315. As to the advantages the painter will find in it when it is required,—

316. We learn by the law of simultaneous contrast of colours, that when we regard attentively two coloured objects at the same time, neither of them appears of the colour peculiar to it; that is to say, such as it would appear if viewed separately, but of a tint resulting from the peculiar colour and the complementary of the colour of the other object. On the other hand, if the colours of the objects be not of the same tone, the lightest tone will be lowered, and the darkest tone will be heightened.

317. The first conclusion from this is, that the painter will rapidly appreciate in his model the colour peculiar to each part, and the modifications of tone and of colour which they receive from contiguous colours. He will also perceive and be prepared to imitate modifications in them, which, if they had not always escaped him because of their feeble intensity, might have been disregarded, because the eye is peculiarly susceptible of fatigue when it seeks to disentangle modifications, the cause of which is unknown, and which are not very prominent.

318. Let us now return to mixed contrast (77 et seq.), in order to make it evident that the painter is liable to see the colours of his model inaccurately. As the eye, after observing one colour for a certain time, has acquired a tendency to see its complementary, and as this tendency is of some duration, it follows, not only that the eyes of the painter thus affected cannot see correctly the colour which he had for some time looked at, but also whatever colour he sees while this modifying influence lasts. So that, conformably to what we know of mixed contrast, he will see,—not the colour which is before him,—but the result of this colour, and of the complementary of that seen previously. It must be remarked, that besides the want of clearness of view which will arise, in most cases, from the want of exact coincidence of the second image with the first—for example, when the eye has seen a sheet of green paper A (Plate III., fig. 4), in the first place, and, in the second place, a sheet of blue paper, B, of the same dimensions, but placed differently, this second image, not being coincident in all its surface with the first, A, as represented in the figure, the eye will see the sheet B violet only in the part where the two images coincide. Consequently, this defect of perfect coincidence of images will affect the outline of the second image, as well as the colour which it really possesses.

319. We can establish three conditions in the appearance of the same object relatively to the state of the eye; in the first, the organ simply perceives the image of the object without taking into account the distribution of colours, light, and shade; in the second, the spectator, seeking to understand this distribution, observes it attentively, when the object presents to him all the phenomena of simultaneous contrast of tone and colour that it is capable of exciting in him. In the third case, the organ, from the prolonged impression of the colours, possesses in the highest degree a tendency to see the complementary of these colours; these different states of the organ being continuous.

I have no doubt that the dull colouring with which many artists of merit have been reproached is partly due to this cause, as I shall show more minutely hereafter.

Utility of this Law to facilitate the prompt Imitation
of Modifications of Light on the Model.

320. The painter, knowing that the impression of one colour beside another is the result of the mixture of the first with the complementary of the second, has only to estimate mentally the intensity of the influence of this complementary, to reproduce faithfully in his imitation the complex effect which he has before his eyes.

321. A painter wishing to imitate a white stuff with two contiguous borders, one red, the other blue, perceives each of them changed by the influence of their reciprocal contrast; thus the red becomes more and more orange, in proportion as it approaches the blue, as this latter becomes more and more green as it approaches the red. The painter, therefore, making the borders of a single red and a single blue, reduced in some parts by white or by shade, will reproduce the effects he wishes to imitate. Whenever it is found that the painting is not sufficiently marked, he is sure of what he must add without departing from the truth, farther than by exaggerating a little (305).

2. A grey pattern drawn upon a yellow ground: the ground may be of paper, silk, cotton or wool; according to its contrast, the design will appear of a lilac or a violet colour (66).

The painter who would imitate this object, can reproduce it faithfully with grey. But if a painter, ignorant of the reciprocal influence of blue and red, convinced that he must represent what he sees, adds green to his blue, and orange to his red; as in the second example, he will trace a pattern more or less violet upon the yellow ground. Now, supposing that the painter had perfectly seized the modifications of the model, and, subsequently, had retouched his copy sufficiently to produce a perfectly faithful effect, it is evident it would have been perfect only after a number of trials, since he must have effaced what was first done.

3. I cite a third example of the influence of contrast, not relating to colours, like the two preceding, but to the different tones of the same colour, contiguous to each other.

