The details are so well shown in the illustration that further description is unnecessary. It measures:-extreme length, 53in.; height of triangular upright above beam, in.; length of folding arms, 24in.; diameter of holes in arms, 3/32in.

Roman bronze balance from St. Margaret's, Marlborough.

Mildenhall.

Mr. R. Butler, late of Stichcombe Farm, now of Marlborough, has coins of the following emperors, which he kindly allowed me to examine. They were found in and around the famous "Black Field." He has at different times filled in several wells which interfered with agriculture. Those which were examined contained pottery, horns, and oyster shells. Some had the tops protected by masonry. They were usually about thirty feet deep. A tessellated pavement was found on the hill side above Stichcombe about 1890, but was destroyed. Mr. Butler remembers that when the "Black Field" was under cultivation, at certain seasons one could stand on the hill above and trace the lines of the old streets by the different colours of the corn. The old Roman road can still be plainly seen on the hill close to the forest boundary, running across the arable land from Stichcombe to the S. It is on the left or east of the modern road at that spot.

From a man who had worked about the village all his life and saved what he found I obtained a collection of pieces of Samian pottery which he had found from time to time. A foot rim, form not determinable, bears the name "CUCALUS," of unknown origin. Another, the stamp of the potter "BUCCUS," who worked at Rheinzabern (?) in the second century. The remainder is mostly of fine quality of Lezoux fabric and of early second century date, bearing figures, mostly on Form 37, of birds, lions, gladiators, cupids and dolphins, and a Gorgon's head.

Purton. Pavenhill,

In 1896 men working on the N. side of the road leading from Purton to Braydon, about 350 yards W. of the Purton Workhouse gate, found a quantity of square brick tesseræ, in. and in. square, part of a destroyed pavement. Two of these are now in my collection. As far as I could learn nothing else was found, and the remainder were destroyed. This find is not to be connected with fragments of mosaic of glass said to have been recently found at this place, which are obviously of foreign make and comparatively modern date.

The following coins have been found at Pavenhill :-Claudius, Decentius, Constantine, Magnentius.

PURTON BRICKYARD.

From the Cricklade Road a lane leads E. from the Packhorse Toll Gate to Blunsdon. At a point under Down Farm are two cottages. Behind and S. of these is a clay pit, now grown over with grass and disused. About 1885 men digging clay came to a small square foundation, near which was found much pottery and a few coins. I have examined several of Claudius, Antoninus, Faustina, and the inevitable Constantine. The only fragment of pottery I obtained is the neck of a bright red lagena of a form characteristic of the first century. Some pottery, &c., from this spot passed into the hands of Mr. Leslie, then of Wootton Bassett, and from him to my own collection. Three little bead rim pots are of late Celtic type.

PURTON CHURCHYARD.

About 1890 a man working here dug up a small Roman pottery lamp of dark brown coarse ware with the usual holes in the top and nozzle, 2ĝin. × 2in. × 1ĝin. high. It has a curious knob on one side, and the spout is much blackened by fire. The lamp is now in my collection.

Ramsbury.

At Eastridge in 1896 were dug up (exactly how or where cannot now be said)five pieces of Kimmeridge shale, waste from the manufacture of armlets. Each piece bears a square hole, by which it was fixed to the lathe chuck, and the size indicates that the rings which were turned from these specimens were 12in., 24in., and 2žin. in inside diameter. Similar remains are common in Dorsetshire, and may be seen in all stages of manufacture in the Dorchester Museum. They were formerly known as "Kimmeridge Coal Money" before their real nature was understood.

Russley, near Baydon (? in Bishopstone).

Five hundred yards east of Russley House, on a patch of newly-broken down, the S. W. corner is covered with the refuse of Roman occupation, pottery of coarse type and all thicknesses being mixed with fragments of fine Samian, one of which bears a characteristic rosette. New Forest thumb pots with metallic lustre, red and black, also occur, and one fragment is ornamented with white slip on a red ground. A large brass coin of the first century was picked up in the middle of the site. This spot is not far from Botley Copse.'

Stanton Fitzwarren.

Four hundred yards S. of the railway station, where a cattle track runs under the line, are patches of rough tessellated pavements with tesseræ of large size. Three now before me are of rough dark brown sandstone, carelessly cut, and roughly one inch square. Traces of rough foundations can be traced for about 200 yards. Three-quarters of a mile S.E., under the lodge gate on the Swindon-Stanton Road, was found a small vase of soft pink pottery, 34 inches high, much like Form 78, but with a high and small foot, of late fourth century date.

