The following survey of Sargent's painting technique as seen through the eyes of his contemporaries is adapted from the cited sources. I have emended or edited where it seemed necessary to clarify or untangle the syntax, and added comments in brackets.

When Mr. John Collier was writing his book on The Art of Portrait Painting he asked John Singer Sargent for an account of his methods. Sargent replied:

As to describing my procedure, I find the greatest difficulty in making it clear to pupils, even with the palette and brushes in hand and with the model before me; to serve it up in the abstract seems to me hopeless.

With the assistance, however, of two of his former pupils, Miss Heyneman and Mr. Henry Haley, it is possible to obtain some idea of his methods.

When he first undertook to criticize Miss Heyneman's work he insisted that she should draw from models and not from friends.

If you paint your friends, they and you are chiefly concerned about the likeness. You can't discard a canvas when you please and begin anew — you can't go on indefinitely until you have solved a problem.

He disapproved (Miss Heyneman continues) of my palette and brushes. On the palette the paints had not been put out with any system.

You do not want dabs of color, you want plenty of paint to paint with.

Then the brushes came in for derision.

No wonder your painting is like feathers if you use these.

Having scraped the palette clean he put out enough paint so it seemed for a dozen pictures.

Painting is quite hard enough without adding to your difficulties by keeping your tools in bad condition. You want good thick brushes that will hold the paint and that will resist in a sense the stroke on the canvas.

He then with a bit of charcoal placed the head with no more than a few careful lines over which he passed a rag, so that is was a perfectly clean grayish colored canvas (which he preferred), faintly showing where the lines had been. Then he began to paint. At the start he used sparingly a little turpentine to rub in a general tone over the background and to outline the head (the real outline where the light and shadow meet, not the place where the head meets the background), to indicate the mass of the hair and the tone of the dress. The features were not even suggested. This was a matter of a few moments. For the rest he used his color without a medium of any kind, neither oil, turpentine or any other mixture.

The thicker you paint, the more color flows.

He had put in this general outline very rapidly, hardly more than smudges, but from the moment that he began really to paint, he worked with a kind of concentrated deliberation, a slow haste so to speak, holding his brush poised in the air for an instant and then putting it just where and how he intended it to fall.

To watch the head develop from the start was like the sudden lifting of a blind in a dark room. Every stage was a revelation. For one thing he often moved his easel next to the sitter so that when he walked back from it he saw the canvas and the original in the same light, at the same distance, at the same angle of vision. [This is explicitly the sight size method of drawing and painting that was traditional in 19th century academic art training and which appears to be the method Sargent passed on to his pupils. It seems to conflict with descriptions below of Sargent's personal studio technique, and it was certainly not the method by which he painted his landscape watercolors.] He aimed at once for the true general tone of the background, of the hair, and for the transition tone between the two. He showed me how the light flowed over the surface of the cheek into the background itself.

At first he worked only for the middle tones, to model in large planes, as he would have done had the head been an apple. In short, he painted as a sculptor models, for the great masses first, but with this difference that the sculptor can roughly lump in his head and cut it down afterwards, while the painter, by the limitations of his material, is bound to work instantly for an absolute precision of mass, in the color and outline he intends to preserve.

Economy of effort in every way, he preached, the sharpest self control, the fewest strokes possible to express a fact, the least slapping about of purposeless paint. He believed, with Carolus-Duran, that painting was a science which it was necessary to acquire in order to make of it an art.

You must draw with your brush as readily, as unconsciously almost as you draw with your pencil.

He advised doing a head for a portrait slightly under life size, to counteract the tendency to paint larger than life. Even so he laid in a head slightly larger than he intended to leave it, so that he could model the edges with and into the background.

The hills of paint vanished from the palette, yet there was no heaviness on the canvas: although the shadow was painted as heavily as the light, it retained its transparency.

If you see a thing transparent, paint it transparent; don't get the effect by a thin strain showing the canvas through. That's a mere trick. The more delicate the transition, the more you must study it for the exact tone.

