CHAPTER XXII.

TRACERY LIGHTS AND ROSE WINDOWS.

Glass in tracery lights and Rose windows cannot consistently be planned on the lines suitable to lancets or other upright shapes; and it is interesting to observe the modifications of design necessitated by its adaptation to circumstances so different. This applies not only to Gothic glass but to Renaissance, the best of which, as it happens, is in Gothic windows. Happily it never occurred to sixteenth century artists to hamper themselves by any affectation of archaism, and their work is deliberately in the new manner. One can understand, too, a certain "up-to-date" contempt on their part for the "old-fashioned" stonework; but it is rather surprising that so few of them seem to have realised how greatly their own work would have gained by a little more consideration of (if not for) the stonework.

Where, as at Gouda, by way of exception, Gothic windows were built to receive later glass, tracery is to all intents and purposes abandoned: the builders would have done away with mullions had they known how otherwise to support such huge glass pictures. It has been explained already, in reference to the influence of the window-shape, and especially of the mullions, upon glass design, how much more formidable these divisions appear upon paper than in the window. That is very plainly seen in many a window where the designer has relied upon them to frame his subjects. The pictures have a way of running together in the most perplexing way, and one has to pick them out for oneself again. The practical conclusion from that is, that the designer is under no obligation to confine himself too strictly within the separate lights of a large window. What he is bound to do is to take care that the mullions never hurt his picture; if they do, it is his picture which is to blame. He may urge with reason that the upright shafts of stone are there merely for the support of the window, and that it is not his business to emphasise them, enough if he acknowledge them. In tracery, however, it is his bounden duty to take much more heed of the stonework. It was designed, in intricate and often very beautiful lines, with deliberately ornamental intent; it was meant to be seen, and it is his function to show it off. The question he has to put to himself is now no longer: does the stonework hurt my design? but: does my design hurt the stonework? And he should not be satisfied unless it helps it. The artist who, at Bourges, having _fleur-de-lys_-shaped tracery to deal with, carried across it a design quite contrary to the lines of the stonework, was guilty of a blank absurdity.

The Early Rose windows, which were habitually filled with rich coloured glass, consisted either of simple piercings, as at Lincoln, or they were made up of piercings very definitely divided by massive stonework. In proportion as mullions become narrow, and form in themselves a design, it seems doubtful how far deep-coloured glass can do them justice. Only strong tracery lines will stand strong colour. At Châlons-sur-Marne, for example, the foils of certain cusped lights surrounding a central circular picture are successfully ornamented with arabesque of deep yellow upon paler yellow ground; and again at Or San Michele, Florence, certain gorgeous wheels of ruby and yellow, or of blue, green, and yellow, and so on, are unusually satisfactory. In such cases not only breadth of effect but definition of the tracery forms is gained by keeping them (more especially in their outer circumference) much of one tone, whilst contrast of colour between one light and another helps still further to assist definition. But this applies only to stonework strong enough to take care of itself. There is a sort of perverse brutality in putting into delicate and graceful tracery deep rich glass which hides its lines. Such lines want sharply defining against the light.

Early windows had, of course, no tracery properly so called. The great Rose windows, and the smaller Roses surmounting a pair of lancets, were rather piercings than tracery; and it was not difficult to adapt the design of a medallion window to suit them. A small piercing was ready designed for a medallion subject; nothing was wanted but a border round it, narrower, of course, than would have been used for a broad lancet light, but of the same foliated character. The individual quatrefoils or other principal openings, which went to make up a great Rose window, were filled in the same way. If the opening were wedge-shaped, as it often was, the obvious thing to do was to introduce into it a medallion (probably circular) of the full width of the opening, at about its widest, and to fill up the space about it with foliated ornament or geometric mosaic, with which also the smaller and less important piercings would naturally be filled. Sometimes the recurring figure medallions were set alternately in foliated ornament and geometric diaper; or the lights might be grouped in pairs, two with foliage and two with diaper. Similar alternation of the two common kinds of Early filling, naturally occurred in minor openings which contained no medallion. Something of this kind occurs at Reims.

