sign in sign up
<<
Home , Bailiwick, Cefnllys, Chepstow Castle, Grosmont Castle, Montgomery Castle, Skenfrith

White and the Dating of the Towers

White Castle, . Built by Hubert de in 1229-31, and the Outer ward circuit towers in 1234-9.

White Castle and the dating of the cir- and 1234-1239. For those who want to cuit towers. know the full references for the facts and figures displayed here they need to refer to Paul Remfrey makes a detailed case for my works on all three of the Trilat- dating the towers to 1229-31, and 1234-39, eral. built by Hubert de Burgh. The first thing to be done is to The dating of castle masonry is always a examine the history of the site, also known hazardous business. In the case of White as Llantilio castle, from 1219 to the end of Castle the surviving evidence appears quite the century by which time most of the clear as to when major building work took refurbishment of the castle would have place in the thirteenth century. This is inter- been completed. During the period 1219 to esting in itself as it shows that the round 1232 the castle was held by the , towers and of the fortress are Hubert Burgh (bef. 1180-1243). His impor- part of the early twin-towered tant career has been much studied and it is that graced Britain in the first third of the certain that from the time of King John thirteenth century - the re-dating of the (1199-1216) affairs of state must have tak- inner gatehouse at Pevensey castle being a en up much of his time. However the de- good case in point. To investigate White marcation between government and private Castle thoroughly it has been necessary to business was not clear-cut in this era. In look at all of the Trilateral and August 1220 the government came to their history. Here the salient points are and there is substantial circum- reiterated to show the most likely conclu- stantial evidence that this castle was mas- sion that the main masonry works at White sively rebuilt between then and 1224. Castle date to the time period 1229-1231 Between 1224 and 1227 it was the turn of

The Castle Studies Group Journal No 24: 2010-11 213 White Castle and the Dating of the Towers

Grosmont castle to have major stone addi- works at Llantilio as previous orders for tions to its defences. By 1227 the Consta- quarrels seem to have occurred at the com- ble's gate at Dover castle, one of Hubert's pletion of work at Skenfrith and Grosmont major projects, was complete and the castles. If Hubert had begun building the building of the FitzWilliam Gate there was new gatehouse and towers at White Castle well under way. Finally in the period 1229 in the spring of 1228, it is possible that they to 1232 and 1234 to 1239 the circumstan- were nearing completion in 1230. tial evidence suggests that the defences of On 29 July 1232, an exasperated Llantilio (White) and Hadleigh castles King Henry III ordered Earl Hubert to sur- were overhauled as Hubert Burgh's final render all his castles to the Crown. Simul- military building projects. taneously Hubert was replaced as justiciar During 1229 Hubert visited the by Stephen Segrave and a few days later Trilateral, the king authorising the expend- the king demanded of him an account of all iture of 1s 6d for the cost of a messenger his expenditures in this reign and the last ‘going to the king's justice at Skenfrith’. and all the money that ‘had been wasted Although this suggests that Hubert was either in war or in any other way’ and also residing at the most hospitable of his three of the liberties by which Hubert enjoyed Gwent castles, it is also possible that Hu- his lands. To this Hubert replied that he bert was here to oversee the initiation of the held a warrant from King John which re- refurbishment of the defences of White leased him from giving any account of any Castle, which would appear to have started monies he had received or used. The bish- about this time. A year after the possible op of Winchester correctly replied that this commencement of the new works at White writ had ended with the life of John. Henry Castle, on 20 November 1230, Hubert was then charged Hubert with various pieces of granted ceremonial permission to construct malicious gossip and Hubert was forced to a castle at Hadleigh in Essex. It would abjure the reins of state. Few it would therefore be most surprising if Hubert was appear were sorry to see him go, until the to upgrade the defences of only three of his colour of the new regime became increas- four main castles. It is even more surpris- ingly apparent. ing to accept this when a contemporary On 19 September 1232 control of chronicle specifically states that Hubert White Castle and the Trilateral were added spent a fortune on all four of these castles to the burgeoning possessions of Peter Ri- and not just on three of them. vaux, the nephew of Bishop Peter les Ro- On 2 May 1230 William Braose ches of Winchester. The next year, on 19 of and , whose father July 1233, the king further granted Peter Reginald had held the Trilateral between Rivaux, the treasurer of Poitiers, the man- 1215 and 1219, was hanged by Prince Lly- ors of Grosmont, Skenfrith and Llantilio welyn. In July 1232, the king, following (Lentilliok) in with their castles for malicious gossip, claimed that Hubert him- the service of two ' fees. By this self had sent the prince information about time war had almost broken out in the William’s intrigues. On William’s death Welsh borders between on the one part the Hubert had added the old Braose estates to followers of Walter Clifford and the new his own holdings and in the increased ten- Earl Marshall and on the other the govern- sion resulting from these acts the constable ment of Bishop Peter les Roches of Win- of St. Briavels castle was mandated to let chester. As Henry III and his army the constable of Skenfrith have 300 quar- marched into Wales at Hay on Wye on 31 rels on 27 May 1230. This may have August, the of was or- marked the completion of some building dered to receive 10,000 quarrels from the

