The western gatehouse has been more fortunate; the staircases, however,

are broken and irregular, and the vaulting injured. Through the floor

of its central apartment a hole has been broken into the vaults of the

portal, and of one of the lodges beneath.

In the floor of the Braose or triforial gallery are two large holes

which open upon a staircase and passage below.

The buildings within the court have suffered severely. The hall is

covered by a temporary roof, but the structure of its ancient roof

is apparent from the remaining corbels. The pavement has been long

removed: the sills of the windows have been cut away, and the tracery

and mouldings which adorned them are broken and defaced.

A window and door at the east end have been shattered into one, and the

vaulted passage leading to the offices is a shapeless and rugged hole.

The vaulted roof of the kitchen is broken, but enough remains to

display its original structure. The steps of the water-gate gallery

have been removed, but the vaulted roof is but little injured.

In the great court a depression in the sward indicates the ancient

well. It has lately been opened a few feet down, but nothing of

importance was discovered.

The four bastion towers of this ward deserve special notice, since it

is the position of one of them which has conferred upon this castle

much of the notoriety it possesses.

That these four towers have been mined and blown up with gunpowder, at

some period when the effects of that agent were well understood, is

evident on inspection. The mine has been sprung near the centre of each

tower, and has produced effects, differing in degree only, upon each.

That on the north-east is altogether levelled, on the outside, entirely

to the ground, crushing in its descent the very bastion on which its

foundation rested—on the inside, the door and a portion of wall, as

high as the curtain, only remain. The destruction of the north-western

tower has not been by any means so complete. Only a third of its outer

circumference has fallen, and the rest, deprived indeed of its floors,

remains as firm as ever. The portion which has fallen lies in fragments

upon the neighbouring bastion.

At the south-western tower the mine has operated outwards; the whole

of the outer portion has fallen upon the bastion and into the ditch,

but the inner strip connecting it with the rest of the building, and

containing the entrances to the several stories, has been protected by

the outbuildings on its southern side, and is unshaken.

The last, or south-eastern, is the celebrated leaning tower, the

obliquity of which has been much exaggerated, and absurdly accounted

for. In the case of this tower the mine has exploded in a contrary

direction from the rest, and the inner portion, with the adjoining

curtain, has been thrown into the court, while the outer portion

remains standing, although the force of the explosion has thrown the

mass out of the perpendicular, so that it overhangs its base, towards

the south-west, nine feet. The parapet at its summit remains quite

perfect, and is the only one in the castle that is so.

The neighbourhood of these four towers, and the intervening gatehouses,

upon which the force of the gunpowder has been chiefly employed, is

a chaos of ruins; subverted masses of the gallery, staircases, the

vaulting of large portions of the chambers themselves, lie in confusion

upon the ground; and the thin mantle of vegetation which has enveloped

them, although it adds much to their picturesque beauty, increases

in no slight degree the difficulty of accurately comprehending their

original disposition.

Throughout this immense building the iron work, even to the staples of

the doorways, has been removed; nor is there any lead to be found in

the sockets of the window-bars.

The hewn stone forming the door-frames, window-cases, newels of the

well-staircases, and in some instances the stairs themselves, have

been rudely wrenched away, with damage to the walls, for the purpose,

probably, of converting them into lime.

Portcullises, stockades, doors, with the roof of the hall, and every

particle of timber in the place, have long been removed. Every

staircase, gallery, and chamber is pervious to the rain, and exposed

to the pernicious force of the frost, yet such and so durable are the

materials, and so firm the mortar with which the whole is cemented,

that time and weather alone have produced but trifling injuries upon

the pile, compared with the wilful destruction of the hand of man.

Before arriving at any general conclusion respecting the age of

Caerphilly, it will be proper to make a few remarks upon certain

details, on which those conclusions in some measure rest.

And first of the doorways. With certain exceptions shortly to be

enumerated, the doorways throughout the building are of the same

general character. The arches are “drop,” that is to say, they are

obtusely pointed arches, whose centres lie below their spring. This

is obviously the best form of the pointed arch for the portals of a

castle, and it is that usually employed in the military structures of

the Edwardian period. With the same exceptions, the arch-mouldings

are composed of a five-sided rib, upon the front and widest face of

which a smaller rib, of the same figure, is placed. This pattern of

rib-moulding is also very commonly employed in castles.

The principal portals, together with the doors leading from the first

story of the towers upon the ramparts, are defended by portcullises,

working in a D-shaped groove. This groove passes up as a chase or

slot into the chamber above; but there is no evidence of the sort

of contrivance employed in raising the portcullis. The portcullis,

however, might have been raised by mere manual exertion, and a bar

thrust across would be sufficient to retain it securely when raised.

The sills are destroyed, so that it does not appear whether the points

of the portcullis were received into, or had worn, small holes in them.

Besides the portcullis, the larger portals are provided with a chase,

but without side grooves, intended, as is presumed, to allow of the use

of a sort of wooden frame. Also, in the main portals are four or five

square holes or _meurtrières_ in the arch, through which beams to form

a stockade might be dropped. It may be observed further that although

some of the portal passages are of considerable length, yet the ribs of

their vaults are all transverse, never passing diagonally from an angle

towards the centre, in the manner employed at Caldecot and elsewhere,

to vault a compartment of such passages.

There appears to have been more than one kind of drawbridge employed

in this castle. In some places, as at the great gate, and at the

passage in its gatehouse tower, the bridge, when drawn up, fitted into

a depression, so as to lie flush with the upper wall, from whence,

therefore, its length may be inferred. In other cases it simply rested

against the wall, making a projection. It seems always to have been

long enough, when up, to cover the gateway.

The method of hanging the bridge also varied. On the sides of some

of the portals a stone has been inserted, into which the horizontal

pivots of the bridge (of iron, from the small size of the pintle or

hole) fitted; but, connected with the place for the gudgeon or pivot

is another groove, which passes up at an angle of forty-five degrees

for a few feet, and then passes on horizontally for a few more. It

appears as though this were a contrivance, when the bridge was raised,

for throwing its lower end upwards and forwards, so as more effectually

to shield the upper part of the door, to present an oblique surface

to missiles, and, by making the bridge lean back against the wall, to

remove the strain from its chains or ropes, and to prevent it from

falling, even should they be broken. It may be, however, that into

these grooves fitted some lever, or other contrivance for working the

bridge; where they occur, there are no holes above for the passage of

the drawbridge chains into the portcullis chamber. Similar grooves are

seen in the upper gatehouse at Chepstow.

The defences of the great, or water, postern are singular. The grooves,

which in other cases form the portcullis slides, here stop abruptly

a little above the arch. They are too deep for the hinges of gates,

and were probably filled by a defence similar to a portcullis, but

which was received into a cavity below. Indeed, as there is only a

lofty wall, and no chamber above the postern, the regular plan was

inadmissible.

There is a further contrivance for the defence of a gate, consisting of

a sort of shoot, opening obliquely downwards from the sill of a window,

employed in two places in this castle; one over the door of the eastern

inner gateway, and the other over the door of the north-west principal

bastion tower; in both cases evidently with a view to the defence of

the towers when the enemy had gained the inner court.

The battlements and parapets throughout the castle are of a very plain

description. They are massy and flat topped, the coping being a rough

slab of sandstone. The height and thickness, together with that of

the rerewall and the width of the rampart walk, may be always deduced

from a careful inspection of the walls or towers against which they

terminate. The parapet and rerewall are usually of the same height, and

nearly as high as the top of the doors leading to them.

The embrasures are contained within parallel sides, and bear a small

proportion to the merlons, which latter are each perforated by a loop.

These details may be seen upon the summit of the leaning tower, or,

more conveniently, upon the northern curtain, toward the north-west

bastion tower. It is evident, from the unfinished character of the

battlements, and the flat undressed coping, that they were intended

to be masked by an exterior gallery, or brattice of wood, resting on

the stone corbels which still remain. For this reason there are no

machicolations in stone, or devices for dropping missiles through the

floor of a projecting parapet—a contrivance which adds so materially

to the grandeur of the towers of Warwick, Raglan, and Cardiff. Over

the eastern middle gateway, the parapet has a false machicolation, or

slight projection, supported upon a table of corbel blocks, but without

apertures, or a projection sufficient to admit of any.

The windows, with certain exceptions, are either loops, or, if larger,

of a very plain character. In the hall, however, and in the large

rooms of the two inner gatehouses, they are very wide and lofty, and

have been highly ornamented. The two latter rooms are so much injured,

and the windows so mutilated, that it can only be said, that what

little remains of ornament are seen resemble in style the more perfect

ornaments of the hall. The oratory attached to the eastern inner

gatehouse has a vaulted roof, divided into two square compartments,

supported by transverse and diagonal ribs. The two windows towards the

south are long and narrow, without a mullion, and trefoiled; their

mouldings are only an exterior chamfer. There are some other windows in

the gatehouses, looking towards the interior, which are much shorter,

but otherwise resemble this. The four hall windows are lofty and well

proportioned; they open to within four feet of the ground.

The exterior moulding of the windows is completely gone; that of the

door was discovered by removing the grass about its base.

The interior mouldings of the windows are extremely rich, owing to the

reduplication of the bands, to occupy the great thickness of the wall.

The angles of the mouldings are, at two depths, removed, and their

place occupied by a hollow groove, in which the pomegranate ornament

is placed at intervals, making up the circle by its projection. Beyond

each of these bands of pomegranates are pilaster strips, filleted at

their angles, and surmounted by small angular capitals: within is a

handsome ogee canopy, enriched with crockets and finials, in a very

pure style.

The door has a good internal drip, but its inner moulding is composed

of only one band of ball flowers. The outer mouldings are rich. There

are three bands of pomegranates, which no doubt were continued, as in

the windows, round the arch; and between them are two rows of small

disengaged columns, with the circular concave pedestal. Of these only

the pedestal remains.

The fourteen corbels upon which the beams of the roof rested are

composed of three short clustered columns, connected by their posterior

half, and separated by a fillet and bold hollow; above they are crowned

with a neat cap moulding, and below, they rest upon three projecting

busts, of which the central is the lowest and largest. A fillet runs

up the centre of each of these columns, and, pausing at the abacus, is

continued up the capital, and finally dies in the astragal. Corbels, of

somewhat earlier date, but in general appearance resembling these, may

be seen in the keep at Chepstow.

There are no decorations remaining about the fireplace. The plain

stringcourse along the east end of the hall, returned from the corbel

of the chapel window, is perfect. A base tablet is seen at the west end

of the north side, but it is destroyed along its length.

