It formed part of the canopy scheme, that the structure should end before it reached the top of the window, so that you could see beyond it into s.p.a.ce. The designers would have been only too happy if they could have done away with the gla.s.s above that. If they had had big sheets of plate gla.s.s, they would certainly have used them to produce the effect of out of doors--there was already a _plein air_ school in the eighteenth century--as they had not, they were obliged to accept the inevitable, and lead up their white gla.s.s; but they went as far as they could to doing away with its effect, using thin, transparent material, which was not meant to appear as though it formed part of the composition. Occasionally they would use pale blue gla.s.s, or tint it in a blue enamel, further to suggest the sky beyond. This (page 222) would commonly be glazed in squares. The pure white gla.s.s also was often glazed in square or, as at Brussels, diamond quarries (page 71).

Subjects themselves, it has been explained, came to be glazed as much as possible in rectangular panes; but it marks, it may here be mentioned, a decline of design, as well as of technique, when these came to interfere in any way, as they did, with the drawing. Having made up his mind that his design is to be glazed in rigid square lines, the artist should logically have designed accordingly. He had only to mark off the glazing lines on his cartoon, and scheme his composition so that it was not hurt by them. Towards the seventeenth century the plain gla.s.s, the extra part beyond the canopy or beyond the picture, would often be glazed in some simple pattern. That, you might imagine, stood for the window _behind_ the picture or the monument. At the church of S.

Jacques, Antwerp, above a picture of the Circ.u.mcision, is a canopy leaded in squares and painted to look like falsehood, beyond which clear gla.s.s is glazed in a pattern.

Occasionally an attempt is made to merge the picture into the plain glazing above, as at S. Paul's, Antwerp, where the yellow sky, against which is shown the distant city, and so on, is glazed in squares, which further off become gradually white, and then at their interstices have smaller diamond-shaped pieces of gla.s.s let in.

Where a subject glazed in quarries is represented against a background of plain glazing of more elaborate design, there is difficulty in joining the two, except by means of a strong lead outline to the figures, or whatever may come next to the plain gla.s.s, which outline the seventeenth century designer was anxious before all things to avoid.

Accordingly, as the plain pattern work approached the margin of the painted work, he replaced the leads by paint, which sham leads, of course, could be made to disappear as seemed good to him. But these little games of his, to judge by results, were hardly worth the candle.

It will be seen how, in the French gla.s.s on page 200, the canopy came to be backed and surrounded by unpainted gla.s.s, quarries in that case.

There the canopy sufficiently occupies the window s.p.a.ce not to strike one unpleasantly; but that is sixteenth century work; later, and especially in Flanders, canopies are represented, as in the cathedral at Antwerp (1615), adrift, as it were, in a sea of plain glazing. Even when the gla.s.s has some quality of gla.s.s the effect of that is not happy.

When the gla.s.s is thin and transparent it is disastrous.

At S. Jacques, Antwerp, again, coats of arms hover unsupported in mid-air, the mere lines of the glazing being quite inadequate to their apparent support. It is different, of course, where the heraldic device, as opposite, is itself little more than plain glazing. That is a very mild form of art; but, in its way, it is satisfactory enough.

Perhaps least fortunate of all in effect are the landscapes at S.

Jacques, which float, without even a canopy to frame them, in an atmosphere of leaded gla.s.s. Antwerp is rich in gla.s.s, much of it very cleverly executed, which would serve very well to ill.u.s.trate how _not_ to design a window.

[Ill.u.s.tration: 165. GOUDA, 1688.]

The place of the canopy was supplied sometimes, especially in later Netherlandish work, by the cartouche so dear to the Dutch. It fulfilled very much the office of the canopy, framing the design; and, had it been kept white, it would have framed it well, affording circular and other shapes which form a welcome variation upon the usual arched opening. But it was not white at all; very much the reverse. Indeed the idea of the Dutch cartouche, with its interpenetration of parts, and curling and projecting straps and bolts, tempts the painter to a heavy method of painting, destructive of the very quality of white. The device depends for its effect far too much upon force of shadow to be of any great use in white gla.s.s. The comparatively early cartouche in the lower half of the window at Gouda, given on page 223, is of the simplest kind, and has none of that too-seductive bolt work; but it is dull and heavy in effect, being painted in heavy brown, with the idea of giving atmospheric effect to the picture supposed to be seen through it.

[Ill.u.s.tration: 166. PLAIN GLAZING, S. GERVAIS, PARIS.]

