Twelfth Century Great Towers - The Case for the Defence - Richard Hulme

The Studies Group Journal No 21: 2007-8 209 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence

TWELFTH CENTURY GREAT TOWERS angles were initially partially countered by po- The Case for the Defence lygonal towers before it was discovered round I towers provided the solutions. Liddiard also claims: ‘Recent work on such as Orford Great Towers: or Symbols? and Hedingham has completely demolished In 1215 King John’s miners brought down a the idea that the donjon was primarily de- 4 corner of the great tower at Rochester after the signed for defensive purposes’. New theories castle’s defenders had retreated there following emphasise symbolism, displays of wealth, and the capture of the . Even then the defend- elements of theatricality or ‘choreography’ to ers fought on from behind the great tower’s in- reinforce their owners’ power. It is becoming ternal cross-wall. It was a dramatic , well a commonplace that towers were not designed documented, with incidental detail such as for defence e.g. ‘the fact that donjons were sel- John’s order for forty fat pigs to help fire the dom designed to be defended tells us more props underneath the tower, and illustrates the about the than (say) whether traditional view of donjons or tower as Rochester tower was mined with a tunnel or a 5 ‘both the castle’s ultimate military strong point sap’ (Coulson). and principal residence’ (Allen Brown).1 Liddiard seeks to extend this argu- The authors of the History of the ment: ‘Revisionist arguments over the are King’s Works (HKW) declared ‘Both (Anglo- critical to our understanding of castle develop- Norman and Anjevin rectangular tower keeps) ment across the whole medieval period, since were designed for passive rather than active if it can be shown that keeps were emphatical- resistance…In an age when the only projectiles ly not raised for utilitarian military purposes were stones, lances, arrows and the like, they (and it should be said that this is still contest- were, however, as nearly impregnable as any ed) then a central – and arguably the most im- form of yet devised’. Thus, ‘the portant – plank of the traditional military 6 keep dominated the twelfth century conception interpretation of castles is removed’. If of a castle’.2 twelfth century towers were designed in mar- tial style (to denote aristocratic rank) rather Recent research has cast doubt on than martial substance it supports Coulson’s these ideas, questioning both defensive capabil- sweeping conclusion: ‘The vast majority of ities and residential use. Some towers, like castles in , , , and Chepstow, lacked latrines and fireplaces, im- have virtually no ‘military history’ of or plying a limited residential role. The Tower of physical conflict across the whole panorama of possessed latrines and fireplaces but more than five centuries. Rather than being seems to have housed only prisoners. At Hed- built for defence, as was once imagined, the ingham the one-way gallery above the grand majority display a refined aristocratic taste…’ 3 hall suggests a design for ceremonial use. (my italics).7 Forebuildings have been interpreted as proces- sional entrance routes rather than defensive fea- Were great towers designed for de- tures (towers with forebuildings often also had fence, residence, or ceremonial? Were these a secondary entrance). different design criteria incompatible? Major sieges like Rochester were rare II and few great towers were used as places of Ceremonial final refuge, raising the question whether defen- sive considerations were important in tower de- Recent research into Chepstow castle, Gwent, sign. Critics point out tower building suggests the great tower was built by William chronology does not support the idea that mili- the Conqueror in the 1080s. Its unusual niche- tary weaknesses associated with exposed right lined first floor room, interpreted as solely for

210 The Castle Studies Group Journal No 21: 2007-8 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence ceremonial use, for the king to receive homage an exterior castle wall 2.4m thick, was effec- from the Welsh, never have been used as tively indefensible: the roof and anyone on intended: ‘Perhaps these events (of 1093 – see the wall top would be exposed to archery below) made the Great Tower at Chepstow from high ground 60m away. With serious redundant’.8 The concept of a dedicated cere- disorder in Gwent in the would monial hall implies there was another domestic Chepstow’s tower have been left uncrenellat- hall in the castle, although versatility and ed? At Colchester temporary multi-functional use was a characteristic of were built on the tower when there was in a structures like Westminster hall, the break in building operations after one storey ‘ceremonial centre of the Anglo-Norman king- was constructed12, probably due to the Danish dom’ (HKW).9 Ceremonial halls, even with invasion scare of 1085. unique features, doubtless functioned as If Chepstow’s tower was intended be ‘normal’ halls. Would Chepstow tower’s hall undefended why are the walls, except the have remained little used? north wall overlooking the river Wye, so thick In 1093 the killed Rhys ap and why are the windows, except one over the Tewdwr, king of south-west Wales, in battle doorway (a useful defensive and practical fea- near Brecon and overran south Wales, estab- ture), concentrated in the north wall? A fur- lishing many castles including Cardigan, Rhyd ther (third) storey was later added, y gors (Carmarthen) and Pembroke. The Nor- necessitating substantial rebuilding of the mans had already occupied Anglesey and the original tower top. Most of the new upper north coast. In 1094 the Welsh fought back, walls on the east and south have disappeared, storming the castle on Anglesey (the motte of so it’s questionable whether any traces of ear- Aberlleiniog) and massacring the earl of ly crenellations would survive. The Chep- Chester’s garrison of 125 horsemen. In the stow researchers accept there are difficulties south ‘the castles of Ceredigion and Dyfed of interpretation in their reconstructions, par- were all taken except two castles, Pembroke ticularly if the two tiny round windows high and Rhyd y gors; and they were all razed to the in the west gable wall are original, as they ap- ground and the spoils carried off’ (Brut Y Tw- pear to be, because they would block a wall ysogyon). In 1095 the Welsh took Hen Domen walk that end at the height postulated.13 (Montgomery) castle, killing the earl of Dixon and Marshall identified Nor- ’s garrison. In 1096 the Normans ham tower’s first phase (of 1121), approxi- abandoned Rhyd y gors but Pembroke with- mately 23m x12m externally with 3m thick stood a long blockade by the Welsh. There walls, within the enlarged mid twelfth century was widespread raiding and counter raiding in structure.14 The upper storey, ‘with little sign Gwent and Brecon. The Brut records two of domestic provision, makes it appear that it Welsh successes in battle, though ‘the castles was a grand ceremonial chamber – a first- 10 were still intact and the garrisons in them’. floor hall rather than part of the living accom- These events doubtless rendered a cer- modation of the bishop’. Although a recon- emonial hall redundant but surely increased the struction drawing suggests a defensive need for defensive effectiveness. Chepstow structure (battlements above a concealed roof) tower’s south wall originally contained no win- they compare it to the Exchequer hall at Caen, dows. Defence would have to be conducted which is thinner walled, had larger windows, from tower top battlements, but because an exposed roof and its main door in a gable ‘Nowhere is there evidence for crenellations in wall at ground level. The Exchequer hall had this phase’, reconstruction drawings assume no defensive pretensions; it lies within the none, postulating a southern parapet some Conqueror’s stone curtain, 50m from Henry 0.7m high.11 Thus the tower’s long south wall, I’s massive great tower.15 At Norham the on-

