Chapter XXI: Part 21

The visible remains of this ancient fortress, so celebrated in border story, are very scanty, and are confined to the central mound, and to some fragments of the rude but substantial keep which was placed upon it. The castle stood on ground rather higher than, and on the north-western edge of, the town, completely commanding it; and it is only of late years that its site has been nearly surrounded by buildings. The church is about half a mile distant to the south, the town intervening between the two. From the mound, the view to the east and south is very extensive; to the west it is shut in by the wooded heights of the Denbighshire border, crowned by Offa’s Dyke, at a distance of about two miles. To the north, a mile or so distant, is the detached hill and camp of Hên Dinas, or Old Oswestry, which is in part obscured by the still nearer but lower eminence of Llwyn. No water enters into the landscape, nor is there any considerable stream near at hand. Wat’s Dyke lies close to the east.

The castle mound, though standing on high ground, is wholly artificial, and rather oval at its summit, which is about 60 feet by 100 feet. It is about 30 feet high, and perhaps 200 feet diameter at its present base. On its table top are some fragments of masonry, composed of large rolled boulders, laid in a thick bed of mortar; very rude but very strong work. One fragment, which in places is about 8 feet thick, is 9 feet or 10 feet high; and near it are two other large masses, one of which at least is overthrown. The keep was of the shell type, and probably polygonal. The slopes are covered with bushes, much obscuring the surface, but there seems to be a further trace of masonry on the south-west side. The moat, out of which the mound rose, has been completely filled up, and all trace of the outer ward, its ditches, banks, or buildings, has been utterly removed. Nothing is known of a well. The entrance was probably on the south-west side. Gallowtree bank—a name preserved in the southern suburb—preserves the memory of the lord’s power, as does Oswald’s well of the ecclesiastical legend.

It is probable that the masonry remaining is the work of William Fitz-Alan, who, after 1155, confirmed a previous gift of the church of Oswestry to Shrewsbury Abbey. He died 1160, in which year the Pipe roll designates Oswestry as “Blanc-Minster.” On his death, Guy L’Estrange became custos of the Castles of Clun, Ruthyn, and Blanc-Minster, and works were in progress at the latter. The livery allowed for the garrison was £18. 5s. per annum. In July, 1165, Henry II. was encamped near Oswestry, but his advance was checked by the Ceiriog river, on which he retired to Chester.

Guy L’Estrange continued to be custos of the Fitz-Alan estates and castles until 1175, during which time very large sums were expended by the Crown upon Oswestry, apparently about £2,000. A well cost £5. 8s., palisades 40s. and £2. 6s. 8d. A house was built within the walls. The regular garrison was a knight, two porters, two watchmen, and twenty men-at-arms, costing £48. 13s. 4d. per annum. In 1188, William Fitz-Alan received Archbishop Baldwin and Giraldus in the castle. On his death, in 1210, King John stepped in, and Robert de Vipont was in charge in 1212, and in 1213 and 1214 John Mareschal and Thomas de Erdington, who had purchased the wardship during the minority of William Fitz-Alan, who, however, died under age at Clun in 1215. In 1216, King John was before Clun, and (August 16) burned the town of Oswestry, and attacked the castle, but probably without success. In 1226, Oswestry was named for the place of conference between Llewellyn and the lords marchers, but the king, distrusting John Fitz-Alan, sent Hubert de Hoese with nine knights to attend it. In 1240, John Fitz-Alan’s death placed the castle again in the care of the Crown.

In 1257, Oswestry was again named for the meeting of a Welsh and English commission, and John Fitz-Alan, the new lord, had license to levy customs dues for five years, to pay for a wall round the town. At his death, in 1267, the castle was valued at £44. 12s. 5½d. per annum, and mention is made of the “walcheria” or Welshery. The wall probably proceeded slowly, for, in 1283, King Edward issued a patent, licensing the bailiffs to levy customs for twenty years, to complete the wall. Probably this was in consequence of an attack by the Welsh, who actually held the town for a short time. In 1302, on the death of Richard, Earl of Arundel, it was found that he held Oswestry Castle and its lands by the tenure of two and a half knights’ fees. There were one hundred and forty acres of demesne land and four water-mills, but the castle was worth nothing; and its maintenance cost £10 per annum. The mills were Bailey, Weston, Cotton, and Cadogan.

