# Popular Science Monthly/Volume 83/November 1913/The Application of the Physiology of Color Vision in Modern Art

 THE APPLICATION OF THE PHYSIOLOGY OF COLOR VISION IN MODERN ART
By HENRY G. KELLER

CLEVELAND SCHOOL OF ART

AND

Professor J. J. R. MACLEOD

WESTERN RESERVE MEDICAL SCHOOL

Introduction

LEONARDO in his treatise on painting says;

Those who become enamored of the practise of the art without having previously applied themselves to the diligent study of the scientific part of it, may be compared to mariners, who put to sea in a ship without rudder or compass and, therefore, can not be certain of arriving at the wished-for port. Practise must always be founded on good theory.

Instead of serving as an incentive to more extensive study of the use of colors in art, these words seem to have marked the advent of an epoch extending over several centuries, during which colors came to be less and less successfully employed. The ideals of art came to be dictated by the academic painter and they were much more mythological and allegorical than founded on the beauty of color patterns. Much of art became black painting, little attempt being made to use pure colors even in landscape painting, and no consideration being given to the effects which could be produced by the influence of juxtaposed colors on one another. With the exception of some masters the ideal of artists was merely to reproduce as closely as possible the color tones and values as seen in nature—to produce a colored photograph without adding to it that mysterious something for which is responsible the peculiar charm and strength of the paintings of the early Italian masters and of the Chinese and Japanese, and which includes some subtle influence of the picture itself quite apart from what it represents; something that endows it with a charm that is all its own, and which no colored photograph can ever contain.

It is true that from time to time in the history of modern art masters have arisen who have, intuitively as it were, produced pictures the color schemes of which have contained this "something." But it is the individual rather than the system that has been responsible, and no attempts have been made until comparatively, recently to evolve new principles for the use of colors which would serve as a guide to all; nor indeed was such an evolution possible until some progress had been made in the scientific interpretation of color. This progress is itself only of comparatively recent date.

At the present day there is an unrest in the world of art, an unrest which has resulted in the creation of innumerable schools, each endeavoring by some peculiar method of its own to inculcate new principles and to establish new ideals. Within a short period of time realism has given place to impressionism, impressionism to post-impressionism, and this again has become parent for so many other "-isms," that, to follow them, has become almost impossible. However unpictorial from our ordinary viewpoint the creations of some present-day artists may appear to be, there is nevertheless in many of them some newly discovered truth; they are the steps in an evolution, and we may hope that some day the evolution will be consummated and. that from out of the apparent chaos, which at present exists, a really compelling picture will be created.

It is most of all in landscape painting that the evolution of modern art can be seen. The old landscapes of Claude Lorrain and Constable are no doubt full of charm, but they entirely lack the atmosphere and force of the so-called impressionist paintings of Monet, Sisley, Pissaro, etc. In the older landscapes an attempt was made to copy everything that could be seen by prolonged study, and the canvas was covered with detail to its very edges; in impressionism, it is merely the flash, the fleeting effect of the landscape which it is attempted to reproduce. There may indeed be considerable detail in certain portions of the picture, but the greater part is merely a color pattern. But after all such an impressionistic picture can occupy our attention for a moment only. We do indeed receive an impression more or less like that which the artist received on viewing his object, but closer study of the picture does not carry us farther; there is something absent from it with which nature abounds, something that compels us, as when viewing a landscape, to keep shifting our gaze from point to point, a restlessness, a constant source of interest and fascination. In post-impressionism the attempt is being made to supply this want, to compel us namely to regard more than the fleeting impression. The closer we study such a picture, if it be successful, the more comes out of it, colors by their influence on one another become changed in hue and saturation, a curiosity develops and, subconsciously, we are compelled to continue our study with the result that we get ever other and other effects. It is kinetic, not static, art; it is a pattern of nature designed to create visuo-psychic impressions expressing an idea rather than an object, subjective rather than objective.