The Castle Studies Group Journal No 21: 2007-8 211 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence ly other early stonework might be the inner assault on William’s siege castle failed. . Rebels holding Norwich* surrendered on Bishop Ranulf of Durham built Nor- terms after three months in 1075. In 1076 ham for protection against robbers and Scots. William abandoned the siege of Dol after six In 1138 the Scots besieged it. Many defenders to eight weeks, fleeing in disarray on the ad- were wounded during a vigorous defence and, vance of a relieving force. In 1085 William despairing of help from Bishop Geoffrey, they ordered the abandonment of a long blockade surrendered. The bishop was criticised for in- of Sainte-Suzanne*, where defenders had 18 adequately garrisoning his castle (there were killed or captured numerous of his men. only nine ) and his knights were criti- William Rufus faced immediate re- cised for their feeble resistance, because ‘the bellion in 1088. He besieged Pevensey *, wall was in good condition ( optimum), which surrendered after six weeks; supplies the tower very strong (et turris fortissima), and were running low after sea borne reinforce- their provisions abundant’ (Richard of ments were sunk or captured by royal ships. Hexham).16 Internally Norham’s tower might Rochester* eventually surrendered after have been a ceremonial hall, but externally it plague broke out in the city and it was clear was a turris fortissima. the rebellion had failed.19 In the rebellion of In the eleventh century halls could be 1095 Rufus took the castle at the mouth of strongpoints. According to William of the Tyne by storm after a two month siege the stone hall (aula) at Brionne served as cita- (while Tynemouth is the obvious site it may del (arx) during the long siege of 1047-9, and have been Newcastle*, which charter evi- Mayenne possessed stone when dence shows was besieged). Bamburgh* on- Duke William attacked in 1063. William’s ly surrendered when Rufus threatened to men fired buildings in the and blind its lord, Robert of Mowbray, who had 20 burst in. The defenders retreated to the been captured on a raid. (arx) but surrendered the next day. The arx In 1105 Henry I abandoned a siege was probably the early stone hall and its asso- of Falaise* when his coalition collapsed. ciated courtyard revealed by recent The following year he defeated and captured investigation.17 his elder brother in battle Whatever their nominal internal func- when he tried to relieve Henry’s siege of tion robust stone structures without vulnerable Tinchbrai. Henry then hurried to secure the openings could be externally strong and valu- ‘almost impregnable castle’ (firmissimam able defensively. All Chepstow tower required munitionem) of Falaise (Orderic). Sieges to have been strongly fortified was a crenellat- were frequent during this period. Sometimes ed wall top. assault was successful. Often defenders sur- rendered, for a variety of reasons, but some III held out while others repulsed besiegers by Early Siege Warfare effective sallies e.g. Dover’s* under strength garrison routed Eustace of Boulogne’s men There were a number of protracted in 1067.21 eleventh century sieges. At Domfront*, be- sieged over the winter of 1051-2, the defenders Castles which defied besiegers for surrendered after Duke William mutilated pris- lengthy periods (and rarely surrendered be- oners captured at nearby Alençon* (taken in a cause the fortifications were overcome) were dawn assault following a night march). Rebels obviously powerful fortresses. Those above in Arques*, 1052-3, surrendered because of marked * possessed great towers by c. 1200 supply shortages; relieving columns were am- – but not, except possibly Sainte-Suzanne bushed and although one reached Arques an and Pevensey, by 1106. Besiegers, due to

212 The Castle Studies Group Journal No 21: 2007-8 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence crags and undergrowth, had difficulty ap- 1147. These possessed rectangu- proaching Sainte-Suzanne castle, and there is lar great towers. no indication Rufus’ troops breached the Ro- In 1136 Exeter only surrendered after nearly man wall at Pevensey, so these great towers, if three months of expensive and diverse meth- they existed in the 1080s, were unlikely to have ods of attack when the wells ran dry. The ag- been important factors in the sieges. Powerful gressive defenders of Wark kept the Scots at early castles usually consisted of strong natural bay throughout 1138, until ordered to surren- sites enhanced with massive ramparts and der by their lord. From 1139 to 1153 Walling- ditches defending an inner enclosure. Rectan- ford consistently defied Stephen, three gular great towers were insignificant in elev- blockades failing (in 1139, 1146, and 1152-3). enth century sieges. This is not surprising; Lincoln castle resisted Stephen twice, in the great towers were rare. winter of 1140-1 (when Robert of Gloucester’s From 1066 castle numbers in England relieving army forced him to battle and cap- and Wales rose rapidly to something over 500 tured him), and in 1144. surren- by 1100 (only approximately 100 of these early dered after a three months siege in 1142, but sites are documented). New castles continued only after had made a dra- being founded throughout the twelfth century, matic escape on foot over the frozen Thames, particularly in Wales and northern England. dressed in white.24 These were motte and bai- Between 950 and 1,150 castles in total were ley castles (usually stonework strengthened), built before 1216, but many were abandoned or except Exeter (a ring work with stone curtains not maintained so the numbers of active castles and gatehouse). probably remained around 500 to mid century, Many fortresses in Normandy submit- perhaps increasing slightly under Stephen. In ted to Geoffrey Plantagenet after Stephen’s the second half of the century active sites de- capture at Lincoln, including Falaise. Geof- clined for a variety of reasons (royal action, frey had failed to take it in 1138, fleeing at changing technologies and increasing stone night on (incorrect) reports of an approaching construction, and economic, social and political relieving force.25 changes, particularly as they affected lesser lords). By 1216 there were probably around The most remarkable siege of 400 active castles (over 300 of which are Stephen’s reign occurred while he was held documented).22 captive in the tower at Bristol. Matilda and Robert of Gloucester besieged Henry, bishop In 1100 there were perhaps 10 to 15 of Winchester’s men in the bishop’s fortified great towers, rising to 50 or 60 by mid century, palace (Wolvesey castle) and his or and over a 100 by 1216.23 Thus, the proportion turri fortissima in the middle of the city. From of castles with towers (excluding shell-keeps), the two castles Bishop Henry’s men flung fire- over 25% by 1216, barely exceeded 10% at the brands, burning down much of the city. Mean- end of Stephen’s reign, nineteen years of wide- while the queen rallied Stephen’s supporters spread warfare including over 100 known at- and organised a blockade around Winchester, tacks on castles. besieging the original besiegers. After seven In 1138 Stephen abandoned a project- weeks, with the blockade causing , Mat- ed siege of Bristol, power base of his main op- ilda and Robert of Gloucester attempted to ponent, Robert of Gloucester, because the town break out but were defeated. Matilda managed of Bristol, between two rivers, was naturally to escape in the confusion, but Robert was cap- too strongly fortified.Corfe withstood a block- tured and incarcerated in the tower at ade by siege castle in 1139. Pevensey surren- Rochester.26 dered after a long land and sea blockade in Rectangular great towers had little impact on Anglo-Norman castle warfare. The

The Castle Studies Group Journal No 21: 2007-8 213 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence

‘traditional castle story’ overstated their mili- (citadel) and turris some 95 times each but tary role and importance. Rectangular great only used mota 3 times (two of which are a towers were neither common nor a necessary place name). Dangio appears 4 times, in the element of a powerful castle. Therefore, even sense of royal citadel; twice of Évreux, once if their defensibility was suspect, it has little of Gisors and once generally (in 1119 Henry I relevance to military interpretations of twelfth restored certain , including Alençon, to century castles. William Talvas except for the dangiones, in 29 IV which he placed his own guards). Évreux castle has disappeared but at Gisors the dan- Mottes gio was probably the though there The charter of Westminster, 1153, recorded the is a small octagonal tower on the motte. Lam- terms ending the wars of Stephen’s reign, in- bert d’Ardres late twelfth century description cluding custodial arrangements to ensure that of Ardres around 1060, ‘a very high motte or on Stephen’s death specific castles were hand- lofty donjon’ (motam altissimam sive - ed over to Henry Plantagenet: the Tower of jonem eminentem), supports Allen Brown’s London, Winchester (castles with towers), observation that mottes, like freestanding Windsor, Oxford, Lincoln, and Southampton great towers, could be referred to as (castles with mottes, the mota at Windsor and ‘donjons’, and both fulfilled the symbolism of 30 Oxford being specifically referred to in a con- lordship. The Round Tower at Windsor was text implying the whole castle).27 On the Bay- regularly called the great tower (magna turris) 31 eux Tapestry the castles of Dol, Rennes, Dinan, in the thirteenth century. and Bayeux are represented by their mottes. Pipe roll references to the turris at The origins of the motte remain a mat- motte castles such as Arundel, Berkhamstead, ter of debate but they were common in England Oxford, Shrewsbury, Southampton, and from the early . Possibly the circum- Worcester appear to refer to the structure on 32 stances of the conquest, hundreds of castles the motte. The Gesta Stephani describes built to protect limited numbers of Normans, extensive defences at Bedford in 1137, when were conducive to a design offering protection Stephen besieged it, including an ‘unshakable to a small garrison. keep’ (inquassabili turri), presumably the shell on the motte (ironically in the 1224 siege Mottes come in many shapes and siz- the tower was brought down by mining).33 es, castle builders preferring natural hills if Mottes and their superstructures were possible. Height made them easy to defend; (cheaper) alternatives to rectangular great attackers were exposed to defenders’ missiles towers, similar in function if not form. There while climbing the steep slopes. Motte top was no contemporary term for ‘shell keep’. towers provided observation posts and could overtop siege towers, which would be prevent- Around 1075 the Welsh plundered ed from approaching by the motte’s bulk. At and burned the bailey at Rhuddlan, but the Berkhamsted, Hereford, Southampton, Durham Normans held out in the tower (twr in Welsh), and Warwick there are references to the domus probably a wooden tower on the motte. In (house) on or in the motte implying residential 1116 the Welsh failed to take Llandovery and use.28 Swansea, only managing to burn the ‘outer castles’. At Carmarthen they plundered and The use of mota in chronicles before burned the ‘outer castle’ in a night attack but the mid twelfth century is rare e.g. Orderic ‘the towers had escaped’ (Brut). Presumably (died c. 1142) referred to castles/ strongholds/ the mottes at these castles held out. Garrisons fortresses about 750 times (employing, in de- of castles overrun in Wales were sometimes scending order of usage, castrum, , massacred. Negotiated surrender in exchange munitio, castellum, municipium etc.) and arx for ‘life and limb’ (i.e. no mutilation) was one