PENMARK CASTLE.

Penmark Castle is about a mile east of Fonmon, and two, three, and four miles from the castles of East Orchard, Barry, and Wenvoe.

It was originally built by Sir Gilbert de Umfreville, soon after the conquest of Glamorgan, but the present ruins are scarcely older than the thirteenth century. It is probably of the reign of Henry III. or Edward I., with some trifling additions of later date; but it exhibits no traces either of a Norman keep or of any of the usual Perpendicular or Tudor additions. It is, and has been for some centuries, a complete ruin.

The castle stands along the brow of a steep bank, about 100 feet above the meads of the Duffryn brook, which, with its marshy banks, formed an excellent defence on the north or Welsh front. It is composed of two courts.

The inner court was oblong, about 70 yards by 50, of irregular figure, with a curtain upon the north side along the hedge of the bank, part of which, at each end, remains. On the south side the wall is gone, but its line is marked by its interior moat which, extending from the bank at one end to that on the other, formed the defence on the east, west, and southern fronts.

At the eastern end the wall remains, about 10 feet high on the outside. Inside is a mass of earth, no doubt covering ruins. At the north-east angle a circular depression resembles the foundations of a tower.

The west end is most perfect. Here is a curtain about 12 feet high, the rampart walk of which remains, with traces of an open stair to the battlements. At the south end the wall returns square; and here was probably a plain gateway towards the church. At its north end the north-west angle of the court is capped by a tower semi-circular to the field, but angular towards the court. It has two loops and a fireplace on the ground floor, and appended to it on the north side is a square building of two floors, containing a garderobe in each floor. This is later than the tower. Also on the west wall, near the tower, is a hollow buttress containing two garderobes. The tower opens into the court.

GROUND PLAN.

Wyman & Sons, Gᵗ. Queen Sᵗ. London.]

The outer court includes the inner one on the east, south, and west sides, terminating at each flank on the steep bank.

It contains the moat of the inner court, and was itself contained within a moat and wall, which probably included the church. Along the west face the wall may be traced to its termination in a flanking tower, evidently built for a pigeon-house.

The position of the castle is bold and striking, and although a mere ruin, and without the usual accessories of ivy and trees, it has a fine appearance from the north.

In common with most of the strongholds of the district, it is reputed to have been ruined by Owain Glyndwr.

From the castle a road descends obliquely towards the mill, placed about half a mile up the stream.

West of the castle is a considerable table-land, defended by two valleys on its north and west; an excellent place for a display of any cavalry that might be included in the garrison.

THE CASTLE OF PENRICE, IN GOWER.

The castle of Penrice, or, as it was anciently called, Penrees, in West Gower, in the county of Glamorgan, is inferior only to Caerphilly, Cardiff, and Coyty, in the area contained within its walls, and is second to none in its strong, commanding, and picturesque position. Penrice stands at the bottom of Oxwich Bay, a mile within the shore, and about 150 feet above the sea. It occupies the rocky crest of a steep slope of greensward, at the base of which is a small lake, and beyond this a sort of lagoon communicates with the bay, here fringed by a line of sand-hills.

The castle, though high, is backed on the north-east by the long and still higher ridge of Cefn Bryn, about half a mile distant. It stands upon the south-eastern end of a platform formed by the crop edges of the mountain limestone, which here is almost vertical, having been lifted up by the old red sandstone of Cefn Bryn. On the south and east the castle is defended naturally by a cliff ranging from 10 feet to 40 feet in height; on the west the ground is broken and strong; on the north it is level, or nearly so; and this, naturally the weakest side, has been converted by works into the strongest and principal front of the place.