The lightness and certainly of his touch was marvelous to behold. Never was there any painter who could indicate a mouth with more subtlety, with more mobility, or with keener differentiation. As he painted it, the mouth bloomed out of the face, an integral part of it, not, as in the great majority of portraits, painted on it, a separate thing. He showed how much could be expressed in painting the form of the brow, the cheekbones, and the moving muscles around the eyes and mouth, where the character betrayed itself most readily: and under his hands, a head would be an amazing likeness long before he had so much as indicated the features themselves. In fact, it seemed to me the mouth and nose just happened with the modeling of the cheeks; and one eye, living and luminous, had been placed in the socket so carefully prepared for it (like a poached egg dropped on a plate, he described the process) when a clock in the neighborhood struck and Mr. Sargent was suddenly reminded that he had a late appointment with a sitter. In his absorption he had quite forgotten it. He hated to leave the canvas.

If only one had oneself under perfect control, one could always paint a thing, finally in one sitting. Not that you are to attempt this. If you work on a head for a week without indicating the features you will have learnt something about the modeling of the head.

Every brush stroke while he painted had modeled the head or further simplified it. He was careful to insist that there were many roads to Rome, that beautiful painting would be the result of any method or no method, but he was convinced that by the method he advocated, and followed all his life, a freedom could be acquired, a technical mastery that left the mind at liberty to concentrate on a deeper or more subtle expression.

I had previously been taught to paint a head in three separate stages, each one repeating — in charcoal, in thin color wash and in paint — the same things. By Sargent's method the head developed by one process. Until almost at the end there were no features or accents, simply a solid shape growing out of and into a background with which it was one. When at last he did put them in, each accent was studied with an intensity that kept his brush poised in mid air until eye and hand had steadied to one purpose, and then ... bling! The stroke resounded almost like a note of music. It annoyed him very much if the accents were carelessly indicated, without accurate consideration of their comparative importance. They were, in a way, the nails upon which the whole structure depended for support.

Miss Heyneman subsequently left a study she had made, at Sargent's studio with a note begging him to write, "yes" or "no," according to whether he approved or not. He wrote the next day:

I think your study shows great progress much better values and consequently greater breathe of effect with less monotony in the detail. I still think you ought to paint thicker paint all the half tones and general passages quite thick and always paint one thing into another and not side by side until they touch. There are a few hard and small places where you have not followed this rule sternly enough.

A few days later he called. Miss Heyneman's usual model had failed, and she persuaded her chairwoman to sit in instead; Sargent offered to paint the head of the model.

This old head was perhaps easier to indicate with its prominent forms, but the painting was more subtle. I recall my astonishment when he went into the background with a most brilliant pure blue where I had seen only unrevealed darkness.

Don't you see it? The way the light quivers across it?

I had not perceived it: just as, until each stroke emphasized his intention. I did not see how he managed to covey the thin hair stretched tightly back over the skull without actually painting it. He painted light or shadow, a four cornered object with the corners worn smooth, as definite in form as it was indefinite in color, and inexpressibly delicate in its transitions.

He concentrated his whole attention upon the middle tone that carried the light into the shadow. He kept up a running commentary of explanation as he went, appraising each stroke, often condemning it and saying:

That is how not to do it! Keep the planes free and simple.

He drew a full, large brush down the whole contour of a cheek, obliterating apparently all the modeling underneath, but it was always further to simplify that he took these really dreadful risks, smiling at my ill concealed perturbation and quite sympathizing with it.

The second painting taught me that all the tonal values of a portrait depend upon its first painting, and that no tinkering can ever rectify an initial failure. Provided every stage is correct, a painter of Mr. Sargent's caliber could paint for a week on one head and never retrace his steps — but he never attempted to correct one. He held that it was as impossible for a painter to try to repaint a head where the understructure was wrong, as for a sculptor to remodel the features of a head that has not been understood in the mass. That is why Mr. Sargent often repainted the head a dozen times. He told me that he had done no less than sixteen of Mrs. Hammersley.

When he was dissatisfied he never hesitated to destroy what he had done. He spent three weeks, for instance, painting Lady D'Abernon in a white dress. One morning, after a few minutes of what was to be the final setting, he suddenly set to work and scraped out what he had painted. The present portrait in a black dress was done in three sittings.

