effect to those windows composed of pieces of varied colours suitably contrasted. Yet I must not omit to instance the bad effect which results from the mixture of coloured glass with transparent colourless glass, at least when the latter has a certain extent of surface in a window; but at the same time I recognise the effect obtainable by mixing ground glass with coloured glass, and also of small pieces of colourless transparent glass framed in lead, so that at the distance at which they must be viewed, they produce the effect of a symmetrical juxtaposition of white parts with black parts. (435.) I conclude we must refer the causes of the beautiful effects of coloured church-windows 1°. To their presenting a very simple design, the dif ferent well-defined parts of which may be seen without confusion at a great distance. 2°. To their offering a union of coloured parts distributed with a kind of symmetry, which are at the same time vividly contrasted, not only among themselves, but also with the opaque parts which circumscribe them. (436.) Coloured windows appear to me to produce all the effect of which they are really capable only in a vast edifice where the differently coloured rays arrive at the eye of the spectator placed on the floor of the church, so scattered by the effect of the conical figure of the rays of light emanating from a single point, that they impinge upon each other, whence results an harmonious mixture, which is not found in a small structure lighted by stained windows. It is this intimate mixture of the coloured rays, transmitted in a vast edifice, which permits of tapestries being placed on the ground floor when the lower walls have no colourless glass windows; it is evident that if tapestries are placed too near stained windows, they must lose all the harmony of their colours, as when blue rays fall upon red draperies, yellow rays upon blue draperies, &c. Thus, when we have to put coloured glass in windows, it appears to me necessary to take into consideration not only their beauty, but also the effect which the coloured lights they transmit will have upon the objects they illuminate. (437.) The coloured windows of a large church appear to me really as transparent tapestries, intended to transmit light, and to ally themselves harmoniously with the sculptures on the exterior, which destroy the monotony of the high walls of the edifice, and with the different ornaments of the interior, among which tapestries must be taken into account. (438.) I shall sum up my ideas on the employment of stained-glass for windows in the following terms: 1o. They produce all the effect of which they are really susceptible, only in rose-windows, bay-windows, or the pointed windows of large gothic churches. 2o. They produce all their effect only when they present the strongest harmonies of contrast, not of colourless transparent glass with the black produced by the opacity of the walls, iron bars, and strips of lead, but of this Black with the intense tones of Red, Blue, Orange, Violet, and Yellow; 3°. If they represent designs, these must always be as simple as possible, and admit of the harmonies of contrast; 4°. While admiring windows a large number of which consist of paintings upon glass of undoubted merit, especially in examining the difficulties overcome, I maintain that it is a kind which should not be much encouraged; because the product has never the merit of a picture properly so called, as it is more costly, and will produce less effect in a large church than a stained window of much lower price. 5°. Windows of a pale grey ground with light arabesques have a very poor effect wherever they are placed. I shall recur to the employment of stained windows in churches when I treat of the relations of the law of contrast with the decoration of the interiors of churches. SECTION I CALICO-PRINTING. (439.) THE object I have in view in this Chapter, is the examination of the optical effects produced by patterns when printed upon woven fabrics, but not the chemical effects which arise between the pigments and the stuffs upon which they are printed. For a considerable period of time printing on textile fabrics was limited, so to speak, to cotton cloths: only of late years has it been extended to fabrics of silk and wool for furniture and clothing. This branch of industry has undergone an immense extension, fashion having accepted these products with much favour; but whatever may be the importance of the subject in a commercial point of view, I must treat it briefly, because this book is not directed exclusively to that branch of inquiry, and, moreover, all the preceding part is intimately connected with it, that to go deeply into details would expose us to the inconvenience of repetition with no compensating advantage. I shall content myself with stating many facts which serve to show that, for want of knowing the law of contrast, the cotton manufacturer and the printers of woollen and silk stuffs are constantly exposed to error in judging the value of recipes for colouring compositions, or rather to mistake the true tint of the designs which they have themselves applied upon grounds of different colours. A. FALSE JUDGMENT OF THE VALUE OF RECIPES FOR COLOURING COMPOSITIONS. (440.) At a calico-printer's they possessed a recipe for printing green, which up to a certain period had always succeeded, when they fancied it began to give bad results. They were lost in conjecture upon the cause, when a person, who at the Gobelins had followed my researches on contrast, recognised that the green of which they complained, being printed upon a ground of blue, tended to become yellow, through the influence of orange, the complementary of the ground. Consequently, he advised that the proportion of blue in the colouring composition should be increased, in order to correct the effect of contrast. The recipe modified after this suggestion gave the beautiful green which they had before obtained. (441.) This example demonstrates that every recipe for colouring compositions intended to be applied upon a ground of another colour must be modified conformably to the effect which the ground will produce upon the colour of the composition. It proves also that it is much easier for a painter to correct an effect of contrast than it is for a calico-printer, supposing that both are ignorant of the law of contrast: for if the first perceives in painting a green pattern on a blue drapery that the green comes too yellow, it is sufficient for him to add a little blue to the green, to correct the defect which strikes him. It is this great facility in correcting the ill effect of certain contrasts which explains why they so often succeed in so doing without being able to account for it. B. TRUE TINTS OF DESIGNS PRINTED UPON COLOURED GROUNDS MISUNDERSTOOD. (442.) In treating of the modifications perceptible in bodies through the medium of light, I have instanced cottons of a coloured ground printed with patterns which the calicoprinter intended making colourless; but, owing to the imperfection of the process, were really of the colour of the ground, but of an exceedingly light tint (292., 293.); we may be satisfied of this by looking at them after they are isolated from the ground by means of a white paper cut out like the pattern. I have remarked that, notwithstanding their colour, the eye judges them to be colourless, or of the tint complementary to that of the ground. (443.) I will now explain the cause of these appearances, because they have been the subject of questions frequently addressed to me by manufacturers of printed stuffs and by |