Fig.1

Whether this theory of the spectrum formation is absolutely correct or not, the fact is established that the colors found in a prismatic solar spectrum are always the same under the same conditions and the order of their arrangement is never changed. By means of the quality of spectrum colors called the wave length, a given color can always be located in the spectrum, and hence if a spectrum color is selected as a standard it can always be determined by its recorded wave length.

Six Spectrum Standards of Color.

Therefore it seems possible to establish certain standards of color by a series of definitely located portions of the solar spectrum and in the system here presented six have been chosen, namely red, orange, yellow, green, blue and violet. These six are more distinctly recognized than the others, and from them by combination in pairs of colors adjacent in the spectrum all the other colors can be very closely imitated, and hence these six are selected as the spectrum standards. In these standards the most intense expression of each color is chosen i.e. the reddest red, greenest green, etc. which by the closest scientific investigation have been located by their wave lengths so that if they are in doubt in future they can be re-determined by individuals or if disputed, may be corrected by any authoritatively established congress, selected for the purpose. The wave lengths of our six standards are represented by the following numbers in ten millionths of a millimeter. Red, 6571; Orange, 6085; Yellow, 5793; Green, 5164; Blue, 4695; violet, 4210. Having thus scientifically established these unchangeable standards the attempt is made to secure the best possible pigmentary imitation of each.

To any one who has ever compared a piece of colored material with a good presentation of a spectrum color, it is unnecessary to say that the result in an attempt to match the spectrum color with the material or pigmentary color is a very weak approximation, but the one thing aimed at is to secure nearly as possible the same kind of color. For example in the red, it is the aim to obtain the same kind of red, by which we mean the same location in the spectrum, i.e. a red neither more orange nor more violet than the reddest spot in the spectrum. This selection must be based on a purely æsthetic perception or impression of color. The same is true of each of the six standard colors, as for example, for orange we select the location which has seemed to a large number of good judges to best represent the feeling of orange as between the quite well defined red on one hand and the equally definite narrow band of yellow on the other, and it is quite wonderful what unanimity of opinion there is on this particular color which would naturally seem to be the one most doubtful in its location. On the other side of the yellow the green seems to offer little difficulty and the pure Paris or emerald green is very nearly the standard. The violet being at the other end of the spectrum is as easily decided as the red, but the blue between the green and violet is not so easily determined, because, from the best blue the hue runs so imperceptibly into the violet on one side and the green on the other. Pure ultramarine blue is the nearest approach to the spectrum standard of blue of any of the permanent pigments, but even this is a trifle too violet.

For educational purposes papers coated with pigments afford at once the purest colors and the most economical and useful material, and on this plan a line of colored papers has been prepared for color instruction in the kindergartens and primary schools in imitation of the above described spectrum standards.

From the pure spectrum standards it is possible by reflected light to combine the two standards to produce a color between them, for example if two small mirrors are held in a spectrum one at the "red" and the other at the "orange" and the two reflected on to the same spot on a white surface, the result is a color between the red and the orange. So also if we mix red and orange pigments together we may produce colors between the two which may be termed orange-red or red-orange; but unfortunately there is no means known by which we can measure the proportion of the red and orange color-effect which is produced by any given mixture of these two pigments, because color-effect cannot be measured by the pint of mixed paint or the ounce of dry pigment.

The Color Wheel and Maxwell Disks.

We, however, have another means for measuring color effect which just in this emergency seems providential. It is a fact well known to every boy that if he rapidly whirls a lighted stick the fire at the end produces the effect of a circle of light, which phenomenon is explained by a quality of the eye called retention of vision, by which the impression made by the point of light remains on the retina of the eye during an entire rotation. It is a fact, based on the same quality of vision, that if one color is presented to the eye, and instantly replaced by another the effect is a combination of the two colors. Therefore if one-quarter of the surface of a disk of cardboard is covered with orange paper and three-quarters with red paper, and then the disk placed on a rapidly rotating spindle, the color effect is a mixture of red and orange, and the effect is exactly in proportion to the angular measurements of the two sectors, so that if the circumference is divided into 100 equal parts the resultant color will be definitely represented by the formula "Red, 75; Orange, 25."

Less than forty years ago an English scientist named J. Clerk Maxwell while making experiments with such painted disks happily conceived the idea of cutting a radial slit in each of two disks from the circumference to the center so that by joining the disks they could be made to show any desired proportion of each and hence they are called Maxwell disks. With such disks made in the six pigmentary standards red, orange, yellow, green, blue and violet, the intermediate pigmentary spectrum colors may be very accurately determined by combination and rotation. If we give to each of these standards a symbol as R. for red, O. for orange, Y. for yellow, G. for green, B. for blue, V. for violet, we then have the basis for a definite nomenclature of colors in imitation of the pure spectrum colors. As all pigmentary or material colors are modified by light and shade thus producing in high light tints and in shadow shades of the colors, we must seek for some means of imitating these effects, and fortunately find them in white and black disks. If with a standard color disk we combine a white disk we may have a line of tints of that color, and with a black disk, shades. Giving this white disk a symbol of W. and the black disk N. we complete our nomenclature. We cannot use B for black because B has already been used for blue, and therefore we use N. for niger, the Latin word for black.

The Bradley System of Color Instruction.

Briefly stated then this system of color instruction is comprised under the six general heads: Spectrum Standards; Pigmentary Standards based on the spectrum standards; Maxwell Rotating Disks in the pigmentary standards and Black and White; a Color Nomenclature based on the accepted standards and their disk combinations; and Colored Papers and Water Colors made in accordance with these standards.

For spectrum standards, six definite locations expressing the natural æsthetic or psychological impressions of red, orange, yellow, green, blue and violet are selected. Six standards are chosen instead of a larger number as for example twelve, because for the purpose of a nomenclature the smaller number is more convenient than a greater number. The six are selected rather than three, four or five, because while in the consideration of colored light alone the smaller number would possibly suffice to form by combinations imitations of all other colors, any number smaller than six is entirely inadequate to form by pigmentary or disk combinations fairly good expressions of the corresponding spectrum color combinations.

In selecting the spectrum standards special prominence has been given to the psychological color perceptions of experts in determining those locations in the spectrum best expressing the color feeling of red, orange, yellow, green, blue and violet, while the purely scientific consideration of these several questions has not been ignored or lightly treated.

For pigmentary standards the best possible pigmentary imitations of the six spectrum standards are secured and to these are added the nearest approach to white and black that can be produced in pigments.

Pigmentary standards on which to base a nomenclature are valueless without some means by which measurements of standards embraced in a given compound color can be expressed.

The Maxwell color disks are the only known means by which we may measure the relative proportions of color effect embodied in a given color, and therefore the eight color disks are the foundation of the original color nomenclature herein advocated.

Colored papers are chosen for primary color instruction because paper is a valuable medium for simple schoolroom manual training and because no other pigmentary medium is at once so economical and affords such pure colors as may be secured in specially prepared colored papers, without a glazed surface.

Before leaving this part of the subject we do well to remember that in the present conditions of chemistry as applied to the preparation of pigments it is not possible to establish any absolutely definite science of such color combinations. Nor is it possible to establish permanent pigmentary standards without great expense, but if the locations of the standard colors in the spectrum are established by wave lengths the pigmentary standards may be re-determined at any time and produced, in the purest pigments available at the time. In art or harmony effects, the purity of the pigmentary standard is not so important as its hue, i.e. its location in the spectrum, which may always be determined by the established wave length. This last statement may be illustrated by the investigations regarding complementary harmonies. Scientifically one color is not considered complementary to another unless when combined in equal quantities they produce white light, or in other words when combined by the rotation of disks each color must occupy a half circle and the result must be a neutral gray. But this is not essential in considering a complementary harmony, as harmonies in different tones and in various proportions are pleasing and as yet the proportions and tones which produce the best combinations have not been determined.

The entire question of harmonies or pleasing color effects is dependent on individual color perception, and the establishment of rules and laws on these points can result only from a comparison of the opinions of many experts in various localities and at different times. This cannot occur without some means for recording these opinions in generally accepted terms. It is too late for any individual opinion to be accepted as authority regarding the relative values of two different harmonies in color and this will be still less possible as we become better educated in color and able to sense finer distinctions in color combinations.

Color Definitions.

MONG other advantages to be gained by a logical study of the psychology of color is the establishment of more accurate color terms and definitions. If experiments and discussions based on accepted standards and methods of comparisons can be carried on we may hope in time to have as definite expressions of color terms as we now have in music and literature.