He did the same with the portrait of Mrs. Wedgwood, and many others. Miss Eliza Wedgewood relates that in 1896 he consented, at the insistence of Alfred Parsons, to paint her mother. She sat for him twelve times, but after the twelfth sitting he said they would both be the better for a rest. He then wrote to Miss Wedgwood that he was humiliated by his failure to catch the variable and fleeting charm of her mother's personality — that looked like the end of the portrait. Some weeks later he saw Mrs. Wedgwood at Broadway, and struck with a new aspect he said:

If you will come up next week we will finish that portrait.

She came to Tite Street, a new canvas was produced, and in six sittings he completed the picture which was shown at the Memorial Exhibition.

Paint a hundred studies: keep any number of clean canvases ready, of all shapes and sizes so that you are never held back by the sudden need of one. You can't do sketches enough. Sketch everything and keep your curiosity fresh.

He though it was excellent practice to paint flowers, for the precision necessary in the study of their forms and the pure brilliancy of their color. It refreshed the tone of one's indoor portraits, he insisted, to paint landscapes or figures out of doors, as well as to change one's medium now and then. He disliked pastel, it seemed to him too artificial, or else it was made to look like oil or watercolor, and in that case why not use oil or watercolor?

Upon one occasion, after painting for me, he saw one hard edge, and drew a brush across it, very lightly, saying at the same time:

This is a disgraceful thing to do, and means slovenly painting. Don't ever let me see you do it....

I have also seen the assertion that he painted a head always in one sitting. He painted a head always in one process, but that could be carried over several sittings. He never attempted to repaint one eye or to raise or lower it, for he held that the construction of a head prepared the place for the eye, and if it was wrongly placed, the understructure was wrong, and he ruthlessly scraped and repainted the head from the beginning. That is one reason why his brushwork looks so fluent and easy; he took more trouble to keep the unworried look of a fresh sketch than many a painter puts upon his whole canvas.

The following extracts from Mr. Haley's account of Sargent's teaching at the Royal Academy Schools, 1897-1900, throw further light on his method.

The significance of his teaching was not always immediately apparent; it had the virtue of revealing itself with riper experience. His hesitation was probably due to a searching out for something to grasp in the mind of the student, that achieved, he would unfold a deep earnestness, subdued but intense. He was regarded by some of his students as an indifferent teacher, by others as a "wonder"; as a "wonder" I like to regard him.

He dealt always with the fundamentals. Many were fogged as to his aim. These fundamentals had to be constantly exercised and applied.

When drawing from the model, never be without the plumb line in the left hand. Everyone has a bias, either to the right hand or the left of the vertical. The use of the plumb line rectifies this error and develops a keen appreciation of the vertical.

He then took up the charcoal, with arm extended to its full length, and head thrown well back: all the while intensely calculating, he slowly and deliberately mapped the proportions of the large masses of a head and shoulders, first the poise of the head upon the neck, its relation with the shoulders. Then rapidly indicate the mass of the hair, then spots locating the exact position of the features, at the same time noting their tone values and special character, finally adding any further accent or dark shadow which made up the head, the neck, the shoulders and head of the sternum.

After his departure I immediately plumbed those points before any movement took place of the model and found them very accurate. A formula of his for drawing was:

Get your spots in their right place and your lines precisely at their relative angles.

On one occasion in the evening life school I well remember Sargent complaining that no one seemed concerned about anything more than an approximate articulation of the head upon the neck and shoulders. The [conventional] procedure was to register carefully the whole pose at the first evening's sitting of two hours. The remainder of the sittings were devoted to making a thoroughly finished tone drawing in chalk, adhering to the original outline, working from the head downwards. Thus the drawing was not affected by any chance deviation from the original pose by the model. Sargent could not reconcile himself to this. The method he tried to inculcate was to lay in the drawing afresh at every sitting, getting in one combined effort a complete interpretation of the model. The skull was to articulate properly upon the vertebrae; the same with all the limbs. A keen structural easy supple, moveable machine, every figure with its own individual characteristic as like as possible, an accomplishment requiring enormous practice and experience with charcoal, but taken as a very desirable goal to aim at, a method he followed in his own painting. To the student it meant a continually altered drawing, to portray the varying moods of the model.