When the shape of the great Rose permitted it--if, that is to say, the circular outline was strongly pronounced--it was possibly further acknowledged by a fairly broad border, following it and disappearing, as it were, behind the stonework; otherwise, except in the case of smaller medallion-shaped openings, it was not usual to mark them by even so much as a border line. Small Roses had sometimes, as at Auxerre, a central figure medallion round which were secondary foliage medallions set in diaper. A certain waywardness of design, already remarked in medallion windows, was sometimes shown by filling the central medallion with ornament and grouping the pictures round it.

As the lights of a Rose window radiated from the centre, features which recurred throughout the series arranged themselves inevitably in rings; and according to the disposition of the emphatic features of the design, the rays or the rings pronounced themselves. This is partly the affair of the architect who sets out the stonework, but it lies with the glazier whether he choose to subdue or to emphasise either feature. It is hard to say why one or other of these schemes of glass design, in rays or in rings, should be preferred; but, as a matter of experience, the sun and star patterns are not among the most happy. Perhaps the stone spokes of a wheel window assert themselves quite enough any way, and the eye wants leading, not vaguely away from the centre, but definitely round the window.

The circular belts of pattern formed by medallions or other features answer to, and fulfil the part of, the horizontal bands in upright windows (page 153), and bind the lights together. The band has it all its own way in a mere "bull's-eye," such as you find in Italy, where there are no radiating lines of masonry. It is strongly pronounced in some circular medallion windows at Assisi, in which an extraordinarily wide border (a quarter of their diameter in width) is divided into eight equal panels, each enclosed in its own series of border lines, within which is a medallion set in foliated ornament. This is fourteenth century work; but, as in thirteenth century Roses, the bars follow and accentuate the main divisions of the window.

Even when it came to the glazing of a Rose window in a later Gothic style, it is not uncommon to find a series or two of medallions running round the window, as occurs at Angers. They hold the design together; but in the nature of the case they are on too small a scale for the pictures to count for more than broken colour. Indeed you may see here the relative value in such a position of small figure subjects and bold ornament. The scrollwork is as effective as the medallions are insignificant. In fact, compared to them, the illegible medallion subjects in the lancet lights below are readable by him who runs. It has to be confessed that quite some of the most beautiful and impressive Rose windows are perfectly unintelligible, even with a good field-glass. This is so with the West Rose at Reims. In the centre it is ablaze with red and orange, towards the rim it shades off into deliciously cool greens and greenish-yellows. It may mean what it may; the colour is enough.

Room for figure work on an intelligible scale is only to be found by a device which verges on the ridiculous. In the beautiful North Rose at S. Ouen, Rouen, figures which should be upright are arranged in a circle like herrings in a barrel. Similar figures on a smaller scale occur in certain tracery lights at Lincoln, two of which are here given. Again in the North Rose at Le Mans there are twenty-four radiating figures. In fact, they were customarily so arranged, even down to the sixteenth century, a period at which one does not credit the designer with mediæval artlessness.

It is obvious that out of a series of twenty or more figures, radiating like the spokes of a wheel, only a very few can stand anything like upright. The designer of the South Rose at S. Ouen has endeavoured to get over the difficulty, as well as to accommodate his design to the exceeding narrowness of the lights as they approach their axis, by giving his personages no legs, and making them issue from a kind of sheath or bouquet-holder. A number of the figures pretending to stand in the radiating lights by a Rose or wheel window must be ridiculously placed. And then there occurs the question as to whether they shall all stand with their feet towards the hub. Where the figures have space to float, it is different. The angels in the Late Gothic Rose window at Angers, with swirling drapery which hides their feet, and makes them by so much the less obviously human, if not more actually angelic, solve the difficulty of full-length figures (on any appreciable scale) in the only possible way.

A portion of a simple and rather striking wheel window of the Decorated period, in which concentric bands of ornament form a conspicuous feature, is shown overleaf. In the small Rose from Assisi (page 278) the glazier has very successfully supplemented the design of the architect, completing the four circles, and accentuating them further by glazing the central spandrils in much darker colour than the rest of the glass, which is mainly white.