The position and outline of the fortress have been governed by the disposition of the ground; the general design is simple, but is by no means of one date, and to the original structure there have been considerable additions. In plan the castle is composed of a principal court, or bailey, of irregular figure, and about 60 yards north-east and south-west by 80 yards east and west. Three sides are curtains only; the fourth, or northern, is composed of the gatehouse and main buildings, or _corps de logis_, and a large drum-tower with its appendages. The curtain skirts the edge of the precipice on the east and south sides. It ranges from 30 feet to 40 feet in height outside, and is about 25 feet within, and from 8 feet to 9 feet thick. It is strengthened outside by a number of half-round buttresses, of about 12 feet diameter; these are solid, and have no projection within; they rise to the height of, or a little above, the walls, and form small flanking places of arms upon the ramparts. Of these buttresses there are none on the east, two on the south, and one on the west face; there is also one larger and loftier, but still solid, capping the south-western angle. There is besides a buttress tower near the south-east angle, of larger dimensions, and hollow; but this, in its present form at least, is probably a late addition built for a dovecote, as which it is still fitted up. The wall near this tower has been broken through and rebuilt, and again broken through to make space for a rectangular building, one angle of which is seen outside the line of the curtain; the rampart wall or battlement remains, and, though more or less ruined, is original. The merlons are about four times the breadth of the embrasures, and each is pierced with a loop; the rear wall remains.

The north, or upper, side of the court is occupied by the main buildings and front of the castle, extending about 70 yards. At the east end is the gatehouse, at the west the drum-tower and its appendages. Between these two was a large square mural tower, placed in the middle of the front. Westward, between this and the drum, was no doubt the hall; eastward, a single curtain connected it with the gatehouse. The gatehouse seems to have been an addition to the curtain wall; it contains a central passage and two lateral chambers, which occupy two flanking towers of bold projection towards the north or exterior front; these are neither rectangular nor half-round, but more near to the latter figure, the angles having been rounded off. The floors were all of wood. There is but one groove, and that of very rude construction, for a portcullis, which defended the outer gate; and there seems to have been a drawbridge which let down between the flanking towers, probably over a pit, for there are no traces of a regular ditch.

The drum-tower caps the north-west angle of the castle, and projects into the court. Its clear diameter is about 36 feet, the walls being 8 feet thick, and the space within 20 feet across. It is of three floors; the basement is entered from the hall side; the first floor by two doors, one of which seems originally to have been a window. Of these doors, one opened from the hall, and one from an appendage on the south-east; there is also a loop towards the court, and a small vaulted chamber and garderobe on the east or outer side, perhaps an addition. The upper floor of the tower is remarkable; it has no entrance whatever, and no opening in the walls; it must have been reached by a trap-door. The floors were all of timber.

On the south-east side of this tower is a concentric addition, a sort of chemise, or rather chemisette, covering less than a quarter of its circumference, and projecting into the court; it is entered by a door from below the hall, and is lighted by two loops towards the court; it ends abruptly by a square wall. There was an upper floor opening into the hall and into the first floor of the drum, and there is a curious curved opening in the jamb of the lower door, evidently for a squint, with a rebate for a wooden shutter. On the north, or exterior side of the drum has been added a rectangular building of three stories, with fireplaces and chambers for garderobes. This building projects and forms the end of the north front, facing somewhat towards the north-west. From it a subordinate building, also rectangular, and containing garderobes, projects towards the west. The space between this group of buildings and the central square mural tower was contained between two walls, one still remaining, which is also the wall of the court, and another now destroyed, which formed the exterior defence on this side. As the square tower is also destroyed to its foundations, the manner in which the space was occupied can only be inferred by the openings in the remaining wall. Probably there was here, on the first floor, a hall and below it smaller rooms, perhaps cellars; there is, however, a door opening on the ground floor through the wall into a very small building projecting into the court, and which appears to have had two lateral and one end or south longer window. This may have been a chapel, but its axis, without any necessity of position, is nearly north and south. Within the court, attached to its east wall, remains a gable end, which, no doubt, belonged to a great barn, storehouse, or barrack; it was not a hall, as it had no large windows.

