Notebook, 1993-


Thompson, Daniel V., Jr., Research and Technical Adviser, The Courtauld Institute of Art, University of London. The Practice of Tempera Painting. New Haven: Yale University Press. 1936. Fourth Printing, 1946.

The Practice of
Tempera Painting - Pigments and Brushes

Pigments and Brushes
The most obvious essential for any kind of painting is a set of brushes. A pigment is a colored solid which can be held in place by a binding medium. It is used to distinguish between pigments and dyes, dyes being soluble colored substances which penetrate the material to which they are applied, and alter its color by combining with it physically and often chemically. Pigments, in the strict sense of the word, are simply held in place on the surface by the adhesive medium in which they are applied. [p. 74]

Function of Pigments
The simplest illustration of a pigment is perhaps to be seen in mosaic. Each little colored cube may be regarded as a pigment particle held in place by the cement in which it is set. In the ordinary sense of the word, of course, the separate particles of a pigment are small; and instead of being set in the binding medium, as they are in mosaic, the pigment particles in most painting systems are completely immersed in the medium, and surrounded by it. Ordinarily a pigment for painting is a very fine powder; but this is not always the case, and even a fine powder, examined philosophically under the microscope, proves to be made up of separate bits which may be compared with the separate tesser² of the mosaic painting. [p. 75.]

Natural and Artificial Colors
In the Middle Ages, it was customary to distinguish between natural and artificial colors. The most obviously natural colors are, of course, the earths and minerals. The most obviously artificial colors are the products of laboratory chemistry. It is not always easy, however, to make such sharp distinctions, and even in the Middle Ages classification into these groups could seldom be carried out completely. Cennino speaks of colors which are natural "but need to [p. 75] be helped artificially." In point of fact, though nature has been generous with colored materials suitable for use as pigments, very few of them indeed occur ready for use without some special treatment; and it is not to our advantage to preserve this distinction between natural and artificial which the Middle Ages took over from Pliny. [pp. 74-75]

The Colored Earths
We may, however, begin our consideration of pigments with a glance at the colored earths. Everyone knows the appearance of colored earths in nature; but the red and gold and purple of new-ploughed fields, the mellow browns of well-tilled soil, are not readily adapted for painting purposes. These colored earths owe their colors partly to the presence of strongly colored salts of iron, manganese, and so on, resulting form the alteration of ores of these metals, and partly to organic matter, resulting from the decay of roots and leaves and wood; but these coloring elements are much diluted in most natural earths with little-colored clays and sand. They look quite dark and bright when they are wet, but they dry out generally quite weak and dull. It is only an earth which contains a very high percentage of the decomposed ores of coloring metals, such as iron, which will yield a good earth pigment. The slow action of air and water and organic acids breaks down the hard structure of the mineral ores, and turns them into powder, which we call ocher. The red earths and the yellow earths, the hematites and limonites of the geologists, are largely made up of oxides of iron, more or less hydrated, and more or less combined with other salts. The sand and clay and humus that they contain must be washed away to make them fit for pigments.

This washing is important, not only to produce a good, strong color, but also to insure its permanence. Some samples of raw sienna contain enough organic matter to cause them to mold in tempera, to swell on the panel, and eventually to come away. Naturally, the less washing a native earth requires, the better it is for the pigment-maker; so the red and yellow ochers of the painter tend to be made from the richest deposits of hematite and limonite-colored soils, from relatively few localities. One such locality, Sinope in Pontus [p. 75] [Asia Minor], was the great classical source of red earth, and this place-name has survived in the English color-name, "sinoper," meaning reddle or ruddle, a red earth rich in iron oxide.

There is enormous variation in the color of these natural earths. The reds range from the warm pink of Pozzuoli red [from the soil of Pozzuoli, in the bay of Naples] to the deep, almost purple red of certain Persian and English ores which are sold under the name of Indian red. What is called "English red" is usually a yellow ocher which has been turned brick-red by roasting. Venetian red usually means a fairly light, warm red ocher. The natural earths are very far from uniform in color, and the color merchant often has to secure the uniformity that his clients demand by blending several earths together. The medieval painter took his natural ochers as they came, and made the most of any special character they might exhibit. If he got hold of a right, warm earth red, he might put it aside for making flesh tones in fresco, and use it for that only, until he found a better pigment for the purpose. If a sample turned out dull and poor, he might still be able to find uses for it in some minor capacities, such as drawing on the wall, or underpainting for other colors.