214 The Castle Studies Group Journal No 21: 2007-8 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence way of avoiding this fate, and could be V achieved after having retreated to the tower, as Defensive Capabilities at Dinefwr in 1213, where the defenders sur- rendered a day after the bailey was captured by Early castle defences were often constructed assault.34 At Le Puiset in 1111 Suger, of earth and timber. However, medieval in- an eyewitness, describes how, after French roy- cendiaries were remarkably effective, and tim- al troops breached the bailey , Hugh of ber defences were frequently destroyed by Le Puiset ‘took himself off to the motte and the fire. Stone structures therefore potentially wooden tower on top of it. He cowered strengthened a castle. 35 there…(and) surrendered without delay’. It was rare for an adequately defend- Towers were also used as refuges. In ed stone castle tower to be burnt. In June 1140 nominally supporting Matil- 1090 the wooden shingles of the roof of the da made a surprise night attack on Devizes. hall (aula) at Brionne were set alight by fire Using scaling ladders they eluded the guards arrows while a successful assault was and captured most of Stephen’s men asleep, launched. As the defenders mustered only six except for a few who ‘took refuge in a very knights (defenders such as archers are fre- high tower’ (turri eminentiori) (Gesta Stepha- quently omitted in chronicles) the garrison ni). Not expecting help they surrendered a few was probably under-strength. Henry I burnt days later. The tower, which John of Worces- the stone building (lapideam domum) in the ter calls an interior tower (turrim interiorem) castle of Le Renouard in 1119, but only after was probably similar to the surviving tower at it surrendered. In 1174 the Scots set fire to Sherborne, also built by Bishop Roger of the tower at Brough after the defenders, with Salisbury.36 In 1202 eighty year old Eleanor of only half a dozen knights, abandoned the bai- Aquitaine was besieged in Mirabeau castle by ley when it was assaulted on the first day of forces hostile to her son, King John. ‘As there the siege.39 was not strength in the garrison to resist them, Stone churches however were regu- the castle was surrendered to them except for a larly destroyed by fire e.g. Bayeux cathedral tower into which Queen Eleanor had thrown (during Henry I’s assault on the town in herself with a few soldiers…They therefore 1105), Évreux cathedral (during Henry I’s as- directed their attacks against the tower’ (Roger sault on the town in 1119), Nottingham (when of Wendover). John, however, made a light- Anjevins looted the town in 1140) and many ning advance and surprised and captured the others. ‘Civilian’ casualties usually resulted, 37 besiegers. because people commonly sought refuge with- Retreating to a citadel (motte or great in churches. Churches defended by knights tower) was a last resort if the bailey was taken, were also often burnt down e.g. St. Pierre-sur- usually used to negotiate an orderly surrender, Dives abbey (by Henry I in 1106) and Wher- not carry on a protracted resistance. Failure to well Abbey, near Winchester, in 1141, where reach a refuge could be fatal. In 1138 King John Marshal evaded capture by hiding in the Stephen’s men assaulted Shrewsbury castle, burning church tower, though molten lead burning and forcing the . Rather than re- dripping from the roof blinded him in one treat to the motte the defenders fled, but many eye.40 were captured. Stephen, ‘because unruly men Westminster hall was one of the larg- regarded his gentleness with contempt’, or- est halls in Europe when built by William Ru- dered ‘Arnulf (the commander) and about 93 fus around 1097. It had one storey, probably of the men who had defied him should be four doors, a dozen large windows in each of hanged on gibbets or put to death in some other its long walls (set in arcaded wall galleries), 38 fashion’ (Orderic). and low, exposed, multiple roofs.41 This was

The Castle Studies Group Journal No 21: 2007-8 215 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence an undefended structure. By contrast the Tow- built on mottes e.g. Bronllys, Caldicot, Long- er of London has thicker walls, three storeys, town, Richard’s Castle, and Tickhill. one (perhaps two) doors on the second storey, Guildford’s rectangular tower was also built on fewer windows (larger ones above the reach of a motte, facing high ground. scaling ladders), and high walls carried up to Robert of Torigni, listing Henry I’s conceal the roofs. These features, common in castle building in Normandy (1106-35), re- great towers, were unnecessary for residential cords the construction of other defences as well or ceremonial use, but are useful defensively. as a tower at most important castles.43 In Eng- Dixon concedes it is ‘hard to find convincing land Henry II’s expenditure on castles from alternatives (to ‘keeps designed as points of pipe roll records was, in descending order: Do- last resort’) for the tall, almost windowless, ver £6,440, Nottingham £1,816, Windsor unheated and insanitary towers…’ though he £1,476, Orford £1,471, Winchester £1,233, defines the class narrowly: Irish monastic Newcastle £1,144, Scarborough £683, Bowes round towers, German Bergfrieden, and the £617. These eight castles account for nearly like.42 Wells, fireplaces and latrines added do- 70% of Henry’s castle expenditure of £21,500 mestic comforts but enhanced rather than de- (the pipe rolls do not record all expenditure but tracted from a tower’s refuge capabilities. it is unlikely major works are not reflected in Protracted and successful castle de- them).44 fences involved retaining control of the main The great tower at Dover overlooks its bailey. The primary defensive role of a great contemporary surrounding bailey, which has tower or motte was therefore supporting bailey regularly spaced rectangular flanking towers, defences. Tower top panoramic views could including two twin-towered , each reveal besiegers’ dispositions. Signals could protected by a . Tower and bailey pro- be sent to allies outside the castle using ban- vided a concentric defence. A partial third ners and trumpets. Arrows, stones and javelins (outer) line of defence incorporated the semi- hurled down assisted bailey defenders. It was octagonal Avranches tower, equipped with pointless attackers attempting to throw javelins dedicated arrow-loops. The HKW’s observa- up at a tower top and arrows shot upwards lost tion that the great tower was ‘obsolete almost much of their stopping power. as soon as it was built’45 obscures the castle’s Rather than being sited at the least many advanced features. vulnerable position in a castle towers were of- At Nottingham Henry II constructed ten placed to directly support bailey defences. curtain walls around the upper bailey Many towers were built on the line of the cur- (sometimes referred to as the mota) and middle tain wall e.g. Barnard, Brough, Chepstow, bailey, a great hall and other domestic accom- Conisborough,Corfe,Helmsley,Monmouth, modation, and possibly the tower (it might be Norham, , Pevensey, Portchester, St. earlier).46 At Windsor he rebuilt Henry I’s Briavels, White, Winchester, Usk, often by a shell keep, spent £550 on ‘the king’s houses’ gate e.g. Bamburgh, Brougham, Coity, Kenil- and walled the upper bailey, the middle bailey, worth, Lydney, Peveril, Nottingham, Ogmore, and part of the lower bailey, including numer- Scarborough, and the Tower of London. ous rectangular, open backed towers.47 Bramber and Ludlow towers were converted gatehouses and Richmond built on the site of At Orford the polygonal great tower the original gate. Other towers were placed was surrounded by an oval bailey wall contain- close (within about 10m) behind the curtain at ing (probably) seven flanking rectangular tow- vulnerable points or on the line of approach ers, one of which was a gatehouse approached e.g. Carlisle, Goodrich, New Buckenham, by a passage flanked by parallel walls.48 The Newcastle, Norwich, Pembroke, and Roches- 15m wide outer was around 20m from the ter. Round or polygonal towers were often curtain. Possibly the original plan was for a