There is a physiological reason for this visual restlessness and before we go into the science of colors it may be well to explain what this reason is. The innermost' layer of the eye, on to which images of exterior objects are focused, is specialized to react to sensation of light, thus setting up nerve impulses which are transmitted to the brain where they are interpreted. This layer of the eye is called the retina and it is very much more sensitive at a small spot in the center than it is over the much larger outer (peripheral) portions, so that, of the image which is focused on it, it is only that part falling on the central portion which is distinctly seen. When we regard a stretch of country, for example, it is only in one part of it that the objects are seen in any detail—namely that part which is focused on the central portion of the retina—the remainder, since it falls on the outer portion, causing only a vague, indefinite impression. We may say indeed that the function of the greater part of the retina is merely to give us a general impression of the environment of the object which is being looked at; an impression, that is to say, which will enable us to judge of its relationship to other things. It tells what else there is to look at, and subconsciously we shift our gaze so that, piece by piece, the whole landscape comes to be focused on the central portion. We regard with the central portion what we know exists to be regarded on account of the duller image thrown on the rest of the retina.

Coming now to the question of color, any attempt to apply the scientific principles of color vision in making a picture must surely fail if it be not granted at the outset that it is only to a limited degree that those principles can apply. Color appreciation is as much a psychical as a physiological process, and indeed it is psychical not only with regard to the objective impression itself, but also with regard to the subjective, the associational mental process. Previous knowledge and training, experience, tradition, the association of color impressions with impressions previously received through other senses and stored away as memories, all play a part in determining the effect which a color or a pattern of opposed colors, has upon us. But even granting all this, there are many of the physiological laws of color vision which must be adhered to before we can expect to produce these effects.

In attempting to show how these laws may be employed in art it will be necessary for us to explain briefly some of the physical and physiological observations upon which they depend. The first of these is a physical one: it is the dissociation of white light into the spectral colors by means of a prism, or better, by means of a diffraction grating.[1] The spectral colors are red, orange, yellow, green, blue (indigo) and violet, the various shades of purple being entirely absent. When we look at such a spectrum we are at once struck with the fact that the colors differ from one another not only in their hue but in their brightness or luminosity, the yellow and the immediately adjacent portions being much brighter than the others. At once then we recognize two physiological properties for each spectral color, hue and brightness. There is, however, another property of colors as seen in nature which is absent in the spectrum, namely saturation. This refers to the degree of white light with which the color is mixed. It is more or less related to the artist's "value" which expresses the translation of the colors into gray.

The most characteristic of these properties of colors is their hue, and for the present we shall confine our attention to this. To understand what the hue is due to we must remember that rays of light exist in space as vibrations of the surrounding ether and that these vibrations occur at right angles to the line of propagation of the light rays. The rate of the vibration varies according to the hue. In other words, the light rays are made up of waves which are small and close together when the vibration is rapid, as at the violet end of the spectrum, and are large and wide apart when the vibration is slow, as at the red end. When these waves strike the retina they create impressions which differ from one another according to the wave-lengths. These differences we interpret as differences in hue. When the rays of the various spectral colors are reunited before striking the retina, the sensation which is created is that of white. This recombination of the spectral colors, which is called synthesis of colors, may in general be brought about in two ways: (1) by causing them to fuse together by means of some suitable optical device (such as a second prism, or reflecting mirrors) before they enter the eye, (2) by causing them to become superimposed upon one another on the retina in rapid succession, in which case the impression created by each color lasts for a sufficient length of time so that it becomes fused with those which succeed it. This result depends on the phenomenon of positive after-images; which can be demonstrated by momentarily regarding some brightly illuminated object and then closing the eyes, when the image continues to be seen for some time. Rapidly succeeding images therefore become fused into one composite impression. This retinal synthesis, as we may call it, is well illustrated in the impression produced by observing the spokes of a rapidly revolving wheel.

For experimental purposes it is brought about by using Maxwell's machine, which consists of circular cards painted in sectors with the various colors and which are caused to revolve around their centers by means of a motor. A spinning top may also be used for this purpose. By revolving a card painted with the seven spectral colors a sensation approaching that of white is produced,[2] by choosing various proportions of the spectral colors this white becomes tinted with all possible intermediate hues.

From these facts we might imagine that the retina contains a special kind of sensory component for each of the seven spectral hues, that equal stimulation of all produces the sensation of white and that varying degrees of stimulation of certain of them, that of the hues which are intermediate between those of the spectrum.