The ruins of the castle, though very considerable, are so devoid of ashlar and ornamental work, and the masonry throughout is so uniformly of a rude and inferior character, that it is exceedingly difficult to arrive at any certain conclusions as to the extent of the original building, or the order of construction of the parts now remaining. All that can be inferred with absolute certainty is that the round tower and the contiguous east curtain are the oldest extant parts, and, though without bond, nearly of the same date; that the north-west lodgings are much later than either; that on the north front were other lodgings now destroyed, and that the _enceinte_ wall of the court, including the gatehouse, are additions to the original building, converting what was at first an exterior curtain into an interior wall looking into the enclosure.

The curtain next east of the tower looks older than the tower, but this is probably not the case. The tower was more probably the original building, and the whole of it. It would be a refuge in case of attack, either from the Welsh or from pirates, capable of holding out until the garrisons of Pennard, Llandremor, Webley, or Llwchwr, or those more distant of Swansea and Oystermouth, could bring up assistance. The tower is probably of very late Norman date, in the transition or pointed style. The curtain eastward was probably the first addition to the tower, and is the remnant of the rectangular appendage upon its north-eastern side. This curtain has been much cut into, and contains several later openings; but the door on the ground floor next to the tower is evidently original, and of early English date. It now leads into the enclosure of the chemisette, but was obviously once an outer door, and possibly the main entrance to the rectangular building. The court was probably next enclosed, and the concentric building, or chemisette, added to the round tower, either as a prison or a cellar, the door leading into it being the old main entrance above mentioned. The gatehouse and the north-west tower are somewhat later. Recently three stones, evidently parts of a window or door-jamb, have been dug up near the round tower; these show a simple half-round bead moulding, six inches diameter, and are evidently parts of a Norman jamb, perhaps of the lower and pointed doorway of the round tower. The north-west tower has a small trefoil-headed window in the third floor, which may be of Decorated date, though the building has rather a Perpendicular aspect. The masonry generally is of a very rough character; the round tower windows within had coigns of ashlar, but the gatehouse is of very inferior workmanship, even the portcullis groove being in rubble.

This is a very curious ruin. A round tower of this early date is a very rare feature; the chemisette may be compared to the more perfect envelope at Tretower and Launceston, but it is evidently an addition. The dimensions are nearly those of Bronllys, which is, however, rather later, and of far superior work. It is singular that the original castle should have been placed so far from the edge of the cliff which constituted a part of its later defences. The excavations now in progress may yet throw a light upon the precise age of some of the parts. It is something to have found mouldings of the Norman period.

The only parts absolutely destroyed belong to the north front, where the wall has certainly been blown down by gunpowder, probably during the occupation of Gower by Cromwell.

This is not the only stronghold within the parish and manor of Penrice. About half a mile south-west of the castle, and west of, and very near to, the parish church, is a circular earthwork, composed of a single mound, with an exterior ditch, and an entrance on the north-west side. The diameter of the interior space may be 100 feet, and the mound from 6 to 10 feet high. This is a very perfect, and evidently a very early work, and no doubt was stockaded with timber.

The adjacent church, though much altered in recent times, has a Norman wall and arch, masked with plaster, between the nave and the chancel, and a remarkable south porch of early Decorated and a south door of early English date. Outside the south door of the church and in the porch, on the right on entering, is a water stoup, hollowed out in a sort of stone seat. The upper part of the tower and a small north transept are said to have been rebuilt in the last century.

The church stands upon a strong position, superior in many respects to that selected for the castle. The adjacent church of Nicholaston has a rude coupled east window of early English date.

The castle, the ruins of which have been described, was for many generations the seat of a considerable Gower family, who no doubt derived their surname from their estate, therein differing from several of their neighbours, the owners of Nicholaston, Reynoldston, Leysanston, and Scurlage, whose names were bestowed upon and are preserved in those manors. Either custom was here common, and there are several examples of each in the Vale of Glamorgan.

No doubt the lords of Penrice were followers of the Bellomonts and early Norman settlers in Gower, where they probably erected, before the close of the twelfth century, that part of the castle to which the lately-discovered Norman fragments may be assigned. The castle descended by a Penrice heiress to the Mansels, and is at this time the property of their descendant, Mr. Talbot.

