The plan of the castle is rectangular, and its earthworks are upon a scale not usual with castles of pure Norman origin, and which, notwithstanding their form, raises a surmise that they may be of much earlier date.
The main ward of the castle is about 200 feet square, level, and contained within a deep and broad ditch, completely surrounding it. A moderate bank of earth crowns the edge of the slope, partly, no doubt, original, but in part composed of the ruins of the curtain.
Beyond the ditch, and forming its counterscarp, is a ridge or bank of earth a few feet lower than the level of the ward, and therefore commanded by it. This ridge, interrupted at the southern angle, so as to communicate with the excavations beyond, is expanded upon the north-west front and again still more considerably upon the opposite or south-east front, so as to form a lesser and a greater barbican, covering the two entrances to the place.
Again, beyond the ridge is a second ditch, also encircling the whole place, and passing, therefore, in front of the barbicans. This, in its turn, is succeeded by a second or outer bank, also interrupted, and in three places, at the north, south, and east angles. This also is of variable breadth: somewhere, a mere ridge; in other places, as before the two fronts, expanded into broad platforms, covering the entrances and the approaches to them. Supposing these ditches to contain either water or mud, the interruption in the banks would very much increase the difficulties of those besieging the place, by breaking the communications, and preventing them from attacking the barbicans by the flanks. There are traces, outside this second bank, of a third ditch, which, however, seems to have been confined to such points as were supposed especially to need further protection.
It is evident that an earthwork such as that described, covering above ten acres of ground, and with ditches 60 feet to 70 feet broad, and deep in proportion, would, in resolute hands, and properly palisaded, be a most formidable stronghold. That such was the nature of the defences contemplated seems certain, since the banks would not support masonry; and although the edge of the inner ditch was of firm ground, the bank or crest thrown up upon it was not, and had masonry been contemplated, would have been superfluous.
When the Norman engineer undertook to fortify the place, he seems to have confined himself to the construction of a curtain of 10 feet thick round the inner area, placing it on the firm ground, and employing the earth-bank as a ramp against the wall; to this he added a gatehouse at each end, and a work of some strength as a barbican beyond the inner ditch; then a second gatehouse, placed upon the barbican; and finally a second or outer bridge. On the west side of the inner area, where the rock was firm, a low cliff of 20 feet to 25 feet was substituted for the slope of the ditch.
Besides the enceinte or curtain-wall, the four drawbridges, and probably four gatehouses, there seem to have been four drum-towers—one capping each angle of the place. It is true that these are no longer to be seen; but a circular heap of rubbish at each angle seems to represent such towers which, indeed, were the usual and necessary constituents of such a work, though whether these towers were of the Norman period may be doubted: probably they were later.
In addition to these works, and reared up high above them all, was the rectangular keep, placed near the centre of the east or town side, and upon and forming a part of the enceinte. Opposite to the keep, and also forming a part of the enceinte, but on the west side, and built therefore upon the edge of the cliff, were the domestic buildings, some parts of which remain mixed up with later works. It may be also seen that a cross curtain between the keep and the domestic buildings divided the ward into a northern and a southern court, and it would seem that a fragment now standing, and which has much the air of having been part of a chapel, was connected, as at Knaresborough, with this wall.