Such a hypothesis could not however be of much practical value in explaining the color phenomena with which we have to deal in daily life. It had to be simplified. This was done by Thomas Young and Helmholtz, who discovered that three of the spectral hues, such as red, green and violet, or certain other triads, are sufficient, when mixed on the retina, to produce the same sensations as those which are produced by the seven spectral hues. These are known as primary colors; when equal quantities of each are used a sensation of white (or gray) results; when only red and green, the sensation is yellow; when green and violet, it is blue; and when violet and red, it is purple. Not only this, but the various intermediate hues can readily be obtained by altering the proportions of the primaries; thus, to produce orange a disc containing a larger proportion of red and a smaller proportion of green is used, and so on.

Fig. 1. Color Triangle.

To represent these fundamental facts and hold them in mind the so-called color triangle has been constructed. At the angles of this triangle are placed the primary hues, the other spectral hues being distributed along its two sides at distances which are proportional to their wave-lengths and the purples along its base, which, since these hues are absent from the spectrum, is represented by a broken line.

But white light can be produced in still another way, namely by retinal synthesis of certain pairs of hues which on this account are called complementary. Thus red and greenish-blue, yellow and blue, orange and blue-violet are complementary. We may express this all important fact by stating that for every spectral hue there is another which when mixed with it on the retina in approximately equal quantities produces the sensation of white. When other than equal proportions of complementary hues are chosen, colors are produced which are of hues intermediate between those of the complementaries and which are mixed with varying degrees of white. They are incompletely saturated colors. These facts may be satisfactorily represented by finding a point, called ${\displaystyle W}$, inside the color triangle, so that any straight line passing through it will on striking the sides of the triangle join two hues which produce white. This method of finding the complementaries necessarily implies that they must be separated from one another by a considerable distance on the spectrum. For representing these facts a circle instead of a triangle may he employed, and for practical purposes, in the use of colors in painting, such a circle has been found more useful than the triangle. Before we proceed to explain its use, however, it may be well to indicate some of the applications which can be made in art of the facts we have already learned.

It is in pointilism that this application is most evident. In this method the pigments are laid down in minute areas or spots or lines so that, when the picture is viewed from a certain distance, the different hues act on the same nerve endings of the retina and therefore produce the same effect as if they had been superimposed, as by the use of Maxwell's discs. Thus, if a white surface be dotted over with red, green and violet, or any other primary colors, or with red and greenish-blue, or any other complementary colors, the surface at a certain distance will appear grayish white. If, in any of the combinations, one hue be in preponderance of the others the gray will become correspondingly tinted, so that a complete picture may be built np of areas which on close inspection are a mosaic of pure colors but appear at a distance as tinted grays.

The impressionists, Monet, Segantini, etc., appear to have laid as the basis of their picture a gray at the brightness (or value) which they desired each portion of it to assume. On these surfaces they then applied color more or less pointilistically. The neo-impressionists, such as Seurat and Segniac, on the other hand, went a step further in that the saturation was made to depend entirely on the synthetic principle. They laid on their pigments strictly in dots on a surface which was as nearly pure white as possible. Some of these neo-impressionists had, however, already begun to apply certain of the principles of color apposition in masses which we shall study later. To build up a picture pointilistically must obviously greatly increase the technical difficulties of the artist, especially with regard to outline and form; his freedom of expression is also seriously curtailed. It becomes necessary therefore that very great advantages should be the outcome of such labor. Among the advantages are the sense of atmosphere, the vibrating, scintillating quality of the color areas and the very satisfactory transitions at the edges between them, all of which are qualities that can be rendered in no way so satisfactory as by pointilism.

There can be little doubt that a great part of the peculiar impression produced by pointilism depends upon the slight movements which the eyeballs are constantly undergoing, even during our most intent fixation. This of course produces a certain amount of overlapping of the colors on the retina just as when they are superimposed by means of Maxwell's machine. In the same way vibrations of the eyelids by moving the eyelashes across the palpebral cleft assist the synthesis, this being made evident by half closing the eyes, a method often used in studying pictures.

The success with which the desired impression can be created in a pointilistic picture often depends upon the purity of the colored dots, its vibrating quality being at the same time much enhanced by leaving a narrow margin of white around each dot. When this is successfully done there comes into play another physiological process known as flicker, which can be experimentally produced by rotating discs with black and white sectors at a speed which is just insufficient to cause a uniform gray. The resulting flicker possesses a glittering quality which makes it appear of distinctly greater brightness than the gray which results from complete synthesis. The same thing may be seen by observing the spokes of a wheel revolving at different velocities. Instead of black and white the sectors may be composed of different hues.

