Few of the books written for artists give an accurate view of the nature of color, and they perpetuate many color myths. These include the notions that there are only three “primary” colors; that these are the yellow, red, and blue of the school paint-box; that cobalt blue is the hue closest to being a “true” blue, and that there is also a “true” red and a “true” yellow; that these primaries are the only colors from which “all” other colors can be mixed; that you should mix your greens rather than using tube greens; that black should never be used (and wasn’t by the impressionists); that the earth colors make your mixes “muddy”; that the color wheel is the best way to represent the relationships between colors; etc., etc.
The Nature of Color
To begin with, the colors we see are not a property of either objects or light; they are a creation of our sensory systems. The electromagnetic spectrum in which we are bathed originates, for the most part, from the sun and it is absorbed or reflected from the objects that surround us. That spectrum covers a range that includes cosmic rays, gamma rays and x-rays at one end and infrared, radar, and radio waves at the other. These rays differ in their wavelengths, and are part of a continuum. The portion of this spectrum from about 4 to 7 Angstrom units is called the “visible spectrum.” This is transformed by our nervous systems into the sense impression we call light, and that is what we see in the rainbow and reflected from everyday objects as well. “Light,” as we understand it, is created by our sensory systems.
We build instruments to generate x-rays, and when we pass those rays through various bodies, the results can be seen on a viewing screen. They may appear green or grey to us, but that does not mean that x-rays are green or grey—it only means that the particular device we have constructed to form the images produces wavelengths of “visible light” that are interpreted by our eyes as the colors green or grey. It was the X-ray device that created those wavelengths. Infra-red (heat) rays can also be imaged by devices we have constructed, and these often produce images with hues of red, orange, yellow, green, and blue representing different levels of heat, but that does not mean that infra-red rays are colored. It only means that the devices we have constructed to form the images produce wavelengths that are interpreted by our eyes as colors. Similarly, the wavelengths of the “visible spectrum” that strike our eyes are not colored—instead, our eyes and the associated nervous system are, themselves, wonderful “devices” that create the sensations that we designate as “colors.”
We really don’t know whether other people have the same sensations that we do, but because our nervous systems are so similar, we assume that their sensations are much the same. At any rate, we are all trained by the cultures we live in to give the same names to our sensations that the people surrounding us do when they are responding to the same things. Consequently, English speakers agree that ripe tomatoes are “red,” grass is “green,” and pumpkins are “orange.” This really means that, when these things are illuminated by the “white” light of the sun, the wavelengths of energy reflected from them to our eyes stimulate our nervous systems to produce the sensations that we have been trained by our culture to call by those color names.
These “psycho-biological” facts of color are often defined by scientists in terms of a hypothetical “standard observer,” which is the average response of many people with “normal” vision who have participated in tests under carefully controlled conditions in which they must discriminate between—and name—different colored-lights and painted-surfaces.
And what good is it for us to know these things? It provides a useful perspective that makes many other things easier to grasp.
Let’s look at colored lights (and I will speak casually in what follows, referring, for example, to “colored lights,” but now you know what I really mean). The so-called primary colors for light are purplish-blue (B), Emerald Green (G), and orange-red (R). By “primary colors” we mean that these lights, when mixed, can produce almost all other colors of light. “Almost all,” because no three primaries, either of light or pigment, can produce all color sensations. But it is also true that any three lights with colors that are widely spaced on the visible spectrum can serve as primaries, although not as well. Our so-called “primaries” are simply the best at it. When you shine a red light and a green light of similar brightness so that they overlap on a projection screen, the result is yellow. Blue and red lights mix to produce magenta. And blue and green mix to produce cyan (a sort of turquoise). We call the light primaries the “additive primaries” because the lights are added together, and this color mixing process is called additive color mixing. (And it explains why TV shows and films are brighter than paintings: the colors are made up of light added to light.) When all three lights are projected together, they produce white.
We can now ask what the “complementary colors” of light mixtures are. If you take blue (B) away from white (B + G + R) you are left with yellow (G + R), blue’s complement when talking about light. If you take green (G) from white, you are left with magenta (B + R), green’s complement. And if you take red (R) from white you are left with cyan (B + G), red’s complement. These complements, yellow, magenta and cyan are the “secondary colors” produced when mixing light. (They are also the “after-image” colors one sees after staring at the light primaries!)
