four or five steps leading to a permanent spring, discovered in August, 1891, three feet below the ground level (see Transactions, 1891, page 244 for plan and section), is situated about 50 yards below the abutments; that an excavation, made parallel with the river bank, cut obliquely across a road from 10 to 12 feet wide, between this basin and the abutments, that this road was buried about 18 inches below the present ground level; that there is sufficient space here for a small building, for instance a small warehouse, or a small villa; and that the ground rises quickly to a commanding elevation about 60 yards distant, from which small landslips, or detritus after heavy rains, may conceal other as yet undiscovered works of men, in connection with this recently unearthed handiwork.

2. Let us now turn to the consideration of the existence of a bridge of timber, or other way of communication, more than half-a-mile lower down the river, at the end of the reach which extends from the New Weir down to Huff Pool at the next bend of the river. A personal observation of the ground, and an inspection of the map, will show that this site is the direct prolongation of the old Roman Road in Madley parish called Stone Street. This road from Abergavenny (Gobannium) to Kenchester (Magna Castra), in the present day in use for five miles from Brampton Hill to three quarters of a mile distant from Huff Pool, extended down to the river at Huff Pool within the memory of living witnesses. James Lloyd, as a youth, less than half a century ago, often traversed it in awe of many a gipsy encampment, amongst which tribe he represents it to have been a favourite settlement. He resided at Canon Bridge, and accounts for its disuse by the fact of its having been ploughed up and annexed by Messrs. Jones and Lee, or Lea, of Canon Bridge, up to the present boundary between their respective properties and that of Sir Joseph Pulley, of Lower Eaton.

The accompanying plate, from a photograph taken by Mr. Robert Clarke, represents the boundary fence where the continuation of Stone Street terminated at the river bank.

On the same day of exploration, namely April 27th, we found at the distance of 20 yards above this boundary fence, where the water at lowest summer level is 12 feet deep, about fourteen piles in tolerably close arrangement, extending to a distance of 15 feet from the bank. Some of these piles were vertical, but generally in an oblique direction, of enormous scantling, two of them, notwithstanding their immersion for an unknown duration, at least 12 inches square; and a few yards further up the river are large timber baulks lying horizontally, like steps.

Could we only have discovered in this situation a row of intermediate piles extending across the river, we should have had proof of a bridge on timber piles. The result, however, of an examination of the river here on another occasion, on May 12th, in company with Mr. Cockcroft, our librarian, proved negative; the only information obtained by fathoming with an iron-shod pole was that the bed of the river was mud, and that its general depth varied from 12 feet at the banks to 13 feet in the deepest part at lowest summer level.

The descendants of the skilful military engineers who overran Gaul, and who (as recorded in Liber iv,, cap. xvii, of Cæsar's Commentaries on the Gallic War),

constructed in the short space of ten days, including felling, preparing, and transporting the timber, a timber bridge over the Rhine, probably at Bonn, where the river is 530 yards wide, would have found it comparatively child's play to span the 60 or 70 yards of the Wye here. Calculating a width of from 8 to 10 yards from pier to pier, for each bay of Cæsar's bridge, five or six piers would have sufficed to cross the Wye, whereas about 60 piers must have been constructed to cross the Rhine at Bonn, the locality of Cæsar's bridge as fixed by Napoleon and others.

We have no records in our own country of ancient bridges reaching to a period so far distant as the Roman occupation-say to 400 a.d. The earliest record of a bridge over the Thames given by the Saxon chroniclers carries us back only to A.D. 1017 (Vine's Cæsar in Kent, 2nd edition, p. 229), when "Cnut the Dane, invading London with a fleet to dispossess Edmund Ironside, found himself unable to pass the bridge over the river at London, which the citizens had strongly fortified. He consequently cut a canal on the south side of the river, deep and broad enough to convey ships above the bridge."

We have adduced evidence bringing the Roman road to the Wye at Huff Pool. On April 27th, our party, having crossed the river in the boat, discovered its continuation on the opposite side of the river, leading directly over the "Wye Meadow" on Mr. Charles Hardwick's land, between his French barn and his residence, the Old Weir, to join the present existing Roman road, where it crosses the Hereford and Hay road, just opposite the entrance gate to the Old Weir. Keen eyes discovered in the Wye meadow two low parallel ridges, distant from each other about 12 or 15 feet. On the day succeeding our visit of exploration, Mr. Hardwick dug a trench transversely across these artificial ridges, with the result of discovering, at the depth of 12 inches below the surface level, a thickness of 12 inches of gravel, extending for a width of 15 feet. As a counter experiment, he dug a hole 10 yards distant. Not a single stone was found in the natural sub-soil loam. The existence of a buried road was so obvious as to preclude the necessity of further excavation.

To conclude; although we cannot go so far as to declare that these fourteen massive piles on the Canon Bridge side of the river at Huff Pool are the foundation piers of an ancient bridge over the Wye, yet the discovery of the buried road upon the opposite bank in the grounds of the Old Weir, in a direct prolongation of the old Roman road called Stone street, enables us to assert that there must have been some method of communication between the two opposite roads.

Nothing has ever been discovered to give a shadow of suspicion of a stone bridge having existed here-in fact, the foundation bed of the river, and its depth here, render it an unfavourable site for a stone bridge—therefore the access must have been either by a temporary pontoon bridge of boats, or what is more probable, over a fairly permanent timber bridge, which has been washed away by floods, here in the present day sometimes attaining a rise of 20 feet, and other vestiges of which have been removed by fishermen. James Lloyd's information comes again to our support. He remembers the fisherman, William Terry of Hoarwithy, who used to net for Mr. Jones of Canon Bridge, occupying all his spare time in sawing and removing timber obstructions at the bottom of Huff Pool. *The average depth of the river between Coblenz and Andernack is 16 feet, and at Xanten 18 feet.

The right bank, or the south side, of the river Wye is represented in the foreground at the termination of the old
Roman road called Stone Street. On the left bank, or north side of the river, is the large Wye Meadow,
in which the buried road was discovered, leading in a direct straight line (between the barn on the right and
Mr. Hardwick's residence, The Old Weir, on the left), to the camp of Kenchester (Magna Castra).