Sir George Gilbert Scott

In the foregoing Lectures, having only brought the history of our Architecture down to the close of the thirteenth century, I have neglected that of the later styles, and, consequently in great measure, the development and progressive changes in window-tracery. This has, however, been so amply treated of in many books and essays that it is not a matter with me of much regret. I confess I had intended to have supplied the omission in subsequent lectures, but circumstances prevented.

It would have been an agreeable task to have followed up the history of window-tracery and the many details which accompanied it, through the remaining two and a half centuries of the reign of Gothic architecture—to have shown how it grew from the purely geometrical system of Westminster, Newstead, and the “Angel choir” at Lincoln into the sweeter tracery of the “Easter aisle” at St. Albans, and of St. Etheldreda’s Chapel in Holborn; and on again into the yet softer loveliness of the Lady Chapel at Chichester, the halls at Penshurst, Mayfield, the gatehouses of Battle Abbey and of St. Augustine’s at Canterbury, and the Chapel of St. Anselm and De Estria’s work at the cathedral there; and then again into the more flowing tracery of Alan de Walsingham’s work, till it fell into debility by its too sensuous ramifications, and was brought back again to vigour by the stern perpendicular work of Wykeham; and how that, in its turn, became softened down, into such works as Crosby and Eltham Halls, and again into the exuberance of the Tudor style. All this would be very pleasant, but would necessitate the treating of all contemporary variations of detail, and would swell my lectures out into another volume: more than this, I have given no such lectures. It has been my task to show the principles on which Gothic architecture was founded, and on which it attained its leading developments, rather than to follow them out to their ultimate results, on attaining which much which led to them was thrown aside, as scaffolding is taken down when a structure is completed.

I feel it necessary, however, while neglecting the more usual course of chronicling the history of window-tracery, to supplement my lectures at this point with some remarks on the general construction of windows—applicable more or less to all periods of MediÆval architecture.

The most normal form of a window in an arched style is simply an opening straight through the wall covered by a barrel arch. This is, however, obviously defective in its fitness for diffusing light in the interior, a deficiency which, though slight in the case of a large window in a thin wall, becomes serious when the window is narrow and the wall thick. The simplest method of meeting this is to splay the jambs and arch of the window, at, for example, an angle of forty-five degrees, so as to allow for the spreading of the rays of light within.

In English architecture of pre-Norman days, this was most frequently done, both within and without, by placing the glass a long way from the outer face, or perhaps in the mid-thickness of the wall (Fig. 154). This had the advantage of splaying the head or arch as well as the jamb, which allowed more high light to enter; an advantage often increased by splaying the exterior of the arch more than the jambs, giving it a bonnet-like shape, and so obtaining still higher light (Fig. 155). Windows thus splayed inside and out, may be seen in the Castle Church at Dover—some few of these are not arched but had oak lintels, splaying upwards at about forty-five degrees (Fig. 156). The bonnet-headed window may be seen at Holy Trinity Church, Colchester; Clapham Church, Bedfordshire and many other buildings.

Fig. 154.

The deeply splayed window may also be seen in part of St. Pantaleon’s Church at Cologne (Fig. 157), which is a work of the tenth century, and in the aisles of the Basse-oeuvre at Beauvais, a church of at least as early a date, so that it may be viewed as a feature common during these early periods of Romanesque which preceded that from which our MediÆval styles were developed. During the rise of the Norman style, a different system was more usually adopted, the splay of the jambs and arch being mainly internal. A series of humble village churches at the back of Dover Cliffs have windows in which the glass was flush with the exterior, and all the splay put inside; many both in Normandy and in this country differ from this only in having a very small external splay, and even when the exterior is shafted the inner splay often comes close to the face of the recessed order.

This excessive flushness is less frequent as the style advances, and in Early English, though sometimes, as in the beautiful chancel at Burgh in Norfolk, (Fig. 158) the glass is sometimes brought extremely close to the outside: it is usual to have at least a few inches of splay around it.