A great cartouche is often used as a kind of base to a canopy extending across the whole width of a wide window, or the base of the canopy may include a very important cartouche, occupied in either case by a long inscription. Here again the oblong patch of white or yellow may have value, in proportion as it is allowed to preserve the quality of gla.s.s.

There is, however, something poor and mean about large areas of small lettering, and it is a pity to see the opportunity which bold inscriptions give quite thrown away. Moreover, the inscriptions are invariably too long. The framers of inscriptions do not realise the mult.i.tude of readers they scare away by the volume of their wording. The design of a window at S. Jacques, Antwerp, consists merely of an inscription label, with a helmet above and mantling in black and white (the black, of course, paint) set in plain glazing.

Up to the very last whole windows were glazed very often in plain patterns, usually all in clear white gla.s.s. A couple of designs, into which a little colour is introduced, are given below and opposite. In spite of the increased facility for cutting gla.s.s, afforded from the beginning of the century by the use of the diamond, patterns were seldom very elaborate; but, by way of ill.u.s.trating what can be done by means of the diamond, there is shown overleaf quite a conjuring feat of glazing.

The thick black lines in the drawing represent the leads; the white s.p.a.ces enclosed are plain white gla.s.s, rather poor in quality; the thinner lines stand for cracks, possibly not, or not all of them, of the glazier's doing, for it would be almost impossible to handle such work without breaking it. It is well-nigh incredible that each of these _fleurs-de-lys_ should have been cut out of a single piece of gla.s.s, the marginal band to it out of a second, and so with the background s.p.a.ces.

Glaziers may be inclined to question the possibility of such a _tour de force_, even in poor thin gla.s.s. Certainly one would not have thought it possible; but there it is, in the museum at Angers, close to the eye, where you can see and examine it. This is glazing with a vengeance. It is not the sort of thing that any one would undertake, except as a trial piece, to show his skill; but if ever a glazier deserved his diploma of mastership here is the man.

[Ill.u.s.tration: 167. PLAIN GLAZING, LISIEUX.]

[Ill.u.s.tration: 168. A TOUR DE FORCE IN GLAZING, ANGERS MUSEUM.]

The composition of some of the windows belonging to the first half of the seventeenth century at Troyes does not follow the general tendency of the period. The better part of this, if not the greater, is attributed to Linard Gontier (1606-1648). But the design of these windows, and the style of them, is so varied, and sometimes so little of the period, that one is disposed to think, either that he was a painter only and did not design them at all, or that he borrowed his designs freely from Italian and other sources. The panel on page 400, the Virgin girt with clouds and cherubs, distinctly recalls the work of the Della Robbia School; and again the figures opposite remind one of late sixteenth century paintings. An unusual thing, however, about some of these windows is the way they are set out. The disposition of the design of the three-light window from S. Martin es Vignes is as simple and severe as though it had been Gothic. The glazing, too, is not in squares, but follows the design. Except for the rather robustious drawing of the figures, and the futile kind of detail which does duty for canopy work, the gla.s.s might belong to the first half of the sixteenth century.

[Ill.u.s.tration: 169. THREE LIGHTS, S. MARTIN-eS-VIGNES, TROYES.]

Again, in the subject of the marriage of SS. Joachim and Anna on page 234, it is rather by the types of feature and the cast of draperies, than by the composition, that the date of the work proclaims itself. It is proclaimed, of course, unmistakably by the use of enamel, not only in the warm-coloured flesh, but throughout, to support, and sometimes to supply the place of, pot-metal gla.s.s. Nevertheless, the effect of much of this gla.s.s is brilliant to a degree almost unprecedented in the first half of the seventeenth century. The painter had skill enough to get the maximum of modelling with the minimum of paint. He could afford, therefore, to use paint sparingly, leaving plenty of gla.s.s clear, and seldom sacrificing its translucency, as was done in the group of donors on page 81, whose black mantles are rendered in solid paint. Those heavily painted figures recall a couple of Donors in a window at Antwerp (1626), equally black robed, against a nearly black screen, all in paint: they would have made a capital votive picture; but they are about as unlike gla.s.s as anything one can conceive.

Exceptionally good seventeenth century work is to be found also at Auch.

