128 Mediceval Military Architeclure in England. They are vaulted, usually in a plain barrel, or equally plain groining, lighted by external loops, and now and then, though very rarely, have, as at Bowes, a fireplace. In breadth they vary from 4 feet to 6 feet, and they are of all lengths. Usually they are more abundant and larger, in the main and upper floors. They were intended for sleeping-rooms, garderobes, oratories, and well-chambers. At Guildford is a very re- markable oratory, at the first-floor level, L-shaped, with a mural arcade. At Brougham the oratory is on the upper floor. At Rochester and Dover the upper gallery does not run all round ; and in the latter ends in a prison cell. Some keeps, even with thick walls, contain but few chambers. The two remaining sides of Norham, a late keep, have none, or, at most, one. Dover, also late, is honeycombed with cham- bers at every level, and even the cross-wall above is threaded by a gallery, a singular example. Newcastle, built in 1080, has very many chambers. Where there are garderobes they have stone seats, and the vent is usually a vertical shaft in the wall, though sometimes there is a shoot upon the exterior surface of it, or, though more rarely, it is corbelled out. At Kenilworth, a fine but late keep, one angle contained a well- stair, and the three others chambers. They are turrets from the ground, and in size larger than usual. They are floored with timber. One angle turret seems to have been wholly occupied by garderobes, and the lower part, at present filled with light soil, was evidently a large cesspool. At Corfe a very large garderobe tower, — perhaps a Norman addition, — is appended to one side of the keep. Most keeps, even early ones, contain fireplaces. One has been discovered in the Tower, long supposed to be without them. At Dover they are in the cross-wall ; at Rochester, where they are very handsome, in the sides. Usually the funnel ascends in the wall ; at Rochester and Colchester it forks, and the two flues open a little way up on the face of the wall, concealed in the hollow angle of a pilaster. It is not easy to tell whether a flue is original when the fittings of the fireplace have been inserted ; thus, in the Tower closets, where the fireplaces are Tudor, there are no marks in the wall as though it had been cut into to construct a flue, so these may represent original fireplaces, though this is not probable. No doubt open hearths were much used, and fireplaces of iron with flues of wood and plaster, which would leave no trace when removed. The fireplaces in the mural chambers at Dover seem all to be of Tudor date. No fireplace has been discovered in Richmond keep ; there is an original one in a mural chamber at Bowes.