216 The Castle Studies Group Journal No 21: 2007-8 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence larger bailey but the wide berm allowed ar- Liddiard points out besiegers could chers at the great tower battlements to shoot shoot arrows through windows, and cites ex- over the curtain into the ditch and beyond (the periments demonstrating attacking archers bank of the ditch is between 60 even had a reasonable chance of shooting in and 80m from the great tower). Orford incor- through arrow-loops. The results (at 25 yards porates the principles of concentric defence, 30% of arrows passed through the slit) were flanking towers, and barbican walls, to be de- obtained at close range without, obviously, fended as a unit. anyone trying to shoot the archers.54 There is Two thirds of Henry’s expenditure on little doubt that in action an archer behind an Winchester comprised houses and chapels. arrow-loop, or shooting from a window, where However, it already possessed a late eleventh wooden shutters could aid protection, had a century stone curtain and an early twelfth cen- significant advantage over a besieging archer tury rectangular great tower, revealed by exca- sheltering behind a mantlet shield. vation. In 1169-71 over £237 was spent Many French towns retained their rebuilding the curtain walls.49 Newcastle tower Gallo-Roman town walls. Surviving exam- and surrounding bailey wall were built 1168- ples, such as Le Mans and Carcassonne, have 78 (building operations were interrupted by the D shaped towers, usually solid to wall-walk Scots in 1173, a siege abandoned due to their height, where rooms have round arched win- lack of engines). 50 Scarborough tower was dows. The Romans used windows to shoot built 1158-68, though other work, including ballistae, anti-personnel torsion catapults, the curtain, was probably undertaken.51 £425 effectively.55 Eleventh century warriors lacked was spent on Bowes between 1171 and 1173 equivalent weaponry (medieval references to though the tower was incomplete when the ballistae usually mean crossbows) but the prin- Scots unsuccessfully besieged the castle in ciple of shooting at attackers while they nego- 1174. The tower occupies the corner of a Ro- tiated obstacles was understood so tower man fort, inside a ditched enclosure without builders probably copied Roman windows. evidence of stone fortifications.52 Fulk Nerra, count of Anjou (d 1040), an early The defensive capabilities of a great builder of great towers, went on to tower need to be considered in conjunction the where he would have observed Byzantine fortifications, which often employed with the strength of the site, the castle’s other 56 defences (commonly upgraded simultaneously relatively wide rectangular loops. with the tower’s construction) and the position The introduction of proper arrow- of the tower within those defences. loops, probably connected to increasingly pow- VI erful crossbows, occurred in the later twelfth century. The Hospitaller castle of Belvoir in Defensive Deficiencies? the (started 1168) has 57 Liddiard regards the existence of large win- many arrow-loops. The Avranches tower at dows in towers as a ‘potentially serious weak- Dover (), designed to block a causeway, ness’, illustrated by an incident at Torrington probably contain the earliest in England, ap- 58 in 1139. Henry of Tracy, learning from scouts parently designed for crossbows. Throughout that part of the garrison was engaged on a the thirteenth century arrow-loops became plundering expedition, made a stealthy night more sophisticated (offset cross-loops, plung- attack, eluded the guards, and flung torches ing loops etc.) but usually provided a restricted through the turris fenestras, setting the interior view. However, for snipers and lookouts seek- aflame. Note the absence of ‘siege etiquette’ - ing to observe specific features without fear of this was a surprise attack on an under-strength being shot they were valuable. garrison.53

The Castle Studies Group Journal No 21: 2007-8 217 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence

We underestimate the usefulness of was at the forefront of the new ideas. Whereas windows for both archery and visibility. An Henry II’s militarily sophisticated Dover, like analogy is helmet design. Eleventh century Hospitaller Belvoir, overwhelmingly incorpo- helmets provided only partial facial protection, rated rectangular designs, Richard’s Château- as a number of high profile casualties hit in the Gaillard, built 1196-8 at huge expense (the face attest, but allowed unimpeded visibility equivalent of about £8,000 - his total expendi- and movement. Later, heavier helms provided ture on English castles during his ten year almost complete protection but limited visibili- reign was some £7,146), employs rounded ty and movement. There were still casualties, shapes. It blocked the French approach route because knights raised their visors to see… to and provided a base from which to Rectangular towers had a weakness; re-conquer the Norman Vexin, lost while the corners restricted visibility and were vul- Richard was imprisoned in 1193. Liddiard, nerable to mining. The ‘traditional castle story’ however, is unconvinced by military explana- held that before discovering the solution, round tions: ‘Explaining some alarming military de- towers, a number of ‘transitional’ towers, of fects at Château-Gaillard (the castle is built on varying shape, were built. However, chronolo- soft chalk and there does not appear to have gy confounds this theory: New Buckenham’s been a well) is rendered unnecessary when it is round tower was probably built in the considered as a visible statement of Anjevin and Henry II’s polygonal tower at Orford intent to reassert lordship over lost territories. (1165-70) pre-dates his rectangular towers at The provision for the probable throne room (in the donjon) underlines the castle’s political Newcastle (1168-78) and Dover (1180s). 61 Speight tracked Allen Brown’s changing pub- purpose…’ lished views from initially proposing a catego- Throughout 1194 and 1195 Richard ry of transitional, polygonal keeps without successfully fought to regain losses in his vulnerable sharp angles to finally thinking French territories using tactics of speed, rapid fashion and rival theories influenced shape assaults on enemy field forces, and taking rather than progress to a superior form. As castles.62 How would Richard reassert author- Speight notes, he moved significantly towards ity over the Norman Vexin from Château- the ‘symbolists’.59 Gaillard; by holding ceremonies or by war? Liddiard argues that analysis of don- VII jons reveals ‘alarming deficiencies’ defensive- Visibility and Mining ly; ‘even when allowing for the reduced defensibility of square towers, some were built Round towers are inconvenient for incorporat- to extremely poor military designs’ e.g. ing complex internal domestic arrangements. Orford’s three buttresses multiplies the number Might non-domestic utilitarian reasons have of corners and creates blind spots; Conisbor- inspired Richard to build round towers at Châ- ough and Trim towers incorporated similar teau-Gaillard? drawbacks. Thus ‘while some aristocrats may The theoretical military weaknesses have favoured rounded keeps…this was cer- of rectangular towers were: (i) restricted visi- tainly not due to the fact that builders had final- bility and limited ability to shoot outwards at ly realised the best ‘military’ form for their the corners, leading to ‘dead zones’, (ii) cor- 60 great towers’. ners were vulnerable to mining, and (iii) flat Building great towers for display or as tower walls were vulnerable to battering from an aristocratic badge were doubtless partial stone-throwing siege engines hitting them at motivations so we cannot completely reject the right angles. The Romans knew the answer: idea of round or polygonal designs as fashion Vitruvius advocated round or polygonal statements, but it is interesting that Richard I towers.63 Note that octagonal was almost as