In connection with the painting, the same principles were maintained:

Painting is an interpretation of tone through the medium of color drawn with the brush. Use a large brush. Do not starve your palette. Accurately place your masses with the charcoal, then lay in the background about half an inch over the border of the adjoining tones, true as possible, then lay in the mass of hair, recovering the drawing and fusing the tones with the background, and overlapping the flesh of the forehead. For the face lay in a middle flesh tone, light on the left side and dark on the shadow side, always recovering the drawing, and most carefully fusing the flesh into the background. Paint flesh into background and background into flesh, until the exact quality is obtained, both in color and tone so the whole resembles as wig maker's block.

Then follows the most marked and characteristic accents of the features in place and tone and drawing as accurate as possible, painting deliberately into wet ground, testing your work by repeatedly standing well back, viewing it as a whole, a very important thing. After this take up the subtler tones which express the retiring planes of the head, temples, chin, nose, and cheeks with neck, then the still more subtle drawing of mouth and eyes, fusing tone into tone all the time, until finally with deliberate touch the high lights are laid in, this occupies the first sitting and should the painting not be satisfactory, the whole is ruthlessly fogged by brushing together, the object being not to allow any parts well done, to interfere with that principle of oneness, or unity of every part; the brushing together engendered an appetite to attack the problem afresh at every sitting each attempt resulting in a more complete visualization in the mind. The process is repeated until the canvas is completed.

Sargent would press home the fact that the subtleties of paint must be controlled by continually viewing the work from a distance.

Stand back — get well away — and you will realize the great danger there is over overstating a tone. Keep the thing as a whole in your mind. Tones so subtle as not to be detected on close acquaintance can only be adjusted by this means.

When we were gathered in front of our display of sketches for composition awaiting some criticism, Sargent would walk along the whole collection, rapidly looking at each one, and without singling out any in particular for comment, he would merely say:

Get in your mind the sculptor's view of things, arrange a composition, decoratively, easy, and accidental.

This would be said in a hesitating manner, and then he would quietly retire. On one occasion, when the subject set for a composition was a portrait, the criticism was: "not one of them seriously considered." Many we had thought quite good, as an indication of what might be tried while a portrait was in progress. That would not do for Sargent. A sketch must be seriously planned, tried and tried again, turned about until it satisfies every requirement, and a perfect visualization is attained. A sketch must not be merely a pattern of pleasant shapes, just pleasing to the eyes, just merely a fancy. It must be a very possible thing, a definite arrangement — everything fitting in a plan and in true relationship frankly standing upon a horizontal plane coinciding in their place with a prearranged line. As a plan is to a building, so must the sketch be to the picture.

Cultivate an ever continuous power of observation. Wherever you are, be always ready to make slight notes of postures, groups and incidents. Store up in the mind without ceasing a continuous stream of observations from which to make selections later. Above all things get abroad, see the sunlight, and everything that is to be seen, the power of selection will follow. Be continually making mental notes, make them again and again, test what you remember by sketches until you have got them fixed. Do not be backward at using every device and making every experiment that ingenuity can devise, in order to attain that sense of completeness which nature so beautifully provides, always bearing in mind the limitations of the materials in which you work.

It was not only students who acknowledged their debt to Sargent. Hubert Herkomer in his reminiscences writes: "I have learned much from Sargent in the planning of lights and darks, the balance in tonality of background in its relation to the figure, the true emphasizing of essentials."

Sargent was well aware of the pitfalls that await the painter of the fashionable world, and as sitter after sitter took his place on the dais in his Tite Street studio he seemed to become more sensible of them. He tried again and again to escape, and he often, in his letters, expressed his fatigue. He wearied of the limitations imposed by his commissioned art. Painting those who want to be painted, instead of those whom the artist wants to paint, leads inevitably to a bargain, to a compromise between the artist's individuality and the claims of the model. Mannerism becomes a way out; that which pleases becomes an aim. Artistic problems give way before personal considerations: the decorative quality of a picture takes a secondary place. Sargent's sincerity, the driving need he had to express himself in his own way, his satiety with models imposed on him by fashion, culminated in revolt. He was forced, now and then, it is true, to return to his portraits, but his Boston work absorbed him more and more. The call of his studio in Fulham Road when he was in London, and of the Alps and the south of Europe in summer, came first. In 1910 his exhibits at the Acdemy, instead of portraits, were Glacier Streams, Albanian Olive Gatherers, Vespers and A Garden at Corfu; at the New English Art Club his exhibits were Flannels, On the Guidecca, The Church of Santa Maria della Salute, A Florentine Nocturne, A Moraine and Olive Grove.