A long window in one of the staterooms resembles, though on a much

larger scale, the windows of the oratory already described. It appears,

however, to have been trefoiled, with a quatrefoil above the head.

There are two small polygonal apartments on either side of the inner

western gate, in the vaulted roofs of which a plain diagonal rib rises

from a corbel at each angle, and meets its fellow in the centre. The

corbels have three flat faces, and terminate in a point, which rests

upon some animal, in every case wantonly defaced. They appear to have

been lodges.

Caerphilly presents as little architectural decoration, in proportion

to its extent, as any castellated building in Britain.

Generally, its series of concentric defences, and the general

disposition of its constituent parts, resemble those of Harlech and

Beaumaris. The plan of these concentric castles is very peculiar. It is

unlike the earlier Norman castles, in which the keep was the principal

feature, and in which comfort was sacrificed to safety; and it is also

unlike the later castles, which possess not only large interior, but

large exterior, windows, as at Sheriff-Hutton, and in which there is no

building to which the name of keep could be attached.

Nor is the style of architecture employed at Caerphilly less

characteristic of its age; the drop arch, the perfectly plain rib,

the general absence of decorations and armorial bearings, the plain

battlements, and the absence of machicolations, indicate generally the

same period.

The columns of the hall doorway, the concave moulding of their

pedestals, the triple cluster of columns forming the corbels of the

roof, their bell capitals, and light cap moulding, are due to the Early

English style, which prevailed from 1189 to 1307.

On the other hand, the pomegranate moulding, the rich, though chaste

and somewhat stiff, canopies of the door and windows, the little

pilasters in the windows with the pentagonal capitals, the ogee arches,

and the plain fillet running up the columnar corbels of the roof, are

marks all belonging to the Decorated style, which prevailed from 1307

to 1377.

The mixture of these two styles, very common in English buildings,

denotes a period varying according to the preponderance of either, and

in the present instance may legitimately be referred to the latter

quarter of the thirteenth century, when the Decorated style was

beginning to supersede the Early English. Instances of this transition,

and of the ball-flower, or pomegranate, moulding, may be seen round the

inside of the choir of Bristol Cathedral, and on the outside of the

south aisle of Keynsham Church.

The earlier alterations at Chepstow, and more particularly the oratory

attached to Martin’s Tower, and the columnar corbels in the keep, may

be cited as of an earlier date than Caerphilly, having been evidently

placed there before the decline of the Early English style.

The internal evidence of the building, which would place its date about

the end of the reign of Henry III., agrees with the evidence of records

cited hereafter, in which the castle is referred to, in the year 1272,

as having been lately erected by Gilbert de Clare, Earl of Gloucester

and Hertford.

Before this period, mention is occasionally made of the castle of

Senghenydd, which, from its having been taken, retaken, and more

than once utterly destroyed, was evidently a place much contested,

but of no great magnitude or passive strength. After the erection of

Caerphilly, Senghenydd Castle is not again mentioned. It is therefore

not impossible that Senghenydd Castle was a rude fortification of

timber and undressed stone, upon the peninsula afterwards occupied by

Caerphilly.

Caerphilly having then certainly been founded by Earl Gilbert a little

before 1272, the question arises as to whether the whole of it was then

built.

The inner ward, its curtains, bastions, gatehouses, with all their

contents and appendages, are of one date. The south wall was always of

its present height, and therefore always intended to support the roof

of the hall, the walls of which are bonded into it. The gatehouses

are evidently part of the original plan, and the long windows of the

staterooms, and those of the oratory in the inner gatehouse are, in

their form and mouldings, precisely similar.

It appears that the curtain connecting the north-west bastion tower

with the west gatehouse was originally as low as the northern curtain,

but that a sort of gallery, and its superincumbent rampart, have been

added. A cluster of buildings has also been added on the outside of the

south curtain, at the angle formed by its junction with the south-west

tower.

The general design of the middle ward, and most of its buildings, are

clearly of the date of the inner ward. The western gatehouse, however,

appears to be of somewhat later date; the false machicolations, the

holes for the portcullis chains, the chimneys rising above the parapet,

and the less durable character of the masonry, seem to indicate this.

The walls, moreover, by means of which this gatehouse is connected

with the curtain of the inner ballium, though of the same age with the

former, are not bonded into, and are separated by fissures from, the

latter—a tolerably sure indication of difference of age.

It is not improbable that the whole exterior line of defence on the

east, and the horn-work on the west, were the last parts of the castle

completed. They form, however, parts of the original design, since, had

the ground on which they stand been left unoccupied, the castle would

not have been tenable.

With respect to the redoubt, it is perfectly evident, from its

appearance, that it was thrown up, not only when gunpowder was in

general use, but when the science of fortification was pretty well

understood. It seems, like the earthworks at Donnington and other

castles, to be of the age of Charles I.

The injuries received by this castle are similar to others at Corfe and

elsewhere, known to be referable to the same period of civil strife in

which the battle of St. Fagan’s, and the occupation of Cardiff, prove

the men of Glamorgan to have taken an active part. Nothing, therefore,

seems more probable than that the redoubt should have been thrown

up hastily by one party for the defence of the castle, and that the

dismantling of the whole should have been perpetrated by the other,

to prevent such a defence being practicable in future. Though history

has afforded no clue as to the one of the contending parties to which

either proceeding is to be referred, there can be no doubt but that the

blowing up of the towers was the work of the Parliament.

There seems no reason to suppose that the works of Caerphilly were

never completed. The flanking towers on either wing rest upon the lake,

and the horn-work is a sufficient defence in the opposite direction.

About three-quarters of a mile from Caerphilly, on the Rudry road, are

the ruins of the “VAN,” or “Ffanvawr,” the ancient and very stately

manor-house of the Lewis family, the direct descendants in blood, and

the heirs in heritage, of Ivor Bach, on whose land Caerphilly was

erected.

Most of the outer walls of the house, and a curious old dovecot,

remain standing. They are of the age of Elizabeth or James, but much

of the hewn stone employed in the windows, doorcases, quoins, and

stringcourses of the lower story, are either of oolite or Sutton

stone, and are very evidently a part of the spoils of Caerphilly. Most

of these stones have been worked up, and their original ornaments

destroyed, but one long stringcourse of Decorated date, evidently much

earlier than the wall in which it is embedded, extends along the west

front of the house.

These stones could not have been removed from Caerphilly earlier than

the reign of Elizabeth, in which reign, or rather in that of Henry

VIII., the castle was used as a prison. Probably the central parts were

so occupied, and the parts allowed to be spoiled were those connected

with the east front. In appropriating the stones of Caerphilly to the

erection of their manor-house, the Lewis family, from whom the ground

was originally wrested, may have committed a breach of taste, but none

had a better moral right to help themselves from that source.

CARDIFF CASTLE, GLAMORGAN.

The castle of Cardiff, though not unknown to border fame, has been the

theatre of no great historical event, nor does it present any very

striking peculiarities of position, scenery, or structure. Its claim

to more than local interest rests upon the character and fortunes of

the great barons whose inheritance and occasional residence it was from

the 11th to the 15th century, from the reign of Rufus to that of Henry

VI. Probably a Roman castrum, and certainly a hold of the local British

princes, it was won, in 1090, by the sword of Robert Fitzhamon, lord

of the Honour of Gloucester, and by him constituted the “caput” of his

newly acquired seignory of Morgan and Glamorgan.

Mabel, the heiress of Fitzhamon, conveyed his possessions, with her

hand, to Robert Consul, Earl of Gloucester, bastard son of Henry I.,

and the reputed builder of the Norman parts of the castle.

They were succeeded by their son William, Earl of Gloucester, who died

1173, leaving daughters only; the inheritance was then for a time

held by King John, then Earl of Moreton, by marriage with Isabel, the

youngest co-heir; and on her divorce, soon after 1199, by Geoffrey de

Mandeville, Earl of Essex, who died, 1216. Isabel’s third husband was

Hubert de Burgh, she died childless.

Upon this, Almaric d’Evreux, who married Mabel, the elder co-heir,

became Earl of Gloucester; but Mabel’s issue also failed.

Amice, the second co-heir, had married Richard, head of the powerful

house of De Clare; and their son Gilbert, Earl of Hertford, thus

finally became Earl of Gloucester, lord of that honour, and possessor

of the castle of Cardiff. He died 1229.

Four earls of the race of Clare possessed Cardiff Castle for nearly a

century; and though chiefly resident at Clare and Tonbridge, did much

to adorn the castle and consolidate the seignory.

In 1320, Eleanor, the elder co-heir of the last De Clare, was married

to Hugh le Despenser the younger, the minion of Edward II. During the

minority or attainder of their son, Hugh d’Audley who had married the

second co-heir, had the earldom, and possibly held Cardiff at his death

in 1347. The Despensers then reappeared in the person of Thomas, son

of Edward, who was son of Hugh and Eleanor de Clare. This Thomas was

created earl of Gloucester in 1397, and attainted and beheaded in 1400.

His son Richard, who succeeded, died a minor and childless, in 1414.

The earldom of Gloucester was not revived, but, including the first

Hugh, five members of this unfortunate race held the seignory and

castle for ninety-four years.

Isabel le Despenser, sister of Richard, and the final heiress, was born

at Cardiff Castle, which she did much to strengthen and embellish.

She married the cousins, Richard Beauchamp, earl of Worcester, and

Richard Beauchamp, Earl of Warwick. The Earl of Worcester married

before 1415-16, and died about 1421, leaving a daughter, whose

descendants became barons Le Despenser in her right, but who did not

inherit Cardiff. Countess Isabel’s chief works at Cardiff were probably

executed after her second marriage, which took place before 1425. Her

charter to Cardiff, as Countess of Worcester, in 1423, confirms those

of her paternal ancestors. Her son, Henry, Duke of Warwick, succeeded

his father in 1439, and died in 1446. His heiress, Anne Beauchamp,

had but a brief and nominal tenure of the seignory, dying in 1449, an

infant of six years.

The castle then descended to the representative of another Anne

Beauchamp, sister and heiress to the duke. She married Richard Nevile,

the great Earl of Salisbury and Warwick, who thus added Cardiff to his

already extensive possessions. One of the town charters, dated Cardiff

Castle, 12th March, 1451, was granted by Richard, Earl of Warwick, Lord

le Despenser, &c., and Anne his wife.

Upon the earl’s death, in 1471, Cardiff Castle fell to Anne, his

younger daughter and co-heir by Anne Beauchamp, in whose right her

husband Richard, Duke of Gloucester, afterwards Richard III., became

lord of Cardiff Castle and the seignory, and in the latter capacity

granted various charters and confirmations yet extant.