Yellow ochers are no more standard in nature than the reds. We can sometimes buy a light ocher and a dark ocher, and sometimes one or two special qualities besides; but these do not by any means reflect the useful range of natural yellow earths. The green earths, too, called "terre-verte" in trade, are far from uniform in nature. They vary from bright apple green to dull olive, and include two distinct mineral classifications, celadonite, which yields the bright Verona terre-verte, and glauconite, the source of the "Bohemian" terre-verte, which run the gamut of yellow greens and olives. [pp. 75-76]

Individual Pigment Characters
In modern painting methods, subtle differences in the characters of pigments tend to be submerged. In oil painting it does not make very much difference in the effect whether the painter uses an earth red in its natural register or forces a bright red down in key to match the earth red in color. Modern oil paint has so much material [p. 76] in it besides pigment that the pigment hardly has a chance to show its individual character. In medieval painting methods, however, the separate pigments tend to be exhibited with emphasis, almost like jewels in a complicated setting. Fine colours were so hard to come by in the Middle Ages that the painter would not willingly degrade them by indiscriminate mixing. The palette was treated almost like a collection of precious stones, to be grouped in the painting with as much regard for their intrinsic beauty as possible. And the media in which they were applied gave full play to the individualities of the pigments; for they were unobtrusive in themselves. Medieval writings assume a regard for pigment characters, an interest in the fatness or leanness, coarseness or fineness, transparency or opacity, of each separate pigment; and even for its source, as a matter of significance to the collector. The medieval painter was as aware of the special qualities of his particular colors as a musician of the special qualities of instruments and voices. He tended to use his pigments in their natural register, modifying them simply and systematically according to established practices based on their known behavior.

Pigments in tempera show their natural characters plainly, and it is worth while to know a good many of them before settling down to the use of any one palette. It is worth while to buy small quantities of respectable pigments from different sources, as opportunities arise, and to select the ones which suit best your personal tastes. It is possible to go a long way with a very simple palette indeed; but it is a pleasure and an advantage to possess a group of pigments selected for their intrinsic beauty and their harmonious relationships. Chance may bring some fine ochers your way, quite different from the ordinary standard trade varieties, like the bright pink or deep purple-red earths which occasionally turn up in small lots; or the unusually bright or rich yellow ochers that one sometimes finds. Vermilions are by no means standard, and you may have use for the scarlet shades and also for the deep, almost violet sort that you can sometimes buy. There is a great range of quality even in cobalt blues and ultramarines, from greenish to violet, and from bright to dull. [p. 77]

It is not wise to buy pigments from doubtful sources unless you mean to test them, or to have them tested; for purity and permanence cannot safely be judged by superficial examination. But among the artists' colormen of unquestionable integrity, the standards of blending for different pigments vary; and many manufacturers are willing to supply ungraded samples of pigment from stock which has not been made up to their standard specifications. A friendly colorman once furnished me with a set of six or eight raw umbers ranging from the deepest olive brown to almost cocoa color, all of them pure and genuine; but in his published list he offered only one raw umber, the same year after year. It is the merchant's pride to maintain these uniformities; but it is not always to the artist's best interest. When you find a pigment that pleases you, it is a good plan to buy a generous stock of it. A difficult taste in ochers cannot be satisfied on the spur of the moment, and when a good chance comes it is wise to take advantage of it. Try colors out in as small samples as you please; and when you are especially pleased with one, buy a pound, or five pounds, or as much as you can, so that when you want more you will not be disappointed. If you buy from doubtful sources, you must get tests made, for safety's sake. It is not worth while for a painter to try to test his colors for himself; and as a general rule his best plan is to avail himself of the stocks and knowledge of the established colormen. They will usually welcome inquiries for choice pigments of special quality. [p. 78]