All color terms used by artists, naturalists, manufacturers, tradesmen, milliners and the members of our households are as indefinite as one might naturally expect from the utter lack of a logical basis for the whole subject.

Without definitions or means for intelligently naming any color, it is not strange that the terms used in speaking of colors and color effects are so contradictory as to lose much of their force, if perchance they retain anything of their original meaning. For example, probably most people apply the term SHADE to any modification of a color, either a hue, tint or shade.

It is true that a concise and reasonably full dictionary of color terms must be the outcome of long experience in the logical study of the science of color and its use in our every-day lives, and at the best only suggestions can be made at present. But as there must be a beginning and some terms seem to be fairly well established, the following incomplete list of definitions is offered, always subject to amendment by the majority vote, for whenever such changes indicate advance they should be welcomed.

Ray of Light.—The finest supposable element of light impression in the eye.

Beam of Light.—A number of rays.

Standard Colors.—As used in this system of color nomenclature, the best pigmentary imitation of each of the six spectrum colors red, orange, yellow, green, blue and violet and black and white. These are more specifically called Pigmentary Standards in distinction from spectrum standards.

Spectrum Standards.—The six colors found in the solar spectrum and definitely located by their wave lengths, as follows in the ten millionths of a millimeter. Red, 6571; Orange, 6085; Yellow, 5793; Green, 5164; Blue, 4695; Violet, 4210.

Pigmentary Colors.—All colors used and produced in the arts and sciences. This is in distinction from colors seen in nature, as in flowers and the solar spectrum. The term refers not only to pigments in the strictest sense but to all surfaces coated, painted or dyed artificially.

Pure Colors.—A pure or full color, also called a saturated color, is the most intense expression of that color without the admixture of white or black or gray. All spectrum colors are pure, while no pigmentary color is absolutely pure, but the pigmentary color which approaches most nearly to the corresponding color in the spectrum must be selected as the pigmentary type of purity of that color. For example, the standard for green must be the best possible pigmentary imitation of the spot in the spectrum which by general consent is called green, and so not only for the six standards but for all their combinations which produce the other colors in nature.

In pigmentary colors the term pure is entirely one of relative degree. As processes of manufacture are improved and new chemical discoveries made, there is good reason to believe that we shall have much more intense colors and hence much better imitations of spectrum colors than are at present possible. Therefore as our pigments become purer those now accepted as full colors will in time become tints or broken colors and new standards will be adopted.

Hue.—The hue of a given color is that color with the admixture of a smaller quantity of another color. An orange hue of red is the standard red mixed with a smaller quantity of orange. With the disks, pure hues are secured only by mixing two standards adjacent in the spectrum circuit.

For convenience in speaking and writing about colors in this system of color instruction, all the spectrum colors other than the six standard spectrum colors are designated as intermediate spectrum hues, and often for convenience in speaking of them they are called simply spectrum hues. To these are also added the colors between red and violet which are not in the spectrum. When so used the term must be considered as purely technical in this particular relation, because a color between the standard blue and the standard green is in the abstract no more a hue than either of these colors. If two standards not adjacent in the spectrum circuit are combined the result is not a pure spectrum hue but always some broken spectrum color.

Local Color.—A term applied to the natural color of an object when seen in ordinarily good daylight and at a convenient distance, as a sheet of paper at arms length, a tree at twice its height, etc.

Tint.—Any pure or full color mixed with white, or reduced by strong sunlight. In the disk combinations a spectrum color combined with white.

Shade.—A full color in shadow, i.e., with a low degree of illumination. In disk combinations a spectrum color combined with a black disk produces by rotation a shade of that color. In pigments the admixture of black does not usually produce as satisfactory shades of a color as may be secured with some other pigments, and each artist has his own preferences in making shades of the various colors on his palette.

Scale.—A scale of color is a series of colors consisting of a pure or full color at the center and graduated by a succession of steps to a light tint on one side and a deep shade on the other.

Tone.—Each step in a color scale is a tone of that color, and the full color may be called the normal tone in that scale. In art this word has had such a variety of meaning as to render it very convenient for Amateur Art Critics, together with such terms as breadth, atmosphere, quality, values, etc., but in the consideration of color it should have this one definite meaning.

Warm Colors.—Red, orange and yellow, and combinations in which they predominate.

Cool Colors.—Usually considered to be green, blue and violet, and the combinations in which they predominate. But it is, perhaps, questionable whether green and violet may properly be termed either warm or cool. The term cool as applied to colors is quite indefinite, except in a general way, but red, orange and yellow are universally considered as warm, and blue and green-blue as cool.

Neutral Gray.—White in shade or shadow. Pure black and white mixed by disk rotation. Black and white pigments mixed do not usually produce a neutral gray, but rather a blue gray.

Warm Gray.—A neutral gray with the admixture of a small quantity of red, orange or yellow.

Cool Gray.—A neutral gray with a small quantity of blue or green-blue.

Green Gray.—A neutral gray having combined with it a small quantity of green. As this color could hardly be classed with either warm or cool grays this fourth class of grays is suggested as helpful in giving definiteness to the more general color expressions.

Broken Colors.—Gray colors, often improperly called broken tints. For simplicity, a tint of a color is described as the pure color mixed with white and a shade as the color mixed with black, and the corresponding broken color is the same color mixed with both white and black or with neutral gray. A tint of a color thrown into a shadow or a shade of a color in bright sunlight gives a broken color. For various reasons a very large proportion of the colors in nature are broken. Broken colors are much easier to combine harmoniously than full colors, or even tints and shades.

In disk combinations when a pure color is combined with both a white and black disk the result will be a broken color. When a color is mixed with both black and white, i.e., with gray, and becomes thereby a broken color, it then belongs to a broken scale and educationally has no place in any pure scale, i.e., a scale in which the key tone is a pure color. Neither has a broken scale of a color any place in a chart of pure scales or spectrum scales.

Neutral Colors.—A term often improperly applied to grays, white, black, silver and gold. See passive colors.

Passive Colors.—A term suggested as covering black, white, silver, gold and very gray colors. The term "neutral colors" is often used in this sense but this is evidently improper if we are to confine the term "neutral gray" to the representation of white in shadow because as soon as a gray has any color in it, it is no longer neutral.

Active Colors.—Those colors neither passive or neutral. Necessarily both the terms "active" and "passive" used in relation to colors must be quite indefinite.

Complementary Colors.—As white light is the sum of all color if we take from white light a given color the remaining color is the complement of the given color. When the eye has been fatigued by looking intently for a few seconds at a red spot on a white wall and is then slightly turned to the wall, a faint tint of a bluish green is seen, and this is called the accidental color of the red, and is supposed to be identical with its complementary color. If with the disks we determine a color which with a given color will produce by rotation a neutral gray, we have the complementary color more accurately than by any other means at present known in the use of pigmentary colors.

Harmony.—Two colors are said to be in harmony or to combine harmoniously if the effect is pleasing when they are in juxtaposition or are used in a composition.

Spectrum Circuit.—If a pigmentary imitation of the solar spectrum with the addition of violet red at the red end and red violet at the violet end be made, and the two ends joined, we shall have a spectrum circuit. This may be in the form of a circle, an ellipse or an oval.

Primary Colors.—In the Brewster theory red, yellow and blue. In the Young-Helmholtz theory red, green and violet are termed primary colors because it is supposed that from these three sensations all color perceptions are experienced. In purely scientific investigations of color perceptions these last three or others which are supposed to serve the same purpose are also called fundamental colors. Practically every spectrum color is a primary, because each has its own wave length.

Secondary Colors.—In the Brewster theory orange, green and purple have been called secondary because it is claimed that they are produced by the combination of primary colors in pairs.

Tertiary Colors.—A term used in the Brewster theory to denote three classes of colors called russet, citrine and olive, made by mixing the secondaries in pairs. These are all broken spectrum colors. The orange and purple produce russet; the orange and green form citrine; the green and purple, olive. There seems to be no good reason for perpetuating the indefinite terms secondaries and tertiaries as applied to color.