When in 1901 Mr. J.B. Manson, then a student, wrote to Sargent for advice he received the following reply:

In reply to your questions I fear that I can only give you the most general advice. The only school in London of which I have any personal knowledge is the Royal Academy. If the limit of age does not prevent your entering it I should advise you to do so. There are also very good teachers at the Slade School. You say you are studying painting to become a portrait painter. I think you would be making a great mistake if you kept that only in view during the time you intend to work on a life class — where the object of the student should be to acquire sufficient command over his material to do whatever nature presents to him.

It is evident that in his student days Sargent shared the apprehension excited in the studio by his brilliant, free spoken teacher Carolus-Duran. "En art tout ce qui n'est pas indispensable est nuisible" — "In art, all that is not indispensable is unnecessary" was one of the precepts which Duran had formulated after his study of Velasquez. It became on of the texts of his studio. He urged his students to make copies of the pictures of Velasquez in the Louvre, not laborious copies, but copies "au premier coup." In painting a picture he would retreat a few steps from the canvas and then once more advance with his brush balanced in his hand as though it were a rapier and he were engaged in a bout with a fencing master. These gestures were often accompanied by appeals to the shade of Velasquez.

Those who watched Sargent painting in his studio were reminded of his habit of stepping backwards after almost every stroke of the brush on the canvas, and the tracks of his paces so worn on the carpet that it suggested a sheep run through the heather. He, too, when in difficulties, had a sort of battle cry of "Demons, demons," with which he would dash at his canvas.

It was, then, to such a workshop and under such a master that Sargent at the age of eighteen was admitted as a pupil, and the question arises, what did Sargent owe to the teaching of Duran? The question is best answered by remembering Duran's precepts and seeing how far they are reflected in Sargent's art. It has already been shown how Duran insisted on the study of Velasquez and the omission in art of all that was not essential to the realization of the central purpose of a painting. He had himself traveled far from the sharp contrast of values by which he had dramatized his picture L'Assassiné. He had got rid of his tendency to be spectacular. From Velasquez he had learned to simplify. His teaching was focused on the study of values and half tones, above all, half tones. Here lies, he would say, the secret of painting, in the half tone of each plane, in economizing the accents and in the handling of the lights so that they should play their part in the picture only with a palpable and necessary significance. Other things were subordinate. If Sargent excels in these respects, it is sufficient to recall the fact that they formed the core of Duran's instruction. There is no need to put his influence higher. Few pupils in painting who have the talent to absorb their master's teaching fail in the long run to outgrow his influence and to progress beyond and outside it on lines of their own.

Sargent himself always recognized his debt to the teaching of Duran. At the height of his fame, when looking at a portrait by a younger painter, he observed to Mr. William James:

That has value. I wonder who taught him to do that. I thought Carolus was the only man who taught that. He couldn't do it himself, but he could teach it.

Again, when Mr. James asked him how to avoid false accents he said:

You must classify the tonal values. If you begin with the middle tone and work outward from it towards the darks and lights — so that you deal last with your darkest darks and highest lights — you avoid false accents. That's what Carolus taught me. And Franz Hals. It's hard to find anyone who knew more about oil painting than Franz Hals. That was his procedure. Of course, a sketch is different. You don't mind false accents there. But once you have made false accents in something which you wish to carry far, in order to correct them you have to deal with both sides of them and that gets you into a lot of trouble. So that's the best method for anything you wish to carry far in oil paint.

Mr. George Moore, in one of the most illuminating essays in Modern Painting, said: "In 1830 tonal values came upon France like a religion. Rembrandt was the new Messiah, Holland was the Holy Land, and disciples were busy dispensing the propaganda in every studio." The religion had no more ardent apostle than Carolus Duran.