In the flicker experiments the gray remains of the same degree of saturation at whatever rate the disc is revolving, provided it is revolving more quickly than is necessary to produce complete fusion, and so in pointilistic painting, when the picture is viewed beyond the distance at which fusion occurs the impression is practically that of the older painting. It must be viewed at a distance just short of that which is necessary to produce complete synthesis. The post-impressionists such as Cezanne, Matisse, etc., realizing this limitation in pointilism, have been searching after a method by which the color scheme maintains its effect on us at whatever distance the picture is viewed. The physiological principle upon which this depends is that known as contrast, and this we will now proceed to study. Being a property exhibited most strikingly in the case of complementary hues, it becomes necessary for us to have, besides the color triangle, some simple experimental methods by which the complementary hues may be determined. Such methods include the experiments of simultaneous and successive contrast, in connection with which many facts of fundamental importance in the use of pigments are brought to light.

Simultaneous contrast is illustrated by regarding a strip of gray against a colored field when the gray becomes tinted with the complementary hue. There are two simple methods for performing this experiment, one is to spin a colored disc, midway between the center and circumference of which is a circle, composed partly of black and partly of white; this synthesizes to a gray which becomes tinted with the complementary hue of the colored field. The other way is to lay a narrow strip of gray paper (cut as a zigzag) on a colored sheet and then to cover the whole with thin tissue paper; the gray will assume the complementary hue. No experiments in color vision are more striking than these, nor are there any that have more direct application in the use of colors in picture painting; thus, a gray wall viewed against a sun-lit background of green is no gray, but like the piece of paper in our experiment it becomes tinted of a purplish hue. Similarly, a shadow cast on yellow sand is blue and one thrown on the skin when this is otherwise in strong light often acquires a striking quality of green.

The phenomenon of successive contrast is elicited by steadily regarding a patch of a certain color for some time and then either closing the eyes, or better still, directing the gaze to a neutral surface, such as a gray untinted wall. A vivid color impression of the same shape as that of the colored patch previously looked at will be seen in both cases, but exhibiting a hue which is complementary to that of the patch.

In the experiments above described the complementary color is demonstrated by the use of a gray surface. It is evident, however, that, if we cause it to be projected against a background which itself possesses a certain hue, the two hues (the complementary and that of the regarded surface) will become blended and will have the same effect as if they had been spun on a Maxwell's disc. For example, suppose we regard for some time a blue surface and then direct the gaze to one of red, the impression will be that of orange, because the complementary of blue, being yellow, fuses with red and produces orange.

Having determined the complementaries by means of these contrast methods we may confirm our results by color synthesis; thus supposing we have determined by the contrast methods that the complementary for a certain yellow is a certain blue, we may proceed to ascertain whether this is strictly the case by preparing discs composed of these two hues and rotating them on Maxwell's machine. If the hues are complementary the greatest possible degree of whiteness will be produced.

Successive contrast finds only a limited application in art, although it is of course conceivable that the intensive fixation of one colored area in a painting, or a design, might, by successive contrast, greatly modify the colored impression created by shifting the eyes to another part. It is improbable, however, that any artist, either intentionally or unintentionally, has laid on his pigments with this object in view. Nevertheless, successive contrast may assist us greatly in the actual determination of the complementary hue. Thus, to take again our example of the gray wall against the green background, we may exaggerate the effect of the green on the gray by regarding the green for some time and then shifting the gaze to the wall, when its purplish hue will be found to be much intensified. On the other hand, simultaneous contrast is of paramount importance in art; indeed it is as important in the final impression produced by a painting or a design as any other quality which this may possess. This importance depends on the fact that when two colored surfaces are placed in apposition each becomes changed as if it were mixed to a certain extent with the complementary hue of the other; or if a gray or a tint of low saturation (see p. 460) is apposed against a saturated color field it will assume a complementary hue of greater or less saturation according to the relative area of brightness of the apposing areas. By applying these principles in picture painting unsaturated hues may be caused to assume much greater degrees of saturation while, if the apposition be false, hues in themselves of almost complete saturation may become dull and subdued.

