The window openings in Romanesque or Norman-French churches were single lights, round or pointed arched, rather broad in proportion to their width. Stained glass, it has been explained, has to be held in its place by copper wires, soldered to the leadwork, and attached to iron bars let into the masonry for that purpose. In the case of a very narrow lancet, such bars would naturally be placed at convenient intervals across the opening. But for the most part windows were the reverse of narrow, and the horizontal bars had to be supplemented by vertical stanchions, so that the window space was divided into rectangular divisions. As a matter of construction the glass was made in panels, corresponding to these, and attached to them. It is not surprising, therefore, that these divisions should often have been accepted as part of the design, or that the design of the glass should to some extent have followed them. On [page 113] is the skeleton of the upper part of a twelfth century window. The strong black lines in the diagram show the bars, the finer ones indicate the main divisions of the design of the glass. It will be seen that the four strips into which the upright bars divide the window are not equal, but that the outer divisions are narrower than the inner, so as to accommodate themselves to the width of the border. Naturally that was determined always by the proportion of the window; such borders measured often one-sixth part, or more, of the entire width. The way in which the central circular shape in the glass breaks across in front of the border is an instance of the spontaneity and unexpectedness of design characteristic of the earliest existing work; later one series of forms would repeat themselves without interruption throughout the length of the window. When, as [above], the centre of a window is occupied by a great crucifix, or, as [below], other such irregularity occurs, it is safe to conclude that the glass, if not prior to the thirteenth century, belongs to its first years. It is characteristic of the very early date of the glass that the bars in the diagrams given do not go out of their way to follow the outline of the circles, vesicas, quatrefoils, and other shapes, but on occasion cut relentlessly across them.
The filling out of such a skeleton as those given would in many respects be much the same in the eleventh, twelfth, or thirteenth century; and in each case it would be in direct pursuance of the traditions of Early Christian design. You may see in Byzantine ivories and enamels precisely the kind of thing that was done in glass; and in the Romanesque Michaelis Kirche at Hildesheim, is a painted roof, the design of which might have been carried out, just as it is, in a giant window.
61 and 62. Border, Angers.
The main divisions of the centre part of such a window would each contain its little “subject” or glass picture; the border and the interstices between the pictures would be occupied with foliated ornament; only, the earlier the work, the more pronounced would be the Romanesque character, alike of the ornament and the figure work. The broad borders from Angers, [above], and the narrower one from Le Mans ([page 327]) differ materially from the accepted thirteenth century type ([page 117]). Witness how in the Angers glass the stalks of the foliage frame little panels in the border, and how in the Le Mans work the stalks take the form of straps, patterned with painted ornament. This elaboration of the stalks with painted zig-zag, pearlwork, and so on, is precisely the kind of thing one sees in Byzantine carving and inlay. The very early spandril from Angers, [below], if not markedly Romanesque in character is yet not of the distinctively Early Gothic type.
63. Angers.
The shape of each medallion would be emphasised by a series of coloured lines or fillets framing it. In quite early work the broader of these would be broken up into blocks of alternating colour; they would be patterned probably (which in the thirteenth century they would probably not be), and altogether the effect of the ornament would be more jewelled. One of these broken and patterned margins is shown in the vesica-shaped framing to the figure on [page 37]—belonging, by the way, to the window given in skeleton on [page 114].