218 The Castle Studies Group Journal No 21: 2007-8 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence effective as round in minimising ‘right angle’ them alight. The tower collapsed during the weak points (polygonal and round towers night but by morning the Turks had blocked could be mined, though with more difficulty). the breach, thwarting an assault.69 The Cru- Using hoards, wooden galleries pro- saders did not employ mining at the sieges of jecting from the battlements, a tower base (taken through treachery after a long could be covered by dropped missiles or ar- blockade and the defeat of several relieving armies) and Jerusalem in 1099 (a bloody as- chers leaning through crenels to shoot down; 70 tower shape was largely irrelevant. Hoards are sault by siege tower). usually depicted, often by Viollet-Le-Duc’s Evidence of mining is scanty in the drawings, as substantial structures on stout numerous sieges of Stephen’s reign. At Exeter beams (like those reconstructed at Carcas- in 1136 the Gesta Stephani reports Stephen sonne). 64 Twelfth century hoards were proba- ‘summoned those who have skill in mining bly simpler. Hides, matting, and wickerwork under ground and ordered them to search into protected siege towers, the apparently flimsy the bowels of the earth with a view to demol- barriers absorbing much of the force of arrows ishing the walls’, but Exeter castle eventually and missiles, and could have also provided surrendered because the wells ran dry.71 In protection at wall tops. At Brough in 1174 a 1144 Stephen abandoned the siege of Lincoln hung two shields over the battlements in disorder following a successful sally by the and threw javelins, killing three Scots.65 The defenders. Apparently engaged in erecting a curious feature ‘hanging’ from the top of the siege castle, ‘nearly eighty of (Stephen’s) tower at Dol on the Bayeux Tapestry may be workmen were buried alive by the enemy’ shields or a fire-protection.66 Such temporary (Henry of Huntingdon); perhaps they were at- defences only required a light framework. tempting to dig a mine.72 protected archers aiming towards An early western use of galleried ‘dead zones’, less significant anyway when a mining occurred during the Second Crusade at tower was defended in conjunction with its Lisbon in 1147. Crusaders from and bailey. In short, visibility problems were not Flanders took a month digging a mine tunnel, insurmountable. with five entrances, into a slope under the Mining was a technique known to the town wall. ‘Thirty cubits’ of wall collapsed ancients, clearly described by Vegetius, whose when the props were fired but the Crusaders’ fourth century military text was influential in subsequent assault up the slope was repulsed. the middle ages.67 There were two types of Lisbon’s Moslem defenders later surrendered mining: digging into the lower courses and after the Anglo-Norman contingent manoeu- foundations under the cover of an armoured vred a siege tower up to the city walls.73 In the roof (direct) and tunnelling under a wall Anglo-Norman world mining was rare, and (galleried), as at Rochester in 1215. In both success elusive. cases when props used to hold up the founda- There was greater familiarity with tions were fired the wall collapsed. Mining mining in the east. In his memoirs, Usamah, a was rarely a complete technique – success usu- twelfth century Arab warrior, describes the ally required a simultaneous assault. underground tunnel at Kafartab in 1115, and In 1068 the Conqueror attempted to the subsequent collapse of the outer part of a ‘undermine the (city) walls’ (subtus murum tower, though the Crusader defenders only sur- suffodere) (Orderic) of Exeter. It was probably rendered after further fighting.74 In 1144 Zan- insignificant because after eighteen days Ex- gi, emir of , undermined the town wall eter surrendered on generous terms.68 At Ni- of Crusader Edessa. After desperate fighting, cea in 1097 the First Crusaders dug into a Zangi’s men broke in and ‘the looting and kill- tower’s foundations, inserted props and set

The Castle Studies Group Journal No 21: 2007-8 219 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence ing began’ (Ibn al-Qalanisi). The citadel sur- to cope with the greater weights and forces rendered on terms.75 involved. Highly skilled specialists were re- 79 In 1179 besieged the new, quired to operate all types of artillery. and probably incomplete, Templar castle of Le Stonethrowers could be fatal to men; Chastellet near Jacob’s Ford, north of the Sea at Mayet in 1099 a knight standing next to of Galilee. It was a great shock when the Sara- William Rufus had his head crushed by a cens quickly undermined a tower and stormed stone. Orderic records engines at Breval in the castle, killing or capturing the garrison. As 1092 hurling great stones (saxa) which could Kennedy notes: ‘It was the first time Muslim demolish (vallum), roofs, and the sappers had shown their effectiveness against a boundary wall (sepes, which also means hedge major Crusader fortification and it was a sign or fence), all apparently wooden.80 Despite of things to come’.76 Saracen sappers had frequent mentions of stonethrowers, damage to rarely been given an opportunity to mine Cru- stone structures was unusual; a slightly confus- sader castles, a time consuming and skilled ing account of Geoffrey Plantagenet’s siege of activity, because Crusader field armies usually Rouen in 1144 provides a rare example. It relieved threatened fortresses. Notwithstand- seems part of the tenth century tower had col- ing the loss of Edessa the Crusaders were gen- lapsed in 1143. Geoffrey aimed his engines at erally expansionist and their castle building that side of the tower, ‘but they could not win aggressive.77 However, Saladin’s rise to pow- the castle, partly on account of the strength of er changed the strategic balance. In 1187 his its position, and partly on account of the stabil- army annihilated the Crusaders at the battle of ity of the building…At length the besieged, Hattin. Saladin overran the Crusader king- finding their victuals were failing them, sur- dom, Jerusalem surrendering after the city wall rendered’. In 1146 Geoffrey ‘repaired the roof was undermined.78 of the tower of Rouen, and the castle, which VIII had been endamaged during the siege’ (Robert of Torigni).81 Stonethrowers Early throwing engines were certainly Machina, mangana, mangonella, petraria, tor- used to hurl firebrands, sometimes in support mentum, instrumenta, ballista, and funda (with of an assault, as at Jerusalem in 1099. Crusad- variations such as Balearic funda) were all ers, besieging Tyre in 1124, recruited a spe- terms used by chroniclers to describe cialist whose expertise was accuracy in stonethrowing engines, unfortunately usually neutralising the defender’s artillery atop the used imprecisely. Machina was a general fortifications.82 term, ballista also meant crossbow, and while funda was a sling it might be a small hand held The trebuchet (a French word) origi- or stave sling. The evidence for Roman style nated in the east, probably invented by the torsion powered engines is limited; most medi- Byzantines. One appears in an illustrated Ara- eval artillery was lever artillery i.e. a long bic treatise presented to Saladin around 1180. beam with a sling attached at one (rear) end Primitive versions of counterweight manjaniqs rotated about an axle raised on a frame, the may have been used in Saladin’s campaigns of 83 motive power being human traction, pulling 1187-9 against the Crusaders. War is a cata- ropes attached to the front end of the beam lyst for the development and spread of new (traction lever artillery). The trebuchet was a technologies. Richard I’s logistical prepara- development incorporating a heavy counter- tions for the Third Crusade, an extremely diffi- weight, which substituted human power as the cult campaign, were comprehensive, and motive force, suspended from the front of the doubtless included the investigation and adop- beam. Trebuchets required much stronger tion of the latest military ideas. His arrival at frames and axles than ordinary lever artillery Acre in early June 1191 soon after Philip Au-

220 The Castle Studies Group Journal No 21: 2007-8 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence gustus, king of France, re-energised a siege in down a considerable portion of Winchester cas- progress since 1189. The Saracens, battered tle walls in 1216. Damage may explain why into submission, surrendered on 12 July, de- Winchester’s tower was demolished around spite the presence of Saladin’s army attempt- 1222 as part of rebuilding operations.89 ing to break the siege. Eyewitness accounts At the Albigensian Crusade siege of report extensive destruction of walls and tow- Castelnaudry (southern France) in 1211 Peter ers by Crusader stonethrowers and miners, of les Vaux-de-Cernay mentions the enemy sometimes both targeting the same tower.84 deploying an enormous siege-engine The level of professionalism in (machina), which another source calls a trebu- Richard’s army is illustrated by the siege of chet. Peter probably wrote this section of his Darum, south of Gaza. The castle, built by the chronicle in 1213. At Toulouse in 1218 Peter Crusaders in the 1160s, was ‘square in form describes the defenders showering Crusaders and at each corner was a tower, one of which attempting to protect their own siege engines was more massive and better fortified than the with a storm of arrows and a hail of stones rest’ (William of Tyre). In 1170 Saladin be- from ‘two trebuchets, a mangonel and numer- sieged it, apparently forcing the defenders to ous sling-staves’ (duobus trabuchetis, mangan- retreat to the main tower until a relieving force ello et pluribus machafundis). Simon de arrived.85 Saladin strengthened the castle be- Montfort, leader of the Crusaders, was killed cause when Richard besieged it in 1192 it had when a mangonel stone hit him on the head.90 seventeen towers. Three stonethrowers, trans- Either Peter’s knowledge of weapon terminolo- ported by ship in sections, were assembled gy improved or he overcame a reluctance to use and commenced a bombardment. Aleppan a French term in his text. sappers captured at Acre (spared when Rich- Armitage, finding no instance of mota ard executed 2,700 prisoners) were suborned in pre twelfth century chronicles, suggested to mine the castle. A tower collapsed, weak- writers of good Latin might have avoided using ened by mining and shaken by repeated blows. ‘vulgar’ French words.91 Similarly, early ap- The Crusaders pursued those fleeing from the pearances of trebuchets may have been unre- ruins, killing about sixty Turks, those on the corded for want of a Latin word. Perhaps ramparts being flung off the walls, while ar- Philip Augustus’ magna petraria called Chad- chers shot Turks running for refuge in the abula at the siege of Château-Gaillard in 1204 main tower. Richard had refused to grant was a trebuchet. Its pounding brought down terms but the Turks in the tower (three hun- the inner gatehouse, already weakened by dred, plus women and children), realising their mining.92 This six month siege featured virtu- situation was hopeless ‘threw themselves on ally every siege technique but mining and the royal mercy and surrendered themselves stone-throwers were particularly destructive. into perpetual slavery’ (Itinerarium). The The defenders surrendered without trying to whole operation took four days.86. hold the fourth line of defence (the donjon) be- In England a trebuchet is first men- cause there was no prospect of aid; a relief ex- tioned at the abandoned siege of Dover in pedition had been repulsed and King John had 1217 but probably the French used one during departed for England. Château-Gaillard was a their unsuccessful three month siege of Dover powerful fortress with a number of flaws (the in 1216, during which they undermined the middle bailey fell after a daring climb up a la- north gate.87 Many castles were badly dam- trine shute), but no castle is impregnable with- aged in the wars of 1215-17. At Carlisle the out field forces to support it. 4 knights were Scots undermined the walls and seriously killed and 36 captured. As scarcely 90 of the damaged a number of towers, implying the garrison survived this implies only 54 sergeants use of siege engines.88 French engines threw and archers survived from a force of 120 ‘and