Stratton St. Margaret's.

North of the railway bridge and S. of the village on the Wanborough— Stratton Road, labourers working in 1910 by the N.E. side of the modern road, came upon a long line of foundation a foot thick, formed of coral rag. I examined this and have little doubt that it was the original Roman road (Spina-Durocornovium), now buried nearly three feet deep.

Swindon.2

In 1904 a small thumb pot of late fourth century New Forest ware was found in a rabbit hole in a field close to and E. of the town. It is 44in. in height, 33in in greatest diameter, lip lin. and base 1țin. in diameter. Of red body coated with a brown wash. It has eight shallow indentations divided from each other by vertically impressed rows of punctures apparently formed by pressing a comb of eleven teeth against the wet clay (ladder pattern).

At various places on Swindon Hill coins of the following emperors have been found:-Antoninus, Gallienus, and Constantine.

During quarrying operations alongside the M. & S. W. Railway and close to the bridge in Westlecott Road, a rubbish pit was cut through in 1907, roughly circular, 14ft. in diameter, and 9ft. deep. This contained numerous fragments of Roman pottery, and at 6ft. an extended skeleton on its back E. and W., head to the latter point, hands by the side, but with no relic in actual contact with the bones.

From the site of the Roman house at Westlecott I have one small piece of rosette stamped red ware of Ashley Rails New Forest type.

2 For other information on Roman Swindon see A. D. Passmore in Wilts Arch. Mag., vols. xxx. and xxxviii.

Winterbourne Monkton. Hackpen Hill.

On the S.W. slope, just below the group of barrows, and exactly E. of the village (O.M. XXXVIII., N.W.), a large urn of hard grey pottery was found in 1898, empty, and smashed into fragments. I examined a large piece of the rim soon after the discovery. Apparently it was an ordinary globular Roman pot, with out curved rim. Near the above, but not in relation to it was a small bronze T-shaped fibula, with solid catch plate, coiled spring and pin in one piece, the spring being held in the T-head by a loop carried from each end of the spring through a hole in a flange on the head itself. Of first century date. Here also turned up a small bronze armour ring, similar to those described above under Medbourne.

WANSDYKE,

ITS COURSE THROUGH E. AND S.E. WILTSHIRE.

By ALBANY F. MAJOR, O.B.E.

The writer's examination of the course of Wansdy ke practically began in 1913, though he had visited portions of it before. In that year, prior to the joint meeting of the Cambrian Archæological Association and Wiltshire Archæological and Natural History Society at Devizes he explored the country between Savernake Forest and Inkpen, and traced the course of the dyke for most of the way. In the following year he went over this ground again, tried to find traces of the dyke in the forest, and followed it from its reappearance W. of the forest, through the West Woods and across the downs to Morgan's Hill, besides seeing it at several points of its course along the Roman Road. Later in the year he followed it from the Wilts-Somerset border to its termination, so far as it can be traced, on Maes Knoll, S. of Bristol. The further investigations which he had in view were postponed owing to the war: but meantime he was informed by Mr. H. C. Brentnall, of Marlborough College, that he and a colleague, Mr. A. R. Gidney, had found the point where Wansdyke ran up to Savernake Forest on the W., besides various lengths of vallum and ditch, which might be remains of its track through the forest. One of these, which is marked in the O.S. maps, the writer had already noted as a possible remnant of Wansdyke.

This year it has been possible for him to again take up the work, and in company with Mr. Brentnall he has followed the supposed course of the dyke through the forest and made a further examination of its course beyond the forest as far as Inkpen; and has given a further two days to an attempt to trace the branch of the dyke which Sir R. Colt Hoare described as running down on to Salisbury Plain from a point near Great Bedwyn. In this he was successful beyond his expectations, and he thinks the time has now come when the results of his investigation, so far as Wiltshire is concerned, should be communicated to the Wiltshire Archæological Society.

Though his conclusions do not always agree with those of Sir R. C. H., the latter's authority stands deservedly so very high that archæologists, especially in Wiltshire, may think it presumptuous of anyone to profess to correct his observations and those of the able surveyors employed by him, considering, moreover, that his investigations were made a hundred years ago, when the remains he was examining were generally in a better state of preservation than they are at present. It must, however, be borne in mind that Sir R. C. H. explored the country on horseback, and, though this may help in covering the ground, the observer on foot is able to make a much closer investigation, and of necessity takes more time about his work. Further, earthworks waste most rapidly in the first few years after their