PENRITH CASTLE, CUMBERLAND.

Penrith Castle stands upon a slight elevation of old red sandstone gravel, about a furlong from, and from 50 feet to 70 feet above, the church and the old town of Penrith, and a few yards east of the modern railway station. Originally and always a simple structure of no particular military strength or architectural merit, its remains are now scanty, and chiefly remarkable for the excellence of their material and workmanship, and well known from their position in full view of one of the great highways of the country.

The castle is in plan rectangular, 130 feet north and south by 153 feet east and west. It is composed of a stout and lofty curtain-wall, within and against which were built the rooms occupied by the garrison, and outside is an artificial and dry ditch, also rectangular in outline, and rather peculiar in the space left between it and the castle wall. On the west side this ditch has been removed in the formation of a deep road leading from the town to the station. On the south it remains, and beyond it are some excavations, which appear to be old, but are not very intelligible. On the east side the ditch is very perfect. About 9 yards intervene between its scarp and the foot of the curtain wall; and beyond, crowning the counterscarp, is rather a high bank, advantageously placed for those who intended to attack the wall with the weapons and engines of the Middle Ages. On the north front is a considerable platform, about half the area of the castle, and on the same level. This is defended towards the town by a natural slope, made somewhat steeper by art, but the ditch is continued in the rear of this, round the north-east angle, just west of which is a roadway indicating the approach to the place. This approach was flanked by the north-east angle of the platform, and the ground beyond the ditch.

The earthworks on this quarter are so sharp and fresh that they seem to have been deepened and strengthened to resist a modern attack from this quarter; and in that case the bank outside the east ditch may have been thrown up to protect the base of the curtain from being breached by cannon. Possibly the intention was to resist the Pretender’s army in 1745.

The curtain, of which the ditch is the outwork, is about 30 feet high and 5 feet thick. It is built of blocks of the red sandstone of the district, of a deep colour and handsome appearance. These are squared, and laid in courses. The wall appears, wholly or in part, to have been crowned by a corbel-table, of moderate projection, upon which rested the parapets. At some special points these corbels are bolder, and form machicolations, but more commonly their object was to add 6 inches in breadth to the rampart walk.

Along the east, and on the adjacent part of the north and south sides, the curtain is tolerably perfect. On the west side are some foundations, and a fragment of wall of full height. At the south-east angle a square buttress, original, of about 3 feet broad by 3 feet 6 inches projection, is set diagonally, and caps the angle. In the middle of the east wall is a similar buttress, but set square. These have no bases or setts-off, and rise to the parapets. At the north-east angle seems to have been a bartisan, or square turret, projected upon corbels; the wall is quite plain below. The other two angles are gone; but the foundations show no trace of mural towers. Grose’s plan, though very incorrect, may be taken to show that late in the last century the walls were tolerably perfect; and that in the centre of the south wall and at the south-west angle were buttresses similar to those described above.

All traces of the gateway are gone; but the earthwork shows it to have been on the north front, rather near the east end. West of this some traces of windows seem to indicate the site of the hall, and part of a large half-round barrel-vault, abutting against the remains of the west wall, seems to have belonged to its substructure. In the east wall, near its summit, are two large window openings, about 5 feet broad. The tracery is gone; but the arch of the recesses is flat segmental, strengthened by two plain bold ribs.

There is no trace of any keep or detached central building. The head of a flat-topped window of two lights, trefoiled, has been dug up and preserved. It seems to be early Perpendicular. The general plan of the building and what remains of its vault and arches, point to the Decorated period, during which both the round-headed and the flat segmental arch were largely used in this district. What remains appears to be of one date. The Ordnance sheet lviii. 4. 24. contains a plan of this castle to a scale of 10·56 feet to the mile, or ¹⁄₃₀₀.

Wyman & Sons, Gᵗ. Queen Sᵗ. London.]

In describing the building, it has been found convenient to call the tower face the north, but really it fronts more to the north-east.