The Castle Studies Group Journal No 21: 2007-8 221 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence many others’ (total garrison strength was prob- around £120.95 The Tower of London has a ably between 200 and 300). Many hundreds of round corner stair and a projection to non-combatants also perished after the defend- accommodate the Chapel’s semi-circular apse. ers ejected ‘useless mouths’ to preserve food, Semi-circular apses were common in but the besiegers refused to allow them pas- Romanesque ecclesiastical architecture. A 93 sage, so they starved in no-man’s-land. fine example is Norwich Cathedral’s early IX twelfth century apse, which has two bi-lobed rounded chapels attached. The circular Chapel Tower Shape of the Resurrection in the Church of the Holy Before 1190 mining was rare and often unsuc- Sepulchre in Jerusalem inspired a number of cessful, and thick walled towers unlikely to be round churches e.g. Cambridge (1130s). The damaged by stonethrowers. The vast majority Temple Church in London, and Ludlow of towers were rectangular, a choice doubtless castle’s chapel (1130s), also with Templar as- preferred for ease of interior space utilisation. sociations, are round.96 Around 160 churches From 1190 into the fourteenth century round, in East Anglia have round towers, including polygonal or D-shaped tower designs predomi- about twenty of Saxon origin, some of which nate. One explanation is a change of fashion, were refuge towers.97 William d’Albini per- followed fairly consistently from Krak des haps adapted local building expertise for his Chevaliers to Conway. Another is that dramat- round great tower at New Buckenham, Nor- ic technical improvements in siege techniques, folk, in the 1140s. Unusually, it had a cross- forged on Crusade, revolutionised tower de- wall at basement level and probably contained sign. Trebuchets could shatter stone buildings only two storeys.98 Another inspiration may and exploit damage caused by mining. Con- have been the oval shell keep at Arundel, stant pounding from stone-throwers by day and probably built by Henry I, but perhaps by night (as at Darum) hampered defenders at- D’Albini99, a social climber who married Hen- tempting to temporarily repair breaches, there- ry I’s widow. by eliminating the need for assaults to be Castle Rising, another of D’Albini’s launched immediately after undermined walls towers, has many similarities to Henry I’s tow- collapsed. Trebuchets and mines, used in com- ers at Norwich and Falaise.100 Rising, in a bination by increasingly skilled engineers, sparsely populated region with poor agricul- shifted the advantage to the attack. Militarily tural land and no strategic significance, more there was no ‘transitional’ phase of tower resembles a palatial lodge.101 Howev- shape; Richard I returned home with new ideas er, Norwich was a royal castle overlooking to thwart siege techniques he had perfected. one of the largest and wealthiest cities in the Round, oval or polygonal structures kingdom, and Falaise a fortress commanding were built before 1190 for reasons other than southern Normandy. The towers of these cas- the threat of mining and stone-throwers. Shell- tles were similar, but the castles differed in keeps were adapted to motte shape. ’s importance and function. Philip Augustus, shell is a regular twelve sided polygon though after conquering Normandy, attached a round many, such as Lincoln and Carisbrooke, are donjon to Henry’s tower at Falaise.102 irregular polygons, probably reflecting motte Whatever the design inspiration for irregularities. Windsor’s ‘Round Tower’ is Orford’s polygonal tower, and its complexity almost flat on the south, where subsidence was invites exotic interpretations103, its buttresses a problem.94 On constricted motte tops roofing and the corners in the rectangular curtain wall the whole structure created a ‘true’ tower, such towers show the elimination of exposed right as the eleven sided tower raised on the high angles for military reasons was not an influ- motte at Tickhill in 1179-8, superficially simi- lar in plan to Orford but much cheaper at

222 The Castle Studies Group Journal No 21: 2007-8 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence ence. Possibly royal builders found the (blind arcades, dummy windows etc.) or dis- tower’s polygonal design unsatisfactory. Or- play windows were not incompatible with mil- ford contains a basement under two halls, each itary effectiveness in an era lacking destructive just less than 60sqm in area. The pro- artillery i.e. before 1190. Nor was display; vide a number of small chambers (a 1m thick Richard Basset, having become wealthy, built turret wall resulting where a kitchen was a well-fortified stone turrim at Montreuil to squeezed in).104 The rectangular tower at make ‘a show of superiority to all his peers Newcastle, commenced as Orford neared com- and fellow countrymen by the magnificence of pletion, also contains a basement under two his building’ (Orderic). When Geoffrey Plan- halls, each just larger than 60sqm. It contains tagenet invaded Normandy in 1136 he twice numerous chambers within its thick walls, two assaulted it but withdrew after some of his of which are quite spacious with fireplaces. men were killed.107 Orford’s short external flight of steps and awk- The presumption that military consid- ward forebuilding (incorporating a nearly tri- erations took priority in castle design has been angular chapel) compares unfavourably with convincingly challenged for peaceful four- Newcastle’s processional forebuilding leading teenth century England.108 Defence was some- into the high-roofed upper hall. Orford has times accorded low priority in the twelfth only one entrance, so reaching both basement century; Woodstock and Clarendon were royal and upper hall involved going through the low- palaces rather than castles because they lacked er hall. Newcastle has a at first floor fortifications.109 Castles were usually centres level leading into the main stairway, useful as of estate management, sometimes featuring a service entrance. Despite Orford’s sophisti- parks, fishponds and gardens in arranged land- cated novelty, Newcastle was more convenient scapes. A lightly defended country house (a to use and, costing around £800 to £1,000, thin-walled two storey building with a cross- probably less expensive. The king of Scots wall) was built at Castle Acre around 1080, interrupted building operations in 1173 but more domus defensabilis than castle, and it is concluded he wouldn’t take Newcastle without unlikely to have been unique.110 From an early siege-engines.105 date the balance between fortification and resi- Expense probably explains why no- dence at individual castles varied due to politi- one copied Conisborough. The six massive cal circumstances, region (Wales and East buttresses increased the external wall surface Anglia experienced differing levels of insecu- and material required by over a third but pro- rity), and strategic importance (most castles vided minimal extra accommodation in a tow- had only local significance). Comparisons of er containing only two well appointed rooms, a castles, or great towers, are therefore fraught tiny chapel and storage areas. Undermining with pitfalls, especially since grandiose proj- one of the buttresses, where combined wall ects by wealthy, individualistic lords could re- thickness is over 6m, might bring down the sult in idiosyncratic towers such as New buttress but would almost certainly leave the Buckenham and Conisborough. Each castle cylindrical core unbreached. Conisborough, was different, builders’ motives varied, and probably built c.1180 to 1200 (by an illegiti- thus military theories are unlikely to encom- mate half-brother of Henry II), is a variant on pass every castle. round, rather than a ‘transitional’ shape. However, twelfth century England X witnessed far more war than the fourteenth Conclusions century. In the 1130s and 1140s the defences of Castle Acre, which has no ‘military history’, Most building projects include elements of dis- were considerably strengthened, the country play or ostentation but we should be wary of house rebuilt as a thicker-walled, taller, nar- overemphasis.106 Decorated tower exteriors rower, great tower and the upper ward curtain

