The Romans, who constructed the outer walls, seem to have been content with a single line of defence; but the Northmen treated the whole area after a different fashion. Within the area, at its eastern end, a mound was thrown up, table-topped, and about 30 feet high; this, though within the area, was upon its margin, and rested against the eastern wall. The material for the mound seems to have been derived from a ditch which surrounded about two-thirds of its circumference, extending from wall to wall, and which thus isolated it from the remainder of the Roman area. This ditch has been filled up, probably to give space, but its line is still marked by a slight depression in the soil. By this means a strong place would be formed very nearly in accordance with the early English practice, having a mound or bank, with its proper ditch, and an appended court. The only peculiarity would be that the court was walled, and thus the ditch of the mound would be traversed by the masonry, and the outer side of the mound supported by it.
The Normans, who at once saw the value and took possession of Pevensey, probably were for a time content with the Roman walls as they stood, and with the palisaded citadel of the mound. At least, there is no certain trace of any very early Norman masonry. Indeed, the only masonry of Norman date at all now to be seen is a fragment of wall with a window, the remains of a superstructure upon one of the northern towers, and some patchwork in flints, and a few courses of stone laid herring-bone fashion, by which the face of another of the Roman towers has been repaired. Had the Normans of the eleventh or twelfth centuries constructed any eastern walls, gatehouses, or mural towers within the court, some trace of them would probably remain. The chapel, indeed, judging from its dimensions, was Norman, and the base of the font decidedly so; and it is possible that the shapeless fragments of rubble masonry which encumber the top and slopes of the mound may be of the same, that is, of late Norman, date. In truth, the castle, as the Normans found it, was a very strong place. The walls only needed a battlement, and even if this were surmounted, the entrenched and palisaded mound would be perfectly defensible so long as provisions held out.
At this time the Roman enceinte contains the remains of a strong and tolerably perfect mediæval castle. This, as usual in such cases, takes the form of an addition to the defences of the mound, shutting it off as a citadel from the rest of the works. Advantage was taken of the broad and deep ditch extending from the east to the south wall, 210 yards in length, curved westwards or outwards, and which shut off the mound, and a part of the great area, and thus formed an inner ward, of about an acre and a quarter, containing the mound or keep. The ditch, which was probably supplied with water from the sea at its south end, gives off a branch northwards towards the Roman tower, called the watch-tower, and this cuts off the north-east corner of the ground, which thus forms a sort of small middle ward between the forks of the ditch. Behind the ditch is a curtain wall, near the centre of which is a gatehouse of some pretensions, and three large drum-towers, of which two flank the gatehouse, and one is placed to the north of it.
The gatehouse points to the west, opposite to that of the Roman area, now the outer ward, at a distance of 184 yards. In front of it are the two solid piers of the drawbridge, 14 feet wide, and approached from without between curved wing walls. The piers were faced with ashlar, now stripped off. The space between them is 12 feet, and may have been 10 feet The gatehouse is composed of two half-round towers, produced backwards to contain the entrance passage. Outside, these towers somewhat resemble those of the Roman or outer gate, which may have served as their pattern. Their loops are of unusual length, one being 15 feet long. The vents of two garderobes are seen, opening flush with the wall. One seems to be the lower end of a loop. They contain a basement, a ground, and an upper floor, looped towards the field, not vaulted, and duly provided with garderobes. The north tower, faced with sound, though rather open-jointed, ashlar, is still standing, though mutilated. In its ground floor is a fireplace. The south tower is quite broken down. The entrance passage is tolerably perfect, although the gateway at each end is gone, as is the upper chamber for working the portcullis, of which part of the grooves remain. The passage, 12 feet broad and 35 feet long, was vaulted with a segmental vault, strengthened with plain broad chamfered ribs, now broken away, and in the vault, a little behind the portcullis, is a large square central hole or “meurtrière” for the defence of the passage. In either wall is an arcade of two arches, a larger and a smaller, low pointed. The larger are closed, the smaller pierced by doors opening into the ground floors of the gatehouse.