It seems that it was proposed (towards 1650) to complete the windows in a way worthy of the splendid beginning in the choir; but the art was not forthcoming; and the Chapter of that day was wise enough to fall back upon comparatively unimportant quarry windows, with borders and tracery in white and stain and blue enamel, which is at least brilliant in colour, and pleasing in effect. That may be said also of the Western Rose. In the Roses of the transepts, the artist goes further, and produces, by means of arabesque in white and stain, upon a ground mainly of blue and ruby, occasionally varied by green, each light defined by a simple border of white and stain, a couple of flamboyant Rose windows with gla.s.s which would do credit to the period of the stonework. They might well (at the distance they are placed from the eye) be taken at first sight for Early Renaissance work. In fact they are really mosaic gla.s.s--so rare a thing by this time that the windows are probably of their kind unique.

Even at its best enamelled gla.s.s is less effective than the earlier work. In proportion as the place of pot-metal is supplied by enamel, the colour is inevitably diluted, and at times it is quite thin. Indeed, it is pretty well proved, by the work of men who are masters in their way, that, in painted as distinguished from mosaic gla.s.s, the choice lies between weak colour and opacity. At Auch and at Troyes we have weak but still often pure and brilliant colour.

The opposite defect of opacity reaches probably its greatest depth in the four great Rubens-like windows at S. Gudule in the chapel of Our Lady immediately opposite that of the Holy Sacrament, where Van Orley's windows are. The design is there absolutely regardless of any consideration of gla.s.s or architecture. Each window is treated as a vast oil picture, without so much as a frame. Here is no vista of distant architecture, nor any such relief of lighter colour as you find at Gouda. Force of colour is sought by ma.s.ses of deep shadow, into which the figures merge. This shadow being obtained by paint, and the glazing being in the now usual squares, there are literally yards of painted quarries, which, except when the sun is at its fiercest, are all but black. And withal the effect is not rich as compared with even the common Gothic gla.s.s, though it is not without a certain picturesqueness when perchance the sun struggles through. A painter might find it an admirable background to his picture; no architect would choose it for his building. Three of these windows were designed, it seems, by a pupil of Rubens, Van Thulden, who worked under him at the Luxembourg, and they have all the character of his work--except that the colour is dull.

At New College, Oxford, are some smaller windows with figures, also recalling the manner of the master, and said to be by pupils of his.

They, too, are dull and heavy in effect. The canopies over the figures are terrible caricatures of the Gothic shrines in the ante-chapel.

Better seventeenth century gla.s.s is to be found at Oxford in the work of the Van Lingen, a family of Dutchmen settled in England, who executed windows in Wadham and Balliol Colleges and elsewhere. Some of these are rich in colour. Apart from the rather interesting use of enamel made in them, they are not of great value; but they show as well as more important examples the kind of thing which did duty for design.

The windows in Lincoln's Inn Chapel, London, ill.u.s.trate not unfairly the dreary level of dulness as to colour and design to which seventeenth century gla.s.s declined. That it could fall still lower was shown, for example, by Peckitt, of York, who is responsible for the gla.s.s on the north side of New College Chapel, Oxford, facing the work of the Dutchmen. These date from 1765 to 1774.

[Ill.u.s.tration: 170. ST. MARTIN eS VIGNES, TROYES.]

The history of eighteenth century windows may, if one may plagiarise a famous bull, be put into the fewest possible words: there were none--worth looking at. To find pleasure even in Sir Joshua's design at New College, you must consider it as anything but gla.s.s.

CHAPTER XIX.

PICTURE-WINDOWS.

The course of gla.s.s design was picture-ward. Picture design, however, did not stand still, and hence arises some confusion in the use of the word "pictorial." It is time to try and clear that up. Stained gla.s.s, it may be truly said, has been from the very first pictorial. The earliest gla.s.s, therefore, and the latest, the best and the worst, may alike be termed pictorial. The difference is in the conception as to what const.i.tuted a picture, say, in the thirteenth century and the seventeenth. It all depends upon the kind of picture attempted.

Archaic art aims already at nature. We probably do not give the early painter credit enough for his intention of rendering natural things naturally. In part at least the stiffness of his design comes from lack of skill, and often where we find him quaint he meant no doubt to be perfectly serious and matter-of-fact. But it was not alone incompetence that held his hand. He was restrained always by a decorative purpose in his work. Here again he was not conscious of sacrificing to any higher rule of art; he bothered himself as little about that as a bee about the way it shall fashion its cell; he worked in the way to which he was born; but the idea had not yet developed itself that a picture could be painted quite apart from the decoration of something, and it never entered his mind to do anything but adapt himself to the decorative situation.