The Castle Studies Group Journal No 21: 2007-8 223 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence built and shortly thereafter heightened.110 Mil- Shell-keeps, ‘towers’ to twelfth cen- itary purpose may be inferred from history as tury writers, were an alternative design func- well as architectural interpretation. Though tionally and symbolically similar to sieges are only one manifestation of castles’ rectangular great towers, though most would role in war, between 1066 and 1300 AD, the have been ill suited for ceremonial purposes. main era of Anglo-Welsh castle warfare, 60% As perceived deficiencies in tower design fo- of castles with great towers were attacked. For cus on exposed right angles and windows, it those in royal possession the figure rises to implies shell-keeps, usually lacking external 80%.111 The possibility of physical threat to windows, were militarily superior towers. these castles was not remote, and the defensive However, before 1190 exposed right angles capability of towers added to castles which had were not practical military weaknesses. After already been besieged must have been a con- siege weaponry became sophisticated enough sideration. to exploit exposed angles builders altered tow- er shape though rectangular designs continued Castle defenders’ primary objective for other castle buildings. was to retain control of their main bailey, from which they could launch sorties to disrupt be- The ‘traditional castle story’ pre- siegers’ operations. Fundamentally, a well sumed military necessity predominated castle provisioned and determined veteran garrison architecture, thus underestimating other influ- was more important than clever design in de- ences. Conversely ‘revisionist’ theories reso- fending a fortress. Wide and deep ditches lutely seek non-military explanations. Higham (double or water filled if possible), steep ram- pondered if ‘In the later twentieth century, parts, difficult approaches covered by barbi- however, in a society experienced in the hor- cans, and robustness and quality of masonry rors of total war and with a sense of post-colo- were more important than tower shape. The nial guilt, historians and archaeologists were ‘traditional castle story’ exaggerated the mili- perhaps inclined to seek “peaceful” (and thus tary importance of great towers and shell- more socially acceptable) explanations of the keeps. They were not intended to be freestand- past where possible, with emphasis on social ing, impregnable fortresses. Although towers and economic matters?’ He also reflected how provided security for small garrisons and were ‘revisionist’ arguments had become ‘familiar available as a refuge of last resort, usually used to and had been largely accepted by most prac- to negotiate surrender, their main defensive titioners’ during their long gestation period.112 role was aiding bailey defences. The question Perhaps intellectual culture and consensus in is not: was a great tower primarily designed for the academic mainstream has muted counter defence?, but: was it capable of providing ef- argument, which is thus emerging from out- fective defensive support? side academia.113 Powerful twelfth century fortresses When war intrudes (because no-one did not require rectangular great towers. denies some castles were besieged) new theo- Therefore military ‘deficiencies’ in tower de- ries characterise it as aristocratic war, the pre- sign (which I do not believe have been demon- serve of the social elite, fought within the strated) would have had little impact on castle conventions of chivalrous ritual.114 Though strength. Re-interpretations of internal tower beyond the scope of this paper, the nature and function as ceremonial (usually vague in nature conduct of medieval war requires deeper un- and for which non-structural evidence seems derstanding because even ‘chivalrous war’ scanty) rather than residential does not neces- provided little protection for its main victims, sarily affect ‘external’ defensive capability. the innumerable non-noble victims killed, kid- ‘Revisionist’ arguments over keeps fail to un- napped, raped, or left starving after their prop- dermine military interpretations of castles. erty was burned or plundered during ravaging raids, the main activity of medieval war.115

224 The Castle Studies Group Journal No 21: 2007-8 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence

Ultimately, the story of castles is the scription of Wales’ translated by Lewis Thorpe p148-9; Wales generally, Brut p19, 20 and history of those who lived in and built them. notes p158. In the twelfth century warriors outnumbered dilettantes. ‘Gentle’ King Stephen was 11. Chepstow: quote p34, plans p33, reconstruc tions fig. 19 p28, and drawings fig. 12 p20, fig. obliged to fight almost every year to retain his 25 p34. throne. Others embraced war; from the age of 12. P J Drury ‘Aspects of the Origins and develop fifteen in 1173 to his death from a crossbow ment of ’ in Archaeological bolt in 1199 Richard I campaigned every year Journal 139 for 1982 p398, 400. except when he was captured and incarcerated 13. Chepstow p24 n8 (p275) acknowledges the in returning from the Holy Land and problem, p29 notes uncertainty about recon- 1180, when his activities are unrecorded.116 struction fig. 19, the west internal elevation. Even on a sick bed at the siege of Acre he had 14. Philip Dixon and Pamela Marshall ‘The Great himself carried forward to a defensive shield. Tower in the Twelfth Century: The Case of ‘There he used his crossbow, with which he Norham’ in Archaeological Journal 150 (1993) p410-32. was skilled, and killed many (Saracens)’ (Itinerarium). He later shot a Saracen flaunt- 15. Edward Impey ‘The Seigneurial Residence in 117 Normandy, 1125-1225: an Anglo-Norman Tra ing himself in captured armour. In 1194 the dition?’ in Medieval 43 (1999) garrison of Nottingham castle failed to submit p45-73, especially 62-4 on the Exchequer hall; to Richard on his return to England. He shot Chatelain p117 plan IV for Caen great tower. dead a knight in an initial unsuccessful assault, 16. Richard of Hexham in ‘The Hexham Histori then had some prisoners hanged to concentrate ans and the chronicle of Jordan Fantosme’ the defenders’ minds.118 Richard consistently translated by Joseph Stevenson (reprint from The Church Historians of England) p45-6, Lat waged war ruthlessly and professionally. He in quoted in King ii p371 n228. was skilled at besieging and fortifying - and had little need of architectural stage settings to 17. Brionne, WP p11; Mayenne, WP p67, Rob Ear ly ‘Le Château de Mayenne’ in Château-Gail command awe. lard 20 p247-262, Richard Hulme ‘William of Poitiers and Castle Warfare in Normandy be Richard Hulme fore 1066’ in Castles Studies Group Journal 19 Notes p247-51. 1. Brown: quote p18, description of Rochester 18. Arques, WP p35-43, OV iv p85, Hulme in Cas siege p180, 191. tles Studies Group Journal 19 p248-9; Nor wich, OV ii p317; Dol, ASC Manuscripts D,E; 2. HKW quotes p71, 77. Sainte-Suzanne, OV iv p49, Chatelain p127 3. Philip Dixon and Pamela Marshall ‘The great dates the tower to the 1080s on his time chart. tower at Hedingham: a reassessment’ in A-N 19. ASC 1088AD; OV iv p127; JW p53. Castles p 297-306. 20. ASC 1095AD; JW p77; HH p38; Newcastle 4. Liddiard: on tower chronology p49, courtly charters HKW p745 n10. choreography p51, quote p53. 21. Abandoned siege of Falaise, C Warren Hollis 5. Charles Coulson, Viewpoint in Castles Studies ter ‘Henry I’ p189-91, OV vi p90/91 for quote; Group Bulletin 17 p115. Dover, OV ii p205. 6. Liddiard p48. 22. Richard Eales ‘Royal power and castles in Nor 7. Charles Coulson ‘Castles in Medieval Society’ man England’ in A-N Castles p47-53 for cas tle p382. numbers; Brown p215 for numbers docu mented by 1216. 8. Chepstow p35, quote p42. 23. Thompson p64-66 contains tables of ‘keep de 9. HKW p491. tails’ of towers, excluding unroofed shells, 10. Early castles, R R Davies ‘The Age of Con built by 1216. Eliminating wooden, non Eng quest’ p34; Anglesey, Gruffud p73; Hen Do lish and Welsh towers, and ‘repetitions’ leaves men, ASC 1095AD; Pembroke 1096, Gerald of 38 hall-keeps, 22 solar-keeps and 15 round and Wales ‘The Journey through Wales/The De polygonal keeps, totalling 75 towers. As ever in castle studies, the great tower status of some