Suppose several bands in juxtaposition, 1, 2, 3, 4, (Plate 1, fig. 3), of different tones in flat tints of the same scale, to form part of an object: to imitate it perfectly, it is evident that it must be painted in flat tints; but this object will appear to the eye a channeled surface, the lines where the two bands touch will appear like a relief by the effect of contrast of tone (9-11); therefore, if the painter is ignorant of this, he will reproduce, not an absolute copy of the model, but an exaggerated one. I the more willingly cite this example, because it gave me an opportunity of enabling a most skilful paperstainer to appreciate the utility of the law of simultaneous contrast. In going with him over his factory, he showed me a chimney-board representing a child whose figure stood out from a ground formed of two circular bands in grey flat tints; 1 and 2 (Plate I., fig. 7); the first was higher than the second; the phenomenon of contrast of tone was manifested at the borders, a a, of the two bands, so that the part of the band 2, contiguous to the band 1, was darker than the rest, as the part of the band 1, contiguous to 2, was lighter than the rest, conformably to what has been stated above (11). This effect, not being what the skilful artist wished to obtain, he inquired of me how it was to be avoided. I replied, that the grey of the band 2 must be reduced with white, in proportion as it approached the border a a; and, on the contrary, the grey of the band 1 must be strengthened with black, in proper gradations, beginning at the same border. And I proved to him, that to imitate the model faithfully, we must copy it differently from what we see it.

322. From the above we educe the six following principles:—

323. From this principle it results that in many cases where grey is near to a pure colour in the model, the painter, if he wishes to imitate this grey which appears to him tinted with the complementary of the pure colour, need not use a coloured grey, as the effect will be produced in the imitation by the juxtaposition of the colour with the grey contiguous to it.

Besides, the importance of this principle cannot be doubted, when we consider that all the modifications which a monochromous object presents (excepting those which result from the reflections of coloured lights emanating from neighbouring objects,) arise from the different relations of position between the parts of the object and the eye of the spectator; so that it is strictly true to say that, to reproduce by painting all these modifications, it suffices to have a colour exactly identical to that of the model, with black and white. In fact, with white we can reproduce all the modifications due to the weakening of the colour by light, and with black, those which are due to the height of its tone. If the colour of the model in certain parts gives rise to the manifestation of its complementary, because these parts do not return to the eye enough colour and white light to neutralize this manifestation, the modification may be imitated by the employment of a normal grey tone, properly surrounded with the colour of the object.

It is necessary, in many cases, to employ with the colour of the object the colours which are near it; that is to say, the hues of the colour. For example; in imitating a rose, we can employ red shaded with a little yellow, and a little blue, or, in other terms, shaded with orange and violet; but the green shadows which we perceive in certain parts arise from the juxtaposition of red and normal grey.

5. To put a dark colour near a different, but lighter colour, is to heighten the tone of the first, and to lower that of the second, independently of the modification resulting from the mixture of the complementaries. An important consequence of this principle is, that the first effect may neutralize the second, or even oppose it. For example; a light blue placed beside a yellow tinges it orange, and consequently heightens its tone; while there are some blues, so dark relatively to the yellow, that they weaken it so much as not only to hide the orange tint, but even to cause sensitive eyes to feel that the yellow is rather green than orange. A very natural result, if we consider that the paler the yellow, the greener it appears.

6. Put beside each other two flat tints of different tones of the same colour, chiaro-’scuro is produced, because, in setting out from the line of juxtaposition, the tint of the band of the highest tone is insensibly enfeebled, while, setting out from the same line, the tint of the band of the lowest tone becomes heightened; thus there is a true gradation of light.

The same gradation takes place in all the juxtapositions of colours distinctly separated.

I believe that attention to these principles, and especially a perfect knowledge of the consequences of the last three, exercises a very happy influence upon the art of painting, giving to the artist such a knowledge of colours as he cannot possess before the law of their simultaneous contrast and its consequences have been developed.