In the elaborate tracery of the Decorated or geometric period the mullions, as was said, ask to be pronounced. This was usually done in the Second Gothic period by framing each light with a border, separated from the stonework always by a fillet of white glass. The exception to this was in the case of trefoiled or other many-foiled openings, in which a central medallion or boss, usually circular, extended to the points of the cusps, and the border round the cuspings stopped short against the border to that. Or again in triangular openings a central boss would sometimes extend to its margin, and the borders would stop against that, or pass seemingly behind it.

A typical form of Decorated tracery occurs in the West window at York Minster, by far the most beautiful part of it. There, every important opening has within its white marginal line a broader band of ruby or green, broken at intervals by yellow spots, within which border is foliage of white and yellow on a green or ruby ground. Some of the smaller openings show white and yellow foliage only, without any coloured ground. A plan equally characteristic of the period is illustrated at Tewkesbury. There again occurs similar white foliage, its stem encircling a central spot of yellow. This also is on green and ruby backgrounds, the former reserved for the more prominent openings; but the border is in white, painted with a pattern. This broader white border more effectively relieves the dark lines of the masonry than the border of colour, which sometimes confuses the shapes of the smaller tracery openings: it does so, for example, in the Late glass on page 200.

For what was said of the difficulty of carrying a broad border round the heads of Decorated lights applies more forcibly still to tracery. The merest fillet of colour is often as much as can safely be carried round the opening, if even that. On the other hand, a broad border of white and stain, even though it contain a fair amount of black in it, may safely be used--as at Châlons, where it frames small subjects in rich colour. Some admirable Decorated tracery occurs at Wells, much on the usual lines, and containing a good deal of pleasant green; but there the white and yellow foliage in the centre part of the lights is sometimes so closely designed that very little of the coloured ground shows through it, and it looks at first as if what little ground there is had all been painted-out. At S. Denis Walmgate, York, the background to the foliage in white and yellow (which last predominates) is painted solid: the only pot-metal colour (except in the central medallion head) is in a rosette or two of colour leaded into it; the border is white. Another expedient there employed is to introduce figures in white and stain upon a ground of green or ruby, diapered. At Wells there occur little figures of saints in pot-metal colour, planted upon the white foliated filling of the tracery lights. Decorated circular medallions occupying the centre of ornamental tracery lights are usually framed in coloured lines; occasionally the inner margin of the medallion is cusped, in imitation of stonework.

An effective plan, adopted at Evreux, is to gather the lights into groups, by means of the colour introduced into them, which grouping may or may not be indicated by the stonework. In any case, it is a means of obtaining at once variety and breadth of colour.

Perpendicular tracery lights are themselves, in most cases, only copies in miniature of the larger lights below, and the glass is designed on the same plan. A good illustration of this is at Great Malvern, where the design consists of the orthodox canopy work in white and stain, with little figures also nearly all in white, colour occurring only in the lower skirts of their drapery, in the background about their heads, and behind the pinnacles above. The effect is beautifully silvery. Often such figures under the canopies are angels, all in white and stain. Sometimes seraphim, in stain upon a white ground, quarried perhaps, fill the lights, without canopies. These are all typical ways of filling the tracery of a Perpendicular window.

It was quite a common thing to fill it with glass wholly of white and stain. In the centre there might be a medallion head in grisaille, or an inscribed label, the rest of the space being occupied by conventional foliage having just a line of clear white next the stonework. Beautiful examples of this treatment occur at Great Malvern; occasionally the foliage is all in yellow with white flowers. Small openings are thus often glazed in a single piece of glass, or in any case with the fewest possible leads. At S. Serge, Angers, there is larger work of a similar kind, a bold scroll in white and stain on a ground of solid pigment, out of which is scratched a smaller pattern, not so bold as in the least to interfere with the scroll, but enough to prevent anything like heaviness in the painted ground. Similar treatment is adopted in the cathedral at Beauvais. Once in a while one comes, in English work, upon figures in white and stain on a solid black ground extending to the stonework, without any line of white to show where the glass ends and the stonework begins. It would be impossible more emphatically than that to show one's contempt for the architecture.

Some disregard, if not actually contempt, is shown for architecture in the practice, common no less in Late Gothic than in Renaissance design, of carrying a coloured ground right up to the stone, without so much as a line of light to separate the two. Comparatively light though the colour may be, it is usually dark enough, unless it be yellow, to confuse the forms of any but the boldest tracery. Something of the kind occurred by way of exception even in fourteenth century glass, as at S. Radegonde, Poitiers, and at Toulouse, where the tracery of the windows is one field of blue, irregularly sprinkled with white stars. The lines of the tracery are lost, and one sees only spots of white.