Values.—This word is very freely used in discussing effects in works of art, both in color and in black and white. At present it seems to be a very difficult term to define, and yet each artist is quite sure that he can "feel" it, although few will attempt to put into words a definition satisfactory even to themselves. When an engraver, who is also an artist, attempts to interpret nature in black and white on the metal plate or wooden block, he endeavors to reproduce the "values" of the various parts of the subject before him. In doing this he, for one thing, attempts to produce a variety of neutral grays which will express to the eye by means of black and white lines the same tones of color effect as are seen in the several parts of the subject under investigation. If this were the whole problem the matter would be easily expressed by the disk nomenclature. For instance, if we are to consider a certain red object which may be represented by the standard red disk, we place a medium sized disk of that color on the spindle, and in front of it, smaller disks of white and black united. By rotation the white and black disks become a neutral gray at the center of the red disk. If this gray is made nearly white all observers will agree that the gray is lighter than the red, and if it is nearly black the opinion will be equally unanimous that it is darker than the red. Consequently there evidently must be a gray somewhere between these two extremes which a large majority of experts may agree to be equal in depth or tone to the red, i.e., neither lighter nor darker. But the artist-engraver will insist that to him the term "value" expresses much more than this and that he must use different lines in the sky or distance from those which he uses in the foreground; and some engravers will also insist that two different colors in the foreground must receive different treatment with the graver in order to express their true values. We know that true values of colors are not expressed in a photograph, as the warm colors are too dark and the blue far too light. If the term "value of a color" is to be used as expressing something more than a neutral gray of such a tone as to seem equal to it, then possibly this latter quality must be expressed by the word tone, and yet this use of that word will seem to enlarge its scope beyond its present limits as it now is used to express the relations between the different localities in one scale of color, while this new use will extend to the comparison of tones in various color scales, including neutral grays.

Luminosity.—The luminosity of a color is determined by comparing it with a neutral gray. When a color seems to be of the same brightness as a given neutral gray, i.e., not lighter nor darker, then that gray is its measure of luminosity.

A noted authority says: "No colored object can have the luminosity of a white object reflecting practically the whole of the light impinging upon it. Therefore if we take absolute reflection as 100 a fraction of 100 will give the relative luminosity of any body." Luminosity is another expression of the quality above described as forming a prominent feature in the term values.

Potentiality.—The ability or strength of a color to affect other colors by combinations with them. For example, white has a greater potentiality than black, yellow greater than red, and violet the least of all the spectrum colors.

It is a pertinent question whether any quality is involved in this term which is not found in value, tone and luminosity, but it expresses a somewhat different phase of a line of color effects.

Quality.—This term seems to be used rather indefinitely when applied to color, but perhaps it is not far removed from the term hue or kind of color.

Practical Experiments

Illustrating the Theory of Color.

Fig. 2.

In the foregoing pages an attempt is made to explain clearly and as briefly as possible the principles on which the Bradley system of color instruction is based, and also to suggest a few definitions necessary to an intelligent discussion of the general subject of Color. Owing to the peculiar nature of the questions involved, demonstration by actual experiment is more convincing than the mere statement of theories can possibly be, and therefore a few of the following pages will be devoted to the explanation of some valuable experiments, all of which may be tried by the teacher in private, while many of them can be shown the pupils with great advantage.

In this system the Maxwell color disks are the means for color combinations and the basis for measurements, and therefore for a color nomenclature. For this reason the present chapter treats largely of the proper use of the wheel and incidentally the theory of red, yellow and blue primaries with combinations to produce secondaries and tertiaries. No teacher using the material connected with this color scheme can hope to meet with success without a knowledge of the principles on which it is based, and in this subject as in all others, it is essential that the teacher shall know much more of it than he or she is ever required to teach.

The Color Wheel.

Fig. 3.

For most convenient use the machine should be clamped to the front of a table and near one end, so that the speaker using it can stand at the end of the table and operate it with the right hand. Fig. 2 represents the Normal School Color Wheel showing the face of the disks as seen by the audience. Facility in the operation of the Color Wheel is rapidly acquired by practice and the exact position is easily determined by the operator after a few trials.

Fig. 3 shows the Primary School Color Wheel, which has only two sizes of disks, while the largest machine has four sizes and is much finer in construction. The smaller machine does not require clamping to a table, but may be steadied by the left hand while being operated by the right hand.

The Color Top.

Fig. 4.

Many of the experiments of the color wheel can be produced with a small toy called a Color Top, which is shown in Fig. 4. It is composed of a thick cardboard disk forming the body of in the operation of the Color Wheel the top and a central wooden spindle on which the disk closely fits. A number of colored paper disks are provided with this top so that very many of the experiments performed before a class can be repeated individually by the pupils and in this way the facts which may have been demonstrated to the class with the color wheel can be fixed in the minds of the pupils by their own experiments with the top. Also as a home toy in the hands of the pupils it can be of value, not only to the children, but to the parents as well.

Use of the Disks.

Fig. 5 shows the method of joining two Maxwell disks and Fig. 6 their appearance when properly joined to be placed on the rotating spindle of the color wheel. In joining two or more disks for use on a color wheel or top, care should be taken to place them in such relation to each other that when rotated the radial edges exposed on the face toward the audience will not "catch the wind." With small disks on the color wheel this is not important, and if there is no whole graduated disk on the arbor behind the slitted disks there is no advantage, but in using the larger disks it is well to put the graduated disk behind the others for this purpose, as at best it is quite laborious to keep up speed when using several of the large disks, even with the best possible conditions. With the thin paper disks of the color in the operation of the Color Wheel top this is an important matter. It will be noticed that the method of joining the disks for use on the Color Top is the reverse of that to be observed with the disks of the Color Wheel as shown in Fig. 5.

Fig. 5. Fig. 6. Fig. 7.

Fig. 7 shows the same two color disks placed in front of a large white disk having its edge graduated to one hundred parts, so that the relative proportions of two or more colors to be combined can be determined accurately.

As the smaller disks offer so much less resistance in rotation than the larger ones they are most desirable in private experiments or before a small class, and the largest disks of the Normal School Wheel are necessary only when more than three expressions of color are required to be shown at the same time. In making experiments before an audience those persons in front should if possible be at least ten feet from the color wheel. From ten to forty feet there seems to be but little difference in the color perception, but for best tests fifteen to twenty feet is the most desirable position.

For private practice with the color wheel a small mirror may be placed five or six feet in front of the wheel in such position as to furnish an image of the disks to the person operating the machine. Owing to a slight loss of light by reflection the closest criticism may not be possible when working with a mirror in this way, but if a plate mirror is used the results are very good and a bevel plate mirror about 7 x 9 inches without frame, can usually be procured at small cost; this method is much more satisfactory for personal experimenting than an assistant to turn the wheel.

These disks have heretofore been used as a curious piece of philosophical apparatus rather than because they have been supposed to have any practical value in color training, but in establishing a color nomenclature based on six spectrum colors the disks at once assume a great value and are indispensable in a system of color instruction founded on the science of color and on the psychological perception of colors.

Let us suppose that the two disks shown in Fig. 7 are yellow and green, 80 parts yellow and 20 parts green; then by rotation we shall have a green yellow indicated by the symbol Y. 80, G. 20. No argument is necessary to prove that when an exact expression of color effect is required this is better than the simple statement that it is a greenish yellow.

How to Begin the Experiments.

For practice it is profitable to commence with the red and orange disks combined on the spindle, with a smaller red disk in front of them, the smallest being preferable. Begin by introducing say five per cent of orange and notice that a change from the standard red at the center is visible. Gradually increase the orange until it seems difficult to say whether the resulting color is more like red or orange, and then exchange the small red disk for an orange disk of the same size, and continue adding orange in the larger disks until the difference cannot be detected between the small disk and the larger combined disks.

The standards may be combined in pairs, as has been indicated with the red and orange, to produce all the intermediate hues throughout the spectrum, but it must be remembered that these combinations are to be made by joining in pairs, colors adjacent in the spectrum, red and orange, orange and yellow, yellow and green, green and blue, blue and violet. We then shall have representations of all the spectrum colors, but there are still the colors between violet and red, known in nature and art as purples, which must be produced by uniting the red and violet disks, thus completing a circuit of colors containing all the pure colors in nature.

In nature all colors are modified by light and shade, strong light producing tints and shadows more or less deep forming shades.

These effects are imitated on the color wheel by the use of a white disk combined with a disk of a standard color for tints and a black disk for shades, and can be tested in the same order as indicated for the hues, by combining each standard disk with a white or a black disk in varying proportions. It will be noticed early in disk experiments that a very small amount of white produces a decided effect in the tone of a color while a comparatively large amount of black is necessary to produce a marked change. As this is exactly the reverse of the effects of white and black pigments it is always a subject of remark. In pigments these effects are imitated by the mixture of white with a color to produce tints, and black for shades, or more generally instead of black some dark natural pigment approaching the hue of the color, may be preferred because a black pigment will too often impart an unexpected and undesirable hue to the color. As for example, in making shades of red some natural brown pigment is better than black, and so various dark browns and grays are used for different colors. Even with the disks it is impossible to imitate purest tints of all the standard colors, because in some of the colors, as peculiarly in red and blue, the rotation of the white disk seems to develop a slightly violet gray, for which effect there has as yet been no scientific explanation. This gray dulls the purity of the tint as compared with that which is found in the color under a bright illumination, but on the whole both tints and shades as well as the hues can be better illustrated with the disks than in any other way, and in addition, the advantage is secured of being able to measure and record the tone by the graduated disk in the same way as the hues are measured and recorded. A further advantage is secured in the use of disks in color instruction because with pigments, the only other method by which colors can be combined, much time must be lost not only in the mixing and applying of the colors but in the delay necessary to allow them to dry before the true results can be seen.