Among the details which the painter endeavours to render, there are many which, due to contrast, either of colour or of tone, must be produced spontaneously. I presume that the Greek painters, whose palette was composed only of black, white, red, yellow, and blue, and who executed so many pictures which their contemporaries have spoken of with intense admiration, painted conformably to the simple method of which I speak; devoting themselves to great effects, many small ones resulted from them.

Utility of the Law in order to Harmonize those
Colours of a Composition which are Inherent
to the Nature of the Object represented.

324. In all, or nearly all, compositions of painting, we must distinguish the colours which the painter is under the necessity of using, and those which he may choose, because, unlike the former, they are not inherent to the model (313). For example, in painting a human figure, the colour of the flesh, the eyes, and the hair, are fixed by the model; but the painter has a choice of draperies, ornaments, background, &c. In an historical picture, the flesh colours are, in the majority of the figures, at the choice of the painter, as are also the draperies and all the accessories, which can be placed and imagined according to his judgment.

In a landscape, the colours are determined by the subject, yet not so arbitrarily but that we can substitute for the true colour that of a neighbouring scale; the artist may choose the colour of the sky, imagine numerous accidental effects, introduce into his composition animals, draped-figures, carriages, &c., of which the form and colour may be so selected as to produce the best possible effect with the actual objects of the scene.

325. A painter may also choose a dominant colour which produces, on every object in his composition, the same effect as if they were illuminated by a light of the same colour, or as if they were seen through a coloured glass (259).

326. Although the law of contrast affords different methods of imparting value to a colour, genius alone can indicate the mode in which this idea should be realized in a painting.

327. Whenever the artist would attract the eye by colours, doubtless the principle of harmony of contrast must be his guide. The law of simultaneous contrast indicates the means of giving value to the pure colours by each other; means which, although spoken of, are but little known, as may be commonly seen in portraits of vivid colours, badly assorted; and in those numerous small compositions in tints broken with grey, where we look in vain for a pure tone; which, however, from the objects represented in them, are eminently adapted to receive all vivid colours.

328. The contrast of the most opposite colours is as agreeable as possible, when they are of the same tone. But if crudity or too great intensity of colours is feared, we must have recourse to the light tones of their respective scales.

329. When the painter breaks tones with grey, and wishes to avoid monotony, or when on the planes which are more remote, yet not so remote as to render their differences of colour inappreciable, he wishes every part to be as distinct as possible, he must have recourse to the principle of harmony of contrast, and mix his colours with grey.

330. This method of bringing out a colour by contrast, in using either light tones complementary or more or less opposed, or broken tones more or less grey, and of tints complementary to each other; or, in employing a broken tone, of a tint complementary to a more or less pure contiguous colour, ought especially to fix the attention of portrait painters. A portrait will have a very poor effect when neither the colour of the dress nor of the background have been well chosen.

331. The portrait-painter must endeavour to find the predominating colour in the complexion he has to paint; and this found and faithfully reproduced, he has to seek whatever accessories at his disposal will give value to it. It is a very common error to suppose that the complexion in women, to be beautiful, must consist only of red and white: if this opinion be true for most of the women of our temperate climate, it is certain that in warmer regions there are brown, bronzed, or even copper complexions endued with a brilliancy, I may say beauty, appreciated only by those who, in pronouncing upon a new object, lay aside habitual expressions, which (albeit unconsciously to most men), exercise so powerful an influence upon their judgment of objects seen for the first time.

See the section in which I have treated of the application of the law of contrast to dress. (Div. V.)

332. In order to make the best use of colours without being under the necessity of multiplying them, as, for example, in draperies of a single colour, recourse may be had to the coloured rays emanating from neighbouring bodies, whether visible to the spectator or out of sight. For example, a green or yellow light falling upon part of a blue drapery renders it green, and by contrast heightens the blue-violet tone of the rest; a golden yellow light falling on part of a purple drapery imparts to it a golden tint, which makes the purple of the rest come out, &c.

333. The principle of harmony of contrast then procures for the painter in chiaro-’scuro the means of realising, with respect to brilliancy of colours and distinction of parts, such effects as are produced in paintings of flat tints.