Now let’s imagine that we are looking at a red barn in front of a green tree. The white light (B + G + R) of the sun is shining on both of them. We see a red barn because the barn’s paint absorbs the green and blue and reflects the red back to our eyes [(B + G + R) – (G + B) = R]. And we see a green tree because the foliage of the tree absorbs the red and the blue, reflecting the green back to our eyes [(B + G + R) – (R + B) = G ]. And exactly the same thing that is true of the red barn is true of the red paint on your palette, and the same thing that is true of the blue flowers on the ground is true of the blue paint on your palette. (The explanation I have given here is somewhat simplified because I have been talking as if the visible spectrum is made up only of the primaries, but in fact, a pigment that reflects primarily the wavelengths that characterize left-third of the visible spectrum does appear to be a blue. Similarly, one that reflects primarily the middle-third appears to be a yellow, and one that reflects primarily the right-third appears to be a red.)
This brings us to the notion that the school paint-box yellow, red, and blue are the only colors of pigment from which the others can be mixed. This is similar to the case of the colored lights: any three pigments that reflect colors which are widely spaced on the spectrum can be mixed to produce the other colors—but not as well as properly chosen primaries. Any useful selection must include yellow because it is the most reflective color and any reasonably bright palette must include it. Once yellow is chosen, we end up with a red and a blue as “widely spaced” colors. But as it happens, the paint-box yellow, red, and blue are not the best primaries in terms of color-mixing. In fact, if we limit ourselves to three primaries, the best choice happens to be the secondaries of additive color-mixing: magenta, yellow, and cyan. That is why they are used in the printing industry for four-color printing (the fourth “color” being black).
The process of mixing paints is called subtractive color-mixing as opposed to the additive color mixing of colored lights, because the pigments we mix together absorb (subtract) some wavelengths and reflect others, just as the barn and the tree did. (Hypothetically, a mixture of red, yellow, and blue would absorb all wavelengths of light, reflecting none: that is, it would produce black. In practice, we call the color produced by such mixtures neutral.) This makes it clear that the more colors we mix together, the less bright our result will be. A useful practical rule is to not use more than three colors in any mixture. The mixing of opaque paints is beset with complications, but with largely transparent paint films we can predict the colors of mixtures fairly well using the color rules we have all learned: blue and yellow make green, red and yellow make orange, etc. There are surprises though. For example, black and yellow make a nice greenish-yellow, the color of young growth on trees in spring; after all, that is what a neutralized yellow looks like.
As I said earlier, magenta, yellow and cyan constitute the best three-pigment mixing-palette. It allows many greens, blue-greens, and purples to be mixed that can’t be made from the paint-box red, yellow, and blue, but neither magenta nor cyan are colors commonly found in the subjects we paint—although cyan would be very useful in the tropics and magenta might be invaluable for flower paintings. General-purpose palettes usually include more widely useful reds and blues. We will look at various palettes in a moment. But first, I would like to explore the claim that the primaries cannot be mixed.
Let’s perform a thought experiment. Imagine that we have three paints with the same tinting strength (which we rarely have), and we mix one drop of yellow with one drop of blue to make two drops of green as summed up by the equation: 1Y + 1B = 2G. Now let’s mix one drop of yellow with one drop of red to get two drops of orange: 1Y + 1R = 2O. Now, let’s mix these two mixtures: 2G + 2O = (1Y + 1B) + (1Y + 1R), and we can rewrite the right half of this equation as 1Y + (1Y+ 1B + 1R). We know that yellow, red, and blue mixed together give us a dark neutral color which we will call N, so we can now rewrite the result of our mixtures as 1Y + 3N. This is simply a darkened yellow, but one we mixed from green and orange. (Remember that there is a loss of brightness in all mixtures of pigments!) While we are playing with this color algebra, let’s mix a neutral from complements: 1R + 1B = 2V and 1Y + 2V = 1Y + 1R + 1B = 3N. In other words, if our paints had the same tinting strength it would take twice as much of the secondary, violet, to completely neutralize any given amount of the primary, yellow. When you think about it for a moment, that is exactly what you would expect, but what you often hear suggests that equal amounts of yellow and violet make neutral. This isn’t true, and all of this is without regard to what might be great differences in tinting strength. (Tinting strength is extremely important when talking about mixtures of colored paints!)
