Another theory is that the high vaults were constructed by Bishop Farnham in 1242, making them contemporary with the Chapel of the Nine Altars—which is impossible. It seems to me that they belong to the episcopate of Hugh Pudsey, _c._ 1160. The bays of Durham are coupled, Lombard-fashion—_i.e._, large and small piers alternate: with what object, is uncertain. There were three parallel eastern apses. The central apse had no ambulatory. The lateral apses were square externally, as at Romsey and Cerisy-la-Forêt, and S. Maria in Cosmedin, Rome (_c._ 780). The central was separated from the lateral apses by solid walls, as at Cerisy-la-Forêt, whose clerestory also seems to have been copied in the original western clerestory of the south transept. Durham anticipates Gothic not only in vaulting its central aisles in oblongs—for the high oblong vaults of Durham, if not so early as 1133, are still the earliest in England—but in the employment of flying-buttresses. These, I think, were built long before the vaults, their object originally not being to withstand the thrust of vaults. In the triforium of the choir they appear in the form of semicircular arches, with a wall on them, which provided a support for the longitudinal timbers of the aisle-roof. In the choir these flying-buttresses are semicircular arches, and therefore, strictly speaking, are not flying-buttresses at all. But in the triforium of the nave they consist of segments of circles tilted up on end. Here they are genuine flying-buttresses, which oppose resistance to any outward inclination of the clerestory wall. The only constructional difference between those of the nave and Gothic flying-buttresses is that the former are placed underneath the roof of the triforium, sheltered from the weather, while the latter are placed outside and above the aisle-roof, and are thus liable to disintegration by wind, rain and frost. But even in Gothic, flying-buttresses are not always displayed; they still remained concealed under the triforium in the early work at Salisbury; and even in the fourteenth-century work of Winchester nave. At the west end the towers project beyond the line of the nave, but only slightly. Still we may well see in this the germ of the grand western transepts of Ely, Wells and Peterborough. [Illustration: CHOIR.] Internally, the one fault of Durham is its shortness in proportion to its great breadth. The nave ought to be two or three bays longer. Ely nave has twelve bays, Peterborough ten, Norwich fourteen, Durham only eight. But the architect could not build much further to the west, for close at hand is the precipice rising above the river; nor did he like to build further to the east, for the ground there is bad.