Palettes for Tempera
The basis of a palette for tempera is a selection of opaque colors. For ordinary purposes, the transparent colors are on the whole less useful, and should not be used except where their special qualities are actually desired. Actually, this is even more true of oil painting than of tempera; but the slovenly habit of mixing up transparent and opaque colors at random, with no thought of their natural characters, has become so common that many oil painters do not regard it as bad practice. In tempera, raw sienna is less generally useful [p. 78] than yellow ocher, burnt sienna than English red. The opaque oxide of chromium is more often useful than viridian; and cobalt blue works on the whole rather better than the artificial "French" ultramarine. The transparent colors, the siennas, viridian, aureolin, the red lakes, Prussian blue [if it be used at all], and so on, should be kept in reserve for use where nothing else will do, and not introduced casually into mixtures. [pp. 78-79]

Almost every mixture in tempera will contain at least a little white, and it follows that white is the most important single pigment in the palette. It must be selected with care. The classic white pigment is white lead, and it is still in some ways the best of all. It is heavy and dense, extremely opaque, and a little goes a long way. For crisp, loaded finishing strokes, it is unequaled by any substitute; and for the ordinary run of painting it works a little more agreeable than anything else. It is perfectly permanent when varnished, and does not at all deserve the bad character that chemists have tried to give it. Under unfavorable conditions, when not varnished, it is sometimes darkened by sulphureted hydrogen in the air; but this effect seldom or never takes place in practice, except with water co ors or gouache. Even when it does occur, it is not unduly serious; for its tends to correct itself automatically by oxidation of the black lead sulphide to white lead sulphate, a process which can easily be encouraged by exposure to hydrogen peroxide.

There is, however, a serious objection to the use of white lead in any water medium, and that is its poisonous nature. In oil, the risk of poisoning is very slight, because the oil holds the particles of white lead together, and only gross carelessness can account for an artist-painter's getting lead into his system from using this pigment. In tempera, however, one deals with white lead powder. It is ground simply in water, and it may dry out and be inhaled. It may get on a cigarette, or under a fingernail. In tiny quantities, it may find its way into the body in many ways. Once in, it tends to accumulate there, and infinitesimal amounts often repeated may in time build up a dangerous total. This danger is not imaginary; and it is very [p. 79] hard to guard against. If white lead is used, even on a small scale, one cannot take too earnest precautions against getting even a trace of it into his system. I consider it somewhat better than any other white pigment for tempera, and indispensable for oil; but I do not feel that its superiority in tempera is sufficiently great to warrant taking the trouble necessary to use it safely.

Instead of white lead, therefore, the tempera painter will be well advised to use a titanium white. Titanium lithopones, compounded from titanium and barium, are brilliantly white, opaque, permanent, and entirely adequate substitutes for white lead in tempera. Zinc white is not. Zinc white is bulky and disagreeable under the brush: it lacks tinting strength, so that you have to use a great deal of it to make other colors lighter, and hiding power, so that you have to use many coats of it to get a solid effect. Titanium white, on the other hand, is only a little more bulky than white lead, and beautifully dense. It give clean, sharply defined value gradations, paints out pleasantly, and leaves an agreeable surface. It requires little or no grinding, unlike white lead, which requires a great deal. [pp. 79-80]

Next to white, in tempera, black is perhaps the most useful of pigments. It has been the fashion for some years to condemn the use of black paint on the ground that there is no black in nature. Actually, I think, the origin of this prejudice is probably the ugliness of black prepared as an oil color in the modern way. There is nothing objectionable in some of Goya's blacks, certainly. But commercial oil paint black has neither luster nor fluidity in itself; and as an element in mixtures often tends to sully the color. In tempera, however, black and white yield fresh and beautiful grays, and the range of neutrals which may be made with them can be used to good advantage. The mixture of complementaries to produce neutrals, so much in vogue among oil painters, may well be forgotten by the tempera painter; for he can produce more satisfactory results by simpler and more economical means. He may use black as much or as little as his color scheme calls for, without fear that it will make his colour muddy. He will find that lampblack, an insufferably dirty color in [p. 81] oils, is an exquisite instrument in tempera, yielding a series of grays quite different from those he gets with ivory black; and those in turn quite different from the series which vine black, blue black, gives. They are all useful. If you want only one, choose ivory black. [pp. 80-81]