What about the notion that you should always mix your greens? You have already seen that there is a loss of brightness in all mixtures, so if there is a bright tube-green of a color you like, you should use it. After all, it is easy to vary a green by adding yellows, or oranges, or reds, or blues to it, and this is generally more satisfactory than beginning with blues and yellows. You should also be aware that many paints are already mixtures of various pigments. That is true of Hooker’s Green and Sap Green for example. I am looking at a tube of Hooker’s Green that lists two pigments: brominated copper phthalocyanine and isoindolinone. These have the Pigment Index numbers PG 36, and PY 110, respectively: a green pigment and a yellow pigment. By contrast a tube of Chromium Oxide Green lists only one: Chromium Oxide, PG 17. (We will look at the use of black and the earth colors later when we discuss palette choices.)
Color Wheels and Triangles
The color wheel is the most common diagram for showing the relationships between colors. The first was actually made by Isaac Newton in 1704. He didn’t apply color to his wheel, but he put the names of the spectrum’s colors in the positions that he thought they should have: Red, Orange, Yellow, Green, Blue, Indigo, and Violet. Indigo is probably not in the spectrum, but Newton was a man of his time, and he wanted to think that there were seven colors, like the seven notes of the musical scale. The color wheel or circle has some value in demonstrating certain color relationships, but there is nothing fundamental about it at all: the wavelengths of red and violet actually lie at opposite ends of the visible spectrum, and that isn’t a circle but a line. It is probably an evolutionary accident that our visual apparatus makes red and violet seem so closely related that we are tempted to create a circle and link them together.
Unfortunately, this circle fragment only goes about four-fifths of the way around, for there is a whole family of colors that we can see but which are not in the spectrum—the purples. Once we have added them, we must fiddle with the spacing if we are to put red, yellow, and blue where we want them. And if we choose the paint-box primaries and make them equidistant, we must distort the relationships of the remaining colors. Such a wheel is said to be unbalanced. One goal in designing a color wheel might be to place colors and their complements directly opposite each other on the wheel, but it has been shown that you cannot do this accurately and also have the various hues spaced around the wheel so that they seem to be visually equidistant one from the other. Thus, there is nothing sacred about the color wheel, and you should regard it as being merely something we created to serve our own purposes—not something that is the “right” arrangement.
Color triangles go back to the physicist, James Clerk Maxwell, who was, like Newton, concerned with light rather than pigment. He took advantage of a peculiar property of equilateral triangles: the sum of the shortest distances from any point within the triangle to the three sides is the same as the length of the altitude (the shortest distance from any of the triangle’s apices to the opposite side). This means that you can place percentage scales on each of the altitudes ending with 100% at the apices, and then identify any point within the triangle in terms of three percentages that sum to 100%. Maxwell used such a diagram to identify any color being studied in terms of a given number of units of green, blue, and red—the light primaries.
Color triangles of one kind or another became very popular in the early 20th century, but not in terms of their mathematical characteristics, and the continued focus on the paint-box primaries in such triangles made them of little real use. The Dudeen color triangle introduced by Charles Allen Winter (and later picked up by John Sloan) is typical. It places red, yellow, and blue at the apices of the triangle and then attempts to place various named paints within the triangle. (Stephen Quiller adopted this triangle as well, but then inflated it in into the The Quiller Wheel ™, an unbalanced color wheel that uses a version of the paint-box primaries, Cadmium Lemon, Winsor Blue, and Alizarin Crimson. He had also made an effort to locate the various paints within the diagram, just as was done in the Dudeen triangle he had borrowed from Sloan.)
I have created my own version of a Maxwell triangle for representing pigment rather than light: it uses the print primaries, yellow, magenta, and cyan. This is essentially a color map of the colors mixable from these primaries, and such a diagram is called a color surface as opposed to the color circle (which is really just one-dimensional). However, it has its limitations. First, the colors are represented in increments of ten percent (screen values), and second, they are not varied by the addition of black or white. One might imagine a prism made by extending quadrangles up from the map surface in ten percent increments of white, and extending quadrangles down from the surface in ten percent increments of black. The resulting prism would be a color solid.