Upon the fall and death of Richard, the claims by heirship were set

aside and the castle and seignory escheated to the Crown. They were

subsequently granted to Jasper Tudor, Duke of Bedford; but on his

death, in 1495, again became crown property. The seignory, with its

“jura regalia” and prerogatives of marchership, was not again revived;

but Henry VII. and his son leased the lordship to Charles Somerset,

who was residing at Cardiff in 1513; and Edward VI. granted or sold

the castle of Cardiff, with much of the landed estate and the manorial

rights of the old seignory, to William Herbert, the first of the new

earls of Pembroke, in whose heirs general the whole has since remained.

The history of this long succession of powerful lords, most of whom

set their mark upon the great transactions of their age and country,

has invested the castle with something of historical interest; which,

however, can scarcely be extended to the particulars of the building

itself, the subject of the present paper.

The castle of Cardiff stands upon the broad gravel plain between the

rivers Taff and Rhymny, upon the left bank of, and two hundred yards

from the former stream, at about the lowest point at which, in ordinary

seasons, it is fordable.

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.

Within the great enclosure, near to, and a little west of, the centre

of the north bank, is the Mound, a grand earthwork, which rises from

a circular base of 60 yards in diameter to a height of 32 feet above

the surrounding ground. Its summit is a platform, also circular, 36

yards across, and crowned by the remains of the keep, called by Leland

the White Tower. This is a shell, or polygon of twelve nearly equal

sides, 80 feet in diameter, the wall being 9 feet thick and 30 feet

high, and constructed mainly of rolled pebbles. This wall is pierced by

the putlog holes, so common in the older masonry, especially of this

district; and in one of the eastern sides is a fireplace of enormous

dimensions, but the chimney of which appears as a mere recess in the

wall, with no present traces of a fourth side or front. Near it is

a sink, so that this was evidently a kitchen: a part, no doubt, of

the buildings which the wall supported. The walls have never been

strengthened by buttresses or pilaster strips; but the exterior angles

of the polygon are capped with ashlar quoins, which appear to be of

the date of the Beauchamp alterations, with some recent additions, one

of which may be the fireplace. The entrance was through a lofty and

strong gate-tower, duly portcullised on the south side, and of which a

part remains. The shell is probably the work of Robert Consul; but the

style and finish of the gate-tower testify to its being due to Isabel

Beauchamp, or her husband, of Warwick.

This gatehouse was connected with a cluster of towers of great

strength, which occupied the southern slope of the mound, and

terminated below in a second gateway, which was the first or outer

entrance from the middle ward into the keep. Here, though in an older

structure, was the reputed scene of the barbarity practised upon

Curthose, and of his subsequent imprisonment. Meyric describes the

rooms in these buildings as “not so fair as strong.” They were probably

barracks. Much of this building fell down late in the last century,

and was removed. It was, no doubt, of Norman foundation, and probably

altered both by the De Clares and by Isabel Beauchamp.

Since this account was drawn up, the ditch of the mound has been

carefully opened out, and once more supplied with water. This laid open

the piles of an early bridge, and the arch of a later one, probably of

late Norman date, with the base of a gateway at the foot of the mound,

and a well close to it. Also were laid bare parts of the steps by which

the mound was ascended, laid alongside of a cross wall as at Tickhill.

The keep does not seem to have contained any central building. A plain

stone stair, of which traces remain against the wall, led to the

battlement, which was also accessible from the gatehouse. Probably an

interior lean-to or shed surrounded the court.

On the face of the opposite fronts of the gate-tower of the keep and

of the Black Tower, are sections of the great curtain wall, which

extended from one to the other. This wall, thus seen to have been 7

feet thick and 30 feet high, and embattled on each face, was probably

not the work of the founder of the castle. It was removed late in

the last century, but the foundations remain, and have been laid open.

From Meyric’s description, and an oil painting preserved in the castle,

it appears that it was pierced near its centre by a gateway, flanked

by one if not two semi-drum towers which projected towards the outer

ward, and which, in fact, formed the real gateway to the strong part

of the castle. There was a second smaller door, a sort of postern, in

the southern drum. The foundations corroborate this view. Probably the

curtain was an addition by an early De Clare, after, and probably in

consequence of, the well-known and successful attack of Ivor Bach.

[Illustration: CARDIFF CASTLE LODGINGS. MAIN FLOOR

CARDIFF CASTLE.

G. T. Clark del J. H. Le Keux sc ]

For the completion of the defence it was absolutely necessary that

the great curtain should have been continued north of the keep. This

portion, however, has been removed. It certainly was never bonded into

the keep wall, but a corbel remaining on the north and outside of this

may have been connected with it.

The lodgings, or habitable part of the castle, form a rectangular pile,

145 feet long by 55 feet deep, which occupies about two-sevenths of the

western side of the court towards the southern end, and thus forms a

part of the outer line of defence. This pile is composed of a tower,

a central part or body, two main wings, and two lesser wings. All are

built against the great Norman wall; the tower and lesser wings outside

of it, the remainder inside.

The tower is a bold and well-proportioned octagon of 10 feet in the

side, three faces of which project from the outer or west front,

while the remainder is incorporated into, and now forms part of, the

older Norman wall. It rises from a square base of 26 feet, passing by

broaches into an octagon, to a height of 75 feet, or 12 feet above the

contiguous buildings. It is boldly machicolated, having five corbel

arches on each face, and a lofty parapet above, with two embrasures

each way, the intervening merlons, eight in number, being pierced with

a cruciform loop or oillet. The four outer or western angles, at the

base of the parapet, are capped with bold grotesque heads of animals as

gurgoyles.

The base is solid. About 6 feet above the ground it contains a rude

chamber, 13 feet square, having a barrel vault, slightly pointed, with

doorways, which seem to be original, in its northern and southern, or

gable ends. The northern door, now blocked up, was, as late as the last

century, a postern; and the chamber was a passage, and seems, from

traces of a wall, to have contained a sort of lodge, subdivided into

two cells.

Above this chamber is the cylindrical interior of the tower, 13 feet

diameter, now a mere shell, occupied by a stair, and vaulted above.

There are six windows in two tiers, the lower 38 feet from the ground.

They are almost loops, small and square headed, but boldly splayed

within, so as to give light and air, and showing the great thickness

of the wall. Their arrangement proves the upper part of the tower to

have been occupied by two chambers. It is difficult to speculate on the

use of the lower part, which must always have been dark, and is rather

large for a newel staircase. The present stairs and the groining above

are very modern, though the latter may cover older work.

The shell is original and untouched. The material is lias ashlar,

backed with rolled pebbles from the Taff. The quoins and battlements

are chiefly of a white limestone, dressed with care. This tower gives

character to the whole mass of the building. It has been compared with

Guy’s tower at Warwick, which, though of smaller dimensions, it much

resembles; and it is, no doubt, the work of the same nobleman or his

wife, possibly of the same architect, and probably was built between

the years 1425 and 1439, during which period Richard, Earl of Warwick,

was lord of Cardiff.[3] Some of the prints of the last century show a

turret rising out of the Cardiff tower; but this seems to have been a

fiction, for no traces of such a turret are found, and in that position

it would have been inconsistent with the internal arrangements of the

tower.

Immediately behind the tower is the central part, or body of the

building, about 70 feet by 30 feet, now composed of a dining-room and

entrance lobby on the main floor; a basement, with cellars and offices

below, and a range of bedrooms above. The tower is divided from this

building by the older main wall of the castle, 10 feet thick and 40

feet high, which runs through the whole, and is much cut about and

mutilated by later communications.

The present dining-room and lobby appear to have composed a hall, 62

feet long by 18 feet broad and 13 feet high, having a flat ceiling,

probably like that of the hall at Warwick. A passage cut through the

wall leads from the lobby into the tower at the foot of the stair, and

is no doubt as old as the tower. In recent times, probably by the first

Stuart, the face of the great wall has been cut away 3 feet, from the

floor level upwards, to give a width of 21 feet to the dining-room.

Also, about five years ago, a passage 32 feet long and 3 feet wide, was

cut like a tunnel through the axis of this wall, to give a way from the

tower to the breakfast-room and offices beyond.

The eastern front of this hall, which looks into the middle ward, forms

the centre of the present façade. It is divided into three compartments

by four octagonal turrets of half projection, about 4 feet in the side.

These rise to the roof. That to the south contains a stair, with an

original door from the court. The other three consist of two stages

of three windows in each, divided by a string-course. In the lower

stage the two central turrets are more ornate than the rest, and have

their angles capped with slender buttresses surmounted by pinnacles.

This tabernacle work is original in the southernmost of the two, but

was added to the other when the windows were pierced in it, and it

was cased, a few years ago. These turrets are battlemented and looped

above, and range with the regular parapet of the building, but they are

not machicolated, their structure being but slight. The stair-turret

is much older than the rest. The stair, 7 feet in diameter, rudely

restored upon an original newel, communicates with the basement as

well as with the court, and by doors, now closed up, opened into the

hall and bedrooms. It is from the position of these doors, and from

the turret windows, that the height of the old hall has been inferred.

The stair is lighted by square-headed windows, and above by a small

quatrefoiled opening. It leads up to the roof.

The three curtains or wall spaces connecting the four turrets are

also pierced by two rows of single windows, six in all—the lower

range square headed, the upper pointed. All are of two lights, with a

transom. The turret and curtain windows are all alike, save that in the

former the lower tier are pointed.

The present entrance door is modern, made by cutting down a window,

and probably all the windows have been renewed during the past

half-century. A drawing of this front in 1776 shows, however, windows

generally resembling the present, excepting that the turret second from

the north, like the stair-turret, has no large windows.

Passing into the interior, the three turrets appear as bays from the

main and upper floors, the middle one opening by a sort of passage, as

though it had been once a mere dark closet, or perhaps a staircase: the

other two open by pointed arches with plain, bold, round, and hollow

mouldings. The wall is 5 feet thick, and the passage through it is

divided by three ribs into two panels, which are continued through the

soffit. The bays themselves, 8 feet wide, with walls only 18 inches

thick, have five faces, of which the two inner ones are blank, and

the three outer pierced by the windows already described. Each of

the six angles, and the centres of the two blank sides are occupied

by a slender pilaster shaft, rising from a tall octagonal base, and

terminating in a delicate cap, decorated with a sort of trefoil. These

shafts are arranged above with some ingenuity, so as to support the

sixteen ribs of a groined octagonal roof, meeting at a central boss. In

the two northern bays and those of the upper story, this boss is a mass

of foliage, probably a very modern restoration; but in the southern

bay of the three, that next the stair-turret, it bears an original and

elaborate armorial achievement. Within a wreath formed of a vine stem,

truncated so as to represent the well-known ragged staff of Warwick,

is a shield, set anglewise, of Newburgh and Beauchamp quarterly; and

in the centre, Despenser, on an escocheon of pretence. The helmet

has large tasselled lambrequins, and upon it is placed the Beauchamp

crest of the siren’s head, ducally gorged. The whole is painted in

colours, probably after the original pattern; and it is obviously the

achievement of Richard Earl of Warwick, and Isabel Despenser, who

therefore built these turrets, as is also evident from their style. It

is, however, probable, that the ribs and groining of the central bay

were copied from the others, and added when its windows were opened,

and its walls cased or reconstructed.