Fig. 8.

The shades of yellow as shown on the wheel will not be generally accepted without criticism, but careful comparison with yellow paper in shadow will prove the substantial truth of the disk results. This experiment may be tried as follows: Join two cards with a hinge of paper or cloth to form a folding screen like the covers of a book as in Fig. 8. On the surface A, paste a piece of standard yellow paper and on B, a piece of yellow shade No. 1. Hold these two surfaces toward the class in such a position that the strong light will fall on B, which is the yellow shade, and thus bring the face A, which is a standard yellow, in a position to be shaded from the light. By varying the angle of the covers with each other and turning them as a whole from side to side, a position will be secured in which the two faces will seem so nearly alike as to convince the class that this color which they may have thought to be green, is not green, but a color peculiar to itself, a shade of yellow; because the darker paper when in full light appears substantially the same as the standard yellow in the shade or shadow.

In our experiments thus far with the wheel we have combined the standards in pairs to produce the colors of the spectrum between the standards, which for convenience may be called intermediate spectrum hues, and also have combined a white disk with each of the standards to produce tints of the standards and a black disk to make shades.

By combining a white disk with an orange and a yellow disk, for example, forming a trio of disks, a variety of tints of orange yellow and yellow orange may be made. Also by the use of the black disk instead of the white a series of shades of the intermediate hues may be produced, and thus a great variety of tints and shades of many spectrum colors shown.

Now if the white and black disks are combined with each other the result will be a shade of white, i.e., a white in shadow, which is an absolutely neutral gray. As the experiments progress it will be seen that this neutral gray is a very important feature in the study of color, and therefore it may be well at this point to make sure that the disk combinations give the true gray of a white in shadow by a test similar to the one used for the shade of yellow, thus disarming criticism. Such a test may conveniently be made by covering the reverse sides of the folding covers with white on one cover and "neutral gray paper No. 1" on the other. As the neutral gray papers are made in imitation of combinations of black and white disks this experiment is as convincing as the one regarding the yellow shade. This is but one of many examples of the value of disk combinations in the classification and analysis of colors.

In an elaborate chart of colors highly recommended for primary color instruction a dozen years ago no correct understanding of the classification of colors is shown, the tints and shades being indicated by a very decided change of hue rather than a consistent modification of tone. For example, in the red scale the standard or normal red is vermilion, i.e., an orange red; shade No. 1 is simply a red less orange in hue than the standard, and shade No. 2 a shade of the standard red advocated in this system; while tint No. 1 is a broken yellow orange and tint No. 2 is much more yellow and more broken than No. 1.

Similar inconsistencies occur in all the other scales, showing that the author had no correct knowledge of the analysis of colors, and yet this was the best and practically the only aid offered for instruction in color at that time.

Neither were there any true standards for neutral grays and the term "neutral" was used in such an indefinite way as to rob it of all actual value, until by the aid of disk combinations it came to be confined to white in shadow as closely imitated by the combinations of white and black disks.

Fig. 9.

With colored papers made in imitation of the six standards and two tints and two shades of each, six scales of colors may be produced by arranging the five different tones of each color in a row, as in Fig. 9, which represents the orange scale with tints at the left and shades at the right. If, in addition to these six scales, we have two scales between each two of the standards, we may have between the orange scale and the yellow scale a yellow orange scale and an orange yellow scale, and if we thus introduce the intermediate scales between each of the other two standards, and include the red violet and violet red, we shall have eighteen scales of five tones each.

The eighteen scales as above named may be arranged as shown in Fig. 10 to form a chart of pure spectrum scales which is very valuable for study and comparison and especially so in the study of the theory of harmonies. All these tones are called pure tones and this chart is therefore called a chart of Pure Spectrum Scales.

The idea that soft, dull, broken colors produce best harmonies when used in combination may or may not be a universally accepted truth, but there is a general belief that it is much easier to make acceptable combinations with broken colors than with pure spectrum colors and their tints and shades, and therefore the temptation has been strong to select a general assortment of colors which easily harmonize because of the pleasing effect, instead of having regard solely to the educational value of colors.

Truth in education requires that when colors are classified as spectrum colors they shall all be the nearest approach possible to the true spectrum colors, and in the spectrum there are no broken or impure colors. Therefore, whenever the spectrum is set up as nature's standard or chart of colors and an imitation is made in pigments or papers, great care should be used to secure the most accurate imitation possible, but in the past this has not been the case, because of the prevailing idea that the colors must all be possible combinations of three primaries, and hence the orange, green and violet have often been very broken colors. While pure colors and their tints and shades may be advantageously combined with various tones of broken colors in one composition for artistic effect, they should be definitely divided when classified for educational purposes, and their differences clearly explained to students.

In a scale of tones in any color the several papers will harmonize more easily if the tints and shades are not too far removed from the standard, but it is thought by many good judges that the educational advantage in learning to see the relationship of color in the more extreme tones is of greater importance in the elementary grades than the facility for making most pleasing combinations. Consequently in the Bradley colored papers the tints are very light and the shades quite dark.

If, instead of adding either a white disk or a black disk to a spectrum color, by which we make pure tints and shades, we add both white and black, a line of gray colors or so-called broken colors is formed. This is most beautifully shown with the disks, and in this way a line of true broken colors is secured, because in each case a true neutral gray has been added to the color, which cannot be insured in the mixture of gray pigments. As an example, this may be shown with the three smaller sizes of the orange disks. With the medium size of these three make the combination Orange, 35; White, 10; Black 55. With the larger size disks make the proportions Orange, 16; White, 5; Black, 79, and with the smallest size Orange, 43; White, 33; Black, 24. Place these three sets of disks on the spindle at one time and you have the three tones of a broken orange scale.

With similar combinations applied to the six standards and one intermediate hue between each two, there will be material for a chart of Broken Spectrum Scales, as shown in Fig. 11, including twelve scales of three tones each. These are the most beautiful colors in art or nature when combined harmoniously. Because of the loss of color in broken colors it is not advisable to attempt so many different hues or so many tones of each hue as in pure colors, for slight differences in either hues or tones are not as readily perceived.

In these two charts of color scales two distinct classes of colors are represented, namely, pure colors and broken colors. The pure colors consist of the purest possible pigmentary imitations of spectrum colors, with their tints and shades, and the broken colors are these pure colors dulled by the admixture of neutral grays in various tones. This distinction is readily recognized under proper training, so that if a broken color is introduced into a combination of colors from a pure scale it will be readily detected, which always occurs when the attempt is made to produce a series of spectrum scales by the combination of the three primary colors red, yellow and blue. By this method, if logically carried out, the orange, green and violet are dark broken colors, and hence to a less extent the intermediate colors also, because each of these is a mixture of a pure color with a broken color. The usual result, however, is that the orange made from the red and yellow seem so out of place in the warm end of the spectrum that it is modified and made much nearer the pure color, usually, however, too yellow, while the greens and violets, which are deep and rich broken colors, may seem more harmonious, but are so dark as to be out of place among spectrum colors.

Fig. 10. Fig. 11.

If light broken colors are properly combined a beautiful imitation rainbow is produced, which is more harmonious than the spectrum made from full colors. A series of such colors combined in spectrum order produce a more pleasing effect when separated by a small space of white, black, gray, silver or gold. The reason for this may be found in the discussion of simultaneous contrasts.

In nature nearly all colors are broken. First, there is always more or less vapor together with other impurities in the air, so that even in a clear day objects a few hundred feet from us are seen through a gray veil, as it were, and in a misty or hazy day this is very evident. In the case of somewhat distant foliage the general color effect is produced by the light reflected from the aggregation of leaves, some of which may be in bright sunlight and others in shadow, with a mixture of brown twigs. All these tints and shades of green and brown are mingled in one general effect in the eye. Also, owing to the rounded forms and irregular illumination of objects, we see very little full or local color in nature.

Therefore the study of broken colors becomes the most fascinating branch of this whole subject, and it also has an added interest because nearly all the colors found in tapestries, hangings, carpets, ladies' dress goods, etc., come under this head. In fact it would be hazardous for an artisan or an artist to use any full spectrum color in his work, except in threads, lines or dots. A considerable quantity of pure standard green, for instance, would mar the effect of any landscape.