For yellow, to begin with, get a good ocher, and make it go as far as possible. That is likely to be surprisingly far; for in tempera yellows look stronger than they really are. This is so true that one can seldom use much of a yellow brighter than ocher without having it look out of key. Have a little middle cadmium on hand, but keep it in reserve, and do not use it until you find that ocher is really not bright enough. You may want a little of the stronger yellow in your mixed greens; but even there be careful! The great beauties of tempera are the cool, pearly, light, blond tones, which one cannot get in oil; and they have the effect of making even a dull yellow seem quite bright. Strong yellow or yellow green is apt to look rank in their company; but this is, of course, a matter of taste and judgment for each painter to settle for himself by experiment. There is no objection to bright yellows on grounds of handling. Chemically, middle cadmium is probably the most permanent.

Cennino had a yellow brighter than ocher, called "giallorino," but we do not know with certainty what it may have been. We know that at least two pigments went by this name, one a sort of yellow glass or frit, and the other, massicot, a yellow oxide of lead; but Cennino's giallorino may not have been either of these. There is only this against its having been massicot, that Cennino says it can be used in fresco; and we should not think it good practice to use massicot on fresh lime. Cennino's pale, bright yellow was orpiment, not unlike pale cadmium in color. Orpiment is a sulphide of arsenic, and extremely poisonous. Cennino did not regard it as very important for panel painting; and a modern painter will certainly do better to use the safe cadmiums instead, as far as bright yellows are needed. [p. 81]

Among your reds, you will almost certainly want a vermilion; [p. 81] and it is worth while to take some trouble to get a sample which suits you. The best Chinese vermilion is cold, pure red or very slightly violet, and rather more generally useful, perhaps, than the scarlet or orange tones of the cheaper vermilions; but they are all worth having. The way to judge a vermilion is to try it out in mixtures with white; for this brings out its character more fully than any other test. Vermilions are made by two different methods, and those made by the process used in Cennino's time are usually called "Chinese" vermilion in modern trade. They are somewhat better than the others, probably a little more inert chemically, and a little less likely to be to be darkened by sunlight. Any vermilion that you buy from a reputable colorman, however, is quite sure to be well made, well washed, and permanent enough to be used with an easy conscience. In America and Europe, the adjective "Chinese" is not intended to deceive. In China, in Tientsin, I bought a vermilion, beautifully packed and supposed to be of the highest quality, which owed no small part of its beauty to sophistication with an aniline dye. It pays to deal with a merchant of known dependability. Vermilions are expensive, and are frequently adulterated by second-class merchants. It is not worthwhile to grudge the cost of the finest qualities; for a little goes a long way, and it may as well be the best. If you have trouble in getting dry vermilion wet with water, add a drop or two of alcohol.

There is a modern red which looks a good deal like vermilion until you mix it with white, called cadmium red. It should really be called selenium red; for if there were no selenium in it, it would just be cadmium yellow. It is a useful color, and perfectly permanent. It does not take the place of vermilion, but makes a useful addition to it. Red lead, too, may be used quite safely if the color is wanted. The modern pigment is made by a different process from that used in the Middle Ages, and is redder, less orange, and on the whole less useful. It cannot be recommended; but it is quite reasonably permanent.