It has been suggested, from the rectangular plan of the earthworks, and from the prevalence of Roman remains in the district, that the castle occupies the site of a Roman encampment, and that its material, said by some to have been taken from Mayburgh, was derived from some Roman building. This latter suggestion is, no doubt, perfectly possible. Certainly the material was not taken from Mayburgh, which never could have supplied ashlar in large quantities.

It is certain that there was no castle here during the early Norman times, or while the manor was the heritage of the kings of Scotland. From them it came to John Baliol, and was confiscated with his other English possessions late in the thirteenth century. The Neviles of Raby then held it. Richard, Duke of Gloucester, had the castle and manor, and seems to have been often here. He probably made any pure Perpendicular additions of which evidence may remain.

There are extant two licences, one to crenellate the town of Penrith, 20 Edward III.; and the other, 22 Richard II., to William de Strickland, with permission to make a mantlet of stone and lime, and to crenellate it. Probably the present castle is the mantlet referred to.

PEVENSEY CASTLE, SUSSEX.

Pevensey is, in some respects, the most interesting place in the south of England. Not only is it closely associated with English history during the eleventh and twelfth centuries, but its ancient and present names, and a part of its material remains in masonry and earthworks, connect it closely with the British, Roman, and early English occupation of our island.

It was, without doubt, the chief place in the great forest of Anderida, which, in remote times, extended over the south of Sussex and the Weald of Kent, and of which the not infrequent remains are seen in the very numerous and wild parks found in these districts. The British name, both of forest and town, is preserved in its Roman form, and the site of the latter is still indicated by its Roman walls and towers, which, like those of Porchester, have been incorporated into a later fortress. The Romans probably left Anderida in good repair. One of the chief strongholds of the “Comes littoris Saxonici” was not likely to have been neglected, and it appears, from the Saxon Chronicle, that the Britons were well aware of its value, and held it against their piratical invaders with fierce but unavailing valour. Late in the fifth century it was besieged and taken by Ælle and Cissa and their followers, and every Briton within it was put to the sword.

Andredes-ceaster, or Andreceaster, from a Roman and British, became an English fortress, and, by slow degrees, the Forest of Andredes-weald became encroached upon by cultivation. The change of masters also brought a change of name, and the island or “eye” in the marsh is supposed to have become the property of “Peofn,” whence its present name is thought to be derived.

Under the English rule, the divisions and names denoting property were gradually introduced, and Pevensey became the chief town of a Rape, an honour it shared with Chichester, Arundel, Bramber, Hastings, and Lewes, in each of which the town was distinguished by early earthworks, and at Chichester by Roman walls. Each of these Rapes had its town and fortress, and each town was placed upon a river. That of Pevensey rises in the wooded ridge about Penshurst and Dallington by a number of streamlets which meander athwart the marshy level of Pevensey, until a little above the castle they form what is still known as the old Haven, and which was in use in the twelfth century. As late as 1317 a grant by Edward II. mentions the marshes as overflowed by the sea, and as in no man’s tenure. The castle then stood on the margin of the sea, from which it is now more than a mile distant, and the whole area of the level, many square miles in extent, seems to have been an impracticable morass, covering the fortress towards the south and east, and in some degree to the north. The knoll must so have presented itself to Julius Cæsar, if, as generally supposed, he here landed, and so, with the addition of the Roman walls, it was certainly seen by Ælle and his followers, and 500 years later by Duke William when he landed between Eastbourne and Hastings, and selected the higher ground to the east of that castle for his march inland to give battle. “Mare transivit,” says the Bayeux tapestry, “et venit ad Pevensæ,” and the chronicle of Battle says he landed “prope castrum Pevensel dictum,” whence the soldiery went to seek victuals at Hastings. Here, then, it was, beneath these very walls, that the Conqueror took seisin of his yet unconquered kingdom.