It is a very interesting diversion to analyze samples of the dress goods sold each season under the most wonderful names. For example:—

"Ecru," a color sold a few seasons ago, is a broken orange yellow with a nomenclature O. 12, Y. 15, W. 17, N. 56, while this year "Leghorn" and "Furet" are two of the "new" colors, the former having a nomenclature of O. 16, Y. 54, W. 19, N. 11, and the latter O. 18, Y. 18, W. 8, N. 56, all of which are very beautiful broken orange yellows.

"Ashes of Roses" of past years is a broken violet red which can be analyzed as follows: R. 8-1/2, V. 2-1/4, W. 15-1/4, N. 74.

"Anemon" of this season is R. 28, V. 7, W. 5, N. 60, which is another broken violet red.

"Old Rose" is a broken red: R. 65-1/2, W. 24-1/2, N. 10.

"Empire" of past seasons is G. 18-1/2, B. 11, W. 16-1/2, N. 54, while "Neptune" of this season is G. 13-1/2, B. 2-1/2, W. 11, N. 73, both being broken blue greens.

"Topia," a beautiful brown, is O. 10, N. 90, a pure shade of orange, while "Bolide" is a lighter yellow orange with a nomenclature of O. 18-1/2, Y. 2-1/2, W. 1-1/2, N. 77-1/2.

We might analyze "Elephant's Breath," "Baby Blue," "Nile Green," "Crushed Strawberry" and others common in the market, but while the names will no doubt occur each season the colors will change with the fickle demands of the goddess of fashion and the interests of the manufacturers and dealers. In writing any color nomenclature the letters should be used in the following order: R.-O.-Y.-G.-B.-V.-W.-N., thus always listing the standard colors before the white or black. For example, never place Y. before O. or R., and never use N. before W. If this order is strictly adhered to the habit is soon acquired and a valuable point gained.

It has been shown that combined white and black disks form neutral gray, which is a white in shadow or under a low degree of illumination. If to such a gray a very small amount of color is added, as orange for example, by the introduction of an orange disk, this neutral gray becomes an orange gray, but unless the amount is considerable it can not be detected as an orange, but the gray may be termed a warm gray, denoting that it is affected by some one of the colors near the red end of the spectrum. If blue instead of orange is added to the neutral gray, a cool gray is produced. When green is added to a gray the result can not fairly be called either warm or cool, and hence we have termed it a green gray. According to this plan we have four classes of grays, Neutral, Warm, Cool and Green grays. As there may be many tones of each, and many intermediate combinations from red to green, or green to blue, the number of grays in nature is infinite, but these four classes with two tones of each in the papers form what may be called standards or stations from which to think of the grays, the same as the six standards in the spectrum constitute points from which to think of pure colors.

A careful consideration of the foregoing pages, accompanied with a color wheel or even a color top, can hardly fail to give a student who will make the experiments a clear idea of the use of the disks in the system of color education in which they form such an important feature, and therefore the old theory of three primaries, red, yellow and blue, and all that it leads to can be very intelligently considered and tested by them in the experiments which follow.

This old theory briefly restated is as follows: It is said "there are in nature three primary colors, red, yellow and blue; and by the mixture of these primary colors in pairs, orange, green and violet may be made." In fact leading educators have said that "in the solar spectrum, which is nature's chart of colors, the principal colors are red, orange, yellow, green, blue and violet; of these red, yellow and blue are primaries from which may be made the secondaries, orange, green and violet." All such statements as heretofore made in any popular treatment of the subject are understood to mean that in a pigmentary imitation of a spectrum the secondaries as enumerated may be produced by the mixtures of the primary pigments, because pigmentary mixtures are the only combinations generally recognized.

This theory has also included the statement that the primaries are complementary to the secondaries in pairs, and that the combination of the secondaries in pairs may produce a distinct class of colors called tertiaries.

It will be the aim of the following pages to demonstrate that in all this there is neither scientific or æsthetic truth nor educational value.

The Old Theories Tested by Mixture of Three Pigments.

Experiments in mixing the three pigments, red, yellow and blue, to produce the secondaries, orange, green and violet, have been very carefully made with interesting and instructive results. All such experiments are valueless unless made with one accepted set of primaries for the three combinations, because it is self-evident that if we select a vermilion red which is very decidedly an orange red, and choose for our yellow one of the orange yellows, the mixture will more nearly approach a true orange than if a standard red and standard yellow are used. Also in making a violet, if we mix a carmine, which is a violet red, with a decidedly violet blue, of which there are many, the result will be a better violet than the combination of the standard red and blue. So also in the mixing of blue and yellow to make green, a greenish yellow and a greenish blue will necessarily produce better results than the standards. Therefore, to test the matter fairly, the same pigments which are used to coat the standard red, yellow and blue papers have been combined so as to produce the best possible orange, green and violet, and these results when analyzed on the color wheel are as follows:—

The orange made by mixing standard red and yellow pigments in the best proportions is equal to O. 46, W. 2, N. 52. The violet is equal to V. 20, W. 1, N. 79, and the nearest approach to a standard green is shown by disk analysis to be G. 37, W. 7, N. 56, which is better than the violet and nearly as good as the orange.

These experiments show that heretofore when a line of standards of six colors has been prepared from three primaries, red, yellow and blue, even though the purest possible colors may have been selected for the primaries, the secondaries have not been in the same class of colors, and that all of them are very dark broken colors. Therefore, in using educational colored papers based on such a scheme, the pupil has received no correct impressions of the relative values of the several colors involved in pure spectrum scales, but has been shown at the outset a mixture of pure and broken colors as standards.

This is not a matter of opinion regarding best harmonies, because it is easy to demonstrate that less skill is required to combine broken colors harmoniously than pure colors, but it is a choice between truth and error in the early education of color perception.

Old Theories Tested by the Color Wheel or Color Top.

While it may be impossible for the reader to secure pigments exactly like the standards, red, yellow and blue, used in the above experiments, and therefore the statement here made can not be accurately verified, any one having a color wheel or even a color top may test the same combinations by use of disks. If it is true, as claimed, that a good standard orange can be made by mixing red and yellow, then it should follow that when a red and yellow disk are combined and a smaller orange disk placed in front of them, that it ought to be possible to so adjust the proportion of red to yellow that by rotation the outer ring of color will match the central orange disk.

A trial of this experiment will show that while the color resulting from the best possible combination of red and yellow is a kind of orange, it is not even an approximation to the standard orange, but is a shade of orange which may be matched by combining the smaller orange disk with a black disk in the proportion of O. 45, N. 55, the larger disks being R. 89, Y. 11.

In combining red and blue disks to make a violet the result is more satisfactory, while if we attempt to produce a green by combining the yellow and blue disks the result will be surprising, but probably not convincing, because the statement that yellow and blue make green has been so persistently reiterated as a fundamental axiom that people who have given the subject but little attention will feel that to doubt it is rank heresy. In a text book treating of color is found the following passage: "Green substances reflect the green, i.e., the blue and yellow rays of the sunlight and absorb all the others." It is a fact, however, that in the mixture of blue and yellow light there is little or no trace of green, as a single experiment with a color top or color wheel will readily demonstrate.

In response to this convincing experiment a colorist of the "old school," (and there are few others) will doubtless say, "Such an assertion seems to be true when applied to these rotating disks, but we see no practical value in experiments of this kind, because in the use of color we must depend on pigmentary combinations, and in pigments yellow and blue do make green." The author of a statement of this kind is always honest in making it, and yet it is absolutely untrue, because as has already been shown, the green resulting from the mixture of yellow and blue can not be placed even approximately in the same class as the yellow and blue of which it is composed.

In accepting the disk combinations of standard pigmentary colors we are assuming a system of color investigation based on the combination of colored light rather than the mixture of pigments, and to an artist who has given the subject little thought this seems quite radical, not to say startling. But, logically, why is it not the most natural as well as the correct basis for this work?

Art in color must be based on the imitation of natural color effects. We must first learn to see color correctly and to know what we see, and after that it is a very simple matter to learn which pigments to combine for producing any desired result which is already clearly defined in the mind. In fact the best selection of pigments must often be based on their chemical and mechanical qualities as much as on their peculiar hues.