To the artist it comes to be of the highest importance that he possess some easily remembered scheme by which he can predict these contrast effects. The color triangle may be thus employed, but a simpler, though perhaps less scientific device, for the same purpose is the chromatic circle of Rood. To construct such a circle we must know the wave-lengths of the various colors which we desire to contrast.[3] The differences in wave-lengths are then calculated so as to correspond to angular differences, these angles being formed by the radii of the circle. As in the color triangle, opposite radii will join complementary colors and the center will represent white light, i. e., the nearer the center the less will be the saturation of the color.

Fig. 2. Rood's Chromatic Circles as used to show the Influence of one Color on the others.

If one such circle, drawn on transparent paper, be superimposed on another, the effect which is produced by contrasting two colors can be readily ascertained. Thus, suppose we desire to determine the influence which red has when contrasted with the other colors. Having accurately superimposed the two circles we move the transparent one so that the point on it which corresponds to red is displaced along the line joining red and its complementary, blue-green. The colors on the upper circle will now stand in positions on the lower corresponding to the changes in hue and saturation which they would have suffered by contrast with red. Thus orange will stand nearer the center and somewhat nearer yellow, whereas green-blue will merely be removed farther from the center, which means that orange will become less saturated and yellower, whereas green-blue will increase in saturation but be unaltered in hue.

In general we may say that the effect produced by contrasting two colors is to move them farther apart on the chromatic circle, thus causing mainly a change in hue in the case of colors that stand near one another, but making a change in saturation in those which are far apart.

In order that the contrast effects may be taken full advantage of, certain conditions must be fulfilled. The most important of these are as follows: (1) The complementary tint which gray assumes is most vivid when it is somewhat darker (i. e., of less brightness, see p. 460) than the hue against which it is apposed, in the case of the warm colors (the reds, oranges and yellows), and when it is lighter in the case of the cold colors (the greens and blues). The dividing line between the warm and cold colors may be taken as that joining the complementaries, 5'ellow-green and violet. (2) When a color of low saturation (i. e., nearly a gray) is apposed to one of high saturation and of complementary hue, the former will become more saturated, and conversely, if two colors which are identical in hue but of unequal saturation be apposed, the paler one may appear gray. When they are not complementary, the hue which undergoes the greater change is that which is the paler. (3) The greatest effects are produced when the color field, whose hue it is desired to alter, is much smaller in extent than that of its complementary and when it is completely surrounded by the latter. By placing a thick black line between the areas the complementary effects may be suppressed. Thus, the complementary hue which a piece of gray paper placed on a colored field assumes when it is viewed through tissue paper becomes much less evident if a thick black line be drawn on the tissue paper at the edge of the gray. When the color areas are large it is at the edge only that the complementary influence is noticeable. On the other hand when a colored area is very small it undergoes no complementary change, but merely blends with the neighboring color. (4) To obtain full advantage of color apposition the colored patterns should be very simple and of similar texture and their surfaces should be broken up by detail to the least possible degree. (5) The most marked complementary effects are obtained when the opposing hues are of equal brightness.

When we attempt to employ the chromatic circle for another purpose, namely for determining what will be pleasing and what displeasing color combinations, we find that its use is somewhat limited. This is because a psychological influence enters into our judgment in such cases. In general however it may be taken as a working hypothesis that good combinations are always more than 80°-90° apart on the circle, that is, they should be separated from one another by about one quarter of the circumference. Even complementaries may form displeasing combinations (i. e., certain reds and greens), in which case, as Rood has pointed out, the hues are usually far removed from the line which separates those that are cold and warm. When we are compelled to appose hues having a hurtful influence on one another, the unpleasing impression which they create may be lessened by certain tricks, such as by assigning one of the hues to a much smaller field, or by decreasing the saturation of one of them, or by adding a third hue whose position on the chromatic circle is as far as possible removed from the others: thus the disagreeable effect of a yellowish-green and yellow is much improved by the addition of some violet, etc.

So far, for simplicity sake, we have regarded but one quality of a color, its hue, although in doing this it has been impossible entirely to neglect the closely related qualities of brightness and saturation. These we shall now proceed to consider.

