Page images



STANDING as it did on the borderland of Morgannwg and Deheubarth, on the shore of a bay into which the navigable river Tawy flowed, the Castle of Senghenydd played a conspicuous part in the history of these provinces and of South Wales, as related in the as related in the pages of Brut y Tywysogion, especially that Gwentian Version of it which Stephens taught us ought to be called the Book of Aberpergwm-but where it stood has been the subject of as much speculation by past antiquaries as perhaps the site of the more celebrated Castle of Rhydy-gors in Carmarthenshire. It has at last been recognised as the Castle of Abertawe, otherwise the original Castle of Swansea. But the exact site of that Castle has been long lost to memory, although it could be located within comparatively narrow limits, for from the description given by Leland it stood north of the present or new Castle built by Bishop Gower in the reign of Edward II.

From this Castle, northward, runs a road called Worcester Place, as far as a cross-road leading down to the river on the east, known by the modern name of Welcome Lane. Nearly parallel with Worcester Place, a little to the west of it, ran, until the spring of 1912, the narrow, tortuous old street called Castle Bailey Street, and to the east of it an old lane called Castle Walls, a cul-de-sac, the only entrance to which is from Welcome Lane. The highest point is on the long quadrangular piece of land between Castle Bailey Street and Worcester Place. The ground falls to the north to Welcome Lane, slowly to the west and south, but much more abruptly to the east down to the river Tawy. The lane is about a third of the way down this declivity. On its western side is a wall, somewhat in the nature of a retaining wall, against the hill side,

which has every appearance of great age and gives the lane its name. But, were it not for the modern cottages on the eastern side of the lane, this wall must be visible from the river and Kilvey Hill beyond it, and it is not shown upon Buck's view of Swansea, taken from that point of view about 1750, nor in an older, very small, view of Swansea: both of these give the hill side unbroken by any wall or lane, and have naturally led observers to believe that the wall is not nearly so old as it seems, and, indeed, could not have been in existence when Nathaniel Buck made his drawing. Another reason why this locality, between Castle Bailey Street and Castle Walls Lane, has never (in spite of these suggestive names) been accepted as the site of the old Castle (though it has often been suspected), has been, that local antiquaries, Leland notwithstanding, have always considered that the new Castle must have been built on the site of the old one.

It has always been recognised that the narrow Castle Bailey Street was entirely within the old Bailey or Base Court, whence its restricted width, for the foundations of the Tower of its Northern Gate, commonly called Harold's Gate, were disclosed in 1845, and the wall and ditch have more than once been found on the southern side of College Street-which runs westward in a line with Welcome Lane. This ditch was last exposed in the course of some town improvements in 1909; it was 15 ft. below the present roadway, its scarp was well preserved and stood at an angle of about 50°. The counterscarp was under the road, so that the width of the ditch could not be ascertained. The undisturbed passage across the ditch showing the site of Harold's Gate was again observed. The ditch bore the appearance of having been filled in, and not of having silted up. A few coins of the early Georges were found not far below the surface. The wall and ditch turned to the south, making the western limit of the Bailey along a road called Bunker's Hill (supposed to be a corruption of Banc y Gaer) where they are now almost entirely destroyed.

The carrying out of an important town improvement, the widening of Castle Bailey Street, necessitating the demolition of all the houses on its eastern side as far back as Worcester Place, in the early part of this year, 1912, has thrown important new light on the old problem. Portions of the walls and ditch of the old Castle itself have been uncovered, and the exact site

[merged small][merged small][ocr errors][merged small][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small]

of the Castle of Senghenydd may with certainty and
accuracy be located. They were found embedded in
débris of ruined buildings and covered with much later
house-refuse forming a high bank, which was often
considered to be the mound of the first Castle of

In the earliest print of Swansea Castle, made about
200 years ago, when there were few buildings on the
ground, a considerable gulley is shown between the

new Castle and the site of the old one. The latter is there represented as a veritable mound. If an artificial mound ever existed it must have been removed when the Stone Castle was built. Apparently, however, it was not required, for now that the great accumulation of débris outside the walls has been removed it is seen that the Castle stood on the highest spot of the natural ground, and on a fairly high plateau above the ground around it. If the mound is correctly shown in the print it must have been the débris of the Castle, not a mote on which it had stood.

[merged small][ocr errors][merged small][merged small][merged small][merged small][merged small]

In the recent excavations practically the whole of the ground between Castle Bailey Street and Worcester Place was opened to a depth of 16 ft. below the level of the present roadway, but a considerable portion had been removed in earlier times for the foundations of cellars to the houses which were erected on the ground. The remains of the old Castle, now disturbed for the first time, were very disconnected. At first it seemed hopeless to piece the "finds" together, and the difficulty was increased by the number of contractors employed on the excavations. But as soon as the importance of the work was recognised, the Corporation of Swansea, the

ground landlords, afforded every facility for continuing it systematically. It will be convenient to commence the description of what was disclosed at the

southern end.

The foundations, composed of large boulders, of a wall about 8 ft. thick, 3 ft. deep, were here found only 3 ft. below the surface. A very good section of the wall and ditch was obtained here. Though little more than foundations, in places portions of the wall, 3 ft. or 4 ft. high were to be seen for a length of 50 ft. Coming into view after (apparently) crossing Worcester Place roadway from east to west, it described a curve of about the eighth of a circle until its direction was northwest. Then it turned more directly towards the north parallel with Castle Bailey Street; here it was 6 ft. thick and 6 ft. high, built of excellent masonry of flat quarried stones, with a well-dressed face in a straight line for about 10 ft. At this point was found the flank of a rectangular western tower, projecting 21 ft. from the wall, with a face of 30 ft. The wall of the (western) face of this tower, owing to the fall of the ground, was over 10 ft. high; it was 6 ft. thick (about 7 ft. at the bottom). On the east lay the undisturbed ground of the interior of the Castle. What remained of the outer-the southern, western and northern-walls of this tower were remarkably perfect. To the north of the tower, the original line of the curtain wall was again resumed in very good condition for 20 ft., beyond that all traces of it were suddenly lost. It was not discovered again, although, as mentioned above, the old wall called the Castle Walls may now be considered as the continuation of this line, forming the eastern, as this was the western, bounds of the Castle.


The floor of the tower was paved at a height of about 10 ft. above the level of the bottom of the wall outside. Beneath the paving was undisturbed gravel of the locality. This paving was in places fairly perfect (though I doubt it being part of the original work), and the

« PreviousContinue »