Page:Archaeological Journal, Volume 5.djvu/133

 OF TIIR MONASTERY OF ST. GAI.L. 91 peii, the opening had a cistern below to receive rain. But in the north, if a' fire-place was below it, the central opening nuist have been covered with a sort of turret or lantern, with open sides, to prevent the rain from pouring down upon the fire. Comparing the two hospitia, I am inclined to think that this is the interpretation of the ciuitral s(piare in most of the examples in which it occurs, namely, that- it represents the central 0})ening of a roof, which roof may either slope out- wards or inwards, as the case may be, and that in some in- stances it was covered with a small roofed turret. But the fire-place was not always in the middle ; on the contrary, the little oval which is seen in the corners of some of the rooms, for example, in the abbot's house and in the bedchambers of the great hospitium, doubtless marks the stove or charcoal pan. In the kitchen and brewhouse the central square fur- nace must have had over it an open lantern. In the great kitchen it is inscribed  foniax super arcus a cooking stove upon arches. Three of the rooms in the plan, namely, that under the dormitory, one in th3 novices' convent, and one in the infirmary, are warmed by a hypocaust, which is shewn by the fire-place being in a different place from the chinmey. For the fire-place, "caminus ad calefaciendum," is at one end of the room, and the chimney, "evaporatio fumi," or " exitus fumi," at the other, and both outside the walls. It is clear, therefore, that the fire and smoke nmst have circu- lated through flues under the floor of the apartment, as in the ancient " hypocaust," and finally found its exit by a chinmey at the end of the series of flues, for it is only thus that we can account for the entire separation of the fire-place from the chimney upon the plan. In the great farm buildings at the south-west piU't of the establishment the small central square nuiy indicate that the central s})ace has an overhanging shed carried round it, leaving the opening in the middle ; or if this appears improbable, we must suppose in this case that it means a pond for water, or, as Keller seems to think, a little cabin or sentry,-box, which I confess does not appear very likely. Most of the buildings seem to be in one story only, but in some of them upper stories are indicated by the inscriptions, as in the three o-reat buildinu's which enclose the cloister, the sacristy and library which are attached to the east end of the church, the abbot's house, and the great stable for horses and cows. The general distribution of the buildings with respect