gray. The matter did not, however, stop here: a slip of pure gray paper was now brought into the same green field, but, instead of serving as a standard to correct the illusion, it assumed at once the appearance of a reddish-gray. The pure gray slip really did approach reddish-gray more than the green field surrounding it, and hence was accepted for this tint.

It has been stated above that the effects produced by simultaneous contrast are due not to retinal fatigue, but to deception of the judgment; now, as the effects of simultaneous contrast are identical in kind with those generated by successive contrast, it is evident that the statement needs some proof. This can be furnished with the aid of a beautiful experiment with colored shadows. In making this experiment, we allow white daylight to enter a darkened room through an aperture. A, arranged in a window, as indicated in Fig. 13. At M we set up a rod

Fig. 14.—Shadows of Rod, using Daylight and Candle-Light.

and allow its shadow to fall on a sheet of white cardboard or on the white wall of the room. It is evident, now, that the whole of the cardboard will be illuminated with white light, except those portions occupied by the shadow, 1. We then light the candle at C (Fig. 14); its light will also fall on the cardboard screen, and will then cast the shadow 2—that is, the candle-light will illuminate all parts of the screen except those occupied by the shadow 2; this portion will be illuminated with pure white light. Instead, however, of appearing to the eye white, the shadow 2 will seem to be colored decidedly blue.

For the production of the most powerful effect, it is desirable that the shadows should have the same depth, which can be effected by regulating the size of the aperture admitting daylight. Now, although the shadow cast by the candle is actually pure white, yet, by contrast with the surrounding orange-yellow ground, it is made to appear decidedly blue. So strong is the illusion that, even after the causes which gave rise to it have disappeared, it still persists, as can be shown by the following experiment of Helmholtz:

While the colored shadows are falling on the screen, they are to be viewed through a blackened tube of cardboard held in such a way that the observer has both the shadows in his field of view; the appearance, then, will be like that represented in Fig. 15. After the blue shadow

Fig. 15.—Blue and Yellow Shadows viewed though a Tube.

has developed itself in full intensity, the tube is to be moved to the left, so that the blue shadow may fill the whole field. The tube being held steadily in the new position, the shadow will still continue to appear blue instead of white, even though the exciting cause, viz., the orange-yellow candle-light, is no longer acting on the eye. The candle may be blown out, but the surface will still appear blue, as long as the eye is at the tube. On removing the tube, the illusion instantly vanishes, and it is perceived that the color of the surface is identical with that of the rest of the screen, which is at once recognized as white. In a case like this, the fatigue of the retinal elements can play no part, as the illusion persists for a far longer time than is necessary for their complete rest; we must hence attribute the result to a deception of the judgment.

The simple experiments of H, Meyer are less troublesome than those just described, and at the same time highly instructive. A small strip of gray paper is placed on a sheet of green paper, as indicated in Fig. 16; it will be found that the tint of the gray paper scarcely changes, unless the experimenter sits and stares at the combination for some time. A sheet of thin white tissue-paper is now to be placed Fig. 16.—Green and Gray Papers For Experiment on Contrast (One-quarter Size). over the whole, when it will instantly be perceived that the color of the small slip has been converted by contrast into a pale red. Persons seeing this illusion for the first time are always much astonished. Here we have an experiment showing that the contrast produced by strong saturated tints is much feebler than with tints which are pale or mixed with white light, for, by placing tissue-paper over the green sheet, the color of the latter is extraordinarily weakened and mixed with a large quantity of white light. In this experiment it often happens that the red, which is due to contrast alone, seems actually stronger than the green ground itself. If, instead of using a slip of gray paper, we employ one of black, the contrast is less marked, and still less with one of white. It is scarcely necessary to add that, if red paper is employed instead of green, the small gray slips become tinted by contrast with the complementary color—i. e., greenish-blue; the same is true with the other colors. By preparing with India-ink a series of slips of gray paper, ranging from pure white to black, an interesting series of observations can be made on the conditions most favorable for the production of strong contrast-colors. The strongest contrast will be produced in the case of red, orange, and yellow, when the gray slip is a little darker than the color on which it is placed, the reverse being true of green, blue, violet, and purple; in every case the contrast is weaker if the gray slip is much lighter or much darker than the ground. We must expect, then, in painting, to find that neutral gray will be more altered by pale tints of red, orange, or yellow, which are slightly lighter than itself, and that the gray will be less altered by these colors when differing considerably from it in luminosity; analogous conclusions with regard to green, blue, violet, and purple, can also be drawn. Saturated or intense colors in a painting have less effect on white or gray than colors that are pale; this was shown in the preliminary experiment when gray was placed on a ground of strong color. In repeating these experiments it will be noticed that the effect of contrast is stronger with green, blue, and violet, than with red, orange, or yellow—that is to say, it is stronger with the cold than with the warm colors.

We must next examine the effects that are produced by contrasting colors that differ in luminosity or in saturation. If the two colors are identical except in the matter of saturation, it will be found that the one which is more saturated will gain in intensity, while its pale rival appear still paler. A slip of paper painted with a somewhat pale red, when placed on a vermilion ground, appears still paler, and may actually be made to look white. If a still paler slip be used, it may even become tinged greenish-blue, its color being in this case actually reversed by the effect of contrast. When the colors differ in luminosity, analogous effects are observed: a dull-red slip was placed on a vermilion ground; the effect was as though a quantity of gray had been added to the slip; it looked more dingy and somewhat blackish. Another slip, still darker and containing less red, when placed on the same ground looked as if it were tinged with olive-green; a still darker slip, with still less red color, when treated in the same way looked black, with a tinge of blue. When, however, this last slip was placed on a white ground, or compared with true black, it was seen that its color was far from black. The general result of contrasting colors which differ much in strength then is, that the feebler one appears either more whitish or grayish, or assumes the complementary tint; the stronger one, on the other hand, appears still more intense.

If the strong and weak colors are complementary to each other, then each of them gains in intensity and appears purer, this gain seeming to be greater in the case of the pale tint. From this it follows that while the juxtaposition of strong with feeble colors usually injures or greatly alters the latter, colors that are complementary furnish an exception, the reason of which is evident at the first glance.

When the pale or dark colors are not complementary to their more intense or brilliant rivals, they undergo the same changes indicated in the table on page 7, the changes in the case of the dull or pale colors being considerably greater. In proportion as the colors are distant from each other in the chromatic circle, do they gain in saturation and beauty; while, as they approach, their character is altered and they are apt to look very pale, or, in the case of the dark colors, blackish or dirty. This is particularly so when the brilliant color is large in surface and surrounds the darker one; with the reversed conditions the effect is not so much felt. Thus, a somewhat dull red near vermilion no longer looks red but brown; a dull orange tint under the same conditions looks like a yellowish-brown.

There is a still more general reason upon which the pleasure that we experience from contrast depends: after gazing at large surfaces filled with many varieties of warm color, skillfully blended, we feel a peculiar delight in meeting a few mildly contrasting tints; they prevent us from being cloyed with all the wealth of rich coloring so lavishly displayed, and their faint contradiction makes us doubly enjoy the richer portions of the painting. So also when the picture is mainly made up of cool bluish tints; it is then extraordinarily strengthened and brightened by a few touches of warm color.

1. From the advance-sheets of a work on "Chromatics, with Applications to Art and Industry."