two weeks ago, when I took up the long-suspended work. No place could have been more favourable for a beginning. We began in no chance place, but in one that furnished the clearest indications. The work was continued in true archæological succession. The results of my Avork I place on the accompanying plan.
The first digging occurred at the point A. The men worked both north and south, the object being to determine the extent of the counter- scarp. Northward it was followed to the point A'. There the work got very deep, and if continued would have passed through the garden in front of a new house, which may be seen to be built against the scarp. The owner of the house told me that for the foundation of the north-east corner he had to dig 25 or 30 feet. This suggested that it had been built out into the ditch. I took him to the point B and told him that there I should find the rock at a depth of a few feet. We sunk a shaft, finding the rock at 9 feet, with a scarped face descending. This would give the depth of the ditch at this point at 15 or 20 feet, if we can trust to the figures of the owner of the house.
The men who worked southwards followed the counterscarp for a few feet only, when it turned a sharp angle to the south-east, and was lost. At K we unearthed a room, built in the rock, with rough walls covered with plaster, and with a Mosaic pavement of a somewhat complicated pattern, with tesseræ in red, blue, yellow, and white. This will be photographed later. It is evidently late work. Could the counterscarp have been cut down to have made place for this house ? Just before this point the counterscarp is only 4 or 5 feet high, with a shallow channel running north along its base for 15 feet or more, when the rock drops another 5 feet. Thus from the Mosaic pavement[1] we traced the counter- scarp north-east for 110 feet to the point A', from that point to K is about 75 feet, where we found it again, making a distance of 185 feet The counterscarp is not exactly parallel to the scarp, the ditch being at different points 40, 49, 54, and 65 feet wide.
At C we found the rock at a slight depth. Here was a curious cutting in the solid rock 5 feet deep, 13 feet long, 8 feet wide at one end, and 5^ at the other. The north end had a platform about 3 feet high, occupying half of the room ; it had been arched once, and contained a fireplace, which had a low rounded roof of its own. Channels for water led into the lower part. Against the south rock-wall of the chamber there was what I must describe as the silhouette of a stairway, as the steps projected only an inch or two from the rock, which was cut away to form the three steps. These, of course, will be drawn when a proper plan of the whole place is made. It looks as if they had been intended as rests for a wooden stairway. Many Roman tiles were exhumed. Near by was one bearing the stamp of the tenth legion. The place was probably a bath connected with the room with the pavement.
We dug trenches along the line D — D', finding a wall of medium sized
- ↑ From the corner at B (?). — Ed.