The Later Gothic plan was to keep tracery light, even though the window below it were altogether in rich colour, and the effect was good; as at Alençon, where a distinctly blue window has in the tracery only angels in white and yellow on a white ground; or, again, at Conches, where white-robed angels, on a ground of rich stain, contrast pleasantly with the cool blue of the lights below.

Unusual treatment of the tracery occurs at Auch (1513). In the main the tracery lights contain figures in colour upon a ruby or paler-coloured ground, which, as in so many a Renaissance window, runs out to the stonework; but occasionally here and there a light is distinguished by a border of white. Moreover, the ground is, as a rule, not of one colour throughout, nor even throughout a single light, but varied; and that not symmetrically or pattern-wise, but so as artfully to carry the colour through. In fact, the artist has taken his tracery much more seriously than usual, and has carefully studied how best he could balance by the colour in it the not quite so easily-to-be-controlled colour of his figure composition below. The result is that the windows are all of one piece--each a complete and well-considered colour composition: the tracery is not merely the top part of the frame to the coloured picture below.

In Renaissance glass the tracery was more often in comparatively full colour, even though the lights below were pale. A grisaille window at Evreux, with practically blue tracery, has a very pleasant effect.

It was not often that the Renaissance glass painters gave very serious attention to the tracery which they had to fill. They were, for the most part, content to conceive each separate opening as a blue field upon which to place an angel (as above), a crown, a _fleur-de-lys_, or other emblem, as best might fit. In very many sixteenth century windows the design consists merely of angels, emblems, labels, or even clouds, dotted about, as suited the convenience of the designer. Sometimes, as at S. Alpin, at Troyes, there occurs in a tracery light a tablet bearing a date,--presumably, but not always positively, that of the window. Such devices were very often in white upon a ground of blue, purple, or ruby. Angels of course adapted themselves to irregular shapes in the most angelic way; and they are introduced in every conceivable attitude--standing, kneeling, flying, swinging censers, singing, playing on musical instruments, bearing scrolls or shields; angels all in white, angels in white with coloured wings, angels in gorgeous array of colour: and more accommodating, still, is the bodiless cherub, beloved of Luca della Robbia.

There is a quite charming effect of colour in a Jesse window at S. Maclou, Rouen, where the tracery lights are inhabited by little cherubs, in ruby on a grey-blue ground, in grey on deeper grey-blue, and in emerald-like green upon the same.

The scroll without the angel was a very convenient filling for smaller openings. Some elaborately twisted scrolls, in white and stain on purple, occur at Moulins.

Larger and more prominent lights often contain a separate picture, or one picture runs through several lights, or perhaps all through the tracery. Worse than that is, where the picture runs through from the lights below; as at Alençon, where the trees grow up into the blue of the tracery, broken otherwise only by white clouds; or at Conches, where the architecture from the subject below aspires so high. It is almost worse still where, as at Alençon again, and at the chapel at Vincennes, it is the canopy which so encroaches. In the exceptional case of a Jesse window there seems less objection to accepting the whole window as a field through which the tree may grow; yet the tracery is not the happiest part of the Beauvais window (page 368). Sometimes the heads of the lower lights are made to appear as though they were part of the tracery.

A happier form of Renaissance tracery design is where medallion heads in white and stain are introduced upon a ground of plain colour--blue at Châlons, purple-brown at Montmorency. These are sometimes most beautifully painted, as are the Raffaellesque little cherubs amidst white clouds, also at Montmorency; but they are much more delicately done than they need have been, and less effective than they might. Very delicate painting upon white does produce an effect even at a distance; at least it gives quality; but there should be some relation between effort and effect; and here the effect is weak as compared with the expenditure of art. In the tracery on page 213, fairly effective though monotonous, the birds are glazed in with such unnecessary avoidance of lead, that the cutting of the ground must have been a work of great difficulty. In glass of every period it has been the custom to put too much into tracery; in Early work too much detail, in Later too much finish. What is wanted is breadth.