All color impressions of material substances are produced by colored light reflected from a material surface to the retina of the eye, through which by some unknown means it is conveyed to the brain. When the white sunlight falls on a material substance a portion of the rays are absorbed and others are reflected to the eye, thereby conveying impressions of color. If on a surface of yellow material we throw a strong orange light through a colored glass, some of the orange rays from the glass will mingle with the yellow rays and the two are reflected to the eye, thereby producing an orange yellow or yellow orange effect where before it was yellow. So in a summer evening landscape when there is a so-called red sunset, everything is illuminated by an orange light and each color in the landscape is affected by the orange rays which mingle with the rays of the local color and are reflected to the eyes of the observer, producing the effect of local colors mixed with orange.

In a room where the windows open on to a green lawn with many trees in close proximity to the house, nearly all the light is reflected from green surfaces, and hence is green light. In such a case a correct painting of objects in that room would have a general green effect.

The afternoon light in a room on the west side of a city street may be nearly all red light, reflected from an opposite red brick wall, and such a room would be ill-adapted to showing fine dress goods, because the hues of the more delicate colors would be entirely changed, and hence would give a false impression as to the relations of the several colors in combination as seen in white or clear daylight. If a piece of light blue silk is illuminated by sunlight passing through a bit of yellow glass, no trace of green effect will be produced, but a gray either slightly yellow or blue, according to the relative strength of the colors in the glass and the silk. This same effect would be secured if the yellow light of the setting sun illuminated the same material, but under such conditions everything else would be similarly affected so that the effect would not be so apparent.

The idea that all color is derived from the three primaries, red, yellow and blue, is so generally believed that our best writers among artists, colorists and educators have repeated it for many years. George Barnard, an English artist, in a very valuable book on water color painting, speaking of the colors of the spectrum which may be re-combined to form white light, says that if the yellow and blue rays are combined they produce green.

Chevreul also states in his invaluable book on color contrasts that yellow and blue threads woven into a texture, side by side, produce green. This statement is the more remarkable because the writer was a very careful investigator and is but another evidence of the strong hold which the Newton and Brewster theory has had on the public mind for so many years.

The story is told of an artist who wished to introduce into a composition of still life a blue vase with a bit of yellow lace thrown over a portion of it, and having been educated to believe that yellow and blue made green, gave a green effect to the portion of the vase covered by the lace. Had he known that blue and yellow light combined make gray instead of green he would have avoided the error.

The fact that gray is the product of blue and yellow light is sometimes taken advantage of in forming backgrounds in lithographic printing, in which a stippling of alternate dots of yellow and blue, very close together but not overlapping, is used to produce a beautifully transparent gray much more pleasing than any one tint of gray. This result is due to the blending of the two colors in the eye with the same effect as the colors of two rotating disks are mingled. The fact that there is a difference between the color effects produced by mixing two pigments and the mixing of the light reflected from similar colored surfaces is a very strong argument for a system of color instruction based on disk combinations, rather than on pigmentary mixtures.

In order to obtain the most truthful effects of color in nature the artist should have sufficient knowledge of the principles which govern the combination of colors by reflected light, so that his reason may aid his eyes.

A little experimenting with the rotating disks and with pigments will convince any one that the disk combinations form the only possible basis at present known for logical color instruction.

Concerning the Complementary Colors.

Having shown that the three colors, red, yellow and blue, can not be combined to make an orange, a green or a violet of a corresponding degree of purity, we will consider the other claim which is set up by the advocates of the Brewster theory, namely, that the secondaries are complementary to the primaries in pairs, the green to the red, the violet to the yellow and the orange to the blue.

As all color is contained in white light, if we take from white light any given color, the color remaining is the complementary. If a small disk of standard red paper is placed on a white wall and the eyes fixed intently on it for a few seconds, and then the eyes slightly moved back and forth, a ring of a bluish green tint will be seen surrounding the red paper, or if the eyes are fixed intently on the disk for a short time and the paper suddenly removed, a disk of the same blue green tint will be seen in place of the red disk. This is called the accidental color and is supposed to be identical with the complementary color, although the image is too faint to give any very exact effect, but it is sufficient to furnish a clue to the complementary, and we may infer that a color between green and blue is that which is required.

Now if we can determine in what proportions red, blue and green must be united to produce white light we may solve the problem. This is not possible in the use of any pigmentary colors, because of the impurity of all pigments as compared with spectrum colors. Although the mixture of colored light reflected from the disks, which are made of pigmentary colors, gives much purer color than the actual mechanical mixture of the two pigments, still, because it is a reflection of pigmentary colors, it is far lower in tone than the corresponding mixture of spectrum colors. Therefore it can not be a pure white, but may be white in shade or a neutral gray, which, as already shown, can be produced by the combination of a white and a black disk.

Fig. 12. Fig. 13.

Therefore if red, blue and green disks of medium size are joined on the wheel and in front of them small white and black disks are combined, we have a means for solving this problem. If these various disks can be so adjusted that when rotated the effect of the three colored disks is a neutral gray, (or white under a low degree of illumination) exactly matching a gray that may be obtained by adjusting the small black and white disks, then one step in the solution is taken, as shown in Fig. 12.

With such an arrangement a very close match is produced, when the combined disks show the proportions to be R. 41-1/2, B. 22-1/2, G. 36 for the larger disks, and for the small disks W. 15, and N. 85. Now if blue and green are combined in the same proportions, as indicated above and in quantities sufficient when added together to fill the entire circle of 100 parts, blue will contain 38.3 parts and green, 61.7 parts, as shown in Fig. 13, and the disks when rotated will give the color which is the complementary of red: namely, a blue green.

In the same way the complementary of each of the other standard colors, and in fact of any color, may be obtained.

The complementary of orange is another color between the green and blue, but more largely blue. The complementary of green is a violet red, and of violet a color between yellow and green, while yellow and blue are very nearly complementary to each other.

These figures furnish the results in a very well-lighted room, with a perfectly white interior. It is a well-established fact that this experiment is somewhat affected by the degrees of illumination, and also that colored light from the walls and ceiling of a room must of necessity have its effect, but all these matters are so insignificant as to be of no material consequence in the æsthetic study of the subject, and they can be very nearly eliminated when necessary by a careful selection of conditions. Whenever accurate experiments in pigmentary color comparisons are to be made, either by the use of rotating disks or otherwise, it is desirable to have a very well-lighted room, with a northern exposure and to select a morning or noonday light from a slightly overcast sky. These conditions obviate the unpleasant effect of direct sunlight in the room and also the very slightly blue effect of the clear sky. These precautions are unnecessary in experiments relating to the ordinary æsthetic consideration of color combinations, but even in such work it is important to exclude all light reflected from neighboring trees or colored buildings. Also the interior of the room should be as free from color as possible; a clean white surface is especially desirable.

A Chart of Complementary Colors, shown in Fig. 14, has been found very valuable in fixing in the minds of teacher and pupils the complementaries of the six standards. In this chart, which is about eighteen inches in diameter, the circles at the ends of the six diameters are colored papers selected from the Bradley coated papers, as approximating the true complementaries. In the majority of cases they are not far from correct, but are least satisfactory in the blue and yellow. Theoretically the complementary of the ideal standard blue is a slightly orange yellow, and of the standard yellow a slightly violet blue. But there is as yet no blue pigment in the market suitable for commercial use which is free from a slightly violet effect. Therefore the standard blue paper is practically as good a complementary for the standard yellow as the violet blue paper. But notwithstanding these slight imperfections which are at present unavoidable, the chart is a valuable aid in fixing in the mind the positions of the complementary pairs in the spectrum circuit.

Fig. 14.

Each of the foregoing experiments furnishes an interesting class exercise, and may be very closely repeated by the pupils with their tops. Also the computation of the proportion of green and blue when raised to the full circle may form a practical problem in proportion for pupils of the higher grades. Taken together, these experiments prove that the complementaries of the old primaries are not found in the secondaries.

The last claim of the Brewster theory is that the secondaries by combination form three lines of colors peculiar to themselves, called citrines, russets and olives. It is asserted that the mixture of orange and green makes citrine; orange and violet russet; green and violet olive. Although these names may be very convenient terms to express three general classes of colors, they must of necessity be too general and indefinite to be of value for accurate expression of color effects, and are in fact so vague that hardly two persons can be found in a large company who will agree as to the best expression of either of them. The following are formulas for a number of colors in each class, as made from analyses of colors coming under these names. It is an interesting exercise to produce some of these colors by means of the rotating color disks and test the opinions of the different members of a company as to which best represents to each one of them a tertiary color, as citrine, for example. For this purpose three different formulas may be shown at the same time, with three sizes of disks.

Citrines.

O. 7. Y .13. W. 3½. N. 76½.

Y. 15. W. 4. N. 81.

Y. 13. W. 5. N. 76. G. 6.

O. 6. Y. 20. W. 4. N. 70.

O. 3. Y. 6. W. 8. N. 83.

Russets.