A picture, then, in mediaeval times was a work of decorative art, designed to fit a place, to fulfil part of a scheme of decoration, in which it might more often than not take the first place, but no more; it had no claim to independence.

In gla.s.s the picture obeyed two conditions which more or less pulled together: as art it subserved to decorative and architectural effect; as craftsmanship it acknowledged and accepted the limitations of gla.s.s painting. In the course of years the ideal of architectural fitness underwent successive changes, and the limitations of the gla.s.s painter grew less; his scope, that is to say, was widened, and his art took what we call more pictorial shape. Still, so long as the pictorial ideal itself was restrained within the limits of mediaeval ambition, gla.s.s painting might safely approach the pictorial. It was not until painting broke loose from traditional decorative trammels and set up, so to speak, on its own account, until pictorial came to mean something widely different from decorative, that the term became in any way distinctive of one kind of art or another. It is in that later sense that the word pictorial is here used.

Artists still differ, and will continue to differ, as to the precise use of the term. There are artists still who contend that, since in old time art was decorative, and since in their opinion all art should be decorative, therefore the picture which is not decorative is not art.

Arguing thus in a circle, they would say (the pictorial including in their estimation the perfection of decorative fitness, and all art which overshot the mark ceasing to count with them) that art was always at its best when it was most pictorial. But that is a species of quibbling about words which not only leads us no further, but hinders mutual understanding. It is wiser to accept words in the sense in which they are generally understood, and to try and see where the real difference of opinion is.

Difficult it may be, impossible even, to draw the line between a picture which is decorative and decoration which is pictorial; but there is no difficulty in drawing a band on one side of which is decoration and on the other picture. You have only to draw it wide enough. If we can succeed in defining a picture as distinguished from a work of decorative art, and can then show how a stained gla.s.s window, in attempting to conform to conditions which we have agreed to call pictorial, fails of its decorative function, it will then not be so difficult to see how, in proportion as gla.s.s aims at the pictorial, it falls short of making good windows. Granted, then, that a picture may fulfil all decorative conditions, and that a decoration may sometimes rightly be pictorial, that the two go, as historically they did, a long way hand in hand, it is contended that there is a point at which decoration and picture part company and take distinctly different ways; thenceforth, if either is led away by the other, it is at the cost of possible success in the direction more peculiarly its own.

Now, the first point at which picture definitely parts company with decoration is where the painter begins to consider his work apart from its surroundings. The problems the artist may set himself to solve are two. "How shall I adorn this church, this clerestory, this chancel, this window, with stained gla.s.s?"--that is distinctly a problem of the decorator; "How shall I realise, on canvas or what not, this thought of mine, this fact in nature, this effect seen or imagined?"--that is distinctly a problem of the painter. Each, it is granted, may be swayed more or less by the other consideration also, but according as a man starts with the one problem or the other, and seeks primarily to solve that, he is painter or decorator. Suppose him seriously to endeavour to combine pictorial and decorative qualities in his work, there will come times when he has perforce to choose between the two. Upon the choice he makes will hang the final character of his work, decorative or pictorial.

We are too much in the habit of laying down laws as to what a man may or may not do in art. He may do what he can. He may introduce as much decorative intention into his picture, as much pictorial effect into his decoration, as it will stand; it is not till he overweights one with the other, attempts more than his means or his power allow him, and fails to do the thing that was to be done, that we can say he has gone wrong.

When the two ideals of decoration and painting were more nearly one, and in proportion as that was so, success in the two directions was possible; when painting aimed at effects, of painting--in proportion, that is, as it became pictorial--it was impossible. It is safe to say, since masters attempted it and failed--since, for example, the finest work in gla.s.s which aims at the pictorial and depends upon painting ends always in being either thin or opaque in effect--that the happy medium was not found. The fact is, the time came when a painter, in order to design successfully for gla.s.s, was called upon to relinquish some of the effects he had come greatly to value in painting: effects of light and shade, atmosphere, reflected light, relief, perspective, violent foreshortening. To seek these at the expense of qualities proper to decoration and to gla.s.s, was to attempt picture; to sacrifice such pictorial qualities to considerations of architectural fitness, to the quality of the gla.s.s, its translucency, its colour, its consistent treatment, was to attempt decoration; and in proportion as the sacrifice is not made, the work of the gla.s.s painter may be characterised as "pictorial." There should now be no possible misunderstanding as to what is meant by the word. It implies something of reproach, but only as applied to gla.s.s. Let the pictorial flourish, in its place--that is, in picture. All it is here meant to a.s.sert is that, pictures being what they are, what they were already by the end of the sixteenth century, the pictorial element in stained gla.s.s is bound to spoil the window.