One picture Sargent exhibited at the Academy in 1896 may be especially mentioned because it elected the warm admiration of Mr. George Moore, who was far from being enthusiastic about Sargent. Mr. Moore wrote of this portrait (Miss Priestley):

"Gradually a pale faced woman with arched eybrows, draws our eyes and fixes our thoughts. It is a portrait by Mr. Sargent, one of the best he has painted. By the side of a Franz Hals it might look small and thin, but nothing short of a fine Hals would affect its real beauty. My admiration for Mr. Sargent has often hesitated, but this picture completely wins me. The rendering is full of the beauty of incomparable skill. The portrait tells us that he has learned the last and most difficult lesson — how to omit. A beautiful work, certainly. I should call it a perfect work were it not that the drawing is a little too obvious: in places we can detect the manner. It does not coule du source like the drawing of the very great masters."

It was a common experience for Sargent, as probably for all portrait painters, to be asked to alter some feature in a face, generally the mouth. Indeed, this happened so often that he used to define a portrait as "a likeness in which there was something wrong about the mouth." He rarely acceded, and then only when he was already convinced that it was wrong. In the case of Francis Jenkinson, the Cambridge Librarian, it was pointed out that he had omitted many lines and wrinkles which ought to be shown on the model's face. Sargent refused to make, he said, "a railway system of him."

His refusal more than once led to scenes. On one occasion the lady who had taken exception to the rendering of her mouth became hysterical and fainted. Sargent was the last man in the world to cope with such a situation. A friend who happened to call found him helplessly contemplating the scene. The model was restored to sense, but the mouth remained as it was.

A sitter has given the following account of being painted by Sargent in 1902:

At one of my sittings during which Mr. Sargent painted my hands I sat motionless for two hours. A certain way in which I had unconsciously put my hands together pleased him very much because the posture, he said, was clearly natural to me. He implored me not to move. We worked very hard — he with his magical brush, I with my determination to control fidgets and the restless instincts to which sitters are prone when forced to remain still for any length of time. For the most part we were silent. Occasionally I heard him muttering to himself. Once I caught: "Gainsborough would have done it! Gainsborough would have done it!"

He worked at a fever heat, and it was so infectious that I felt my temples throbbing in sympathy with his efforts, the veins swelling in my brow. At one moment I thought I was going to faint with the sense of tension and my fear to spoil the pose which had enthused him.

At the end of two hours he declared that the hands were a failure, and he obliterated them.

"I must try again next time," he said in a melancholy tone. At the next sitting he painted the hands quickly as they now appear, a tour de force in the opinion of some, utterly unsuccessful in the eyes of others.

My husband came several times to the sittings. On one occasion Mr. Sargent sent for him specially. He rode across the Park to Tite Street. He found Mr. Sargent in a depressed mood. The opals baffled him. He said he couldn't paint them. They had been a nightmare to him, he declared, throughout the painting of the portrait. That morning he was certainly in despair.

Presently he said to my husband: "Let's play a Fauré duet." They played, Mr. Sargent thumping out the bass with strong, stumpy fingers. At the conclusion Mr. Sargent jumped up briskly, went back to the portrait and with a few quick strokes, dabbed in the opals. He called to my husband to come and look: "I've done the damned thing," he laughed under his breath.

My sister, on the occasion of her visit to the studio during my last sitting, remembers seeing Mr. Sargent paint my scarf with one sweep of his brush.

What appeared to interest him more than anything else when I arrived was to know what music I had brought with me. To turn from color to sound evidently refreshed him, and presumably the one art stimulated the other in his brain.

Adapted from John Sargent by Evan Charteris (New York: Scribner's Sons, 1927).

To see one of Sargent's watercolours in the making always reminded me of the first chapter of Genesis, when the evening and the morning were the first day, order developed from chaos, and one thing after another was created of its kind. Having chosen his subject and settled himself with the sunshade, hat and paraphernalia all to his liking, he would make moan over the difficulty of the subject and say, "I can't do it," or "It's unpaintable," and finally, "Well, let's have a whack at it."

Perfect absorption would follow, and after what looked like a shorthand formula in pencil was on the block, the most risky and adventurous technique would come into play, great washes of colour would go on the paper with huge brushes or sponges, and muttering of "Demons! Demons!" or "The devils own!" would be heard at intervals.

All the time the picture was growing surely, swiftly; he worked through to the end, only stopping when it was a subject where light and tide changed before he could get it all in, and two "goes" were necessary.

Mary Newbold Patterson Hale, in The World Today (November 1927).