R. 37. O. 8. W. 8. N. 47

R. 79. W. 10½. N. 10½.

R. 33. O. 20. W. 6. N. 41.

R. 36. O. 4. W. 9. N. 51.

R. 47. O. 7. W. 8. N. 38.

Olives.

G. 19. B. 11½. W. 10½. N. 59.

G. 13. B. 6. W. 12. N. 69.

G. 14. B. 12. W. 8. N. 66.

D. 10½. B. 15. W. 8. N. 66½.

G. 12½. B. 5½. W. 4. N. 78.

The term citrine theoretically covers all possible combinations of orange and green, but as generally understood those colors which are so near the orange or the green as to very decidedly approach either the one or the other are not included, and, as shown in the above analyses, a citrine is a very broken color ranging from an orange yellow through yellow to a green yellow.

Although the russets would theoretically range from violet to orange, yet the general conception of russet will hardly accept a violet red, but will cover only the red and orange reds as above indicated, while olives are confined to blue greens and green blues.

These tests are based on combinations of the Bradley standard orange, green and violet pigments, and therefore are far stronger in color than those colors usually termed citrine, russet and olive, made by mixing the pigmentary secondaries. For example, if a yellow and blue pigment are mixed to form a green, and red and yellow pigments to make an orange, and then this green and orange are mixed to produce a citrine, the result will be very much darker and more broken than the mixture of the purer orange and green colors used as standards.

Restricted to these limits these names may become very useful terms for general color expressions, as covering three different classes of broken colors. If any one believes that these color formulas do not correctly represent the three classes of colors indicated, a series of experiments with even the small color top will prove very convincing.

When the subject of standards as a means for identifying colors is mentioned artists frequently express the feeling that the names of pigments are good enough for them, such as Ultramarine Blue, Prussian Blue, Vermilions, the Siennas, Indian Red, etc. The following are the analyses of several samples of Vermilion, Burnt Sienna, Raw Sienna, and Indian Red of the best tube oil colors in the market:—

Vermilion.

R. 80. O. 14. W. 6.

R. 87. O. 8. W. 5.

R. 50. O. 24. W. 26.

Burnt Sienna.

R. 1¼. O. 6. W. 3. N. 89½.

R. 22½. O. 11½. W. 2. N. 64.

R. 25. O. 12½. W. 5½. N. 57.

Raw Sienna.

O. 18½. Y. 6½. N. 75.

O. 17. Y. 14. W. 1. N. 68.

R. 8½. Y. 3½. W. 2. N. 86.

Indian Red.

R. 11½. O. 7. W. 4. N. 77½.

R. 13½. O. 13½. W. 2½. N. 70½.

A careful examination of these formulas and a reproduction and comparison of the colors on the color top will convince any one that in no case does the commercial name determine the color with a degree of accuracy sufficient for any valuable nomenclature.

Classification of Harmonies.

The theory of the harmonies of colors is a subject which awaits very careful investigation and a general discussion by artists and expert colorists. Such investigations must include many experiments based on common standards and uniform methods of measurements and records.

Harmonies naturally seem to fall into a few general classes which are convenient for comparison and discussion as well as for elementary instruction, but no one person can set himself or herself up to decide which are the best harmonies. The practices and recommendations of noted artists who have appeared to be gifted with intuitive perceptions regarding color combinations have frequently included those for which there seemed to be no recognized authority, and yet their beauty could not be questioned. As the rules of grammar are but the correlation of the practices of the best scholars, so the rules governing color combinations must be the summary of the practices and recommendations of the best artists, if they are to be generally accepted as final, and hence we must patiently await the growth of similarly established laws by the comparison of the opinions of critics of acknowledged ability in various departments of the world of art. This has not been possible in the past and can never occur until there is a language of color through which color facts can be somewhat accurately expressed in verbal and written language, and this language cannot exist until there is an accepted alphabet of color on which it can be based. This alphabet is now in part furnished by the spectrum standards and completed by the pigmentary standards and the rotating disks made like them. Together they form the basis for a nomenclature by the use of which the questions involved in harmonies can be discussed and the results expressed in written language.

In the investigation of any subject with a view to elementary instruction, classification is an important factor, but one which heretofore has been almost ignored as regards color education. Consequently at present the more definite division of harmonies into classes is very much a matter of personal opinion, but Mr. Henry T. Bailey, State Supervisor of Drawing in Massachusetts, has suggested a very useful classification in which he arranges all harmonies under these five heads: Contrasted, Dominant, Complementary, Analogous and Perfected.

Contrasted.—The contrasted harmonies are those in which color is contrasted with non-color, or more accurately in which an active color, that is a tone from the spectrum circuit, is contrasted with a passive color, white, black, gray or silver and gold; for example, a blue green tint with white, or green blue with warm gray No. 1.

Dominant.—By dominant harmonies we mean those in which are combined different tones from one color scale. For example, red tint No. 1, and red shade No. 1, or a green blue tint, green blue, and a green blue shade. A dominant harmony composed of grays, or white, gray and black, is sometimes called a neutral harmony.

Complementary.—This term refers to those harmonies in which are combined opposite or complementary colors in the spectrum circuit. The best of them show not only opposition in color but also opposition in tone. Thus, tints of one color with shades of its complementary produce a more pleasing effect than do complementaries of equal value. The best complementary harmonies contain one or more passive colors.

Analogous.—This name is applied to those harmonies in which are combined tones from scales of neighboring colors in the spectrum circuit.

For example, in a composition of colors from that part of the spectrum containing yellow, green yellow and yellow green the following simple combination may be made: Yellow tint No. 1, green yellow and yellow green shade No. 2.

Perfected.—By perfected harmonies we mean those in which the general effect of one analogous harmony is complementary to that of another.

The above classification of harmonies is very valuable for fixing in the mind the various effects of color combinations, and yet they may seem to somewhat merge into each other in their application, until the underlying principles which govern them are understood. It is unwise to suppose that because the above classification of harmonies is based on the science of color we can infer that it furnishes definite rules for producing best effects.

The Work of Chevreul Reviewed.

The good or bad effect of two or more colors in combination in decorative designs or fine art depends very largely upon phenomena which are elaborately explained in a book entitled "The principles of Harmony and Contrasts of Colours" by M. Chevreul. [A] The first edition of this book was prepared in 1835 and published in 1838. The author had at that time been employed for a number of years as superintendent of the manufactory of Gobelin Tapestries in Paris under the control of the French government.

In this book are described in detail the results of a great number of experiments which were instigated by complaints regarding certain colors produced in the dyeing department of the manufactory, and which afford the most elaborate exposition of the subject ever published.

One of the first things which led Chevreul to make his investigation was the complaint that certain black yarns used as shades in blue draperies were not a full black but more or less gray.

The author says in his preface, "The work I now publish is the result of my researches on Simultaneous Contrasts of Colours; researches which have been greatly extended since the lectures I gave on this subject at the institute on the 7th April, 1828. In reflecting on the relations these facts have together, in seeking the principle of which they are the consequence, I have been led to the discovery of the one which I have named the Law of Simultaneous Contrast of Colours."

The closing sentence of the preface to the first edition and dated 1835 is as follows:—

"I beg the reader never to forget when it is asserted of the phenomena of simultaneous contrast, that one colour placed beside another receives such a modification from it, that this manner of speaking does not mean that two colours, or rather the two material objects that present them to us, have a mutual action, either physical or chemical; it is really only applied to the modification that takes place before us when we perceive the simultaneous impression of these two colours."

It was not till three years later that a publisher could be found for this book, which is still a standard.

The English translation comprises over five hundred closely printed pages with many engraved and colored plates, and yet, it has been of comparatively little value in popular instruction because of the lack of a generally accepted color nomenclature or list of well defined color terms, by which the readers might have understood and repeated for themselves the experiments described.

Unfortunately Chevreul was fully impressed with the Newton-Brewster idea of three primaries, red, yellow and blue, and therefore some of his deductions from his experiments seem to have been more or less influenced by the attempt to make them harmonize with this theory, and yet the subject which he has treated so exhaustively and intelligently is one of the most important in the æsthetic study and use of colors. In all expressions of colors in combination with each other, whether in nature, fine arts or the decorative and industrial arts, every color is affected by its surrounding colors, a fact which is exhaustively treated in this book.