214 The Castle Studies Group Journal No 24: 2010-11 White Castle and the Dating of the Towers constable of St Briavels and have them a modern educated guess, which, as I hope taken to White Castle. Such an is to show, is unsupported by an historical instructive as to where the king thought so evidence. many quarrels might be necessary. In time On 14 October 1241, Walerand of peace only 2,000 quarrels were sent to Teuton, the royal keeper of Skenfrith, Skenfrith in 1224 and 4,000 probably to Grosmont and White Castle, was allowed Grosmont via Skenfrith in 1226. This per- various expenses from the lands he held in haps shows that White Castle was seen by bail. These included 7s 9d on repairs car- contemporaries as the major fortress of the ried out on the three bridges in the outer three, the other sites being more residential. bailey of White Castle; 5s 7½d for two This would suggest that White Castle was brass pots and pans for the king's kitchen of both defensible and powerful at this time. White Castle and finally £1 for ten wooden Five years after returning White crossbows bought and placed in the garri- Castle and the Trilateral to Hubert Burgh, sons of the Three Castles. These moneys King Henry III again moved against his old were on top of the sixty marks (£40) allow- mentor. This resulted on 29 October 1239, ance Walerand had to maintain the castles. in Earl Hubert Burgh of Kent submitting Over two years later on 20 May 1244 the himself to the king's liability and grace and king allowed Walerand, from the issues of surrendering to him the three castles of the lands of the Trilateral, £15 10s spent in White Castle (Blancum castrum), Gros- making a hall for the king at White Castle mont, Skenfrith and the castle of Hadleigh together with a pantry and a buttery. This (Haetfeld) with its and park. In return probably marks the construction of the for this the king granted Hubert and Marga- buildings along the internal face of the ret his wife, for the duration of their lives, castle's west curtain. These internal build- all the other lands they held. One contem- ing works could well be taken to mark the porary chronicler commented that this was final completion of the castle after the the greatest blow to Hubert as he loved all building of the round tower defences. four castles and had spent a fortune on A few weeks after the allowance rebuilding them, though he bore this tribu- Prince Dafydd rebelled once more and lation with his customary patience. White Castle was again placed on a war He resigned four castles into the footing. By 1245 Prince Dafydd and his king’s hand, namely White Castle and supporters had generally been beaten back Grosmont in Wales and Skenfrith and into North Wales. As the war ground to a Hadleigh (Hauefeldiam). In the rebuilding halt, on 6 September 1246, Walerand was of these castles Earl Hubert had spent a allowed a further £5 which he had spent on boundless sum of money. building a chamber anew within White castle. This statement helps confirm that it was Hubert who was responsible for For the next few years White Cas- building the bulk of White Castle, its round tle seems to have remained a quiet backwa- towers and outer ward, and not the impecu- ter. Then, on 20 June 1248, the king nious Lord Edward twenty years after him ordered the sending of eighty marks (£53 in his relatively short and often interrupted 6s 8d) to Walerand Teuton, the keeper of tenure between 1254 and 1267. Otherwise Trilateral. A further order on 23 November why would Matthew Paris state that he had 1249 allocated Walerand a further £20 rebuilt all four castles rather than just two from the issues of the Trilateral bailiwick. or three? In short is it best to accept an The final payment in this series occurred original highly informed source, or accept on 20 May 1251 when Walerand was

The Castle Studies Group Journal No 24: 2010-11 215 White Castle and the Dating of the Towers granted forty marks (£26 13s 4d), ‘his year- castle, in totally pacified land, had been ly king’s gift for his maintenance’. This allowed to fall into decay, even though it seems to be the last mention of White Cas- was no more than twenty years old. tle under royal jurisdiction until 14 Febru- The castles of the Trilateral in the ary 1254 when the king finally disposed of seem to have been kept in the castle as part of the apportionment he good repair under the Lord Edward. In- made to his eldest son and heir, the Lord deed, just as the new Welsh war begins, we Edward. are lucky to have the steward's account for The other lands passed over to the lordship of White Castle for the year Edward's control in Wales included the from 29 September 1256 to 29 September whole of Chester with its castles and 1257. This deals mainly with economics of together with the king's conquest of the lordship, but it also contains some ex- Wales in these bounds, to wit Rhuddlan, penditures which throws further light upon Dyserth (Dissard) and Degannwy (Gunnoc) the status of the castle. After accounting for and the other land of the Perfeddwlad; Bris- various farming activities in the lordship, tol, the Trilateral, Montgomery, Carmarth- Reeve Adam claimed moneys for iron en, Cardigan and the castle of Buellt. The bought to repair the at 3s 9d; same summer, on 23 July, the constable of three pairs of manacles for the prison and St Briavels was ordered to send 6,000 quar- iron bars for the window at two shillings; rels to Gilbert Talbot, Edward's constable, roofing the barn and other buildings at 4s to munition the castles of Grosmont, Sken- 5d; lead and tin for the tower (turris) at 3s frith and White Castle. Such an action 2d and for making a portcullis for the castle would seem to have been intended to let at a cost of £1 7s. Next there was a charge Edward have some of the king's stock for of one shilling for making an ‘outer gate' his own usage in his now semi-independent and a further shilling for the carriage of state. It hardly seems likely that this could timber from . These minor ex- have been related to the completion of any penditures brought the total expenses at defensive building works at the castles as White Castle to £23 5s 4½d, to this was we have already seen how any major royal added £30 0s 6½d which was handed over works would have appeared in the pipe to John Briton who at Christmas 1256 had rolls as these lesser works at White Castle been sent to Abergavenny to organise the already had. We are therefore left with the resistance to Llywelyn ap Gruffydd in the same question of did the great rebuilding . In total it was discovered that happen earlier under Hubert Burgh or later Reeve Adam had spent 2s 7¼d more than under the Lord Edward? he had received and that consequently he It is to be presumed that the three personally was owed this amount. In all this castles of the Trilateral were handed over to suggests that the lordship of White Castle the young Lord Edward in good condition. was well-ordered and running at a useful Certainly the order for the 6,000 quarrels to profit. be sent to Gilbert Talbot to munition them After Reeve Adam's account suggests that they were defensible. Howev- comes the account of John Briton which er in 1256 their contemporary, Hadleigh ran from 8 September 1256 to 30 Novem- castle in Essex, was reported by its royal ber 1257. This gives some idea of how the constable to be in a very bad state of repair. Lord Edward was arranging his military Its buildings were unroofed and its walls affairs in the district. What is very noticea- broken. Presumably this had occurred be- ble is that the garrison troops seem con- cause Hubert Burgh had not been in resi- fined to Abergavenny and to a much lesser dence since 1239 and without a lord the degree Grosmont. From this we can see that

