rant-shape, and were probably used for the erection of brick pillars.

Professor Haverfield found composite columus either in brick or stone, or in brick and stone combined, at Corbridge. They have been found on other Roman sites.

On several of these bricks are the marks of dogs' feet, and the finger-prints of the workmen who took hold of them before they were burnt. The same marks occur on the bricks at Holt.

A few flue-tiles and roofing-tiles, mostly in fragments, and a quantity of highly-glazed bricks and building-stones have been found in various parts of the Prætorium, in the opinion of Professor Bosanquet vitrified by the action of fire.

The subsoil of the camp is clay, which was used for placing under the pavements. The clay and the river gravel could be obtained in the immediate neighbourhood, but the red clay used was probably brought from the same place as the imported bricks.

The outside walls of the Prætorium are 2 ft. 6 in. thick, and the inside walls generally 2 ft. thick, with the exception of the side walls of the Sacellum, which are built on the top of stamped clay. All the walls have deep foundations, and consist of stones placed on end, so as to act the part of drains. Occasionally river stones and boulders out of the subsoil were used when making the foundations.

The main drain of the camp evidently passed under the Sacellum; there, as in other places, the drains have been apparently wooden ducts, consisting of one piece of timber placed on the top of two supports, so as to form a kind of pipe, and were backed with clay. They seem to have been about 8 in. high.

At the door of the Sacellum, and to a distance of 9 ft., there is a paving of large flag-stones, the remainder being open. I should suggest that formerly this spot was the site of a manhole, and that it was originally paved like the remainder of the building.

In the outer court we found a trap door 1 ft. broad 2 ft. long, the hinges and the ring evidently soldered with lead.

By the buttress on the N. side we found a large circular hole or gully, backed with brick and stone 3 ft. wide and 3 ft. deep; a drain enters this from the wall of Room 9, here it is paved with brick and stone. From this gully, on the opposite side, runs a wooden

duct, which passes under the E. wall and down the N. corridor.

On the opposite side we found a drain leading from the stone square, and at a depth of 3 ft. running down the S. corridor and Via Principalis. In this was found a large amount of timber, with what seemed to be the remains of decomposed brick. Possibly it belongs to an older wooden Prætorium, as it runs outside the line of post-holes, and would drain away the water from the roof.

THE FINDS.

The pottery found in the Prætorium was all in a very fragmentary condition. Owing to the work being carried on so late in the year, and want of accommodation, none of the finds have as yet been carefully examined, but a later report will appear, with drawings and photographs. All that can be said now is that the pottery was mostly of a late date.

Samian.-Only one fragment of decorated ware, in the shape of a head of a lion, which was evidently the spout of a vessel. It belongs to a late period. A fair amount of undecorated Samian, mostly vessels of about 5 in. high.

Black Ware, sometimes called Upchurch Ware.This was the most common pottery found, and was perhaps for kitchen use. This ware was found in two shapes

(a) Jars (or Olla). We find that these were mostly of the second and third centuries.

(b) Shallow dishes, 2 in. to 3 in. high, and about 6 in. across, possibly cooking-pots for vegetables, etc.

Coarse-grained white ware in shape of shallow dishes. Shape called "mortarium," mostly of an uncommon shape, the ridges at the top being different from those usually met with.

Iron. The most common objects of iron found are nails, especially in the Prætorium, with the exception of the outer court. They vary in length from 1 in. to 6 in., and have all large heads. What seems to be the bolt of a door was found in the inner court; also the blade of a knife or dagger, and the handle possibly of a sword.

Lead.-A quantity of lead was found scattered all over the Prætorium. It is possible that a great deal was used in the construction of this building, but it may have been removed when the paving and dressed stones were taken away for building materials and road making.

Bronze. A certain amount of bronze in a very decomposed state, a few fibulæ, one in the shape of a dolphin, five large coins (two of Antoninus Pius) and several ornaments in the form of bracelets.

Silver.-One silver coin of Antoninus.

Wood.-A large quantity of decomposed wood, especially in the post-holes. There were two kinds of post-holes.

(a) Those in N. and S. corridors consisted of one large stone, whilst stones on end were placed round to keep the post in position, at Gellygaer. These are not sunk on to solid, but occur in the stamped clay.

(b) The remainder are holes sunk from 1 ft. to 2 ft. into solid clay, with a large stone at the bottom.

Glass.- A certain amount of window-glass, especially in Rooms 10 and 11, in fairly large quantities; also various portions of glass vessels, generally of a square pattern. A portion of a fluted beaker, similar to the one described by Mr. Ward, found at Gellygaer, and a portion of a ribbed vessel. Also small glass counters or draughtsmen.

Stone Implements.-A portion of a quern and a mortar; an implement resembling a pestle, and a large quoit-like stone, two smaller ones, presumed to be weights, and numerous sling stones.

At the entrance to the Prætorium is the usual open paved court surrounded with corridors, with a large number of brick and pillar-stones. Beneath the pavement there was 8 in. of gravel and 4 in. of clay. In the centre of the court, on the northern half, are the remains of a drain running from the centre to the main drain. This does not appear on the other side of the court, but, owing to the extreme difficulty of tracing these drains, it is probable that it existed on that site

in Roman times.

The walls on the N. side of the road leading to the Sacellum were in a more or less good condition, those on the S. were good, but the W. wall had been robbed down to the foundation. We came upon a foundation

of one pillar in the S.E. corner, and what was perhaps a square pilaster on the S. side of the road.

On the S.E. corner is an excellent example of the base of a pillar. It is built on good foundations.

At

Castell Collen: Base of Pillar South-east Corner, Outer Court

the base is a square of stones 2 ft. square; on this is laid a 2-in. cottise of tiles, and then follow two courses of pillar-stones, ten to twelve in a course. The wall seems as if it had been built into the pillar, and it is probable that the pillars formed a colonnade: then,