Flanking the gatehouse, at a distance of 33 yards north and 54 yards south, are two grand round towers, each capping an angle of the curtain. The north curtain has a base or plinth slightly battering. The wall is vertical. There is no cordon between them. The north-west tower is 30 feet diameter, and has a basement, ground, and upper floor. The basement, though below the inner ward level, is on the level of the ground outside. It is arcaded, having six arches in its rounded sides, and one in its flat end or gorge. These arches have a drip of the double-scroll pattern, and between each pair springs a moulded rib, and one from each of the two right angles, eight in all. They are broken away, but their profile is seen, and the plan of the vault may be inferred. The entrance to this chamber is by a straight staircase from the inner ward, and at the foot of the stairs is a lobby on the left or west side leading to a postern doorway placed at the junction of the tower with the curtain. In the gorge wall is a fireplace, the hood of which seems to have been of timber. It is difficult to understand what this chamber can have been intended for, with its ornate details and a fireplace, and yet half under ground.
The ground floor is entered from the inner ward by a separate entrance, in the right or east wall of which an opening passes into an oblong mural chamber, vaulted, and contained within the curtain. This chamber has a water drain, and above it, in the wall, three bold corbels, and above these a small segmental-headed doorway, now blocked up. This is a very peculiar arrangement, and it looks as though there had been a wooden structure, perhaps a garderobe, bracketed out upon the face of the wall, over the ditch, at about 10 feet from the ground. The upper floor of the tower was entered from the battlements, the tower rising above the wall. Only the basement floor was vaulted. Each stage is lighted, or rather ventilated, by loops towards the field.
The southern flanking tower is nearly upon the pattern of that last described, save that the basement is not arcaded, and none of the floors vaulted.
The third tower, in the north wall, 36 yards from the north-west flanker, is of the same pattern, with the same exceptions. The staircase into the basement has a side door opening upon a postern in the east wall, with a segmental head, and from the ground-floor entrance there opens, westwards, a long mural chamber, the counterpart of that described as attached to the north-west flanker, having also brackets and corbels, and a small door in the wall, 10 feet above the ground, as though for a timber garderobe. These are all the regular towers, but in the south wall, where a tower might be expected, is a postern, which pierces the wall at its junction with the Roman wall, and outside and in front of it is a fragment of a Roman tower, which has slipped forwards a few yards, and forms a sort of bulwark concealing and protecting the postern. It is evident that the displacement of the tower is older than the Norman period, and was taken advantage of by the later builder. In the north wall of this inner ward, beneath the north and north-west towers, is a large fireplace, perhaps that of a hall. The kitchen was probably in the north-west tower, and the large mural drain was connected with it.
The mound, which occupied the east end of the castle and carried the keep, remains tolerably perfect, though much encumbered with ruins, produced evidently by gunpowder. Against its east side, and supporting the mound, is a Roman tower, which was worked into the keep, its solid top being battlemented. The mound has a spur of earth towards the north, probably connecting it with the north wall, but nothing definite can at present be ascertained, though the foundations upon it, if laid open, would probably disclose something.
As the mediæval castle is placed within the eastern end of the Roman area, its eastern side osculated with the Roman boundary, which is here common to both areas. About four-fifths of the enclosing wall of the castle is mediæval, but the remainder, that towards the east, is Roman, which is thus common to both fortresses. Commencing at the south postern, where the two walls are in contact, to the great disadvantage of the masonry of later date, the Roman wall extends, partially propped up by a later buttress, until it reaches a Roman tower that connected it with the keep mound. Beyond this, passing southward and eastward, to where the mediæval wall springs from the Roman enceinte, the Roman wall has been left to support the mound, but about 6 feet in front of it a mediæval wall, 9 feet thick, has been built, probably to afford more space above, and to assist in supporting the earthwork. The castle has been attacked on this side, or else those who dismantled it, thinking this the strongest part, have mined and blown it up, for the glacis for many yards is covered with enormous masses of masonry, which have evidently been displaced by gunpowder exploded in large quantities. At one part, abutting upon the Roman tower, the two walls are seen. About 20 feet of the mediæval wall, 9 feet thick and 10 feet high, stands undisturbed, though above this height its superstructure has been blown off. Behind it is the Roman wall of about the same height, not only reduced in height by the explosion, but tilted forward. This is what has happened in this quarter, and the history of it is clear, even in the midst of so great a confusion.