216 The Castle Studies Group Journal No 24: 2010-11 White Castle and the Dating of the Towers

John's account contains many details which are not strictly relevant to the history of White Castle, so only the salient points will be picked out here. John accounted for £1 12s in his taking seven hostages from the land of Buellt and taking them to White Castle where they were no doubt incarcerat- ed. It might be interesting to know what became of them during the repeated attacks upon Buellt until the castle's fall in July 1260. Also five shillings were expended in carrying quarrels from St Briavels to White Castle. In the presumed local White Castle garrison £8 6s was spent on two men-at- arms with barded horses staying there from 28 October 1256 [the outbreak of rebellion in Gwynedd] to 12 January 1257. To aug- ment this force between Christmas Eve and this set of trifling repairs to the fortress. 13 January 1257 Osbert Giffard was paid 27 We can see from these figures, and espe- shillings for remaining at the castle. A fur- cially the paltry sum spent on lead and tin ther two buckets of quarrels were sent from for a tower roof, that the castle was in a St Briavels to White Castle, which obvious- good state of in the mid-1250's. ly indicates that defensive preparations were in hand. An account was also rendered In 1302 merely repairing and for £2 spent on making 'a certain new amending one gutter between the great hall bridge' at the castle. These figures, with and the prince's old chamber in Chester those for the tower roof and drawbridge castle cost £1 6s 3d in the payment of work- repairs, are altogether minuscule when con- men, carpenters and plumbers. This includ- sidering that Briton in this period spent ed the cost of working with the boards and £507 15s 2½d on preparing his position at timber and soldering tin for the lead. We Abergavenny for war. The defences of can therefore see that the expense of just White Castle must have seemed satisfactory over three shillings in 1257 was merely for and an attack on the castle seems to have repairing a small leak somewhere in one of been considered quite unlikely. the tower roofs, which were then probably around twenty years old. The statement Altogether these figures supply us should not be read as showing that the old with an interesting view of life in a medie- Norman (K) was still standing as has val Marcher in the mid-thirteenth been suggested. Also in 1302 at Chester century. It also strongly suggests that there castle the cost of totally rebuilding the was no decay at White Castle as had been chapel and roofing it with over five sheets witnessed at Hubert Burgh's other castle at of lead after a major fire cost £13 14s. Hadleigh. Here in 1275 it was declared that Similarly the costs of setting seven great the castle houses were in ruins and the beams in various rooms in Chester castle fortress itself was badly built (castrum de- and strengthening the walls and setting biliter aedificatum). The survey from White stone corbels within the structures cost £5 Castle clearly shows that little rebuilding 15s 2d. In the late 1270's the building of was going on and certainly that no great Builth castle in the Middle Marches cost refortification had occurred since 1254, oth- over £1,600. In the amount of masonry built erwise these would surely have shown up in and towers and gatehouses constructed

The Castle Studies Group Journal No 24: 2010-11 217 White Castle and the Dating of the Towers

