is notoriously difficult. The coins are commonest in the late 3rd century, but one unworn specimen sealed by the destruction level is of Constantine I minted for Constantine II (312-337). In other words, destruction occurred between then and the end of the Roman period. The date of construction is equally vague—the Great Mosaic in the main house is likely, on stylistic grounds, to be of the 4th century. And it appears to be an original feature.
In plan, the whole villa appears to be of one period. But this is probably an illusion, and it may be that the very geometry of it contains the vital clue. The farmyard is an accurately measured parallelogram, which we thought at first had been due to bushes or trees (in Roman times) making the measurement of diagonals too difficult. But a better hypothesis, which must be checked by excavation, is this: the aisled building, with its long history, may be the original villa. To this was added the main house at an arbitrary angle, the aisled building being kept as a barn and bathhouse. We have proved that the perimeter wall was added to the main house, and its odd shape is
Evidence for demolition. Below three feet of rubble, the broken bathhouse floor and the "ghost" impressions of the lost hypocaust pilae. Scale in feet.
then an attempt to produce the appearance of a rectangle where the angles could not be changed, although the sides could at least be accurately matched. The barn (if it is) and any other buildings to the east would then belong to this final development. The time-scale of this is quite uncertain: the coins suggest a mere eighty years, and details of the buildings imply that the same architect or contractor was employed. Here, then, is the main question to be answered in future seasons.
The aisled building, probably the most important part of the site, has not been completely excavated. But all the evidence suggests a complicated history. It was originally a large aisled hall, some 45 ft. wide, in which the farmworkers could assembly socially, and where they probably lived alongside their livestock and equipment. Doubtless it was divided by partitions, but there was no real privacy until the west end was converted into three comfortable dwelling rooms, one with a red tessellated floor, another with a mosaic. Part of the south-east aisle was walled off to make a hot drying (or bath-) room with walls and floors of chalk. The T-shaped flue was stoked from inside the main building. When the whole south-east corner was later converted, this room was superseded by a full bath suite built largely of brick, and stoked in the opposite direction at a higher level from the same stoke-hole. At the western end was room to undress and a tiny cold bath ; one then entered the warm room, and then the hot room, where the apse contained the hot bath. The hot water was piped from a cylindrical hot tank outside the room (whose setting, in the furnace flue, survived) which was in turn fed from a cold cistern at a higher level beyond. This, too, had survived, although badly wrecked.
Two photographs illustrate the thoroughness with which all this had been demolished. The concrete floor was broken up and the brick pilae removed from under it,