334. Having treated of the utility of the law of simultaneous contrast in the intelligent use of pure opposite colours, and of colours broken by grey similarly opposed when it is required to multiply pure and varied colours, it now remains for me to treat of those cases in which the painter, desiring less diversity in the object, less variety in the colours, employs sparingly the harmony of contrast, preferring the harmony of scale and the harmony of hues.

335. The greater the variety of colours and accessories in a composition, the more the eyes of the spectator are distracted, and the more difficulty is experienced in fixing attention. If then, this condition of diversity of colours and accessories is obligatory on the artist, the more obstacles there are to surmount in drawing and fixing the attention of the spectator upon the physiognomy of the figures, whether they represent the actors in a single scene, or whether they are simply portraits. In the latter case, if the model has such an ordinary physiognomy, as recommends itself neither by its expression nor its beauty, and still more, if he must conceal or dissemble a natural defect, all that is accessory to this physiognomy, all the resources of contrasted colours, well assorted, should come to his aid.

336. But if, fervently inspired, he appreciates the purity of expression, the nobility and loftiness of character pertaining to his model; or even if a physiognomy, to most eyes commonplace, strikes him by such an expression as he judges to belong only to men animated by noble ideas, it is to such a model that he will address himself and fix his chief attention; so that in giving it life upon his canvas, no one can mistake either the resemblance, or the sentiment which directed his pencil. Everything being accessory to the physiognomy, the draperies will be of black or of sombre colours; and if ornaments relieve them, they will be simple, and always in keeping with the subject.

337. When, in this point of view, we examine the masterpieces of Vandyke, and trace the beauty of their effect to the simplicity of the means which produce it,—when we consider the elegance of their attitudes, which always appear natural, the taste which presided over the selection of all the accessories, we are struck with admiration of the genius of the artist, who has not had recourse to those means, so much abused at the present day, of attracting attention, either by giving to the most vulgar person an heroic attitude, to the most commonplace physiognomy pretension to profound thought, or by seeking extraordinary effects of light, such as filling the figure with a strong light, while the rest of the composition is in shade.

338. These reflections indicate the course which an historical painter must take, when he would particularly fix the attention upon the physiognomy of the persons in a remarkable scene. The more he employs allied scales, the more care he must take to select such as do not lose too much by their mutual juxtaposition.

339. There is another important direction to give, which is to avoid as much as possible the same kind of images on different objects; thus figures clothed in draperies with large flower patterns, in a room where the carpet and porcelain vases repeat the same images, are never free from objection, for it is troublesome to the eye to distinguish those parts of the picture which the similarity of ornaments tends to confound. Upon the same principle, the painter must generally avoid placing beside the faithful copy of a model the copy of an imitation which repeats this model. For example, when, he paints a vase of flowers, the artist produces most effect, other things being equal, in painting a vase of grey or white porcelain, instead of a vase upon which a profusion of similar objects are its ornaments.

340. When it is required that a certain colour shall predominate in a composition, or to speak more correctly, when the scene is illuminated by a coloured light, shed over every object, we must not only take simultaneous contrast into consideration, but also the modification which results from the mixture of colours (172), comprising the recomposition of white light by means of a proper proportion of the differently coloured elementary rays.

341. We must here attentively study the article which treats of the principal cases of the modifications of light resulting from coloured rays falling upon bodies of various colours (261, et seq.), when, although the coloured light chosen imparts value to certain colours of the objects upon which it falls, it also impoverishes and even neutralizes others. Consequently, in employing any predominant colour, we must renounce the advantages of others, or the effect produced will be false. For example: if orange colour predominates, for the colouring to be true it must necessarily follow—

• 1. That the purples must be more or less red.
• 2. That the reds must be more or less scarlet.
• 3. That the scarlets must be more or less yellow.
• 4. That the orange must be more intense, more vivid.
• 5. That the yellows must be more or less intense, and orange.
• 6. That the greens lose their blue, and consequently become yellower.
• 7. That the light blues become more or less light grey.
• 8. That the deep indigo becomes more or less maroon.
• 9. That the violets lose some of their blue.

Thus we see that orange light heightens all the colours which contain red and yellow, while, neutralizing a portion of blue in proportion to its intensity, it destroys wholly or partially this colour in the body which it illuminates, and consequently disturbs the greens and the violets.