Builth would have been somewhat smaller lordship of Brecon was overrun. As a con- than White Castle. Such costs again show sequence of this the government of Eng- with a fair degree of certainty that White land ordered Gilbert Talbot to make good Castle was not modernised by the Lord the munitions of the castles of Skenfrith, Edward, but that it was constructed with Grosmont and White Castle ‘by all means some ‘of the infinite amount of money’ lest disaster occurs’ on 24 December. Gil- spent by Earl Hubert Burgh before 1239. bert seems to have succeeded in his task and the defeat by the Marchers of 10,000 The year 1258 saw a bloodless Welshmen at the battle of Abergavenny on revolution in and the curtailing of 3 1263 brought the Welsh threat to the power of the monarchy for a time. One the district around White Castle to an end. of the results of this was the limiting of the During this period the Lord Edward was power of the Lord Edward. Subsequently seeking loans in the towns of Shrewsbury, on 15 May 1260, through the advice of the Ludlow and 'for certain urgent king's council, the Lord Edward transferred business of the king's in the Marches'. the custodianship of to Philip Bas- However we may presume that this had set, Montgomery to John Lestrange Junior, much more to do with the relief of Degan- and the Trilateral to Gilbert Talbot. Indeed nwy and Dyserth castles, than for the un- with the deteriorating relationship between necessary refortification of the Trilateral. father and son about 6 June 1262 the Lord In May 1264 Edward was taken prisoner at Edward was persuaded to exchange his Lewes and as has been seen there can be hold on many of his lands. These included little doubt that no major building work had Grosmont, Skenfrith and White Castle. In been undertaken by him at White Castle return he was to receive the king's Jewry of during his years of lordship. Indeed no England with all the issues, profits, debts more is mentioned of the Trilateral during and customs for the next three years. In the rest of the disturbances which climaxed other words Edward was only in charge of with the defeat and death of Earl Simon White Castle from 1254 to 1262 and for six Montfort at the battle of Evesham on 4 of these eight years Edward was totally August 1265 and ended with pacification cash-strapped fighting a losing war in after the dictum of Kenilworth at the end of Wales. It is therefore quite ludicrous to 1266. suggest that his energies were spent in re- fortifying an isolated castle in South Wales It therefore remains to be seen while it was his North Welsh possessions whether White Castle was rebuilt under the that were being attacked and overrun. Fur- rule of Edward's brother, Edmund Croach- ther it is impossible that the castle was back. On 28 December 1266 Edmund was rebuilt in the period 1254 to 1256 as we granted Builth castle and its appurtenances, partially have the account of the lordship which had previously been part of the pat- which show incontrovertibly that no major rimony of his brother the Lord Edward rebuilding was going on. until it was seized by Prince Llywelyn in 1260. To retake this fortress Prince Ed- On 18 July 1262 Gilbert Talbot, in mund would obviously need a base and his capacity as constable of the Trilateral, what was more natural than to give him the was ordered to help defend South Wales for reasonably nearby and conveniently vacant King Henry III. That same winter of 1262 old lordship of his elder brother at Gros- disaster overwhelmed the Marches of mont? This of course included White Cas- Wales. Roger Mortimer and his allies were tle. Consequently on 30 June 1267, just defeated at in late November and three months before the treaty of Mont- towards the end of December were forced gomery was signed with Prince Llywelyn to retreat headlong out of Wales, while the

The Castle Studies Group Journal No 24: 2010-11 218 White Castle and the Dating of the Towers ap Gruffydd, the Lord Edward restored the family home at Grosmont. There seems honour of Monmouth and the castles of little evidence that any of it was spent at Grosmont, Skenfrith, and White Castle to White Castle, which it has been argued the king so that he could in turn give them above was the most recently refurbished of to Prince Edmund and his heirs. Prince his three local castles (1229-32, 1234-39). Edmund seems to have made Grosmont his The Masonry Defences caput at an early date. Consequently he may not have spent much time or effort on The bulk of the masonry at White Castle the great military fortress of White Castle, belongs to five distinctly traceable phases. although it is apparent that Grosmont was The earliest is the keep (K), of which just remodelled around this time. Prince Ed- under half of its foundations remain. Flank- mund accompanied his brother on crusade, ing and surrounding this are the remains of being out of the kingdom from soon after the curtain walls probably constructed un- 13 February 1271 to early in 1274. At this der the aegis of Ralph Grosmont in the two time an extent was made which described building years of 1185 and 1186. Next the state of White Castle's contemporary, came the addition of four great drum towers Hadleigh castle in Essex. This was found to with the gatehouse in the northern half of have been badly built and its houses were the inner ward probably in the period 1229 consequently much ruined. The state of the to 1231. Around the same time the keep (K) poor workmanship may have been over- was demolished and a new curtain wall with stated, for no repairs were carried out until (P) built over its site. Its seems the end of the 1280's and then only some different from the rest of the masonry and £41 was needed to patch the fortress up. this may have been the final major work at The castle then paid host to the king on two the . Despite this it seems like- occasions in 1293 and 1305. This hardly ly that the construction of the two southern- made it a royal favourite. Despite this the most towers and the outer ward occurred castle was much used by King Edward II last, probably between 1234 and 1239. and surveys of the time mention two tow- These later three phases, the bulk of the ers, a main gate and barbican and a postern. present remains at White Castle, has gener- Hadleigh was later massively rebuilt be- ally been stated to be the work of the Lord tween 1360 and 1370 with two great round Edward, or his brother, Edmund Planta- towers being added to the older castle genet, when they were lords of the Trilateral which appears to have had rectangular tow- between 1254 and 1296. However there is ers like those found at Dover castle. In total no documentary evidence to back up this over £2,000 was spent on the works. The claim. Conversely there is evidence, as has large round tower by the gate and two been seen above, to place the work within square towers may have been part of the the ambit of the ‘boundless sum of money' original design of Hubert Burgh. Earl Hubert Burgh had spent at the Trilater- al before 1239. The similarities between After returned from crusade work at Dover castle and the Trilateral real- Prince Edmund, on 6 June 1275, claimed ly clinches the case on the grounds of prob- that he had only received £600 of the 2,600 ability. marks (£1,733 6s 8d) granted to him by the king for his pilgrimage to the Holy Land. The Gatehouse (G) Consequently the king ordered the Excheq- It would seem that some 28 years passed uer to inspect the relevant rolls and pay any after White Castle first came into the hands balance due. Such large amounts of money of Hubert Burgh in 1201 before he began might always be used to start the refurbish- his refortification of the site in the period ment of what Edmund was making into his 1229 to 1232. Since the 1950's it has be-

