Page:Archaeological Journal, Volume 9.djvu/227

 THE ABBEY CHURCH OF DORCHESTER. 167 terminate at a point not liiglicr than the level of the side walls. It follows then that some third figure must occupy the gable, just as in the smaller examples just mentioned. Unfortunately the gable has been destroyed, so that we cannot recover the exact nature of the original arrangement. But certainly that best adapted to the position would be a single window, rather smaller than those below^, and forming a triangle with those below^ The front would thus exhibit, in a later style, and on a larger scale, the same principle as the west end of Llanbadarn-ftiwr in Cardiganshire, or the east end of Banning in Kent. That such was the original compo- sition, I will not positively aflEirm ; I only say that it would be much the most appropriate one, and that I cannot think that the small square-headed openings on each side, at all prove that it was not really that emplo^^ed. Now within it is clear that such a composition Avould not have the same good effect as without ; a gable window is something essentially external, in no wise calculated to form any part of an inside view ; if it were merely because, in a building of this size, it proclaims itself as being over a vaulted or other ceiling. Hence, instead of the high-pitched open roof, rendered necessary in the choir by the nature of its east window, the aisle must be vaulted, so as to exclude the gable composition. But it w^ould be hard to i&nd any of the ordinary forms of vaulting which would appropriately cover so wide a space w^itli tw^o windows at the end. Some- thing would have been wanting in the head, wdiich the external arrangements could not have permitted ; and it may be doubted whether, with any sort of roof, the two window^s, side by side, with no such provision as the buttress provides without, could ever have been an agreeable arrange- ment.^ This difficulty was avoided by using a single bay of sexpartite vaulting — sexpartite at least as far as the east wall is concerned — over the eastern bay ; by this means flatness is avoided, and no space left unoccupied, each window fits into its own cell, and the vaulting-shaft runs up between them within, just as the buttress does without. The arrange- ment is the same which is adopted, and apparently for the same reason, over the eastern bay of the choir of St. Cross. We can there judge of its actual effect, and, though decidedly open to the objection that it is a sort of mimicry of an " See tlie next note.