It appears from Meyric’s description, and the drawing, two centuries

later, by Grose, that the entrance to the hall in the sixteenth,

seventeenth, and eighteenth centuries, was at its southern end, at the

south-eastern corner, near which were the kitchens. The turret-stair

seems then to have been either closed, or used only for communication

between the basement, bedrooms, and roof. Grose shows its outer door,

as now, partially below the level of the soil.

At the upper or northern end of the hall, on the site of the present

drawing-room, was “a fair dining chamber,” and two other rooms; and,

above these, two other stories, this being the part of the castle in

which the lord and his immediate family and attendants were lodged.

It will be observed that all these arrangements, about the existence

of which there can be no doubt, leave the most highly decorated bay

window, or oriel, and the staircase, at the lower end of the hall. This

could never have been intended when the bay was constructed, and this,

therefore, indicates an earlier and reversed arrangement.

It is well known that Earl Henry, son of the Herbert purchaser, made

considerable alterations in the castle lodgings; but what they were has

not been recorded, nor, as yet, inferred. It may, however, be safely

asserted that he actually reversed all the internal arrangements. It

is clear that in the hall of Richard Beauchamp, the south, with its

heraldic oriel, was the dais end; and this will account for the group

of buildings convenient to this end, which he raised outside of the

great wall; and thus, also, the entrance to the great tower would

open, as was proper, upon the lower end of the hall. Meyric says the

Herberts removed the flower-garden to the north from the south end of

the building, where, no doubt, it had been placed for the convenience

of the occupants of the dais. This also accounts for the turret-stair,

which gave a ready egress into the lord’s private garden, and an access

downwards into the cellar, and upwards into the first floor, where

naturally the safest and best bedrooms would be placed. Also, the

drawing of 1776 shows certain broken walls about the southern end of

the building, on the side of the modern offices, as though the entrance

of the Herberts had been accomplished by the incomplete removal of

old buildings, an inference which is strengthened by a tower shown

in the old oil painting in the castle. It is, therefore, I think,

incontestable that the dais of Richard Beauchamp, and probably of the

De Clares, was at the southern, as that of the Herberts was at the

northern end. The north wall of the former hall is modern, built by the

Stuarts; but the south wall is original, and, from the considerable

distance between it and the oriel, it is possible that there was a

small withdrawing-room cut off from the hall, into which the staircase

turret opened.

The basement is composed of one spacious chamber, or cellar, 62

feet long by 18 feet broad, and spanned by a rather highly-pointed

and four-centred vault, without ribs or groins, but of good

workmanship, and as perfect as when first constructed. This proves,

incontrovertibly, the dimensions of the ancient hall above it. It is,

of course, of Beauchamp date. At its north end is an original doorway

of 4 feet 6 inches opening, with irons for double doors, and holes,

showing that these doors were barred from within. One end, possibly of

the original oak bar, remains in its hole. There seems to have been a

similar door at the southern end; and it is evident from the old work

that the turret-stair opened into the south-west angle of this vault.

There are also two other openings, each in the old and enormously thick

wall, at the south-west angle, which may be original.

The two greater wings are evidently the work of the first Stuart owner,

sixty or seventy years ago, when, no doubt, the central wall was first

cut into longitudinally. The northern wing is an entire rebuilding of

the Herbert residence, of which nothing now remains. The wall between

this wing and the central part was built with it, and replaces the

original northern end wall of the hall of the De Clares and Beauchamps.

The southern wing is of the same Stuart date, excepting that the wall

between it and the central part is original. It is the southern end

wall of the old hall. The wing itself covers the old entrance used by

the Herberts, and stands on the site of a group of towers, shown in

the oil painting in the castle, and of which the entrance was a part.

Probably these towers were Clare and Beauchamp work, and demolished by

the Herberts. They are shown in the painting, but they are wanting in

the drawings of 1776.

There remain to be noticed two lesser wings, or groups of buildings,

placed outside the great wall, one on each flank of the building, and

each connected with the central octagon tower by a sort of corridor.

Of these wings, the southern appears to be coëval with the tower. Its

two lower stories, 14 feet square, are vaulted. The southern corridor,

which connects the wing with the octagon tower, communicates in its

first floor with the basement story of the tower, and seems to have

been the passage from the tower postern towards the interior of the

building. The present wall, however, is in part, if not altogether,

modern, and presents difficulties which have not as yet been cleared up.

The northern wing and its corridor are Stuart work. This wing is

occupied, on the one floor, by the back drawing-room, to enlarge which

above 7 feet have been cut away from the face of the great wall.

Seldom has an old wall been so severely treated, or stood the

mutilation so well. The gain, however, has been an excellent back

drawing-room and study, accessible, by means of these tunnellings, from

both the great drawing-room and the lobby.

Altogether, notwithstanding many faults of detail, the general result

of the alterations and additions made since 1775 has been decidedly

successful. The great court has been cleared, and the keep, though

deprived of its ditch and lower gatehouses, still presents a venerable

aspect, and in summer, when its surrounding thorns are in bloom, one of

singular beauty. The Black Tower has been restored nearly after the old

pattern. In the main building, the great tower which once capped the

north-western angle, has, by additions on the north, been placed in the

centre of its front; and within the court, the addition of the south,

and reconstruction of the north wing, and the opening of windows in the

blind turret, have added much to the completeness of the whole.

In place of the old hall is an entrance lobby and a dining-room, which,

with a breakfast-room and drawing-room, all of large size and excellent

proportions, a back drawing-room, study, and some smaller apartments,

form a suite, quite equal to the aspects and pretensions of the

building. The antiquary, indeed, may be permitted to regret the extent

to which the internal features have been removed or masked by modern

plaster-work and upholstery.

In speculating upon the age, absolute or relative, of the different

parts of this castle, our only guide, down at least to the age of

Elizabeth, is the internal evidence of its structure, and especially

the plans of its basement and main stories. From these it may be safely

inferred that the great west wall of the enclosure, the work of Robert

Consul or his successor, was originally continued in an unbroken line,

the Norman buildings having been in some other part of the court as

well as upon the mound. The castle of Robert Consul was probably a

rectangular enclosure, 216 yards by 84 yards, contained within three

very substantial walls, and possibly a wooden palisade, on the line

of which stood the mound and its keep; and east of this enclosure was

a second rectangular space, the outer ward, 216 yards by 116 yards,

contained within three banks of earth, strengthened by a moat on the

north, by a moat and the town gate on the east, and by a moat and the

town itself on the south, and perhaps further defended by a palisade of

timber or a mere breast-wall along the crest of the bank. The entrance,

even then, from the town was, probably, where it now is, by an archway

in the curtain opening into the outer ward; and that from the outer

into the middle ward was probably in the centre of the intervening

defence. There was certainly no tower at the south-west angle, and

probably none at the north-west; and the Black Tower also seems of

rather later date. The castle was, in fact, in two parts, the one a

mere enclosure of strong walls, and a palisade, with a circular mound;

the other, and larger part, an enclosure within earthworks.

Such seems to have been the Norman castle, calculated from its enormous

passive strength to defy any military machines likely to be brought

against it by the Welsh. The next additions were probably the Black

Tower and the cross curtain wall; and the next, with a view to the

occasional residence of the De Clares, the older part of the present

lodgings, built within and against the western wall.

The extent of this structure cannot now be determined; but it is

probable that it included the present front or east wall of the centre

of the building, the south or cross wall connecting this with the

great wall, and a corresponding north wall, destroyed by the Herberts

or Stuarts, and rebuilt by the latter. There would thus be a clear

space of about 61 feet by 18 feet for the hall; and no doubt there were

besides kitchens at the northern, and some additional buildings at the

southern end. This would give a moderate hall and lodgings, and, with

the Black Tower and the keep, afford very fair accommodation for a

baron and his train. The southern stair-turret was, probably, an early

addition to this work.

Whether the great curtain wall which divided the castle proper from the

mere earthen enclosure be regarded as coëval with the outer wall, or of

later date, the gateway in it, with the drum towers, of which a sketch

and the foundations remain, were evidently later, and probably De Clare

insertions.

The extinction of the De Clares, the division of the inheritance,

the construction of Caerphilly, and the gradual pacification of the

country, were causes which, with the long-continued misfortunes of the

Despensers, no doubt led to the partial neglect of Cardiff, or at any

rate, for a time, checked any additions to its buildings. Caerphilly,

however, once so magnificent, seems to have been found too heavy a

burthen, and to have been neglected, and the heiress who closed the

line of Despenser was born at Cardiff. Probably this fact, and the

ambitious designs of the Beauchamps, led to the partial reconstruction

of the castle; which, moreover, had, no doubt, suffered from Owen

Glendower. Richard and Isabel Beauchamp evidently built the great

octagonal tower, bonding it securely into the old wall. Connected with

this, and at the same time, they added, also outside of and bonded into

the wall, the southern lesser wing, or that towards the town gate.

Within the court they probably remodelled the lodgings, constructing

a grand vault below the hall. Also, they added three turrets to the

east wall, groining the interior of, at least, two of them as bays from

the great hall, and embellishing with their armorial shield that which

opened upon the dais. A tower, containing retiring-rooms at the south,

and probably kitchens at the north end seem also to have been additions

of the same epoch.

Within the court, upon the line of the eastern curtain, and up the

slope of the mound, the Beauchamps also seem to have constructed or

reconstructed the cluster of buildings of which a fragment only is

left. This is that ruined tower which rises considerably above every

other part of the enclosure, and adds as much to the picturesque

appearance of the castle as it formerly did to its material strength.

The Herberts, in their day, made considerable changes. They seem to

have pulled down the kitchens, or whatever buildings existed at the

north end of the lodgings, and to have replaced them by an Elizabethan

building with large mullioned bay windows. They also pulled down the

buildings on the south, and established a kitchen garden on their site.

It is probable that the Herbert work was of a much less solid character

than that which preceded it, since it has all disappeared.