While with our present knowledge of the subject it does not seem that the material use of color can be reduced to an exact science, this should not prevent us from accepting all the natural and scientific aids which have been or may be discovered toward this desirable result. Because of this lack of scientific knowledge in Chevreul's time much of the worth of his experiments is lost to us, yet there is very much of value in his work, suggesting as it does experiments which may be tried with present standards and modern methods.

If the use of Maxwell disks had been known to Chevreul his deductions from his experiments would have been quite different in their details. For example, in accepting the proposition that there are three primaries, red, yellow and blue, which may be combined in pairs to make the secondaries, orange, green and violet, he states that owing to the impurities of the pigments the secondaries are not as pure as the primaries. Consequently he believes that this may account for many of the shortcomings which he was too observing to overlook; but notwithstanding such an error in theory this wonderful investigator made many practical experiments and established very valuable facts regarding color contrasts.

The term Simultaneous Contrast seems rather restricted for a title covering such a range of effects, and the author subdivides the subject into simultaneous contrasts, successive contrasts and mixed contrasts, which he defines as follows:—

Simultaneous Contrast.

"In the Simultaneous Contrast of Colors is included all the phenomena of modification which differently colored objects appear to undergo in their physical composition and in the height of tone of their respective colors, when seen simultaneously."

Successive Contrast.

"The Successive Contrast of Colors includes all the phenomena which are observed when the eyes, having looked at one or more colored objects for a certain length of time, perceive, upon turning them away, images of these objects having a color complementary to that which belongs to each of them."

Mixed Contrast.

"The distinction of Simultaneous and Successive Contrast renders it easy to comprehend a phenomenon which we may call the mixed contrast; because it results from the fact that the eye, having seen for a time a certain color, acquires an aptitude to see for another period the complementary of that color, and also a new color, presented to it by an exterior object; the sensation then perceived is that which results from this new color and the complementary of the first." These last two effects may be shown very clearly in simple experiments.

There are various phenomena which may be classed as successive contrasts sometimes called "after images." The phenomena which Chevreul groups under the term "Simultaneous Contrast of Colors" belong to a class of physio-psychological effects termed after images, and more definitely to the subdivision called negative images. The positive after images are not important in the consideration of the theories of color and therefore are not described here. The specific effect most directly involved in the subject of harmonies may be observed if the eyes are fixed upon a small disk of red paper on a white wall for a few seconds and then the paper is suddenly removed, as there will appear on the wall in place of the full red disk a faint tint of a blue green. This is called an after image, and is nearly or exactly a tint of the color complementary to red.

For making this experiment mount a circle of red paper, say three inches in diameter on a square white card, four or five inches across, and grasping the card by one corner hold it in front of a white wall or a sheet of white paper pinned on any support. Tell the observer to look intently at the red disk for a half minute, and then without giving any notice suddenly remove it and ask what color is seen in place of it. At the first trial the result may not be entirely successful, because the eyes of the observer may naturally follow the red spot when it is removed instead of remaining fixed in the original position, but a second trial will bring the expected result. To illustrate mixed contrast, fasten on the wall a piece of red tint No. 2 paper four or five inches square. This may be very conveniently done by using a bit of beeswax on each corner of the paper, which will not soil the wall. Then having the three-inch circle of standard red paper mounted on a white card somewhat larger than five inches square hold the card in front of the red tint on the wall and repeat the experiment as before. The effect now should be a three-inch disk of very light gray in the center of the pink square, which is a "mixed contrast" according to Chevreul. The reason is simple. The after-image or successive contrast of light blue-green is projected on the red tint and being complementary the resulting effect is a gray. If the red tint could be exactly graded to the complementary effect in the eye the resulting gray circle would be a true neutral gray. Another illustration of the same physical effect by which the complementary is induced may be shown by substituting for the tint of red a light tint of the blue-green paper retaining the full red disk as before. The same blue-green after image is now projected on to the light blue-green paper and hence a circle of more intense blue-green is produced. Thus it is seen that Chevreul's successive and mixed contrasts are both due to the same physiological effect, the only difference being in the ground on to which the after image is projected.

It probably is unnecessary to state that these experiments may be made with any color and its complementary and that red and blue-green are used here merely as an example.

Another phase of the same physical effect is seen under other conditions which may at first seem to be quite different from those described, but which on examination appear somewhat similar.

It is a well established fact that when two surfaces approximating each other in color, as red and orange for example, are placed side by side, both are rendered less brilliant, an effect which might be reasonably expected because in order to see both the eye is naturally directed first to one and then to the other, and in each case the after image induced is a green-blue or blue-green, which being approximately complementary to both, dulls both. Or in other words, it is as though one examines for a long time a line of goods of similar colors so that the eye becomes fatigued and the color dulled. It is said that a good salesman of colored materials will endeavor to occasionally attract a customer's attention for a few moments to some other colors approximating a complementary, so that when the attention is again directed to the goods under consideration the full effect of the color may be secured.

If it is true that the phenomenon of the after image is the cause of the peculiar effects expressed by the terms simultaneous, successive and mixed contrasts, and that by these effects all harmonies in color are governed, it is certainly profitable to understand them while using color material with the children, for their good as well as our own pleasure.

Contrasted Harmony.

Returning to our classification of harmonies, already stated, we find the first to be Contrasted Harmony, which covers those combinations in which a positive color, as a spectrum color for example, is combined with white, black or gray, leaving out for the present silver and gold, which may be confusing, and can at best be used only as outlines.

The simplest combinations of colors are found in this class, all of which are not equally harmonious, and some may not perhaps be entitled to be classed as harmonies, although not positively inharmonious. In this class, as in all others, there is involved contrast of tone and contrast of color, which may best be considered in several divisions.

Color with White.

According to the results of Chevreul's elaborate experiments the effect of a combination of an active color with white is to render the color more brilliant and to give to the white the effect of the complementary of the active color. He admits that the modification of white is very indefinite, but claims that, knowing what to expect, a complementary effect may be seen which otherwise would not be noticed. There is also a contrast of tone which in all cases tends to strengthen a color when used with white.

Black with White.

White and black are both intensified by combination with each other, and this is the type of "contrast of tone." Contrast of tone is very clearly shown when two or more grays of different tones are placed contiguous to each other. This experiment is easily tried by mounting side by side several strips of gray papers of different tones. If more than two are used they should be arranged in order from lightest to darkest. In this case each band will appear to be graded in tone from one edge to the other, each being lighter at the edge next to the darker paper.

This effect is plainly shown on the color wheel by producing several rings of grays with white and black disks of several sizes graduated from light at the center to darker at the circumference.

Color with Black.

In consequence of this law of contrast of tone the contrast of black with active colors generally tends to intensify the black and lower the tone of the color, i.e., to weaken it as though white or light gray was mixed with it, but this effect is modified by contrast of color. Contrast of color is perceptible in black when combined with color simply because the black is not perfectly black but a very dark gray, and hence there is the same complementary effect which shows in white and the lighter grays, but in a smaller degree. This effect is most clearly seen when the color used in combination is blue or blue-green, which induces in the black, yellow or red complementaries and gives the black a "rusty" appearance.

On the other hand, for example, red with black adds the complementary green-blue to the black, which improves it. The orange and yellow have a similar effect by their blue complementaries to relieve the black from any rusty appearance and a green yellow induces a violet effect in the black.

Colors with Gray.

When a color is contrasted with white the light from the pure white surface is so intense as to very largely obscure the complementary effect on the white, while on the other hand the feeble light from the black is not favorable for the exhibition of a complementary. So it might naturally be inferred that some tone between the white and black would be much more favorable than either for the observance of this effect, which is proved by experiment to be the case. This fact is illustrated in the familiar experiment of placing a white tissue paper over black letters on a colored ground, by which the black is practically rendered a neutral gray and the color a light broken color, and in appearance the gray letters receive a color complementary to the color of the page on which they are printed. Each color has its own tone of gray most susceptible to this complementary effect. The truth of this proposition can be perfectly shown on the color wheel by forming with three different sizes of disks a gray ring on a colored surface. For example, select small disks of orange and white of equal size, then a black and a white disk of the second size and an orange and a white disk of the third size. First place the large orange and white disks on the spindle, then join the two medium-sized white and black disks and put them in front of it, and lastly add the small orange and white disks. By rotation the result is the required neutral gray ring on a light orange surface. By the joining of the white disk with each of the orange disks the orange surface may be changed to a variety of tints for trial with the different grays which may be made from the black and white disks, so that the best tones of both orange and gray may be secured. When the best proportions are obtained the effect will be surprising, because when such disks are properly adjusted the complementary effect is so strong in the gray that it appears as a very definite color, a broken green-blue. It is said that the tone of gray should have the same relation to the tone of the color that its complementary would have in order to get best results.