[22] The great breadth of the footings of the rectangular part of Edgar's Chapel--about 6 feet 6 inches on the north and south walls--might easily have inspired a wrong opinion as to the substance of the walls themselves. But students of the work of this date are familiar with the fact that the flat and heavy fan-vaulted stone roofs of the Tudor period require, in addition to their external buttresses, a certain amount of interior support, which is given by building the walls as a series of hollow bays, the windows occupying the recesses, and the intermediate masses being brought out inwards, as piers or counterforts, the same being architecturally treated, so that the description elsewhere given of "piers as panellæ" is quite probably accurate, as a description of such features.--F.B.B. "Search far for the est end of Edgar's Chapel. It is but little damaged. S. Mary and S. Andrew's Chapels, over the ends of ye choir aisles." The only reminiscence of a retro-chapel is to be found in the plan of the Abbey embodied in Phelps's _Somerset_, reproduced by Warner. Phelps shows a dotted-line extension of the east wall of the retro-quire of nearly the same projection as Willis's plan, but in this case it ends with a semicircle--a feature impossible for anything but a Norman chapel, which here would be out of the question. In a corner of Phelps's plan appears a similar diagram, but with the rectangular part much lengthened. Both are lettered "F," and the reference table gives "F" = Lady's Chapel. Warner copies this plan, calling this latter one "the Chapel according to its original dimensions." Neither F.B.B. nor J.A. had at that time seen Warner's copy of Phelps's plan, because the copy of Warner accessible to them in the Bristol Public Library had lost this sheet. But owing to the general haziness of the plan, and its numerous inaccuracies, and the entirely impossible suggestion of a semicircular apse (in a dotted-line figure), as well as owing to the suggestion "Lady's Chapel," these records had proved the reverse of illuminating, save as inspiring a guess that the original and shorter quire might very likely have been furnished with a Lady Chapel, for use in the remote days, before the western chapel of St. Mary had been opened up and thrown into the general series by the inclusion of the "Galilee"--a work of the fourteenth century, and that this Lady Chapel had been ultimately shorn of the greater part of its length by its absorption into the new retro-quire, what time Abbot Monington caused the quire to be lengthened by two bays (about 40 feet) _temp._ 1344 or thereabouts. Such an hypothesis would entail the deduction that after the lengthening of the retro-quire in the fourteenth century, there would remain at the east end a relatively short projection--say, of 12 feet or so--and some such reasoning may well have inspired Willis's scheme (see his plan in the _Architectural History of Glastonbury