In transitional work, as in Norman, the internal splay, often of very great width, usually runs round the arch concentrically; but in developed Early English,

and in subsequent styles, a special variety of internal arch is introduced suited to those numerous cases in which the glass-plane is far nearer (as it is in a majority of instances) to the outer than the inner face of the walls. The simplest form of this internal window-arch takes the form of a barrel (pointed) arch, springing so much lower than the spring of the outside arch as to allow it to span the increased internal width without rising unduly higher than the outside arch, as was the case when the splay was continued round the inner arch. This arch of necessity formed an intersection with the inside splays. Its edge was usually in the plainest specimens, relieved by a chamfer (Fig. 159), which was often exchanged for a moulding (Fig. 160); but a far more agreeable finish was a rib dropping down a little from the arched soffite, its edges being either chamfered or moulded with or without a label over it (Fig. 161). This, if the arch were made slightly segmental, would die into the jamb-splay, or it might be carried on a corbel (Fig. 162) or a shaft (Figs. 163, 164), thus forming a very agreeable and picturesque internal finish to the window.

This rib is usually termed a rere-arch.

Professor Willis, in his paper on the Architectural Nomenclature of the Middle Ages, calls it a “Scoinson Arch,” from a French word “escoinsons.” He also quotes the term “arriÈre voussure,” probably meaning the arch behind the rib.

Professor Willis’s general description, which I had not referred to when I wrote the above, is as follows:—“An arch is placed so as to carry the inner surface of the wall. In simple examples, like the present, this rib is plain, and dies against the jambs, but in superior buildings is richly moulded, and a shaft, with base and capital and side-mouldings, are added to the edges of the jamb. But this arrangement is mostly distinct from the window-tracery. This arch is of different and larger span from that of the window-head, because the spreading or embrasure of the jambs increases the opening inwards. It is also of a different curvature, and the decoration of the two disconnected and separated by the plain splayed sides of the window-opening, connecting the two, and resting at one end on the tracery, and at the other on the rib, is a narrow vault or voussure, which again is not necessarily of the same curvature as the sustaining arches, but which carries the core of the wall above.”

He says farther on:—“We may therefore call the said vault, rib, and shaft; the rere-vault, rere-rib, and rere-shaft of the window.” He also remarks that, “in the thick walls of MediÆval structure, the tracery and its glazing are commonly placed much nearer to the outer surface of the wall than to the inner.” This last observation calls our attention to a great and important distinction by which nearly all MediÆval windows may be classified—viz., those which have their glass-plane at or near the mid-thickness of the wall, and those which, as the Professor says, have it “much nearer to the outer surface than to the inner.”

This distinction was, as I have shown, of early date; being in its earlier ages rather distinctive of “Saxon” from Norman windows. The class, however, in which the glass was nearer the outer than the inner side had, up to about the year 1200, its inner arch concentric with its outer one; but the invention of the rere-arch and its accompaniments obviated this, and established a hard and obvious distinction between these two great classes of windows.

The custom of sometimes placing the glass at the mid-thickness of the wall was in no degree given up, but, on the contrary, was continued through all styles; but, when adopted, the older system of making the inner concentric with the outer arch was nearly always continued, marking more distinctly the great difference between the two classes of window. The choice between them became a mere matter of taste and of outlay; all styles acknowledging both as equally admissible and correct.

The two systems may be distinguished as rere-arch windows and through-arch windows—i.e., those in which the inner is distinct from the outer arch, and those in which the same arch runs through the wall, showing itself more or less similarly on its outer and inner faces.