Towards the close of the last century, in 1778, soon after the Stuarts

came into possession, under the advice of “Capability” Brown they

pulled down the Herbert buildings, cleared the great court, filled up

the moat of the keep (then called the Magazine), constructed the two

wings, modernised the interior of the lodgings, and left everything, in

general features, as lately seen.

It is singular, that in so important a castle as Cardiff no traces

should remain of a regular gatehouse. Leland speaks of two gates,

the Shire Hall and the Exchequer; of which the former was, no doubt,

the present gate, and the latter, probably, that from the outer to

the middle ward. That the present occupies the place of the original

entrance is pretty certain. Where else could it have been? If cut

through the earthworks, or through any other parts of the wall, traces

would certainly remain. Probably, therefore, as already stated, and as

was sometimes the case, the entrance was a mere gateway in the curtain;

and the real barrier was that from the outer into the middle ward,

which was certainly of great strength. The outer ward must have been a

place of common resort for exchequer and other public business; and the

knights’ lodgings were occupied regularly by some persons, though not

usually by the owners. The traffic attendant upon this state of things

would have made the formalities of gates, portcullis, and drawbridge,

inconvenient, and may have been a reason for the usual regular defences

of the gateway having been dispensed with. The foundations of these

buildings, public and private, have been laid open in the great court.

The only remaining difficulty relates to the defences of the

circumscribing embankment. Buck’s general view, published in 1748, and

an engraving by Ryland, show an extensive wall, covering the great

wall and the earthwork at the north-west angle, and prolonged upon the

present course of the feeder. This can scarcely be one of Buck’s common

errors in perspective, since he shows also the present wall capping the

earthen bank. It is, therefore, possible that there was, on the north

front, a wall between the bank and the moat. But, however this may be,

it must be remembered that the enemy who surmounted the earthworks

still had before him a fortress which for thickness and height of wall

was equalled by few in Britain. The chapel of the castle, the Shire

Hall, and the knights’ houses stood in the outer ward, and might be

burned or destroyed; but the knights themselves, and their followers

and effects, would take refuge and be in absolute security in the

interior parts of the castle.

Rees Meyric, writing about 1578, has left a minute, and, on the whole,

a very intelligible account of the castle as restored by the Earls of

Pembroke for their occasional residence, before the building of Wilton.

From his description it appears that the principal entrance was from

the town by “a fair gate,” having the Black Tower with its prisons

on the left, and opening into the outer ward. This ward occupied the

eastern part of the general enclosure, being separated from the inner

and middle wards by the mound and the strong curtain that extended from

the Black to the White Tower or keep.

In this outer ward Leland saw the lodgings of the twelve knights of

Glamorgan who held their lands by the tenure of castle-guard and the

payment of ward silver. In Meyric’s time there remained but one, held

by Sir Edward Mansell, and which had belonged to the Bassetts. Here

also, near the north-eastern corner, as drawn by Speed and seen by

Meyrick, stood the lord’s courthouse, used as the Shire Hall, and in

which the lord’s court for the borough was held until late in the last

century. This was protected by a special wall, upon which the knights’

lodgings stood and formed a part. A small chapel completed the group.

This chapel was granted, with the parish church of St. Mary in Cardiff,

by one of the early Norman earls to the monastery at Tewkesbury, and is

mentioned in a general charter of confirmation by Nicholas, Bishop of

Llandaff (“Dug. Mon.,” ii. 67). The space between the Shire Hall court

and the adjacent bank was occupied by gardens and orchards.

The middle ward was entered a little north of the centre of the curtain

wall, by a gateway between two drum towers, with a postern in that on

the left.

Entering the middle ward, in front was the lord’s lodging, and on the

left a stair led to the battlement, and a roadway to the Black Tower,

which road was divided by a wall from the woodyard, which, as now,

occupied the south-east corner of the court.

On the right, a way, rising rapidly, led to the keep across the ditch

of the mound and up its side. This way passed through, and was defended

by, two gatehouses duly portcullised, and was further protected by

the great curtain, under and along the rear of which it ran. The

correctness of this description has been proved by recent excavations.

The middle ward occupied all the space south of the mound, a cross wall

dividing it from the inner ward. Its west side was chiefly occupied

by the lodgings. The original plaisance, or lord’s garden, was in the

south-west corner of this ward, and was by the Herberts converted into

a kitchen garden.

The inner ward lay next, north of the middle ward, communicating with

it by a door in the cross wall. This ward was also bounded by a part

of the ditch of the keep; but it seems to have been of small area, and

not to have extended to the north outer wall, but to have been limited

by a wall which extended from the north-west angle of the keep, down

the slope, towards the north-western angle buttress of the general

enclosure. This ward contained the Herbert flower-garden, no doubt

placed there for privacy, and to be under the windows of the private

apartments, with which it seems to have been connected by an ornamental

stone staircase shown in one of the drawings of the last century. A

postern opened from this garden, through the great west wall, just

outside of the west gate of the town, and not far from the postern of

the octagon tower.

The narrow space north of the mound must have been shut off in some way

of which there are now no traces.

The main building in Meyric’s time looked, as regards its central

part, much as it did ten years ago. The south wing and the present

kitchens were wanting, and the entrance was up a few steps and by

an open terrace to the south-west corner of the pile. Entering, the

visitor stood in the hall, 61 feet long by 18 feet broad, and 13 feet 6

inches high, with a flat ceiling. On the right was a door opening into

the stair-turret; on the left another, opening through the great wall

into the Beauchamp lesser wing. Walking up the hall, the fireplace was

on the left, and beyond it a passage leading into the Beauchamp tower;

on the right, two bay windows and three ordinary windows lighted the

room, and near the centre was a closet occupying the middle turret.

At the upper or north end of the hall, doors led into the private

rooms, of large size, on both the hall and upper floors, and lighted

with large mullioned windows, looking north into the garden, and

west through a bay cut in the outer wall. In a second floor were the

bedrooms, and, above all, a flat leaded roof, commanding one of the

most lovely prospects in Britain.

This description was drawn up in 1861. Since that time the castle has

undergone great alterations; several new towers have been added, and

the interior has been completely remodelled. Over the vault of the

basement is the library, and above that the hall. The details are in

that semi-Italian style in which Burges was so great a master, and

all that a very refined taste, on the part of both the owner and the

architect, could devise, has been executed in a manner and with a

profusion which more than rivals Alnwick.

CARLISLE CASTLE.

The city of Carlisle appears first early in the ninth century, in

the history of Nennius, as Cair-Luadiit, or Luilid, or the Castra

Luguballia, one of the “octo et viginti civitates ... cum innumeris

castellis ex lapidibus et lateribus fabricatis,” enumerated by that

respectable authority. The fame of Carlisle, however, is due neither

to this early mention, nor to the subsequent gift of the place by King

Ecgfrid to St. Cuthbert, but rather to its name as a centre of the

early cycle of Arthurian romance, well supported by its subsequent

celebration in Border tales and ballads. Indeed, whether in fable or

in fact, Carlisle enjoys no mean reputation. It played a part in the

British, Roman, Saxon, and Danish occupations of the island, and, after

having been held as a frontier fortress, by the Scots against the

English, became, in its turn, the great stronghold of northern England

against the Scots, and the scourge of the wild tribes of the debatable

land.

City and castle are naturally strong. The castle occupies a bluff,

projecting towards the north, in a position which no doubt created its

early, and caused its long-continued importance. Across its front

flows the deep and rapid Eden, here seven miles from the sands, and a

score from its final disappearance in the Firth of the Solway. Above

the city, and covering its eastern flank, the Petterill comes down to

reinforce the Eden, which river, close below the city and beneath its

walls, receives, by two branches, the waters of the Caldew, which thus

covers the flank of the post towards the west.

The city, so protected on the three sides on which it was most liable

to be attacked, is built upon ground about 60 feet above the Eden,

and which, slightly rising, terminates to the north in the prominence

occupied by the castle, and which no doubt is the site of the “Caer,”

whence, whether of British or Roman origin, the city derives the first

half of its name.

Luguballia, or Caer Luel, does not stand upon, but about a mile within,

the line of the Roman wall. This great work, coming from the direction

of Wallby, and in the line of Linstock and Drawdykes, passed by

Stanwix, across the river, to terminate on the Solway, at Dykesfield,

near Burgh-upon-Sands.

The castle occupies the northern, highest, and strongest part of the

city, about 60 feet above the river. It is built upon the New Red

Sandstone rock, and to the east, north, and west the slopes are very

steep towards the meads, which fringe to a considerable breadth the

left bank of the Eden, and the right of the Caldew. In plan the castle

area is nearly a right-angled triangle, of which the right angle is to

the south-west, and the long side, somewhat convex, and 256 yards in

length, is presented towards the north and east. Of the other sides,

that towards the city on the south is 200 yards, and that to the west

143 yards in length. The space within the walls is rather under three

acres.

The outer defence towards the city is an artificial ditch, 240 yards

long, 30 yards broad, and about 10 yards deep, cut across the high

ground from slope to slope, and stopped at each end by the wall

connecting the city with the castle, which thus, though an independent

work, is made to form part of the general _enceinte_. Between the

castle and the city is an open space, about 78 yards broad, which

contains the ditch and a broad glacis, and which, with the castle,

lies outside the municipal boundary. On the west side, about 45 yards

within this boundary, is Irishgate Brow, the site of the Irish gate of

the city. The Scottish gate stood on a somewhat similar position on the

east side; and the third or English gate, guarded by the citadel, was

at the opposite or south end of the city.

The castle is composed of an outer and an inner ward, the keep standing

in the latter. The inner ward forms the eastern end or apex of the

area, of which it occupies about a fifth, and it is divided from

the outer ward by a cross wall, 90 yards long, upon the low salient

of which is the inner gatehouse. The other sides of this ward are

the east, 96 yards, and the south, 73 yards. The keep stands in the

south-west angle, about 20 feet from the two adjacent curtains, of

which the south is thrown out about 18 feet to gain space, and to form

a shoulder flanking the outer gate.

The main entrance is from the city in the middle of the south front,

40 yards west of the keep, through the great gatehouse. The drawbridge

across the ditch was removed in the last century, and is replaced

by a bridge of stone, which crosses the ditch and leads up to the

_gatehouse_, called John de Ireby’s or Irby’s tower. It opens into the

outer ward.