This triangle’s real advantage is that it approximates a mixing triangle which shows a large and representative sample of the possible mixtures, not only of the primaries, but of any colors which appear on its surface. These are real mixes from the primaries spread out according to a rational system, not by guess-work, and the screen values for these primaries are shown around the triangle’s margin. A glance at it will reveal the wealth of purples and blue-greens that cannot be made from the paint-box primaries. This is easily shown to be the case: just identify the basic paint-box primaries on the margins of the triangle and draw lines connecting them. The colors that fall outside of these lines cannot be mixed with those primaries. In the same way, the colors made possible with any limited palette can be shown approximately by locating the colors within the triangle and connecting them with lines. And you can get a general idea of the gamut of color produced by mixture of any two colors by locating them on the triangle and drawing a straight line to connect them. (My triangle is shown at the end of this essay.)
[You should know that some pigments are brighter than the colors that can be mixed with the Process Primaries (these would fall outside of the triangle), but for general purposes the triangle gives a satisfactory approximation of the colors available for painting.]
The colors on your palettes should reflect your needs, and this doesn’t merely mean your choice of subject matter. It also means the expression of your personal color sense and other aspects of your style. (I am going to talk exclusively about water-colors, here.) Anyone who pays attention to color could easily distinguish paintings by Arne Westerman, John Yardley, John Pike, Rowland Hilder, Ron Ranson, and James Fletcher-Watson. John Pike used New Gamboge, Burnt Sienna, Burnt Umber, Cadmium Red Light, Alizarin Crimson, Ultramarine Blue, Thalo Blue, and Thalo Green. You will note that several of these paints are opaque, but that did not keep him from painting lovely, glowing washes. Ron Ranson recommends Lemon Yellow, Raw Sienna, Burnt Umber, Light Red, Alizarin Crimson (rarely used), Ultramarine Blue, and Payne’s Grey. A very different palette.
Rowland Hilder remarks that he, like most water-colorists, is inclined to recommend a small palette of a few colors, but was somewhat surprised to find on counting them that he actually uses twenty-six. This includes the Earth Colors, Phthalocyanine Green, and Lamp Black. He makes extensive use of black since he has noted that the colors of the English landscape are founded on gray. And Eliot O’Hara recommends several large palettes, one consisting of Lemon Yellow, Aureolin, Cadmium Yellow Medium, Cadmium Orange, Cadmium Red, French (Ultramarine) Blue, Cobalt Blue, Cerulean Blue, Viridian, Indian Red, Burnt Sienna, Burnt Umber, Raw Umber, Davy’s Gray, and Ivory Black.
We have already discussed the question of whether you should mix all of your greens from yellow and blue or from a tube green, and have found that the argument for using only yellow and blue is not a very good one. What about the earth colors which are so prominent in some of the palettes we have just looked at. Some people claim that they can mix them, just as they mix their greens, but is that really the most sensible approach?
The notion that the Earth Colors dirty your mixtures won’t bear close examination. “Dirty” mixtures, “mud,” come from placing colors together in a way that is not harmonious—not from the colors themselves. We might actually suspect that Earth Colors from the tube are as likely to be as nice and clear as those produced by mixing. But whether the Earth Colors are mixed or not isn’t really the issue. Eugene Delacroix, a great colorist, said, “Give me mud and I will make the skin of Venus if you allow me to surround it as I wish.” (In fact, raw umber is one of the most useful pigments for mixing the color of Caucasian skin in oil portraits, and Yellow Ochre is useful there as well.) Some artists actually dip into the “mud” that accumulates in the corners of a mixing well and use it as a neutral.
Eliot O’Hara used the Earth Colors to darken and give greater depth and body to the warm colors of his palette, to “strengthen” them, as he put it. (The warm colors are generally lighter in value than the cool colors and it is difficult to darken them without losing their character.) The idea is simple. You could create a darker color by mixing it with its complement, thereby neutralizing it, or by mixing it from adjacent colors on the color wheel, which, as we have seen, also neutralizes it. Or you can do it by adding a darker version of the same general color. Thus, you can deepen a warm red by adding a little Light Red to it, or a cold red, by adding a little Venetian Red (Indian Red). In almost every case, this will be better than adding Neutral. Raw Umber will serve the same purpose for yellows, and Burnt Sienna or Burnt Umber will serve for oranges. For neutralizing colors, O’Hara used Davy’s gray, but with caution. (You should be aware that modern tubes of Davy’s Gray are mixtures that may include some dye colors, and you may not want to use them.)