Pevensey, under the Normans, became once more a place of consequence, and one of the havens through which the sovereigns kept up their communications with Normandy. It was hence that the king embarked on his return to Normandy in 1067. William granted it to his half-brother of Moretaine, who is said to have built a castle there. What he actually built is unknown, and the only existing masonry that can possibly be of his date is the broken superstructure of one, or perhaps two, Roman towers, and some rude repairs executed on the face of one of them. But whatever he did he held, and continued to hold the Roman castle till his death, and he so held it against William Rufus, in 1088, being supported and encouraged by Bishop Odo of Bayeux, and by the hope that Duke Robert would come to their relief from beyond the sea. This hope was fallacious. Rufus having stormed the mound of Tonbridge, laid siege to Pevensey. The garrison was brave, the earl-bishop and his brother count were skilful generals, the walls were high and strong, and for six weeks the king, at the head of a numerous army, assailed the place in vain. A tardy force sent, not led, by Duke Robert, strove to land beneath the castle wall, but, though the king’s English soldiers were thus placed between the double attack from the castle and the ships, they overcame and, for the most part, slew the invaders. The castle was surrendered, and Odo, transmitted under a guard to Rochester, contrived to enter that castle, and to encourage its garrison to hold out.

Pevensey next comes under notice in 1101 as the muster-place of the army led by Henry I. to repel the expected invasion of Duke Robert. He was so far successful that the invaders were driven to land at Porchester. Either under Earl William of Moretaine, who was taken at Tenchbrai and lost his lands, or under his successor, Gilbert de l’Aigle, the lordship or Lewy of Pevensey was erected into an Honour, and finally became the Honour of the Eagle, “Honor de Aquila.” In 1144, the castle was attacked by Stephen, and defended by Gilbert de Clare, then holding it for the Empress, to whose son, Henry, it had been granted by Henry I. It next came to King Stephen, and about 1216 became the property, under the Crown, of William de Warren, and after various confiscations and restorations was finally lost to the De Aquila family in the reign of Henry III. John, Earl Warren, took refuge here in 1264, after the Mise of Lewes, and in the following year it was held by Peter of Savoy for the king against the younger De Montfort. Very little of the present mediæval masonry could have been standing during these various sieges. The strength of the place must have then mainly depended upon the Roman exterior wall, furnished, no doubt, with a mediæval parapet, and dominated by the early English mound, with probably a shell keep of timber or masonry upon it.

About 1269 it fell into the hands of Prince Edward, and remained awhile in the Crown. In 1309 it was in a ruinous condition, “Confractum et male custoditum,” Edward I. having declined to repair it. It must have been soon after this, judging from the evidence of the existing masonry, that the present additions were made, that is, either at the close of the thirteenth or very early in the fourteenth century. The towers are attributed to Edward II., in 1309.

Pevensey, “La Ville et la Lewee de Peuense,” was included in the extensive grants made by Edward III. to John of Gaunt, under whom and the Duchy of Lancaster the Pelhams became hereditary constables. In 1399, Lady Pelham distinguished herself by holding the castle for Richard II. against the combined _posse_ of the counties of Kent, Sussex, and Surrey. After this event it was mainly used as a state prison. Edmund, Duke of York, was confined here in 1405, and in his will bears testimony to his good treatment; and here also were imprisoned James I. of Scotland in 1414, and in 1419 Joan of Navarre, the last queen of Henry IV. In 1650, the castle had a narrow escape from the claws of the Parliament, the materials having been actually sold for building purposes.

The history of the building, though aided by passages in the public records, is mainly to be established by the study of the material remains. Those of the Roman period have fallen under the searching and very accurate notice of Mr. Roach Smith; the present paper deals mainly with the mediæval additions both in earthworks and masonry.

DESCRIPTION.

The Roman fortress is in plan a rounded oblong, 220 yards north-east and south-west by 115 yards, and contains from 8½ acres to 9 acres. It is included within a wall strengthened by towers, and here, as at Lyme, the outline of the plan was evidently governed by that of the ground on which the castle stands, and which rises 8 feet to 10 feet above the sea level and that of the surrounding marsh or meadow. The wall is from 10 feet to 11 feet thick throughout, and at this time from 20 feet to 30 feet high. The length in circuit is about 580 yards. At the bend of the enclosure, towards the south-west, is the main entrance, preserving very nearly its original form. Two half-round towers, 28 feet apart, and 20 feet diameter, with produced flat sides, giving them a depth of 30 feet, and 30 feet high, project 20 feet from the curtain, and were connected by a cross wall, of which only the foundation remains, and in which was the outer gateway. These towers are not quite parallel, but their longer axes radiate slightly outwards. They are solid and have no internal projection, but from each of them a wall ran backwards 18 feet, terminating in a cross wall, in which was the second gateway, the foundations of which show that the base had an opening of 9 feet. This rectangular gatehouse must have resembled those of Porchester, and, like those, had evidently been altered to suit the Norman entrance, of which there remains one jamb of the outer gateway.