The Castle Studies Group Journal No 24: 2010-11 219 White Castle and the Dating of the Towers

tle is heavily fortified. First came a portcul- lis of unusual design. Most portcullis are built into the gatehouse, but at White Castle it seems to have been added to the design as an afterthought. This is strange for one appears to have existed at the earlier Gros- mont castle, unless of course this is a later insertion. An internal portcullis also existed within the vaulted gatechamber of the Con- stable gatehouse at Dover castle. At White Castle a row of projecting quoins were added in front of the outer gate arch. Nor- mally these would have been within the first section of wall. Further, if partially external, it would have been expected for the quoins to form an arch above the gate- way so that the portcullis would have been come acceptable to date the additions to the hidden from view when raised. At White inner ward to the period after 1254, though Castle the quoins continued apparently to there are problems with this which seem battlement level. Here there must have been not to have been adequately considered. a projecting chamber from where the port- Contemporary evidence shows that an out- cullis could have been operated. This is a er gatehouse was standing in 1257 when it singular and unusual design in Wales and needed a new gate. Unless the castle had the Marches. been recently attacked, or the workman- The gatehouse at White Castle is a ship was defective, of which our record unique structure and contains several fea- shows no evidence, this would suggest that tures which suggest that it is an early twin the building was of some antiquity. Further D-shaped entrance. At Dover castle the the towers of both the inner and outer north gate looks quite similar to White wards show every indication of being more Castle in many respects, although the Do- in line with the smaller towers built by ver castle gatehouse was heavily rebuilt Hubert Burgh at Skenfrith and Grosmont, after its destruction by the rebels and than those of the larger ones of the later era French in 1216. Like White Castle it has an of the Lord Edward at Caerphilly, Rhudd- impressive batter spreading a considerable lan or Flint. Similarly the work of Hubert distance down the scarp. Unlike the Gwen- Burgh at Dover can be seen to be similar to tian castle though, it has two or three long, that at White Castle in more than a simple, unsighted loops, without oillets, in both its superficial manner. Therefore it can be remaining lower floors. It also has an exter- reasonably accepted that Hubert was re- nal offset at floor level. This is quite similar sponsible for the refortification of White to the round towers at castle Castle, and that this was carried out after he which are now accepted as late-twelfth cen- had finished rebuilding the other two cas- tury. The FitzWilliam gate at Dover, which tles of the Trilateral as the historical evi- replaced the north gate around 1227, also dence recounted above suggests. bears some comparison with White Castle White Castle gatehouse is re- gatehouse. Both are rectangular structures markably similarly to the one at Whitting- with twin protruding towers set on sloping ton castle which was built in the early plinths. However the FitzWilliam gate led 1220's. The gate passageway to White Cas- to a covered postern which passed through

The Castle Studies Group Journal No 24: 2010-11 220 White Castle and the Dating of the Towers

As mentioned above, many of the rectangu- lar Angevin towers of King Henry II at Dover, built in the period 1179 to 1189, appear to be of a similar design. These much altered towers of the inner ward seem to have three ground floor embrasures and no openings in the upper storeys. Again these features taken together at White Castle seem to point to this structure dating to the work of Hubert Burgh in the 1230's. The suggestion that the portcullis and draw- bridge apartment, like the hoardings around the walls, were wooden offers interesting possibilities. What is immediately apparent is the sheer height and inaccessibility of the machinery set on top of the gatehouse (G). There could have been no comfortable con- stable's chamber set here to vet those enter- the ditch and the next bank of the castle ing the inner ward of White Castle. The defences to an outer gate which is now mechanism set in a probably wooden hoard- mostly destroyed. Internally the gatehouse ing would also have been incredibly vulner- had two windows on first floor level over- able to artillery fire. We are therefore left looking the bailey. On the floor above at with an early twin-towered gatehouse which Dover were a set of twin loops in each is very lacking in respect to later Edwardian tower, with a further loop set above the examples such as St Briavels, Goodrich, gate archway between them. Both gates Rhuddlan, Caernarvon or Harlech. Indeed should also be compared with Criccieth the gatehouse is virtually devoid of external castle gatehouse. This gatehouse, which features except for four crossbow loops in probably dates to before 1239, has three each tower, the battlements and hoardings. loops on its ground floor and none on the As such this again suggests that the refur- floor above. These two early thirteenth bishment of White Castle dates to the first century structures again suggest that the half of the thirteenth century rather than to gatehouse at White Castle dates to the peri- the second half as has been previously ac- od 1229 to 1232 and this enhances the cepted. historical evidence suggesting this era di- vined from the study of both Skenfrith and The Inner Ward Grosmont castles and the history of the The two flanking towers of the inner ward Trilateral. have similarities to the two gatehouse tow- At White Castle the upper floors ers although both lack stairways. The exteri- of both gatehouse towers and the east (E) or of the west tower (W) is perhaps the most and west (W) towers are externally blind. photographed part of the castle, giving fine Presumably this was the received wisdom reflections as it does in the below. It after the building of the towers at first currently consists of three storeys, though Skenfrith and then Grosmont. This tower there may be a deep circular basement as has (M) should also be compared with the twin been found in the east tower (E) and at towered Fitzwilliam gate built under Hu- several of the towers at Skenfrith and Gros- bert Burgh's constableship and the older, mont castles. The ground floor of the west but much rebuilt north gate, both at Dover. tower (W) is currently entered through a broken doorway cut into the earlier curtain