There are two respects in which a stained gla.s.s window differs from a picture: first, in that it is a window; second, in that it is gla.s.s.

Suppose we take these two points separately. It scarcely needs showing that the designing of a window is a very different thing from the painting of a picture. In the first place, the architectural frame of the window is there, arbitrarily fixed, whereas the painter chooses his frame to suit his picture. The designer of a window has not only to accept the window-shape, but to respect both it and the architecture of the building. The scale of his work, the main lines of its composition, if not more, are practically determined for him by architectural considerations, just as the depth of colour in his scheme is determined by the position of his window and the amount of light he desires "or is allowed" to shut out. Moreover, he has to accept the window plane, to acknowledge it as part of the building, to let you feel, whatever he does, that it is a window you see, and not something through the window or standing in it. That was tried, as we have seen, at Gouda and S.

Gudule; but, even if the illusion had been achieved, it would have been destructive of architectural effect. The idea of a picture seen through the mullions of a window is one of the will-o'-the-wisps which led gla.s.s painters astray. They did not succeed; and, had they done so, they would have given a very false, and to some of us a very uncomfortable, impression of not being protected from the outer air.

Mullions are in any case a very serious consideration. It has been shown already (page 197) how the artist sought continuity of subject through the lights of his window, and gradually extended his picture across them. And if he is at liberty to occupy a four-light window with the Virgin and Child and the Three Kings, and if it is lawful to introduce more than one figure into a light, why may not each king be accompanied by an attendant, holding his horse or bearing gifts; why should not the Kings kneel in adoration; why should not Joseph be there, the manger, and the cattle; why should there not be one landscape stretching behind the Magi, binding the whole into one picture? So with the Crucifixion.

If the Virgin and S. John may occupy sidelights, why not introduce as well in a larger window the two thieves, the Magdalene at the foot of the cross, the good centurion, the soldiers, the crowd? Obviously there is no reason why the subject should not be carried across a window; and from the time that windows were divided into lights it was done, at all events in the case of certain subjects, such as the Tree of Jesse, which spread throughout the window, or the Last Judgment, for which the available s.p.a.ce was yet never enough.

But there is a wide difference between designing a subject which extends through the whole width of a window and designing it so that it appears to be seen through the window. In the one case the mullions are seriously taken into account; in the other they are ignored. If you were looking at a scene through a window, of course the mullions would interfere. Why, therefore, consider them if you wish to produce the effect of something seen through? Naturally you would not allow the stonework to cut across the face of a princ.i.p.al personage, or anything of that kind; but, apart from that, its intervention would only add to the air of reality. The problem of dealing with the mullions is thus rather shirked than solved. Its solution is not really so difficult as would seem. Mullions count for much less in the window than one would suppose. The eye, for example, follows naturally the branches of a Tree of Jesse from one light into another, and it is not felt that the stonework interferes with it at all seriously, whilst the scheming of the figures, each within a single light, is a very distinct acknowledgment of its individuality. So in the case of a subject. If the design is so planned that the important figures are grouped in separate lights, the landscape or other continuous background helps to hold the picture together, and is not hurt by the mullions.

The important thing is that mullions should be considered; only on that condition do they cease to interfere with the design. There is no reason always to put a border round each light, or even to keep every figure within the bounds of a single light. A reclining figure, such as that of Jesse at the base of the window (below), Jacob asleep and dreaming, or the widow's son upon the bier, may safely cross two or three lights, if it be designed with reference to the intervening stonework.

Further, it seems desirable that the shape of each separate window opening should be acknowledged by at least a narrow fillet of white or pale colour next the masonry, broken, it may be, here and there by some feature designed to hold the lights together, but practically clearing the colour from the stonework, and giving to the division of the window the slight emphasis it deserves. It is not worth while dividing a window into lights and then effacing the divisions in the gla.s.s. Given a window of four or five lights, the decorator has no choice but to design a four or five-light window. He must render his subject so that the constructional divisions of the window keep their proper architectural place; if his subject will not allow that, he must abandon his subject, or give very good reason why not. The reason of mere pictorial ambition will not hold good. The test of a good picture-window is, how the mullions affect the design. If to take them away would make it look foolish, then it has probably been designed as a window, decoratively; if to take them away would improve it, then it has been designed pictorially; and, however good a picture it might have been, it is a bad window design.