342. With reference to the true imitation of colouring, it appears to me that painters of interiors, have, other things being equal, more skill than historical painters in faithfully reproducing the modifications of light. Historical painters, attaching more importance to the attitudes and physiognomy of their figures than to the other parts of their composition, attend less to small details, the faithful imitation of which is the essential merit of the painter of interiors. Besides, the historical painter is never in a position to see the whole of the scene he would represent, while the painter of interiors, having constantly his model before him, sees it completely, as he imitates it upon the canvas. Hence, therefore, in every small composition the colours, as well as the objects represented, must be distributed with a kind of symmetry, so as to avoid being what I can best express by the term spotty. In fact, for want of a good distribution of objects, the canvas will not be filled in some parts, or, if it is, there will be, in many places, evident confusion; so also if the colours be not properly distributed, the picture will be spotty, because they are too far isolated from the others. (See 249-251.)

343. I believe that those painters who will study the mixed and simultaneous contrasts of colours, in order to employ rationally the coloured elements of their palette, will perfect themselves in absolute colouring as by studying the principles of geometry they perfect themselves in linear perspective. I have no doubt but that the difficulty encountered by painters ignorant of the law of contrast, of faithfully imitating their model, has been with many the cause of a colouring dull and inferior to that of artists, who, less careful than they in the fidelity of imitation, or not so well organized for seizing all the modifications of light, have worked more by their first impressions, or, in other words, seeing the model more rapidly, their eyes have not had time to become fatigued; and thus, content with the imitation which they have made, they have not returned to their work too often to modify, to efface, and afterwards to reproduce it upon a canvas soiled by the colours first put on. There are, indeed, many painters, to whom the maxim, “Let well alone,” is peculiarly applicable.

SECOND DIVISION.—
Imitation of Coloured Objects by Materials
of a Definite Size, as Threads, &c.

344. The tapestries of Gobelins and of Beauvais, and also the carpets of Savonnerie, and certain very elaborate mosaics, may all be considered as works which resemble the method of painting in chiaro-’scuro; while the windows of Gothic churches correspond more or less exactly to painting in flat tints; so also with tapestries for furniture and carpets, which, instead of being fabricated with scales of at least sixteen or eighteen tones, as they are in the royal manufactories, are composed with scales consisting of three or four tones only, and, far from imitating the effects of chiaro-’scuro, the coloured objects produced present to the eye only small monochromous bands of a single tone.

345. There are also some works whose coloured designs are upon a kind of mixed system, being the result of the juxtaposition of monochromous single tinted parts, of a visible size, but in which the effects of chiaro-’scuro are sought by using gradations of scale or a mixture of hues; such are ordinary mosaics, carpets, embroidered tapestries, &c.

346. In the patterns of tapestries and carpets, when the principal effects to be aimed at have been determined, it may be seen what points of ordinary painting may be sacrificed to obtain them, and what must be done for perfecting the special portion of their imitation. Beginning with the physical condition of the coloured elements the weaver employs, and the texture of the tapestry, we see the necessity of representing in this kind of work only large, well-defined objects, and especially such as are remarkable for the brilliancy of their colours. The patterns for hangings must recommend themselves more by the opposition of their colours than by the minute finish of their details. To pretend to rival painting in these manufactures is to establish a confusion most detrimental to the progress of arts, absolutely distinct from painting, both in their purpose and their means of execution.

347. The principles truly essential to these arts of imitation being once deduced from the peculiarities of each, it becomes easy to point out by what efforts we may hope to obtain true perfection.

CHAPTER I.
ON THE ELEMENTS OF
GOBELINS TAPESTRY.

348. To make mixtures of coloured threads intelligently, we must be guided by the three following rules:—The first two resulting directly from observation of facts; the third being the natural deduction from the facts comprised in the two former.

Rule I.— THE BINARY MIXTURE OF PRIMARY COLOURS.

When we unite Red with Yellow, Red with Blue, Yellow with Blue, the threads must not reflect a perceptible quantity of the third primary colour, if we would have Orange, Violet, and Green as brilliant as it is possible, by this method, to obtain.