The plan of this gatehouse is peculiar. It may be conveniently

described as a plain structure, 44 feet square and of 20 feet

projection in front of the line of the curtain. It is composed of a

basement and upper floor, but the entrance, instead of passing, as

usual, through the centre of the building, is at its east end. The

south-east angle of the building is hollow, forming a nook or recess

of 18 feet each way, the two outer sides being walls 6 feet thick, and

about half the height of the main building. These walls are provided

with parapets, front and rear, so as to form a covered way, which

communicates with the east curtain. In front of this inclosure is

the outer gate, of 11 feet opening, with a drop-pointed arch, placed

in a sunk square-headed panel, intended to lodge the drawbridge when

lifted. This entrance leads into an open chamber 12 feet square,

commanded by its outer walls. It is, in fact, a barbican, niched in

a hollow angle of the gatehouse, with outer walls the height of the

curtain. The barbican leads to a second archway, with a portcullis in

a square groove, and a gate. Then follows a vaulted passage ending in

another gate which opens into the ward. In the passage, on the left,

is a lancet doorway opening upon a rising well-staircase, and beyond

it a drop-arched door opening into the lodge. On the right hand is a

shoulder-headed door, which leads, or did lead, into a staircase. In

the front wall of the gatehouse are two corbels, which seem to have

carried a small oriel or bartisan, commanding the approach. Appended

to the east side of the gatehouse, but entirely within the ward, is a

smaller building, fitted on obliquely, as though an addition.

Entering the outer ward, the well is seen at 40 yards’ distance. The

buildings within the ward are modern, of various degrees of ugliness,

and painfully substantial. Some are detached and harmless; others are

built into the old curtain, so as to conceal and more or less injure

it. The curtain, which is extremely curious, and most of it original,

is best seen from the outside. Besides the gatehouse, it carries but

one mural tower,—an original one, open in the gorge, in the centre of

the west front.

The _gatehouse_ of the inner ward is placed upon the salient and

central point of the cross curtain. It is called the Captain’s Tower.

It is rectangular, or nearly so, about 32 feet each way, with a

projection from the curtain of 18 feet. There is one floor above the

portal, which is central. The gateway is a low drop arch, flanked by a

pair of buttresses. The passage is vaulted, and has a door at each end,

and at the inner end also a portcullis. Over the outside of the inner

gateway is a ring of tracery, unusual, but effective. Much of this

gatehouse is Decorated, but the buttresses seem Norman.

[Illustration: CARLISLE CASTLE.

_Face p. 352, Vol. I_ ] There was originally a ditch in front of this

wall, and a gate with a drawbridge, all now gone. In later days, a

small half-moon battery was thrown up about 8 yards in advance of this

gate, and protected by a ditch of its own. This battery was connected

by a light field-work, which extended from it to the outer gatehouse,

laid in a zigzag form, so as to cover and protect a communication

between the two gates, supposing the outer ward to have been breached

and entered from the north-west side. These works have been removed and

the ground made level.

The cross wall of the inner ward is original, strong, and well built,

and backed by a ramp of earth and masonry, containing casemates, one of

which has a Perpendicular doorway. These were no doubt added, perhaps

by Henry VIII., to enable the wall to carry cannon. In the front of

this wall, a little north of the gatehouse, is a large pointed arch of

late Norman aspect, now walled up, and which may have been the original

entrance.

The _Keep_ is rectangular, 66 feet north and south, by 61 feet east

and west, and at present only 68 feet high. It is very plain. There

is the common high and stepped plinth, from which rise pilasters, 12

feet broad and 1 foot projection, two on each face, meeting at and

covering each angle, which is solid. The walls do not batter, but are

reduced slightly by one set-off, at a different level on each face.

The window-cases, though in the original positions, are not original.

The parapet has been removed, and the summit thus lowered, vaulted and

converted into a platform for guns laid _en barbette_. The south wall

is 8 feet thick, the west rather more, and the east and adjacent part

of the north wall, 15 feet. The interior contains a basement and two

upper floors. It is divided by the usual cross wall, laid north and

south. The presumption is greatly in favour of this wall being a part

of the original design; it has, however, been so much altered that it

is difficult to speak positively as to its age or original height. The

entrance is at the ground level, at the north end of the east face.

It has a portcullis, probably the work of Edward I.; from its jamb on

the left, a straight stair ascends in the east wall to the south-east

angle of the first floor, as at Chepstow and Ludlow. The basement has

been subdivided into four compartments, which are vaulted in stone. One

is a passage against the north wall, terminating in a well-stair in

the north-west angle. This stair, now disused, is original, and led to

the first floor, but probably no higher. From the passage doors open

on the left into two vaults, parted by the cross wall of the building.

The east vault is one chamber; the west subdivided by a cross wall

into two, the inner entered through the outer. The vaulting is a plain

pointed barrel, very evidently an insertion. In the smaller vaults are

late stone seats. One of the doorways is of Perpendicular date. These

vaults were evidently prisons, intended no doubt for the custody of

Border reivers, and probably a late addition. In the north wall is a

recess connected with the well.

The first floor, about 16 feet high, is vaulted in modern brick, and

used as a mess-room. In its north side was a very large fireplace,

flanked by Norman columns; but this is now walled up. This floor is

now entered solely from the south-east angle, but formerly had also a

door near the north-west corner, from the well-stair. There is also a

door near the south-west angle, which opens into a second well-stair,

which probably led to the upper floors and the battlements. This is now

closed, and there is no direct way from the first to the upper floors.

The second floor is reached, at this time, by an exterior door in the

west wall, approached by an exterior stair on the north face, and from

the rampart on the east. This door is not original, and has been broken

through at the place of a recess, probably looped, which led from the

second floor into a mural chamber and garderobe in the east wall, and

which are seen on the right hand of the door on entering. This second

floor is about 16 feet high, and has a timber ceiling. In the east

wall, over the present entrance door, is a mural chamber, on the walls

of which are some curious carvings by prisoners. One represents the

Percy crescent and fetterlock, and another a coat of arms. From this

floor a ladder leads through a trap into the upper floor—a modern

arrangement.

The third or upper floor is vaulted in modern brick to support the gun

platform above. This platform is formed of large slabs of stone, laid

down in 1812, which may also be the date of the vault. The walls above

are 11 feet thick all round.

The well of the keep is reputed to be Roman, though this is quite as

likely to be true of the larger one in the outer ward. When the keep

was built, the well, whether new or old, was included within the north

wall, between the doorway and the north-east angle, and its pipe was

carried up in the wall, no doubt with a lighted chamber at each floor,

as indicated by a line of loops still seen in the wall. To make the

well available when the keep was shut up as a prison, a hole was cut in

the outside of the north wall, near the ground level, into the pipe of

the well, and through this the water is still drawn up. The well is 78

feet deep, and its present cill is 92 feet above the sea level.

A curious external stair, probably Edwardian, has been built against

the north face of the keep, and leads up, by the well, to the ramparts

of the curtain, and so to the door of the second floor of the keep. No

doubt its original use was to lead to the ramparts only.

The keep, though much disfigured to make it carry artillery, and much

obscured by its conversion into prisons, a mess-room, and store-rooms,

is for the most part original, and if cleared, as it should be, of the

vaultings of the upper floors, would be a tolerably perfect specimen of

a Norman keep, with a full share of mural chambers and appendages.

The hall and other domestic buildings, including what was called Queen

Mary’s Tower, most of which were standing at the close of the last

century, were ranged at the south-east angle, upon the adjacent walls,

as shown in Grose’s view. All are now gone save a fragment of panelled

work, part of the shell of a grand staircase of early Edwardian date,

which led to the chief apartments.

It appears that in the east wall, near its south end, and, therefore,

under the midst of these apartments, was, in the last century, a Norman

postern with chevron mouldings and a portcullis groove, leading from

the inner ward into the field, independent of the city. These details

are shown in some of the late views of the castle. Grose, in 1774,

shows, obscurely, the position and size of the gateway, and the Norman

pilasters by which the adjacent wall was strengthened. The upper part

of this curtain seems to have been Early English. All about this angle

is now modern.

The space between the keep and its adjacent curtains has been filled up

with earth, kept off from the keep by a sort of area wall, and thus the

ramparts here, as well as along the cross wall, are made wide enough

for cannon. The upfilling is modern, and should be cleared out.

A walk, called the “Castle Walk,” or Castle Bank, has been laid out at

the foot of the curtain outside, whence its details may be conveniently

studied. The south-east angle, as has been said, is modern, but

proceeding north and westward the old Norman part comes into view,

and on the north side of the inner ward the Norman pilasters are seen

rising from a plinth, but partly concealed by six enormous stepped

buttresses of great projection, and Decorated or Perpendicular date, no

doubt a great support to the wall and very curious, but, in a military

point of view, very much in the way.

From near the centre of the north front there was a spur work, composed

of a strong curtain wall, carried down the slope and ending in a

round tower. This was of course intended to annoy the enemy should he

attack on the west side. The whole is now removed: possibly it was an

Edwardian addition. The north face of the curtain and its north-west

corner have been much restored in the Decorated period, but most of

the west wall is original. Near its centre is a small tower, like the

Alnwick Garret Tower, 28 feet broad by 18 feet deep, and about 9 feet

projection, and open at the gorge. It is wholly Norman, of the date of

the keep. It has a stepped plinth about 10 feet high, with six sets-off

of 2 inches each, and on the front face is a central pilaster, dying

into the wall at the base of the original parapet. In its north face,

high up, is the shoot of a garderobe flush with the face of the wall,

and lower down a stone water-spout. About 10 yards south of this tower

are traces of a small postern. The wall connecting the castle with

the city on this side is of Norman origin; but has some buttresses

apparently Edwardian, in one of which is a garderobe shoot, similar to

that of the garret tower. Upon this wall, south of the ditch, is King

Richard III.’s, or _Tile Tower_, 26 feet broad by 20 feet deep, of no

internal projection. This looks Edwardian, but probably is altered from

Norman. It stands about 30 yards north of the city boundary. It is

said that a few years ago a subterranean passage was discovered between

this tower and the keep, and was at once closed up. This is stated

in the “History of Carlisle, 1838,” and should be true, but it seems

improbable, for the passage must have dipped deep to pass under the

ditch.

The south wall of the castle is for the most part original, the Norman

pilasters being seen east and west of the gatehouse. East again of

these, as far as the keep, the wall seems Edwardian, but beyond that,

where it belongs to the inner ward, it is Norman.

The wall between the south-east angle of the castle and the city, about

90 yards long, is mixed Norman and Edwardian, and as it crosses the

ditch it makes a zigzag or shoulder, in which is a large round-headed

postern gateway, either original or in the place of an original

opening, intended, probably, to allow of cattle being driven on to

the esplanade and ditch from the meads on the approach of an enemy.

Probably there was a way from this gate along what is called the Lady’s

Walk, at the foot of the south wall, as far as the great gate. South

of this postern, near what is called the city stone, are traces of a

large bastion, probably a part of the city defences. The postern is now

walled up, and a bank of earth raised against it behind.