In thick walls and rich work there is often another order of through-arch within the tracery order, or rather the outer order re-appears within. The rere-arch is occasionally cusped, as in a window at Broughton, Oxfordshire (Fig. 165), and the intervening space is sometimes groined, as in some windows at Salisbury and Christchurch (Fig. 166), or richly panelled, as in some at Westminster. In some instances the place of the rere-arch is occupied by distinct tracery, like a second window in advance of the real one. This consists in most instances of perfect bar tracery, while the window itself is of plate tracery; as may be seen in some of the windows at Stone Church, Kent (Fig. 167), and as once existed on a much larger scale in the chapter-house at Tintern. I may here mention that the tracery of a window is always viewed as an arch-order; and, though the corresponding order in the jamb is in the solid with the jamb up to the springing, the tracery, like other arch-orders, is severed by a continuous joint from the order above it.

The most normal type of the through-arch window is that in which the glass is placed in the middle of the thickness of the wall, and the interior of the window is a mere repetition of its exterior. This is not, however, by any means necessary or constant; for the glass is often either less or more recessed, and the inner mouldings, etc., are not always similar to the external ones, so that the existence or non-existence of a separate internal arch is the more clear distinction. Some, however, of an intermediate character, are to be found in which an inner arch, separate in design, is nevertheless concentric with the outer arch. In others the separate existence of the inner arch arises from the existence of a triforium passage, which in clerestory windows leads to some changes of design from the normal type. In others the rere-arch is not only concentric, but is so close upon the outer arch as to be almost one with it. The two classes are, however, for the most part easily distinguished.

One of the earliest instances which I recollect of the rere-arch is in the eastern part of Tynemouth Priory.[63] This is in the transitional style, and the strongly-marked separation of the inner from the outer arch is largely owing to the vast thickness of the walls. The glass plane is perhaps four times as far from the inside as from the outside.

A fine series of specimens is in the chancel of Brecon Priory (Fig. 168), where the separation between the outer and inner arch, and the depth of the glass from the inner face are also very great. Most of the early English windows found in churches of an ordinary type are of this class. Among Early English buildings in which the windows are mostly of the “rere-arch” variety, may be mentioned Salisbury Cathedral, Whitby Abbey, the Temple Church (eastern part), the Chapel of the Nine Altars at Durham,[64] Trumpington’s work at St. Alban’s, the choir of Brecon Priory, the eastern Chapels at Winchester (Fig. 169), the chapter-house at Oxford, the choir of Fountains Abbey, etc. Among those of the same style in which the “through-arch” window prevails, may be mentioned the transepts at York, the choir aisles at Carlisle, Rievaulx Abbey, the chapter-house at Furness Abbey (Fig. 170), much of the work at Lincoln, Kirkham Abbey, etc.

Among buildings transitional between Early English and Decorated, or very early Decorated, may be named as having mainly rere-arch windows, Westminster Abbey (excepting the chapter-house), Tintern Abbey, the eastern parts

of St. Alban’s Abbey, the beautiful Templars’ Church at Temple Balsal, the Chapel of the Palace of the Bishops of Ely in Holborn (Figs. 171, 172), the choir of Dorchester Abbey, the Bishops’ Hall at Wells, the choir of Merton Chapel at Oxford, etc.

Among those of a like period in which the through-arch window prevails, may be named the chapter-houses at Westminster and Salisbury (Fig. 173), the later parts of Lincoln, the choir aisles at Selby and Guisborough, the choir of St. Mary’s Abbey at York, most of the Decorated work at York Minster, Exeter, etc. In later Decorated work the same freedom of choice prevails, as it does also in “Perpendicular” buildings, though, as we come down to later dates, the “through-arch” becomes, on the whole, more prevalent.

Taking all styles together, the rere-arch, or in earlier works the wider internal splay, is greatly more frequent, probably because less costly than the other form; and though, when the “through-arch” is used, the glass is usually set deeper from the external face than when there is a rere-arch, and is frequently near the centre of the wall, such is often not the case, as in the eastern windows at Kirkham, where the internal depth is much the greater, and, in a few instances, where it is less, than the external. On the whole, it may be said that the rere-arch system tells most internally, while the other offers greater freedom for external depth of jamb and arch mouldings. Both are equally at the choice and command of the architect, who can use both, if he pleases, in the same building, and to condemn either would be like blotting out an essential element of architecture.