Brightness is most marked, under ordinary conditions of illumination, around the yellow portions of the spectrum. It is a property which is exhibited in marked degree by different grays. Indeed it is measured by finding a gray which appears of equal brightness to that of a given color. Such measurements may be made with considerable accuracy by finding a gray background against which the color becomes indistinguishable when viewed by the very outermost portions of the retina which are color blind, that is, which see no hue in a color but only a grayness, the degree of which is proportional to the brightness of the color.[4] To make such comparisons, the person must regard a dot in the center of a plain black surface and must then gradually move a small piece of colored or of gray paper, mounted on a suitable handle, from the periphery towards the center of the surface. At a certain position the colored paper will be seen as gray because the rays of light from it are striking the color-blind areas of the retina. Various grays are used until one is found which matches exactly with that created by the colored paper. A still simpler method consists in rotating the color on a Maxwell disc along with a synthetic gray. In this case judgment of equality may however be somewhat confused, on account of the gray assuming the complementary hue.

Brightness plays a most important part in the phenomenon of contrast, for not only is the simultaneous contrast of hues obtained most strikingly when these are of equal brightness, but we constantly experience brightness contrast itself. Thus pieces of the same gray paper placed on gray backgrounds of varying degrees of brightness do not look at all alike. It is particularly at the border between the two grays that contrast brightness is most evident. This subserves the function of creating a sharp border between the grays, and it can be demonstrated by causing strips of different gray papers to overlap one another like the tiles of a roof or, still more strikingly, by rotating a disc on which when spun appear three circles of different grays, each synthesized from black and white. In both experiments the grays, though really perfectly uniform, will appear as if shaded from their edges.

Since we measure brightness in terms of grayness, and since it is most marked at the yellow portion of the spectrum, it follows that if we desire, for successful contrast effects in picture painting, to appose yellows with blues or deep reds, we must employ some artificial means either to increase the brightness of the blues or reds or to decrease that of the yellows. This can be done by mixing the pigments with white (or black), that is to say, we may alter what the artist speaks of as the value of the color but which in so far as white is used for producing the alteration is more correctly called the saturation.

It may indeed be said that the object sought in mixing pigments with white (i. e., changing their saturation) is to give the impression that their properties of brightness have been altered.[5] When it is desired to raise the brightness of a given color, we can succeed only to a limited degree by using more pigment; to obtain it further, we must, as already explained, employ the property of simultaneous contrast. These methods used by the artist to alter the brightness of his colors are however liable to have a dulling effect on the whole composition unless they are used with great care and judgment. When he is compelled to lower the saturation of one color he must be careful to apply those neighboring on it in such a manner as to give the impression that the whole of that portion of the picture is of the same brightness. This he may do, either by making his pigments of similar saturation or by assorting the size of the colored areas, so that they appear by contrast to be of similar saturation.

It is a well-known fact that our judgment of the relative brightness of colors, and to a certain extent of their hues, becomes altered when the conditions of illumination are changed. A picture viewed in broad daylight may create a very different impression from that which it produces in dull illumination. For example, its hues may be dull and muddy under the conditions of illumination that are ordinarily present in a dwelling, or even in a gallery, whereas when viewed in broad day light it may sparkle with brilliancy; or there may be very little change in the actual hues, but the portions of the picture which appeared to be of greatest brightness in broad daylight may in dull light actually shift to some other part. These changes are due to what is known as adaptation of the retina. The most striking illustration of this is furnished by observing the colors of a flower-border after sundown. Let us suppose that the border contains geraniums (scarlet), lobelia (blue), and coreopsis (orange). As darkness approaches it will be noticed that the red geraniums become duller and duller until at last they turn black; that the orange coreopsis also becomes more neutral, but that the blue lobelia maintains the same color qualities as it possessed in daylight. The most remarkable change of all occurs, however, not in the hues but in the relative brightness of the colors, for it will be noticed that the sensation of greatest brightness has gradually shifted from the reds and yellows to the blues and greens, so that the foliage and the lobelia may actually come to appear brighter than the coreopsis and the geraniums. It is needless to point out how important an appreciation of these adaptations must be to the artist, how careful he must be to paint his picture in the degree of illumination in which he expects it to be viewed. The physiological explanation of this adaptation is that the outer portions of the retina assume a much greater degree of sensitiveness in dull light, indeed they come to be more sensitive than the central portion itself. This curious change explains why without directly looking at it we may be conscious of the presence of a small light in the darkness—a star for example—which however disappears when we direct our gaze to it. The ability of the thus sensitized outer portions of the retina to judge colors differs from that of the central portion.