Among the earth reds there is so much useful variety that it is difficult to chose. One might start with a Pozzuoli red, a Venetian red, and an Indian red, and add others from time to time. It is [p. 82] extremely convenient in tempera painting to be able to mix a given tone with as few different pigments as possible, so as to be able to match it easily later on; and the less one's pigments are mixed together, the more their intrinsic beauty appears. So there is some advantage in having a palette selected from your special taste and needs. One painter may want a dozen earth reds, from the pinkest Pozzuoli to the deepest caput mortuum, with burnt ochers, like "light red" an English red, and the Mars browns and violets as well. All these are good, and the selection is purely a matter of taste. Differences among these colors are much more clearly marked in tempera than in oil. [pp. 81-83]

Browns are not likely to be wanted in very much variety; but the umbers, both raw and burnt, may be included, and burnt sienna for a redder tone. Raw umber, particularly, is a useful pigment for some color schemes. Cennino mixed his browns, out of black and white and yellow and red; and called the mixture "verdaccio." The word verdaccio means a nondescript greenish color, and that is just what Cennino's mixture produces. Black and white and yellow ocher give olive greens, and adding a little earth red turns them into greenish browns. Adding more red kills the green, of course; so presumably Cennino's verdaccio covered a range of brown to olive. It was not a definitely fixed color. A good olive raw umber saves the modern tempera painter a good deal of mixing. [p. 83]

The blues that Cennino speaks of are not generally available for modern painters. It is possible to buy genuine ultramarine, made as Cennino describes, from one or two English manufacturers; but it is extremely expensive, and few painters nowadays have the medieval enthusiasm for costly materials. It is rather significant that they have not; for it reveals as clearly as anything else the decline of the pride of craftsmanship in painting. A goldsmith could execute most beautiful designs in tin or imitation gold; but he would consider it [p. 82] beneath his professional dignity. Nickel could be substituted for platinum in jewelry, and no one would know the difference. Artificial ultramarine is almost as much like genuine ultramarine as nickel is like platinum. Time was when painters shared with gold-smiths feeling of pride in the nobility of the raw materials with which they worked; and this feeling was, of course, fostered by the noble purposes that their paintings were to serve. Cennino, in Chapter LXXXXVI, argues that it is good business to be liberal with precious materials, as well as becoming to one's dignity as a painter, and acceptable in the eyes of God and the Virgin. I am not at all sure that his argument is not still fundamentally sound. I am not at all sure that it would not be good business nowadays, and bracing to his self-respect, for a painter to use materials of some intrinsic worth, not because they were necessary, but because they were valuable. It would not make bad painting good; but it might have some good effects.

Cennino's other chief blue, azurite, is not offered in trade nowadays. It is no good in oil or water color, and so it has been dropped by the manufacturers from their lists. In tempera it is extremely beautiful, and highly individual. It is absolutely permanent in itself in tempera; but if it is surrounded by oil or varnish, as it may become, if it is insufficiently tempered, it turns black or green as the medium darkens. It is a troublesome color to paint with; for it must not be ground too much. And to get it at all, one must buy the mineral azurite, pound it, ground it, wash it, and grade it himself. There is no difficulty about doing this. Azurite stone can be obtained through Ward's Natural Science Establishment in the University of Rochester [N.Y.] Ward's supplies the very beautiful Arizona azurite to color manufacturers in Japan; but as far as I know, no artists' colorman in America or Europe offers the finished pigment for sale. Few modern painters will wish to take the trouble of grinding and washing the mineral; and cobalt blue makes a fair substitute.

Cobalt blue is an admirable pigment, perhaps, as I have said, more useful in tempera than artificial ultramarine. It is important to get a good quality, and the best grades are expensive as modern pigments go; but the cheap grades are worthless, or adulterated. A good deal [p. 84] of ultramarine passes for cobalt in cheap colors. There is no objection to cerulean blue as a pigment, and there are some other blues which may be used if they are wanted. Indigo is one of these. Its color is not likely to make it widely popular in these days; but anyone who wishes to use it in tempera may do so without much fear for its permanence. In thin washes, in water color, it fades; but in tempera it stands up well. [pp. 83-85]

The terre-vertes are good colors in tempera, and their transparency makes them peculiarly valuable in underpainting flesh. Verona terre-verte is the choicer, and a bright sample will prove extremely useful, alone and in mixtures. Another mineral green, malachite, a brilliant green copper ore, is most beautiful in tempera. It has, unfortunately, shared the fate of azurite. It can be bought as a pigment from one or two firms; but is invariably too finely ground. If you want it, you must buy the mineral, grind it, and wash it, saving the medium-fine particles, those which settle out of the washwater in five or ten minutes. It is hard to paint with, like azurite; for it is rather coarse and gritty. But, like azurite, it is inimitable. Azurite is useful even when finely ground; but malachite turns very pale and dull if it is reduced to anything like the fineness of other pigments.

