Now, nothing could be more rational than this development. The mode of glazing in use was most conveniently applicable to spaces of moderate width. It is true that by the more extended use of iron it was then, as it has often been in modern times, applied to openings of 6, 8, or even 10 feet in width; but narrower spaces were much more convenient. The lights, however, at Westminster Abbey (which is one of the earlier buildings in which this kind of window is systematically used in this country), are 4½ feet wide, and in France they are generally much more. The prevailing practice of placing a massive pier between each of such lights was obviously imperfect. The concentration of pressure upon the buttresses now allowed of openings of almost any size; what, then, was more reasonable than to make extensive openings, and then to subdivide them by light mullions into compartments at once sightly and convenient? That this practice has sometimes, from caprice, been carried to a vicious excess in no degree militates against its _rationale_; indeed, with all our modern facilities for glazing and opening our windows, we continually find the same expedient resorted to for convenience, and invariably so when any extraordinary amount of light, and consequent width of window, is needed. The next question which would arise is, how is the arch to be filled in? This we find done at first by a plate or tympanum of stone as thick as the depth of the mullions, each light being arched, and the tympanum pierced at pleasure with such openings as suited the builder’s taste; and, later on, we find these piercings connected together into those systematic groups which we call tracery; thus converting the window into a perfectly novel and most beautiful architectural feature.