The Castle Studies Group Journal No 24: 2010-11 221 White Castle and the Dating of the Towers wall at current ground level. Of this struc- The Southernmost Towers ture only a few of the lower jambs now The southern (S) and south-eastern (C) tow- remain and these have the Hubert Burgh ers are similar to one another and seem to style doorstops that are found so often at belong to a slightly later phase than the the castles of the Trilateral. other two pairs of towers just described to The first floor of the west tower the north. These towers differ from the oth- (W) was unlit and it seems possible that ers by being D-shaped in their entirety rath- there was no entrance into it, except possi- er than boldly projecting circular towers bly from the floor above. The next floor of with D-shaped upper floors. They are also the tower was accessed from north and slightly larger than the other four towers south via the wallwalk, though there were being some 32 feet in external diameter with no loops or any other feature at this level. walls seven feet thick above the batter. The It is very noticeable that by the second D-shaped chambers within the tower have a stage the tower has assumed a D-shaped diameter of eighteen feet. Like the east and appearance, rather than the round structure west towers (E&W) they form a butt joint of the floors below. Again there are no with the curtain, but in this case the crafts- lights in the three surviving sides of this manship seems less certain and there has tower. To the east the straight side of the been a certain amount of slippage. D-shaped chamber appears to have been The history related above suggests either open or wooden backed. Certainly with some exactitude that they were con- the surviving sides of the opposite east structed by Earl Hubert Burgh, probably tower (E) is in finished masonry. The final early in the period 1234 to 1239 and before level of the tower was a fighting platform the building of the outer ward (O). This is reached from the wallwalk level via a doz- suggested because their loops do not end in en or so straight steps up from the wallwalk top and bottom ball oillets like those of the to the south. This is a similar layout to that outer defences. These towers also differ found at Skenfrith. from the east and west towers in having Opposite the west tower (W) is the loops on two, or, in the case of the chapel east tower. The only real difference be- tower (C), three floors. That the towers are tween the two buildings is that the east of a similar age to the rest is suggested by tower had a deep Skenfrith and Grosmont their loops and battlements which seem sim- style basement. Like these comparable ilar to all the rest that survive. The towers towers the ten feet deep basement of the are also dissimilar to the Lancaster modifi- east tower (E) also has no obvious access cations at Grosmont which were probably point. The first floor was entered some built within twenty years either side of three feet above current ground level by a 1300. However, the approach (J) to the up- now mostly destroyed slightly pointed per storeys of the south tower (S) at White doorway inserted in the older curtain. Castle bares some similarity to that to the Again the similarities with Skenfrith, Gros- approach to the great tower at Grosmont. mont and the rear of the gatetowers at There seems little doubt that the White Castle are to be noted. battlements followed a similar arrangement In all, these four round towers on all six of the inner ward towers. This seem the larger twins of those found at the probably indicates the towers were built neighbouring castles of Skenfrith and reasonably contemporaneously, or that all Grosmont. It therefore seems unnatural to the wallwalks and battlements were refur- ascribe them to different eras of ownership bished near-simultaneously at a later date. It and construction as has recently been done. should also be noted that the lower layout of

The Castle Studies Group Journal No 24: 2010-11 222 White Castle and the Dating of the Towers