The _enceinte_ wall of the castle, being built against the natural

slope, is outside about 28 feet and inside 18 feet high. It varies from

8 feet to 10 feet thick.

The plan of the castle—a headland converted into a detached camp by

a cross ditch—may be British or English, but the general outline of

the masonry, which follows the line of the earthworks, is Norman. The

Norman engineer evidently built the _enceinte_ wall along the edge of

the slope, and planned the inner and outer ward, and the keep. The

castle is generally attributed to William Rufus, who was here in 1092,

when Carlisle, from a Scottish, became an English frontier place. The

see was created by Henry I., and the first bishop consecrated in 1133,

when probably the city wall, of which a part may still be seen below

the deanery, was built. Carlisle was taken by the Scots, and besieged

by Stephen and by John. The latter sovereign was here four times in the

years 1201-6-8, and 1212. In 1204 the constable of Chester was ordered

sixty marcs for fortifying the castle. In 1205 certain grass cut in

the neighbourhood was to be stored there. In 1215 Robert de Ros was

custos of the castle, but in 1216 Robert de Vipont seems to have been

in charge of the repairs and the garrison.

In 1222 Henry III. ordered the houses within the castle to be repaired,

and two ballistæ of horn and two of wood were to be sent there. Walter

Mauclerc was in charge. In 1222 the garrison was continued, and in the

king’s pay.

Edward I. used Carlisle in the Scottish wars, and was here in 1293,

after the great fire, which much injured both city and castle in the

preceding year. Between 1293 and 1307, he was here seven times, often

for many days. He kept his last birthday here in 1307, and went forth

hence to die in the immediate neighbourhood. To his reign are to be

attributed most of the Edwardian additions, repairs of the wall and

keep, the gatehouses, and the domestic buildings, of which only traces

remain. In 1302, Bishop Hatton, then governor, expended £275. 4s. 11d.

in works. The Great Hall, supposed to have been then erected, needed

repairs in 1344.

Camden says that Richard III. repaired the castle, and the six

marvellous buttresses may be of that date, though they look earlier.

Henry VIII. appears to have much altered the castle, probably to make

it carry artillery. He built a block-house or citadel at the south

end of the city, and armed it with cannon, and he repaired the city

walls. His work was probably done in haste, for, in 1563, the whole was

in great decay, as appears from a survey made by the queen’s order,

printed by Grose. Three sides of the keep were in a dangerous state.

The Captain’s Tower wanted parapets, as did much of the inner curtain,

and all the glass of the great hall and great chamber was decayed. In

the outer ward was an open breach, 70 feet long, where the wall had

fallen in 1557. The result of this survey was the building a chapel and

barrack, and no doubt the reparation of the wall and keep.

Mary Queen of Scots found some sort of accommodation here when she fled

from Scotland, and gave name to the lodgings lately pulled down. The

castle suffered somewhat during the great rebellion, but escaped being

dismantled. It was battered from the west, and taken by the Duke of

Cumberland in 1745.

Probably the greatest and most destructive changes are those of modern

date. The hall was taken down in 1827, and the chapel and other

buildings in 1835.

There are traces of two light field-works in the meads north of the

castle, the smaller in the rear of the other, evidently prepared for

the reception of the Scots, in 1745, as they approached over the brow

at Stanwix.

The castle is far too confined and too much a part of the city for the

purpose to which it is applied. The military should be removed, the

modern buildings cleared away, the keep restored, and the area laid out

for the pleasure of the people of Carlisle, and so as to show off the

remains to the greatest advantage.

In the neighbourhood of Carlisle are other military works deserving

notice. Such is, at Hayton, Castle Hill, a mound 12 feet high, and 100

feet diameter at the top. Linstock Castle was built before 1133, but

is now little more than a farmhouse, into which it was converted in

1768. Scaleby Castle was built by Robert de Tilliol, who had licence

to crenellate it in 1307. It was largely repaired in 1596, but retains

much of its original character, and has always been inhabited. Naworth

Castle was the chief seat of the English Barony of Gillsland, at the

Conquest granted to Hubert de Vaux, from whom it descended through the

Dacres to the Howards. The present structure was the work of Ralph Lord

Dacre, in 1335, and is a good example of the quadrangular castles of

that date. Rose Castle was in the Barony of Dalston, and is attributed

to 1336, when Bishop Kirby had a licence to crenellate. It was also a

moated quadrangle. Highhead Castle, on the Ive, is drawn by Buck. Here

was a castle in 1326, but the licence is dated in 1342. The castle was

rebuilt in 1714. Dalston Hall is a castellated house, probably of the

middle of the fourteenth century.

The Roman wall may be traced at various points, both east and west of

the passage of the Eden. It is well seen in a field close to Drawdykes,

a stiff, square farmhouse, built on the site of a Roman Castellum. The

inscription “Diis Manibus,” built into its walls, is said to have been

dug up in Carlisle, near the old citadel.

CASTELL COCH, GLAMORGAN.

The river Taff, from its origin under the Brecon Beacons, after a

course of about 26 miles through the northern and mountain district

of Glamorgan, escapes by a deep and narrow ravine across the last

elevation, and rolls its course, unfettered, to the Bristol Channel.

The ridge which it thus finally cleaves, and which divides the hill

country from the plain, is part of the great southern escarpment of

the coal basin of Glamorgan, supported there by the mountain limestone

rising from below, and in its turn reposing upon the old red sandstone,

the denuded surface of which forms, under the later horizontal rocks

and drift gravel, the basis of the plain. The escarpment, extending for

many miles along the contiguous counties of Monmouth and Glamorgan, is

traversed, in this immediate neighbourhood, by the three passes of the

Ebbw, the Rhymny, and the Taff. The heights bounding the latter river,

though in actual elevation below some other parts of the chain, produce

a very striking effect, from the abruptness of their rise from the

plain.

These heights, on each side of the pass, must always have been regarded

by the inhabitants of the country as places of great security. On the

right bank of the river, the huge lumpish sandstone mass of the Garth

rises to 981 feet above the sea, and is crowned by two remarkable

tumuli, well known as landmarks in the vale, and visible even from the

distant shores of Somerset. The lower Garth, which rises to the front

of the great Garth, opposite to Castell Coch, forms the right bank of

the pass.

The elevation on the left bank, though lower, is more precipitous.

It presents, in the lichen-stained crags about its summits, and the

rich verdure which clothes its sides and base, all those features so

well known to geologists as characterising the scenery of the mountain

limestone.

Nature has rendered the west and south sides of this height— those

exposed to any foe from beyond sea—nearly inaccessible. Across the

north-eastern side, lines of circumvallation have been hewn out of the

rock, the dimensions of which show the value attached to the place, as

a fortress, by the Cymry.

There was reason in the choice. From hence the long ships of the Danish

rovers could be seen while yet distant from the shore, and timely

notice be given, and protection afforded to, the people of the plain,

should the ravagers extend their sweep far inwards from the coast. A

beacon fire upon the headland of Penarth—celebrated in Anglo-Norman

verse for its ancient oak, and long marked by its white church—answered

here, or on the opposite Garth, would be repeated from the summits of

the distant mountains of Brecon and Caermarthen, and would at once

spread the tidings of invasion over the whole of the southern coast.

The Normans, within a century and a half after the conquest of

Glamorgan, had completed a chain of castles along the plain country,

from Chepstow to Pembroke, and were only exposed to the invasions of

the Welsh from the mountain tracts upon the north. To check these they

threw up a number of fortresses, either upon, or within the verge of,

the hill country, of which Morlais and Castell Coch may be cited as

examples.

The site of the Cymric camp was far too difficult of access to allow of

the ready transport into it of provisions, or munitions of war, or of

a constant and rapid communication with the chief castle at Cardiff.

Lower down the scarp, though still high above the plain, the Norman

engineer selected a natural platform on the limestone rock, separated

from the main scarp by a natural depression, and sufficiently removed

from the summit to be out of the reach of any military engines with

which the Welsh were likely to be acquainted, or which they were likely

to be able to bring, with their forces, against the castle. There

is an easy approach to this platform from the east, which probably

communicated with the old road, called Roman, and no doubt Cymric,

which leads direct from Cardiff to Rubina, and close upon which is

the low circular mound, which was the site of a tower of the time of

Henry III., at Whitchurch, and the Celtic tumulus of Twmpath. Upon this

platform was erected the fortress which is here to be described.

Castell Coch, so called from the red tint of its materials, is, in

general plan, a triangle, each angle being capped by a drum tower. Its

general divisions are the _south_, _east_, and _northern towers_, the

_gate tower_, the _curtains_ and _hall_, and the _outworks_.

The platform occupied by the whole is about 200 yards long by 70 yards

broad, and the principal works of the castle occupy its west end.

The south face is, in part, precipitous, and from 20 feet to 30 feet

high. The north face, towards the upper hill-side, is deepened into a

formidable moat, and the east end was defended by a fosse, cut deeply

across the rock, and beyond this by two towers, connected by a curtain

wall.

The _North Tower_ rises from a square base to a cylindrical

superstructure, the north and south angles terminating in buttresses,

each the half of a pyramid cut vertically and diagonally across, after

a fashion very common in Welsh castles, and well seen in Marten’s tower

at Chepstow. The cylinder is 40 feet in diameter. It contains three

stories, of which the middle one is on a level with the inner court, or

_terre-plein_, of the place.

The lower story may have been a cellar. It is vaulted, and has two

great cross-springer ribs, and two windows opening high above the

floor. A narrow passage, vaulted, with steps, leads into it from the

court. Its internal diameter is 18 feet, its walls upwards of 10 feet

thick. The windows were mere loopholes.

The middle story is also circular and vaulted, with similar ribs. Here,

however, the windows open nearly on the level of the floor, though also

loops. There is a fireplace, with a flue carried up in the wall. The

flue is backed with stone. The entrance to this chamber is also from

the court, and, on the east side of the vaulted passage, a gallery

passes off in the thickness of the wall, and leads to what was a small

garderobe, occupying a square projection on the east side of the tower,

at its junction with the curtain. The general dimensions of this story,

and the thickness of the walls, correspond with those of the room below.

The upper story contains one chamber, the south and east sides of which

are flat, the rest circular. Here are no less than three fireplaces,

each of large dimensions, with funnels in the thickness of the wall.