Moreover, the Earth Colors are actually quite beautiful and can be used for their own sake. Thin washes of Yellow Ochre are a gorgeous pale yellow. Raw Sienna is semi-transparent and can substitute for New Gamboge. Burnt Sienna is a lovely color, and it is also wonderful mixer that makes splendid greens for pine and fir trees when mixed with Viridian. In fact, Burnt Sienna or Burnt Umber, together with Ultramarine Blue, make a surprisingly fine limited palette of only two colors for scenes that do not include bright reds, yellows, or greens. Ted Kautzky gives us amazing examples of this in Ways with Watercolor. A famous limited palette used by the early English water-colorists was the “Holy Trinity,” which consisted of Yellow Ochre, Cobalt Blue, and Light Red. The “Velasquez Palette,” Yellow Ochre, Burnt Sienna, and Lamp Black, is also interesting and fun to experiment with. An example is provided by Rex Brandt in Watercolor Techniques and Methods.
So, should you be willing to use the Earth Colors? I think so. It makes sense to ban a pigment from your palette if you have fallen into the trap of over-using it, or if it performs badly given the way you use it. People have dropped Burnt Umber because it is addictive and sometimes doesn’t mix well, and some have dropped Payne’s Gray because it becomes a repetitive, dominant color-note in their paintings. Personally, I would keep Burnt Umber and drop Payne’s Gray, but for another reason: Payne’s Gray becomes so much lighter when it dries that it is difficult to learn to use it well. Lacking a strong reason for banning an Earth Color, I think that you should have them all. They are useful, permanent, and inexpensive.
As for Black, despite the propagandistic slogan, “There is no black in Nature” (which isn’t true), the impressionists did use black. Scientists at the National Gallery in London have found black pigment particles in Impressionist paint films under microscopic analysis. A brief list of paintings containing black will make my point: Manet’s “Music in the Tuileries Garden” (1862), Bazille’s “Self Portrait” (1865), Monet’s “The Beach at Trouville” (1870), Pisarro’s “Fox Hill, Upper Norwood” (1870), Renoir’s “The Theatre Box” (1874), Sisley’s “The Watering Place at Marly-le-Roi” (1875), Morisot’s “Summer’s Day” (1879), and Cezanne’s “Hillside in Provence” (1886). This list includes eight painters and covers twenty-four years—and it could be made very much longer. But black should generally be used in mixtures, not in unrelieved blocks. (The Velasquez Palette mentioned above shows how useful it can be as a mixer: and as I mentioned earlier, it makes a very nice yellow-green when mixed with yellow.)
The hostility to some of these colors may stem from the emergence of a group of painters who value brilliance and transparency above everything else. Suffice it to say that the interplay between opaque and transparent paint films can be very beautiful, and some painters choose to use certain colors because they are opaque. I like what Roland Hilder says, and he uses them all, “I use color to produce a harmony and to convey a mood.” This doesn’t mean that we use them as decoration. Quite the contrary. In my opinion, many modern paintings are spotty or garish, and lack a solid value structure because the artists over-value brilliance and transparency.