White Castle, Gwent . Gatehouse into the Outer Ward , probably built by Hubert de Burgh in the 1230s. three loops in the basement and two loops there are similarities between this gate and against the curtain in the floor above mir- the smaller water gate at Skenfrith which rors the west tower of Hubert Burgh at may date to 1187. Internally at White Cas- . This arrangement has tle, immediately west of the site of the many similarities with the corner towers at postern gate (P), is the thinning of the cur- . There are no towers of a tain marked by the remains of a fine similar nature to the south (S) or chapel (C) quoined joint. This matches similar features towers at Dover castle and those at Ha- on the east side of the gatehouse (G) pas- dleigh have been to heavily damaged to sageway and externally by the north-east make useful comparisons. tower of the twelfth century inner enceinte The Postern Gate (P) and South-Eastern at Dover castle. The purpose of these fea- Curtain Wall tures are uncertain, but they would appear to be decorative. It certainly makes a sym- East of the south tower (S) is an external metrical match with the gatehouse (G) round-headed segmental relieving arch. quoining and having two such features built Under this was the apparently round-head- in such a manner seems deliberate. Similar- ed postern (P) of which the jambs have now ly the example at Dover seems to have little totally disappeared, though two drawbar military logic about its construction. slots remain. Internally is a short four foot gate passageway and behind this is another The Outer Gatehouse (g) robbed-out gateway, topped by a segmental To the north of the inner ward lies the relieving arch. It is possible that this gate- irregular outer ward. This too has been way was the original entrance to the castle, dated to the Edwardian era with little refer- now encased by the later wall. Superficially ence to the castle's history as described

The Castle Studies Group Journal No 24: 2010-11 223 White Castle and the Dating of the Towers

ture which at this time would have easily cost between £300 and £500. All this actu- ally proves is that a new portcullis, outer gate and a bridge were made somewhere at the castle. Indeed the large sum of £2 paid for the bridge might well suggest that it was partially the stone bridge abutment to the outer gate (g) that was constructed now. If it were, it would again suggest nothing, except that the outer gatehouse (g) was standing in 1256 and needed a new gate and portcullis. Indeed the money spent on re- pairing the drawbridge could well also re- late to the outer gatehouse and suggest that above. The castle is currently entered the two rounded turrets had been recently through the outer gatehouse (g). The east added to the older gatetower and that this and west walls of the gatehouse (g) have consequently required the addition of a new mostly gone leaving only the north and portcullis and gate. south walls as two jagged fragments. With- The North Curtain Wall of the Outer in this stood the gate passageway. The west Ward exit into the ward (O) once had an archway in which was a gate. This was probably At the end of the mostly collapsed east similar to the layout at curtain wall was a backless circular mural inner gatehouse which was built in the tower (a) of two storeys. This is about twen- 1220's. Only the lower stops of the White ty feet in external diameter. The tower is Castle gateway remain. At the eastern, tur- joined to the curtain on the west side by a ret end of the possibly earlier rectangular short section of wall which is obviously part gatetower, was a further gate of which the of the original plan. Probably it indicates door stops also remain. None of these door that the architects did not plan the wall and stops would look out of place at Skenfrith tower properly with the result that there was or Grosmont. As at Grosmont castle, the a gap of several feet between them which early drawbridge pit seems to have been had to be filled with this little section of later encased by two projecting walls added wall. Externally a sloping plinth is still evi- to the front of the older gate-tower. At dent around the tower. Such plinths are Grosmont, however, the facing of the pro- lacking on the outer ward curtain wall. Ap- jections were square, while those at White parently this is because the walls stand on Castle have rounded faces. These solid tur- top of the bank of the outer ditch (d), while rets show some similarities to the twin-tow- the towers protrude down the slope into the ered gatehouses at Montgomery and ditch. On the ground level of the corner Longtown castles. It has been suggested tower is yet another 'blind' room. Above is that White Castle outer gatehouse (g) was a room with three crossbow loops. This is only constructed in the mid-1250's when, the standard layout in the outer ward tow- money was spent on making a portcullis, an ers, a reversal of what is seen in the inner ‘outer' gate and a ‘new bridge'. ward both at White Castle and Dover castle, but similar to the Skenfrith towers. There However the shilling spent on a appears to be no doorway into the ground gate was surely for a wooden replacement floor, or the floor above. There is merely an of an earlier castle gate rather than the aperture, which may once have been building of this expensive masonry struc-