Example A.—Red and Yellow.
Plate 11.

3	Red	threads with	1	Yellow	thread,
2	”	”	1	”	”
1	”	”	1	”	”
3	Yellow	”	1	Red	”
2	”	”	1	”	”

yield mixtures which appear to the eye in proportion to the two colours mixed. There is no appearance of grey in any of these mixtures, when we employ a red more inclining to orange than to crimson, and a yellow more inclining to orange than to green.

B.—Red and Blue.

3	Red	threads with	1	Blue	thread,
2	”	”	1	”	”
1	”	”	1	”	”
3	Blue	threads with	1	Red	thread,
2	”	”	1	”	”

yield mixtures which appear to the eye in proportion to the two colours mixed. If we use a red and a blue inclining to violet, the mixture will contain no grey.

C.—Yellow and Blue.

4	Blue	threads with	1	Yellow	thread,
3	”	”	1	”	”
2	”	”	1	”	”
1	”	”	1	”	”

give mixtures which appear to the eye in the proportions of the two colours mixed. If we use yellow and a blue inclining to green more than to red, the mixture will contain little or no grey.

Experiment on all the preceding mixtures demonstrates the rule above; or rather, this rule is but the expression of a generalization of facts.

Rule II.— THE MIXTURE OF COMPLEMENTARY COLOURS.

349. When we mix Red with Green, Orange with Blue, Yellow with Violet, the colours are more or less completely neutralized, according as they are more or less perfectly complementary to each other, and as they are mixed in proper proportions. The result is a grey, the tone of which is generally higher than that of the colours mixed, if the latter are of a suitably high tone.

Examples.—Red and Green.
Plate 12.

• 3 Red threads with 1 Green thread give a dull Red.
• 2 Red threads with 1 Green thread give a duller and a deeper Red.
• 1 Red thread with 1 Green thread give a Reddish-grey.
• The tone a little higher than the preceding.
• 3 Green threads with 1 Red thread give a Green Grey,
• the tone higher than the Green or the Red.
• 2 Green threads with 1 Red thread give a Grey, less
• Green, and of a higher tone than the two colours.

In repeating the same mixtures with higher tones of the same scales of Green and Red, the tone of the mixture of 2 Green with 1 Red is higher relatively to that of the colours mixed, than it is in the mixtures above.

1 Red thread and 1 Yellowish-green thread give a Carmelite-brown or an Orange-grey, the tone of which is equal to that of the colours mixed.

1 Red thread and 1 Bluish-green thread give a copper-coloured mixture or catechu-brown of a higher tone than that of the colours mixed.

Hence we may conclude that red and green threads, properly assorted, and in suitable proportions, yield Grey.

Orange and Blue.

3	Orange	threads with	1	Blue	thread give	a dull Orange.
2	Orange	”	1	”	”	a duller Orange.
1	Orange	”	1	”	”	Chocolate-grey.
3	Blue	”	1	Orange	”	Violet-grey.
2	Blue	”	1	Orange	”	Violet-grey.
		Redder than the preceding.

The results are the same with deeper tones than the preceding, except that the corresponding mixtures are browner.

3 orange threads with 3 blue threads present a remarkable phenomenon, according to the intensity of the light and the position from which it is observed. The tapestry being placed in a vertical plane before the incident light when the warp is horizontal, we perceive blue and orange stripes; but if the warp is vertical, we may then see the upper part of each blue stripe violet, and its under part, as well as the upper part of each orange stripe green, while the rest of each of the latter will appear red, bordered on the lower part with yellow. We may also see the upper part of each blue stripe, violet, and its under part, as well as the upper part of each orange stripe, green, and the rest of each of these stripes red, bordered on the lower part with green, and in the upper part with yellow. We say that they may be seen in this manner, because if the light were strong enough for distinct vision, we should not see the horizontal blue and orange stripes.

Yellow and Violet.

3	Yellow	threads with	1	Violet	give	a Greyish-yellow.
2	”	”	1	Violet	”	Yellow-grey.
1	”	”	1	Violet	”	Grey,
		much nearer normal grey than the preceding.
3	Violet	threads and	1	Yellow	give	a Greyish-violet.
2	”	”	1	Yellow	”	dull Violet,
		greyer than the preceding.