Modern chemistry has given us a most valuable pigment in the opaque oxide of chromium, which supplies an important lack in the medieval palette. Opaque oxide of chromium is much less known and used than its near relative, viridian, the transparent oxide; but its importance will eventually be recognized. Viridian is enjoying nowadays much the same sort of vogue that Prussian blue enjoyed eighty or a hundred years ago. It is a very strong, self-assertive color, and so beautiful in itself that it is often allowed to dominate a great deal of modern painting. Like Prussian blue, it tends to infect the whole picture in which it is used; and, also like Prussian blue, it becomes monotonous. It is a most valuable invention, replacing verdigris; but it has been greatly abused. Much of its beauty is its transparency; and this is often wasted. Viridian is often use in opaque mixtures which could be produced more economically [p. 85] and more soundly with the opaque oxide of chromium as a basis. Both these chromium greens are useful in tempera; but the tempera painter will do well to reserve viridian for positions in which he really wants it. [pp. 85-86]

Transparent Pigments
Besides viridian, there are several other transparent colors which are sometimes wanted: first, the red lakes. Cennino believed in lac lake, a violet red called "Indian Lake" in modern trade, and seldom offered for sale. He knew Brazil wood lakes and grain lakes, and did not think highly of them. He did not, apparently, know the madders, or their synthetic counterparts, the alizarines. For the modern painter, these are the soundest red lakes to use. They are not absolutely permanent, but they are quite permanent enough for use. The madders age a little more agreeably than the alizarines: and as the latter are very strongly colored, the true madders are often more pleasant for glazing. Rose madder may be taken as the normal representative of the natural lakes, and alizarine crimson, of the artificial. Red lake glazes are useful and beautiful in modeling on white, yellow, and green. Transparent yellows, such as aureolin, which corresponds roughly in modern practice with Cennino's arzica and saffron, are useful chiefly for taking the chill out of these green glazes. Gamboge should not be used in tempera. Raw sienna is not likely to be of any special service, unless perhaps in painting hair, and may be omitted from the tempera palette without much hardship. Burnt sienna is quite transparent, and may be used. There is no serious objection on grounds of permanence to Prussian blue, if an absolutely transparent blue is really needed. As we shall see, even the most opaque colors can be made to give effects of relative transparency in tempera, and the transparent pigments are generally best kept in reserve. [p. 86]

Shell Gold
Among Cennino's pigments we must count powdered gold. This is used pure, over grounds of color, and sometimes mixed with the color in varying degrees. Used by itself as a pigment, it can be made to yield most brilliant and decorative effects. The lights on a blue drapery, for example, can be modeled up entirely with the pure gold, and the effect may be very fine. For small quantities it is convenient to buy the powder ready-mixed with gum, like a water color. It is sold as shell gold. One, two, or three little drops of gold paint are placed on a mussel shell, reminiscent of the shells in which the painters and illuminators of the Middle Ages used to mix and keep their colors. If you plan to use much of it, it is a good idea to buy the gold powder from a gold beater, and mix it yourself as you need it with a little week solution of gum arabic or with white of egg prepared as described [in the chapter on Gilding, p. 55] for gilding. [pp. 86-87]


[Thompson, Daniel V., Jr., Research and Technical Adviser, The Courtauld Institute of Art, University of London. The Practice of Tempera Painting. New Haven: Yale University Press. 1936. Fourth Printing, 1946.]



The contents of this site, including all images and text, are for personal, educational, non-commercial use only. The contents of this site may not be reproduced in any form without proper reference to Text, Author, Publisher, and Date of Publication [and page #s when suitable].