The Castle Studies Group Journal No 24: 2010-11 224 White Castle and the Dating of the Towers blocked with a wooden wall similar to that quoins, unlike the south-east wall of the of a ‘black and white house'. inner ward. Presumably this suggests that the south-east curtain post-dates the outer On the first floor are three stag- ward (O). There is a similar unflanked gered cross loops set in wide embrasures. angle as this at Wigmore castle which was These have angular roofs and are quite probably built in the early to mid-thir- shallow, due to the thinness of the walls. teenth century. There are no features in the Externally the embrasures supported three rest of the outer ward curtain that differ similar split sighted crossbow loops with significantly from the rest of the structure ball top and bottom oillets. This layout, in style or date. apart from the open back, is reminiscent of the towers at Skenfrith castle. Indeed the Conclusions embrasures with their pointed segmental One feature that ties the Trilateral castles arches, three to a floor above a basement, together, and therefore suggests that they are pure Skenfrith. It is uncertain how ac- were constructed as part of a single build- cess was gained to the upper floor. Possibly ing scheme, is the access to the wallwalks. it was via wooden steps to the rear, or even This would appear to have been gained a wooden stair down from the wallwalk at mainly via the two individualistic stair- second floor level. It is interesting that the ways in the gatetower, although there were ground floor would have been left literally likely stairs in the chapel tower (C) at undefended by crossbow fire. This is the White Castle and in the west tower at opposite to the inner ward at White Castle. Grosmont. The profusion of wooden steps Two-thirds of the way along the north cur- to wallwalk level which are occasionally tain stands a half round tower (b), other- suggested would simply seem to be imagi- wise similar in design to its corner nary, especially when the number of build- companion to the east (a). This is the most ings clustering against the curtain walls are ruined of the towers of the outer ward. The taken into consideration. At Skenfrith bulk of its first floor loops have been there is no obvious method of reaching the stripped from their embrasures and looking wallwalk, though it must be suspected into this tower you could be forgiven for from the rest of the remains, that the origi- mistaking it for one of those at Skenfrith. nal stairway was in the gatehouse. At Within the tower the ball base oillets of the Grosmont a straight stair can still be made north and west loops are still in situ show- out in the gatehouse, although a short stair- ing that the tower originally had similar way in the west tower appears to allow loops to the others. The rear wall of this access to the mid-floor of the west tower, tower has been gouged out at ground level if not the wallwalk. and the basement walls robbed of their facing. The upper floor, however, is exter- Finally, at all three castles of the nally perfect at the rear and this shows that Trilateral, the lack of sanitation is immedi- the tower was either open to the air or ately obvious. At White Castle garderobes wooden backed. Although the top of the only certainly exist in the outer ward and tower is now gone, it presumably had a these were obviously built with a fair fighting platform similar to the other tow- amount of people in mind, with two or ers at the castle. three in the outer gatehouse (g), two in the west tower (c) and one against the curtain From the north tower (b) the cur- (f). The pit (V) in the inner ward is of tain makes three irregular sweeps, leaving uncertain purpose. The latrine block by the a vulnerable unflanked angle, round to the side of the great tower at Grosmont may largest tower in the outer enceinte (c). The also be an addition to Hubert Burgh's unflanked angle is not buttressed with

The Castle Studies Group Journal No 24: 2010-11 225 White Castle and the Dating of the Towers work, while at Skenfrith the only gard- eral began as the castles of the Bilateral, erobe, a corbelled out latrine, is set in the and that Newcastle/Skenfrith was a late- second floor of the king's tower. Perhaps comer to the scheme. Hubert had taken to heart the lesson of the Secondly the northern four towers loss of Château Gaillard by a common of the inner ward at White Castle can be soldier scaling the undefended latrine shaft. reasonably assigned to the work of Hubert Certainly his provisions for sanitation and Burgh, probably in the era 1229 to 1231, light for the common soldier seems singu- and that the two southern towers and also larly lacking at all three inner fortresses. much of the outer ward probably date to his This study has revealed much new work in the period 1234 to 1239. The work data about the history, building and opera- carried out to ‘the tower’ of Walerand Teu- tion of White Castle. It is also interesting to ton and his immediate successors in the see how the perceived history of the site has mid-thirteenth century was far more likely see-sawed over the years. The great G. T. concerned with the adjustments to the larg- Clarke believed that the inner curtain was er south tower (S) than to the obsolete keep built during the late 1180s and that Hubert (K) which had most likely already been Burgh was responsible for the rest of the demolished by Hubert Burgh when the two masonry. In 1961 this view was overturned great southern flanking towers and joining and the dating of the later masonry was curtain were constructed. Certainly parts of adjusted to the 1250s or 1260s. As I hope the old keep (K) appear to be reused in the to have shown, this reinterpretation was chapel tower (C). It can therefore be seen based upon slim evidence that does not that the dating of the construction of White stand up to detailed scrutiny. Certainly the Castle is reasonably secure and supported singular reference to roof repairs of ‘a tow- by the works of Hubert Burgh at other for- er’ in 1257 cannot be taken as evidence that tresses. the old keep (K) was still standing, nor can In total it can be seen that White the inexpensive works possibly carried out Castle as it now stands is primarily an early on an outer gate be tied to the refortification to mid-thirteenth century enclosure castle of the outer ward (O). built on the site of an earlier keep and bailey What can be firmly deduced is structure. Its great size shows the determi- that the rectangular keep (K) could have nation of the Crown and later Earl Hubert been built at any time between 1067 and Burgh of Kent in denying this district to the 1160, though perhaps the first push into native Welsh. As such it stands as a master- Gwent by the earls of Hereford in the late piece of military engineering to the second 1060s and early 1070s or the Anarchy of quarter of the thirteenth century, just as the Stephen’s reign would have provided most Edwardian castles of North Wales stand motivation. It also seems likely that this testament to the abilities of the last quarter keep (K) was built simultaneously with the of the century. hall block of Grosmont Castle, Grosmont Paul Remfry being the caput and White Castle the major fortress of the honour. With this in mind it This paper is an abbreviated summary of his should also be noted that the only early essay on White and readers should consult Norman pottery to come from all three Paul’s full document: White Castle, 1066 to 1438 (ISBN 1-899376-42-9), a 310 page de- castle sites is from Skenfrith where a single tailed history and analysis; also review his on- sherd came from a silted ditch which was line essays at: www.castles99.ukprint.com said to have lain under the Hubert Burgh /Essays/white.html masonry castle. It is therefore eminently possibly that the later castles of the Trilat-

The Castle Studies Group Journal No 24: 2010-11 226