Besides the towers flanking the entrance are eleven others upon the curtain wall. These also are solid, of the height and age of the curtain, and without internal projection. They vary from 14 feet to 20 feet in breadth, and from 14 feet to 16 feet projection from the wall. They stand at irregular distances of from 14 yards to 35 yards. Besides the thirteen standing towers, there are two displaced and broken down, and the fragments of two others, making seventeen in all. They are closest along the north-west and north fronts, where the ground outside is highest. Towards the south and east there are but few, the shallow, muddy sea and the marsh being found a sufficient protection. Besides the main entrance there were three posterns, of which that to the north-east is still in use. That to the north is broken down, but its remains show it to have passed obliquely, on a slight curve, through the wall. The southern postern was probably a water-gate. In the south wall the mouth of a drain, about 18 inches square, and opening towards the sea, has been laid open. Although the wall is for the most part thickly covered with ivy, it is pretty evident that in parts it is still capped by a later battlement, and one of the northern towers, originally 32 feet high, bears a superstructure of 18 feet, which, from a window in its side, appears to be of the Norman period.

The wall has been breached on the north front for 65 yards, and its mass, thrown forwards, still encumbers the ground outside. To the south is a much longer breach, at least of 200 yards. This seems to have been produced by a slip of the soil, by which the foundations have been moved forward bodily for several feet. At the east end also the wall has been broken down, but here its fragments, which are of considerable bulk, are mixed up with others of later date in wild, but not absolutely inextricable, confusion. Most of the north wall has been picked into to a considerable depth, at the ground level, but the foundation remains uninjured. This is said to be the work of the purchaser of the ruins in 1650, but may possibly be the work of some early assailants, though these are more likely to have worked below the surface of the earth, and to have inserted props of wood beneath the base of the masonry, which, when set on fire, would have caused the destruction of the whole superincumbent mass. No gunpowder has here been employed.

It is evident that there was formerly a ditch at the foot of the wall on the south front, full of water, where, indeed, it may still be traced. Along the north, west, and east fronts, the wall is bordered by a road, to make which the ditch, probably never very wide or deep, has been filled up. There can be no question as to the authors of the exterior _enceinte_, both wall and towers. They are all of one, or very nearly one, date, and distinctly Roman, and, which is not always the case, the towers are bonded into the walls. The substance of the masonry is very rudely-coursed flint-rubble, chiefly composed of flints and pebbles from the adjacent sea-beach, mixed with angular fragments of stone, the whole held together by mortar very freely employed. This mortar is remarkably white in colour, and contains numerous small pebbles, little if any broken tile, and a preponderance of sand. On the whole the mortar has set firmly, and holds well together the rather heterogeneous mass of materials. The face of the work, both inside and outside, is composed mainly of squared stones from Eastbourne or Beachy Head. They are generally about 6 inches by 4 inches on the face, but sometimes as large as 12 inches by 6 inches. The lowest courses at the ground-level are composed of darker and far larger stones, and in the wall above are occasional double bands of tile, and sometimes of stone nearly of the colour of tile. A good deal of the facing at the lower part of the wall has been stripped off, but inside this stripping is confined to the part of the wall just above the ground-level, which is raised artificially higher than the level of the natural soil. In some places this addition is high enough to convert the lower part of the masonry into a retaining wall. It has been thought that the earth thus employed was derived from the inner ditch, an early English work. It may be so, but more probably the contents of this ditch went to form the mound, and it is possible that the raised soil may be derived from the ditches of a British camp preceding the Roman occupation.

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.