Note It is easy to demonstrate that our sensation of color is produced by our nervous systems and is not a perceived property of light or "colored" objects. If you press your finger upon the corner of your closed eye, near your nose (or on the eyelids, but be very gentle), you will see colored lights called phosphenes, which are the eye's response to the physical stimulation. You can do this in a completely dark room. It is impossible to recommend a single comprehensive book on color and light that will be of particular value for artists. The information I have given can be found in many sources, but much of it is buried in references to tristimulus values, standard illuminants, etc. Other books are beautifully illustrated, but very weak on reliable content, and many books spend a great deal of time discussing the physiology of the eye and various theories about vision. Most of the books that are written by artists repeat old myths—and I would particularly advise you to not become sucked into the ideologies of the Bauhaus theorists who created new myths. I will mention just three books. Ralph M. Evans, An Introduction to Color, New York: John Wiley & Sons, Inc., 1948. As you can see, this is a very old book, and a technical one (Evans was the Head of the Color Control Department of Eastman Kodak), but is well-written and clear. Most of what he covers is still valuable. It remains my favorite single book on the subject despite its age. I'm sure that it has been long out of print. Rolf G. Kuehni, Color: an Introduction to Practice and Principles, New York: John Wiley & Sons, Inc., 1997. This is a brief but authoritative survey of the field by an expert. Kuehni is the Adjunct Professor of Color at North Carolina State University, but when he wrote Color he was Vice President of DyStar LP, a versatile company that supplies dyes and other products to the textile and leather industries. Trevor Lamb and Janine Bourriau, eds, Color: Art and Science, Cambridge, United Kingdom: Cambridge University Press, 1995. This is a rather uneven collection of essays, as might be expected since it is composed of the Darwin Lectures. Nonetheless, it has interesting essays on the perception of color, the history of pigments with reference to famous artists, color in nature, the role of color in culture, color in language, etc. Some of the books in the following list are quite old, and many are out-of-print. You may ask why I am including them, and the answer is simple. The very best books of the past eighty years (a large collection that I am barely touching) are almost always better than the new ones that appear in any given year unless there is essential new information available in these, which is seldom true of books on art techniques. (When posed this way it seems self-evident.) David Bomford, Jo Kirby, John Leighton, and Ashok Roy, Art in the Making: Impressionism, London: The National Gallery in association with Yale University Press, 1990. This is an absolutely splendid book which has a sister publication, Art in the Making: Rembrandt. Fifteen paintings were examined with X-radiography, infra-red photography, and analysis of pigments and media used. This is the source of the information I have provided on the use of black by the Impressionists. Rex Brandt, Watercolor Techniques and Methods, New York: Van Nostrand Reinhold Co., 1977. The so-called "California Watercolor School" now appears very dated to my eyes, too much defined by its idiosyncrasies, especially the predominance it gives to calligraphy. Nonetheless, some of their paintings are very fine. Rex Brandt is an excellent teacher and the twenty lessons in this comprehensive book are insightful. The Velasquez Palette appears on p. 35. Rowland Hilder, Painting Landscapes in Watercolor, New York: Watson-Guptill Publications, 1985. A great craftsman, Hilder's lovely paintings are instantly recognizable as his. His palette is described on pp. 36-7, and pp. 38-39 are devoted to a large photograph of his palette, which is a striking but silent answer to those who say that all of the paint wells must be scrupulously cleaned after each use, etc. But then, one must remember that he is painting the English landscape, which he has observed to be based on grey. Ted Kautzky, Ways with Watercolor, New York: Van Nostrand Reihhold Co., 1963. Another fine craftsman, Kautzky's palette is shown on p. 13, and his examples of two-color limited palette paintings are shown of p. 47. Kautzky discusses the principles of composition more fully than most artists, and he gives us some fairly detailed demonstrations. Eliot O'Hara, Making Watercolor Behave, New York: Milton Balch and Co., 1932. O'Hara is a giant figure in the history of watercolor-teaching in American. Trained as an engineer, he had more technical knowledge of light and pigment than most artists. He was dogmatic and opinionated—but almost always right. He discusses palettes beginning on p. 54. Two other small books of his are interesting as well, Making the Brush Behave, and Watercolor at Large. He discusses paint-mixing beginning on p. 34 of the latter, and "strengthening" the warm colors on p. 37. John Pike, Watercolor, New York, Watson-Guptill Publications, 1973. This book is filled with beautiful watercolors and excellent demonstrations (although most of the steps are shown in black and white). Pike writes well with a personal touch that is very appealing. He was the official artist for NASA and the National Gallery of Art for the Apollo 10 launch. His palette s discussed on p. 44. Stephen Quiller, Color Choices, New York: Watson-Guptill Publications, 1989. Quiller's style is instantly recognizable, but it is based on color choices driven by his own interpretation of color theory: so much so that I don't think anyone should attempt to imitate it. In general, I believe he misinterprets the notions of color-schemes and the use of complements. He discusses Sloan and the Dudeen Color Triangle beginning on p. 10, and his unbalanced color wheel is a foldout between pp. 16 and 17. John Sloan, John Sloan on Drawing and Painting, Mineola New York: Dover Publications, Inc., 1944 (reprinted 1977). This was originally published under the title, Gist of Art. Sloan was a member of the famous "Ashcan School," and this book covers drawing and painting and includes his "personal views" of life, art, and teaching. The Dudeen color triangle is discussed beginning on p. 119.