It contains also two small recesses, one a sort of sink, and has two

windows. There are also two doors, one, on the south side, opening upon

the roof and ramparts of the hall and west front, the other, eastwards,

leading to the ramparts of the great or northern curtain. Access to

this chamber, from below, seems to have been obtained by an exterior

stair between the tower and the hall. This story, within, is about 26

feet mean diameter, and the walls vary from 2 feet 3 inches to 4 feet

thick. It was roofed flat, with timber, and above were ramparts and a

parapet, probably reached by means of a trap-door in the roof. It was

evidently the kitchen, here, as at Morlais and Coningsburgh, placed in

an upper floor.

This tower is the most perfect of the whole, and in tolerable

preservation, although the lower chamber is half full of rubbish; the

small apartment connected with the middle story is broken down, and the

roof and ramparts are wanting on the summit. This tower, however, is

evidently the type of, and has served in the present instance as a clue

to, the original plan of the others.

The _South Tower_ corresponded nearly to the last, and, like it,

appears to have contained three chambers, and at its junction with

the west curtain, a square projection, containing in the middle

story a small garderobe, and in the upper, probably a communication

with the battlements of the hall. The lower chamber is entered by a

vaulted passage, down steps, from the courtyard. The middle or main

chamber probably was entered on the level, by a passage from the

courtyard, and a mural gallery seems to have led from this passage to

the window or opening in the south end of the hall. The upper chamber

was accessible from the hall battlements, as it probably also was from

those of the gateway curtain. It is uncertain whether this tower rose

from a square base—probably it did. Its upper part was cylindrical,

40 feet diameter. The walls are 8 feet thick, and the chambers do not

appear to have been vaulted. This tower is in a ruined state. The two

outer thirds of its circumference have been blown away by a mine, but

the part connected with the hall, including a door below, two windows

in the lower and middle story, and the small chamber in the wall,

remain tolerably perfect, and remove all doubts as to the original

elevation and particulars of the whole.

The _East Tower_ corresponds in altitude and general arrangements to

the other two, like them containing three stories. It is cylindrical

from the base, and 40 feet diameter; but, towards the courtyard, it

presents a flat face, with two shoulders projecting at its junction

with its curtains. Like the other towers, it has a square projection

for a small chamber, here found at its junction outside with the great

or northern curtain. The lower story, like those of the other towers,

is below the level of the court, but instead of being entered directly

by a distinct staircase, a gallery branches off from the passage to the

middle chamber, and descends, winding in the thickness of the wall,

to a vaulted prison below. This lower chamber is filled up, but its

existence is evident enough, and the staircase is seen through a great

rent in the wall. The diameter is 18 feet 4 inches, and the thickness

of the wall 10 feet 10 inches.

The first floor is entered by a passage from the court, on a level.

This chamber had two loops. There is no fireplace, and no trace of a

vault, although the walls are above 10 feet thick. On either hand,

opening out of the passage leading to this chamber, are galleries in

the wall, that on the right descending to the chamber below, that on

the left running on a level, to open into a small chamber in the square

projection between the tower and the great curtain. Here seems to have

been the well. The upper chamber appears to have been entered from the

ramparts by a long pointed doorway in the gorge; and over the lower

door, leading from this, on the right, a passage leads to a spiral

stair in the wall, which evidently gave access to the battlements of

the tower. This tower has been rent asunder by a central explosion, but

the outer part has only shifted a little.

The _Hall_ stands upon a vaulted chamber, now filled up with rubbish.

It occupies the space between the north and south towers, which it

connects, its outer wall forming the curtain between them. It is

rectangular, 30 feet 8 inches by 17 feet 8 inches, vaulted, with a

pointed arch, and having its outer wall 7 feet, and its inner wall

6 feet, thick. In the former are three loops, splayed towards the

interior, and having pointed heads. They are high above the base of the

wall, and command a fine view. The door was near the north end of the

opposite side, and possibly there may have been a fireplace on the same

side with the door. At the south end is a window, which opens into a

sort of gallery in the south tower.

Above the vaulted roof was, probably, a platform, with a low battlement

towards the court, and a high one towards the exterior of the castle.

This platform communicated with the north and south towers directly,

and with the court by a narrow stair, already noticed as leading to the

upper chamber of the north tower. The hall is now much mutilated, the

vault and part of the east wall being destroyed.

The great curtain is a large irregular segment of a circle, about 80

feet exterior face, and with a chord of about 60 feet. It originally

was a wall 3 feet thick, which appears to have been found of

insufficient breadth for the use of military engines on the north and

north-east battlements, upon these, the weakest sides of the fortress;

wherefore a parallel wall was built within and against it, 6 feet

thick, extending the whole way from the north to the east tower. The

old wall contains seven loops at the courtyard level, and to preserve

these an arch, 6 feet diameter, is turned in the new work, opposite

to each. Above, there is, of course, a rampart walk of ample width,

entered from the tower at either end. The exterior of this wall, below

the level of the court, is strengthened by a stone facing, forming the

scarp of its moat. This curtain remains tolerably perfect. There is a

breach near its junction with the north tower, and the new and inner

wall is wanting opposite to the four loops, though traces of it are

discernible in the mortar upon the old wall.

The _Gatehouse Curtain_ is much less perfect. It appears to have

been slightly convex in plan towards the exterior, and about 28 feet

in length between the south and east towers, from both of which its

ramparts were, no doubt, entered. It is about 5 feet thick. One loop

remains, about 6 feet above the courtyard level, which could only have

been used by means of a platform, perhaps of timber. Twenty-one feet

from its junction with the east tower, a small half-round tower seems

to have projected from the curtain, serving, no doubt, to defend the

gateway, which seems to have lain between this and the south tower, and

probably consisted in a simple archway and passage, with a portcullis

and doors. That the entrance was here, and between these towers, is

certain from the causeway leading to it, but the gate-tower, and most

of the curtain, are utterly gone.

Thus much of the castle. We next reach the _Outworks_.

The south and north tower, and the hall curtain, needed no exterior

defence. They rise from a very steep bank, and their foundations are

of scarped rock and solid masonry. They are quite unassailable from

below. The other two sides are more exposed. In front of the south

tower is the commencement of the moat, broken by a causeway opposite

to the inner gateway, and leading from it to the outer court. Beyond

the causeway the moat deepens, and is carried round the east tower and

great curtain, steep and deep, and hewn in the rock, so as to render

this, the naturally weaker side, very strong. The moat, which must

always have been dry, ends, opposite to the north tower, in a curious

excavation, resembling a water-tank, which, however, it could scarcely

have been.

The outer court of the castle occupies the remainder or east end of the

natural platform. Its dimensions are about 100 feet long by 40 feet

wide. Its southern side, a continuation of the line of the same face of

the castle, was defended by a precipice, partly natural, partly scarped

by art, though now broken down and filled up. There are no traces of a

wall on this side, but probably there was a parapet.

The opposite, north, or landward, side, is defended by a branch from

the moat, which, after being interrupted and traversed by a causeway,

sweeps round the east end of the works, and terminates in a deep and

broad excavation, which is carried to the brink of the cliff, and thus

defends also the east end of this outer court.

The west end of the platform, or that towards the castle, is cut off

from that building by its proper moat, traversed, as already mentioned,

by the causeway leading to the inner gateway. There is no evidence of

any walled defence to this court, and yet, without such, the moat on

the land side would scarcely have been sufficient to delay an enemy, so

as to expose him to the fire from the east tower and gateway curtain,

upon which the defence of this side depended.

As the principal object was to command the regular approach from the

eastward, the defences were prolonged in this direction. Outside, and

on the counterscarp of the moat of the outer court, and 6 feet from

the edge of the south precipice, there are traces of a _tower_, about

30 feet diameter, with what may have been a sort of buttress on its

southern side, extending to the precipice. Opposite, on its northern

side, and at its junction with a lower curtain, is what appears to have

been a well-stair, or the foundation of a distinct turret. There is no

moat to the east of the tower, but the ground falls in a natural scarp.

This _Lower Curtain_, indicated, like the towers, by a mound of earth

only, sweeps round, so as to cover the counterscarp of the outer

moat, and ends in the _roadway tower_, about 40 feet diameter, the

foundations of which are very distinct, and which must have completely

commanded the approach, at a point much in advance of, and below, the

outer causeway and the eastern tower. The regular approach, it is

clear, lay from the east, and between the precipitous height crowned

by the old Cymric camp and the level platform of the castle, and,

approaching it by the side least strongly defended by nature, would,

at 150 yards from the body of the place, be flanked by the fire of

the lower tower, then of the lower curtain, and then of the roadway

tower. Supposing these silenced, and the outer causeway reached, the

besieger came directly below the east tower, and a part of its adjacent

curtains; and as he crossed the outer court, and reached the second

causeway, he would be opposed by a fire from the east and south towers,

their curtain, and the gateway tower.

Even if the place were surprised and entered, each of the three

towers, and the platform of the hall, admitted of being defended for

a few hours, until aid had been signalled for, and sent from Cardiff.

No doubt, before brave men, all defences fail; and the Welsh, who

certainly were not wanting in courage, did, according to tradition,

more than once take this castle, probably by surprise and escalade;

nevertheless, it was a very strong fortress, both by nature and art,

and must have been a sore thorn in the side of the mountaineers of

Glamorgan.

The present condition of each part of the castle has already been

described. It may be added, generally, that the whole is very thickly

grown over with brushwood and weeds, so much so that it is difficult to

make out the details of the plan of the building. Although no ornaments

remain, yet it is clear from the general plan, and from the doorways,

that no part of the castle is in the Norman style. It is, probably,

Early English, and may very well be of the reign of Henry III., and

probably a little earlier than Caerphilly.

Here and there, especially in the outer court, are some handsome Scotch

firs, and a number of venerable beeches, the peculiar green of whose

foliage marks, from a great distance, the line of the old approach.

These, of course, with the wood clothing the hill side, and the ivy

upon the walls and towers, should be left untouched; but it is much

to be wished that the castle itself, and the works of the Cymric camp

above, were cleared of timber and underwood, and a little care taken to

encourage fair greensward in their stead.

This castle has descended with the rest of the De Clare estates, and is

now the property of the Marquis of Bute.

There are various traditions concerning it, but a great deficiency of

recorded information. Being the key of the upper country, it must have

witnessed many an inburst of the native Welsh, from the Norman Conquest

to the days of Owen Glendower, who is supposed to have descended by

this pass when he burned the episcopal palace of Llandaff, and ravaged

Cardiff.

Since this account was written the castle has been cleared of rubbish,

and completely restored, so as to be again habitable. A well has

been discovered in one of the towers, and cleared out, as has also

been a very strong dungeon. The restoration is very complete indeed,

in excellent taste, and in strict accordance with what has been

ascertained of the original structure.

CASTLE RISING, NORFOLK.