When we come to apply many of the principles of chromatics in art, we are met with difficulties which at first sight may appear to be insurmountable. In most instances, however, this is by no means the case, and we shall now endeavor to show how certain of these difficulties can be explained. First of all, with regard to the mixing of pigments as compared with the mixing of colored lights, of course the two processes yield very different results: for example, mixing yellow and blue lights, as we have seen, produces almost pure white, whereas mixing these colors as pigments, as every artist knows, produces green. The entire want of similarity in the results which follow the mixing of colors by the two methods has had the effect of making some artists conclude that the laws of chromatics are useless as guides in the practical use of pigments. But this is wrong, the apparent difference being really due to a very simple cause, namely to the fact that by mixing pigment we subtract color rays from entering the eye, whereas we add such rays when we mix colored lights. To make this clear let us return to our example of blue and yellow. When we use these as pigments, we must remember that the pigment particles have a certain degree of transparency so that light partly penetrates them, certain rays being then reflected and certain absorbed according to the hue. A blue pigment, for example, absorbs all constituent rays of white light except the blue and the hues which border on blue in the spectrum, it being impossible to procure pigments which are so pure that they do not let some other hues besides their own characteristic one pass through them. Similarly with yellow, it absorbs all the spectral rays save the yellow, the orange, and the green. Adding these two pigments together, we get every spectral ray absorbed except green, a certain amount of which both pigments have allowed to pass. In a similar way we can explain why blue and red give purple and why a mixture of all the spectral colors as pigments produces a dark gray of uncertain hue.

The above applies to a matt surface; when there is any trace of glaze there comes into play another factor which we must now consider, namely, surface reflection of some white light which has not penetrated the pigment particles at all and which therefore causes the color to be more or less unsaturated. It is by diminishing surface reflection of white light that the colors of a picture may be raised in saturation by subjecting it to alcohol vapor, which softens the medium and removes surface cracks. Reflection of white light also takes place at the surface of the pigment particles themselves and is greatly diminished when these are extremely small, hence the importance in the manufacture of pigments of thorough grinding. It is further minimized by suspending the pigments in oil, because this causes the light before it strikes the surface of the pigment particles to pass through a medium which is of approximately the same density as that of the particles themselves. This reduces the reflection, because the greater the difference of density between two media the greater the reflection of light at the interface between them.

The quickly vibrating (blue) rays of the spectrum tend to be reflected more readily than the slowly vibrating (red) rays, hence we often find that a substance is bluish by reflected light, whereas it is reddish when the light passes through it. It is indeed for this reason that during the day the sky looks blue, the light being reflected from the fine particles of dust and moisture which are constantly suspended in it, whereas after the sun has set it is red because the slanting rays come to be transmitted through these particles.

Artificial illumination alters the hues of pictures mainly because of mixture of colored lights, that is to say, of the hue of the light reflected from the surface of the picture and of the hue due to the particular pigments employed. Thus, if we regard a picture in yellow light (gas, carbon filament, etc.) the pale blues may appear white (mixing of complementary colors), the deeper blues assume a greenish hue, and the reds turn to orange.

In the colors which we see in nature influences of a similar kind are constantly at play, for every object, besides being illuminated by the prevailing light, has thrown on to it colors which are reflected from near by objects. In analyzing these influences there are, as Rood has pointed out, at least three factors that must be borne in mind. These are: (1) the natural or "local color" of the object, the cause for which we have already explained; (2) the colored light which is reflected unaltered from its surface, just as we have seen white light to be; (3) the portion of this colored light which is not entirely reflected but which penetrates the surface and is then reflected. Let us suppose that we are regarding a red wall of glazed brick at the edge of a grass lawn: the local brick-red of the wall will be materially altered by surface reflection not only of the white light but also of blue-green which, being approximately its complementary, tends to lower its saturation and pull it towards neutrality; at the same time, the green rays which have penetrated will on reflection assume a yellowish orange hue. The total effect is therefore that the red is somewhat removed towards neutrality and at the same time made to assume an orange hue. But it is by no means always possible to analyze these color effects, so that we must depend rather on the accuracy of the impression which we receive, at the same time bearing in mind that even objects with which we usually associate the most positive of hues may under certain conditions become entirely altered in this regard. In their use of colors, the post-impressionists are most careful to allow for these influences, although they may employ hues to produce them which at first sight appear to be entirely out of place.

