The position, having a river in front and rear, and the sea close upon the southern flank, is such as would be selected by a commander skilled in the art of war, and enclosed in an enemy’s country; and such as, with disciplined troops, would be impregnable.
These conditions, the name of the place, and its position upon the well-known “via maritima,” are suggestive of a Roman origin; an opinion, indeed, but moderately supported by scanty discoveries of Roman remains, but in unison with the form and character of a part of the earthworks which enter into the composition of the present castle.
These appear to have been a single lofty bank raised from an exterior ditch, and enclosing much of three sides of a quadrangular space, of which the fourth lies open towards the river; a practice by no means unusual in Roman or “quasi-” Roman encampments. It is possible, though scarcely probable, that the earthwork was once complete on the three sides down to the river, and that the south-western part was destroyed to form a very considerable mound, which still remains towering over the banks of the enclosure, and is crowned by the shell of an early multangular keep. Wallingford, Tamworth and Wareham are instances in which a quadrangular earthwork has been in part retained and in part removed by the builders of a Norman castle, and in each of which also there is a mound.
Whatever may have been its origin, the castle of Cardiff covers a plot of ground nearly square in plan, being 200 yards east and west by 216 yards north and south; and bounded on the north and east, and partially on the south, sides, by banks of earth, and on the west and remainder of the south side by a wall. The banks are about 30 feet high, 90 feet broad at the base, and 12 feet at the summit, along which runs a light embattled wall about 6 feet high and 2 feet thick. This wall is mentioned by Meyric, in 1578, as in decay, and it was rebuilt from its shallow foundation of only 2 feet in 1861. It cannot have been intended as a serious military defence, and the recent excavations did not reveal any traces of an older or more substantial work. At the south-east, north-east, and north-west angles, the banks are strengthened, possibly to carry towers, of which, however, no foundations have been discovered. The earthwork is returned about 70 yards along the south, and about 30 yards along the west, fronts, to give support to, and cover the commencement of, the walls of those sides which, with an inconsiderable exception, are evidently very ancient, and were probably executed by Robert, Consul or Earl of Gloucester.
These walls are magnificent works, being 40 feet high and 10 feet thick, and perfectly solid.
The main buildings of the castle are in the line and form a part of the west wall. In the centre of the south side is the gateway, once of the outer ward, now of the general inclosure, a mere arch in the curtain, and, in its present form, probably of the age of Henry VIII., whose arms may have occupied a square stone frame remaining above and on the outside of the gate.
Close west of the gate is a lofty tower, apparently of early English or early Decorated date, and restored by the late Lord Bute. This is the black tower. Though so near the gate it is clear that it never had any direct communication with it, nor was intended as a gatehouse. Meyric describes it as a great tower, some stories high, and covered with lead, with two chambers in each story, the lowest being prisons, known as Stavell-y-Oged and Stavell Wen. One, the larger of these prisons, now disused, has a pointed vault and a small loop, high up in the east wall. This tower has four original entrances: one to each of the basement prisons, one from the remains of the great curtain, and one from the ramparts of the castle wall. All are on the north and western faces, and there is no doorway opening towards the great gateway. It is thus clear that this tower, though placed close to the gateway, was never used as a gatehouse.
The northern bank of the general enclosure presents a slight angle outwards, and near the salient a tunnel has been cut through it. This was done about thirty years ago, to give a carriage drive towards the Senghenydd road.
Outside, at the foot of the bank, along the north, south, and east fronts, was a wet ditch, anciently fed by the Taff through the intervention of the mill leat. This moat covered the three fronts, extending as far as the north or Senghenydd gate of the town. In the time of Meyric it was dry and silted up. More recently the eastern portion has been to some extent superseded by the Glamorganshire canal, and the northern, at a lower level, is now a part of the feeder by which the river water is conveyed under the canal to the Dock reservoir. The southern arm has been filled up and built over for many centuries, and its existence is only known from the soft, black soil found in occasional excavations. The mill leat which supplied the lord’s mill, at which the people of Cardiff were bound to grind, and which was occasionally used to flood the low ground for purposes of defence, still runs along the west front of the castle. In the reign of Elizabeth there were three grist-mills and a tucking-mill dependent upon the castle, and one grist-mill was standing in 1660, but was afterwards replaced by a tanyard, removed in 1861, when the new lodge and town bridge were erected. The bridge then destroyed was built about 1796, and replaced a structure of four stone arches, probably of the age of Elizabeth (and referred to in certain Acts of Parliament), placed rather above the castle; so that the high-road from it, towards the west gate of the town, crossed the marsh by a causeway and the leat by a bridge of three arches, and defiled close under the main buildings of the castle. Part of the old gate of the town, with its iron gudgeons, may still be seen in the brushwood a few yards in advance of the great tower, showing how the road entered the town under the castle wall. A small Tudor archway still remains on the right of the old entrance, nearly in a line with the Red House, a building of considerable antiquity, now much altered, and known as the Cardiff Arms Hotel. Recently some excavations between this building and the castle disclosed the foundations of the old town wall, and a large arched passage which might be either a postern or a sewer. It was not followed up.
The area within the castle wall is about 10 acres, and within the counterscarp of the moat about 13 acres.