It is remarkable that in the mixture of a yellow with a violet thread, seen from a greater distance than that at which they appear neutralized, the yellow is so much weakened in proportion to the violet, that the mixture appears of a dull violet.

Yellow and blue afford an analogous result.

Rule III.— THE MIXTURE OF THE THREE PRIMARY COLOURS IN SUCH
PROPORTIONS THAT THEY DO NOT BECOME NEUTRALIZED,
BECAUSE ONE OR THE OTHER OF THEM IS IN EXCESS.

350. When Blue, Red, and Yellow are combined in such proportions that they do not neutralize each other, the result is a colour much greyer or more broken than if the proportion of complementary colour were more equal.

As Red mixed with a Greenish-yellow gives a Carmelite mixture, I shall add the following:—

ON THE PRINCIPLE OF CONTRAST IN CONNEXION
WITH THE PRODUCTION OF TAPESTRY.

351. The tapestry-weaver should therefore thoroughly understand the effects of contrast, to know the influence which the part of the copy he proposes to imitate receives from the colours surrounding it, and so judge what coloured threads to choose. The following examples will explain better than the most profound reasoning the necessity for the tapestry-worker to possess a knowledge of the law of contrast.

First Example.

352. If there be two coloured stripes in a picture, one red the other blue, touching each other, the phenomenon of contrast between two contiguous colours would have arisen, had not the painter sustained the red by blue, and the blue stripe by making it red or violet next to the red stripe (320).

353. Suppose a weaver wishing to imitate these two stripes, but ignorant of the law of contrast of colours, after choosing the wools or silks, he is sure to make two stripes, which will produce the phenomenon of contrast; for he will select his wools or silks of only one blue or one red, to imitate two stripes of different colours, each of which appears homogeneous throughout, but which the painter has only succeeded in making so by neutralizing the phenomenon of contrast, which would undoubtedly have resulted had each been painted of a uniform colour.

354. Suppose the painter has painted the stripes with uniform colours, then contrast will arise, so that the red contiguous to the blue will appear orange, and the blue contiguous to the red will appear greenish.

If the weaver be ignorant of the law of contrast, in attempting to imitate his model he will be sure to mix yellow or orange with his red, and yellow or green with his blue, in those parts of the stripes which come in contact. Hence the contrast will be more exaggerated than if he had worked the two stripes with homogeneous colours.

355. Suppose a weaver has to copy the series of ten grey stripes in flat tints (fig. 3, Plate 1) described (11), it is evident that if he is ignorant of such effects he will exaggerate it in the imitation; for instead of working ten tints of the same scale so as to produce ten bands in flat tints, he will make ten bands, each graduated conformably to what he sees; he will probably also have recourse to lighter and darker tones than those which correspond exactly to the model, and thus he will require a greater number of tones than would have been necessary had he been aware of contrast, rendering the copy an exaggeration of the model.

When we attentively observe the rosy flesh tints of many pictures, we perceive in the shadows a more or less apparent green tint, resulting from the contrast of rose with grey. (I presume that the painter has made his shadows without using green, and that he has not corrected the effect of contrast by using red.) Now a weaver, ignorant of the effect of rose upon grey, in imitating the shaded part will have recourse to a green-grey, which will exaggerate an effect that would have been produced naturally by employing a scale of pure, not of greenish grey.

QUALITIES WHICH PATTERNS FOR
GOBELINS TAPESTRY MUST POSSESS.

356. To determine the qualities which model pictures for tapestry must possess, we must decide what are specific qualities peculiar to this kind of imitation.

The weaver imitates objects with coloured threads of a certain diameter. These threads are applied round the threads of the warp. Their surface is not uniform but furrowed, those which are parallel to the threads of the warp being lower than those which are perpendicular to it; the effect of these furrows being such as would be produced upon a picture, by a series of dark parallel lines, cut at right angles by another series of finer parallel lines, less dark than the preceding.

357. There are these differences then between tapestry and painting:—