Finally, we must say a few words about the relative refractability of different colors, that is to say, the ease with which the different spectral hues are brought to a focus on the retina. The rays of slow vibration, as at the red end of the spectrum, are less readily focused than those which vibrate quickly, as at the violet end. Consequently, when red rays are in focus, violet rays are over-focused and vice versa. The application of these principles in art depends on the fact that our judgment of distance is partly associated with the amount of effort which we must make in order to accommodate our vision. At rest the optical apparatus of the e.ye is accommodated for distant objects so that when these come nearer than a certain point an effort is required to make the focusing stronger. From the amount of this effort we judge in part of the distance of the object. Now it takes more effort to focus red than green or blue rays so that we always tend to locate a red object as being nearer than one that is blue or green. These facts can be very beautifully demonstrated by looking at red and green lamps placed side by side; the green light appears to be behind the red. And in picture painting the same principles can be applied, and seem to be so in many of the post-impressionists' paintings; objects are brought forward by being colored in the reds and they are pushed back by the use of blues and violets.

These facts bring us to a discussion of the influence of the blue-violet line which so many post-impressionists are using to outline objects to which they desire, without shading, to give the impression of rotundity, or more correctly, of projection. The effect of such a line is perhaps best demonstrated in still life studies where its existence at the edges of, say, a vase, will, when the picture is viewed at such a distance that the line just disappears, cause the vase not only to stand forward from its background but also make it appear rotund, as if shaded towards the edges. The line is sometimes used in landscape pictures with the object of holding the pattern together. These effects are most marked when the object is painted in hues that are considerably removed from blue on the chromatic circle, or are of much less saturation (more removed towards neutrality). Similar effects can sometimes be obtained by the use of a black line, but none of the flaring hues can be successfully employed for making it. It is difficult to explain the action of these outlines, indeed it is almost certain that several factors play a role in producing the illusion which they produce. When the line is a blue one and the prevailing hue of the color field which it borders tends towards yellow a synthetic gray will result at a certain distance, thus creating the impression that some space exists between the object and its surroundings. When a black line separates two colored areas there occurs a certain amount of irradiation on to it of the neighboring hues, which therefore undergo a more or less sudden lowering of intensity at its edges, which becomes more and more pronounced towards the middle of the line until the hues finally meet and partly overlap, thus producing a certain amount of synthetic gray. This phenomenon of irradiation is well illustrated by comparing two squares of equal size, one being black on a white field and the other white on a black field; the white square looks distinctly larger than the black one. The reason is that the stimulus produced by white, mainly because of imperfect focusing, spreads on the retina somewhat beyond the margin of its image.

In this account we have not essayed to explain all of the peculiar effects which are produced by some of the most modern creations of the so-called post-impressionists. We have merely indicated some of the physiological truths of color vision upon which certain of their color illusions depend. To go further would require consideration of many optical illusions for which at present there exists no satisfactory explanation. These are not illusions of color but illusions of line, indeed many of the latest post-impressionistic pictures are produced almost entirely in black and white and the peculiar emotions which they arouse depend on metaphysical processes whose explanation we can not undertake to expound. Their aim is "to create an illusion of the fact" rather than the fact itself; to write "a visual music which shall in itself arouse the emotions."

1. In the light decomposed by a prism some hues, such as those of red and yellow, occupy much less space than others, such as blue, although they do not correspondingly differ in wave-length. When light is decomposed by a diffraction grating (a glass plate ruled with very fine equidistant lines) the spaces occupied by the various hues are proportional to their differences in wave-lengths.
2. It would be pure white were it possible to obtain artificial pigments that reflected none other than their own characteristic hues.
3. This can be done by comparing the colors with those of a highly magnified spectrum of white light alongside of which is a scale of wave-lengths.
4. The power to judge hue depends on the presence in the retina of peculiar nerve endings called cones. These are absent from the peripheral portions and only gradually make their appearance towards the center. There is, therefore, a region between the periphery and the center of the retina which is partly color blind, blue and yellow being perceptible, but red and green still appearing as gray.
5. Brightness must be distinguished from color intensity which is purely a physical property and depends upon the amplitude of the wave-lengths.