The Castle Studies Group Journal No 21: 2007-8 225 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence

is debatable e.g. Berkhamsted, Grosmont, Ox- teau-Gaillard XVII 1994 p119-26, especially ford (St. George’s tower), and West Malling. p124, 126 (reprinted in A-N Castles). Other towers survive e.g. Coity, Colchester, 35. ‘Suger The Deeds of Louis the Fat’ translated New Buckenham, Pevensey, Usk, and Win- by Richard Cusimano and John Moorhead p89. chester. Some ‘royal’ towers have disappeared e.g. Bristol, Gloucester, Horston, Northampton, 36. GS p69-70 (ch50); JW p286-7; Bishop Roger’s Nottingham, and St. Briavels. Adding vanished castles at Devizes and Sherborne, WM p45. baronial towers a total exceeding a 100 by 1216 was probable. Approximate numbers by date 37. ‘Roger of Wendover’s Flowers of History’ were obtained by tentative individual dating, translated by J A Giles Vol 2 Part 1 p203 erring towards early dates. (1995 reprint). 24. Bristol, GS p37 (ch27), 43-4 (ch30); Corfe, 38. OV vi p521-3; JW p251. GS p56 (ch39); Pevensey, John Goodall Eng 39. Brionne, OV iv p207-11; Le Renouard, OV vi lish Heritage guide p22, GS p134-5 (ch106); p217; Brough, JF p111. Exeter, GS p25 (ch18); Wark, Richard of Hexham (see note 16) p46-7, 51, 54; Walling 40. Bayeux, C Warren Hollister ‘Henry I’ p185-6; ford, GS p61 (ch42) for 1139, 122 (ch94) for Évreux, OV vi p229; Nottingham, JW p291; St 1146, 150 (ch117), 156-7 (ch120) for 1152-3; Pierre, OV vi p83; Wherwell, Sidney Painter Lincoln, HH p75 for 1140/1, 82-3 for 1144; ‘William Marshal’ p7-8. Oxford, GS p94-5 (ch70-2), WM p133, HH 41. HKW p45-7; building date, ASC 1097AD. p82. 42. Philip Dixon ‘The Myth of the Keep’ in SRWE 25. 1138 siege, OV vi p 527; surrender in 1141, RT p9. p55. 43. The castles are listed in: RT p34, 45; HKW 26. Nicholas Riall ‘The New Castles of Henry of p35; Richard Hulme ‘Henry I and Norman Blois as Bishop of Winchester: The Case Castles’ in Castles Studies Group Journal 20 against Farnham, Surrey’ in Medieval Archae p220-1. ology 67 (2003) reproduces the Winchester an- nal of 1138 concerning the turri fortissima on 44. Allen Brown ‘Royal Castle-Building in Eng- p118. Siege of Winchester: GS p86-89 land, 1154-1216’ in Charters. ch65-7); JW p209; HH p81; Robert’s impriso- 45. HKW p75. nent in Rochester turrim, WM p117. 46. Christopher Drage ‘Nottingham Castle – A 27. English Historical Documents Vol II 1042- Place Full Royal’ p38-9; Armitage p176/7 for 189 edited by David Douglas and George re- references to the mota. enaway p438-9 for the charter; HKW p772, 65 for references Oxford and Windsor mottes; 47. HKW p865. rmitage p181 and St John Hope ‘Windsor Cas 48. John Rhodes ‘Orford Castle’ English Heritage le’ p13 for Latin. guidebook. 28. HKW individual castle entries, p562, 674, 40; 49. Martin Biddle and Beatrice Clayre Durham, H&B p118-9; Warwick, Armit ‘Winchester Castle and the Great Hall’ espe age p 232 n1. cially p5, 8, 14; expenditure, HKW p855. 29. OV mota refs i p328; dangio refs iv p114, vi 50. HKW p746; abandoned attack, JF p45. p148, 224, 342. 51. HKW p830; ‘Scarborough Castle’ English 30. Brown: quote p32, p35, 86; H&B reproduce Heritage guide by John Goodall p12 attributes and analyse Lambert’s description of Ardres the east curtain to Henry though p115-6. it may be earlier, as William of Au 31. St John Hope ‘’ many referenc- male built a wall, ‘The History of William of es, particularly p60-1. Newburgh’ translated by Joseph Stevenson p445-6. 32. HKW individual castle entries, p554 n3, 562, 772, 835, 840, 888. 52. HKW p574. 33. GS p32-3 (ch23); Brown p191-4 for the 1224 53. Liddiard p50-1; GS p55 (ch38), where fenes- siege. tras is translated loopholes. 34. Rhuddlan, Gruffud p62; 1116 attacks, Brut 54. P Jones and D Renn ‘The military effective p40-41; Dinefwr, Brut p87-8; generally John R ness of arrow-loops’, in Château-Gaillard IX- Kenyon ‘Fluctuating Frontiers: Normanno- X p445-456. Welsh Castle Warfare c1075 to 1240’ in Châ-

226 The Castle Studies Group Journal No 21: 2007-8 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence

55. Stephen Johnson ‘Late Roman Fortifications’ Michael the Syrian in ‘Chronicles of the Cru p40-41, reconstruction of tower p34, illustra- sades’ edited by Elizabeth Hallam p118-121. tions after p112. 76. Kennedy: quote p57, 106. 56. Clive Foss and David Winfield ‘Byzantine Fortifications’, esp. photos of eighth to tenth 77. R C Smail ‘Crusading Warfare 1097-1193’, century walls of p248, 250; especially p209-13. Stephen Turnbull ‘The Walls of Constanino- 78. The Old French Continuation of William of ple’ (Osprey) for illustrations of the walls. Tyre in ‘The Conquest of Jerusalem and the 57. Denys Pringle ‘Secular Buildings in the Cru Third Crusade – Sources in Translation’ by sader Kingdom of Jerusalem’ p32 with plan Peter W Edbury p56-8. and references; Kennedy p58-61. 79. Rogers, Appendix III p254-273, discusses 58. Derek Renn ‘The Avranches Traverse’, in Ar- terms and twelfth century artillery. chaeologia Cantiana 84 (1969) p79-92. 80. Mayet, OV v p259-61; Breval, OV iv p289- 59. Brown ‘English Castles’ 2nd ed 1962 p52; 91. Brown ‘The Architecture of Castles’ 1984 81. RT p58 for siege, p60 repair, p34 for attribu p51; Sarah Speight ‘British Castle Studies in tion of the tower to Duke Richard I (died the late Twentieth and early Twenty First 996). Centuries’ at www. history-compass.com. 82. Rogers p83, 114, 247. 60. Liddiard: ‘traditional story’ p47, deficiencies and quote on round towers p50. 83. Suggested by Kennedy p107; David Nicolle ‘Medieval Siege Weapons (2) Byzantium, 61. Liddiard p54. the Islamic World & India AD 476-1526’, 62. John Gillingham ‘Richard the Lionheart’, (Osprey), discusses Saracen machines, p22 purpose of Château-Gaillard p262-5, wars of illustration of the manjaniq. 1194-5 p245-259. 84. Howden Vol 2 Part 1 p209-14; Itinerarium 63. Kennedy p114. p201-18, especially 208-9, 213; Baha al-Din p155-61. 64. e.g. Brown ‘English Castles’ 3rd ed. p184; Brown ‘The Architecture of Castles’ 1984 85. T S R Boase ‘Castles and Churches of the p54, 55. Crusading Kingdom’ p42-3 for quote and siege. 65. JF p113. 86. Itinerarium p316-9, executions at Acre p231; 66. H&B p151. Baha al-Din, for Aleppan sappers p203, exe cutions p165. 67. ‘Vegetius: Epitome of Military Science’ trans lated by N P Milner p128. 87. John Goodall ‘Dover Castle and the great siege of 1216’ in Château-Gaillard 19 p91- 68. OV ii p213; ASC, D manuscript for 1067AD. 102, trebuchet p95. 69. ‘Gesta Francorum’ edited and translated by 88. M R McCarthy, H R T Summerson, and R G Rosalind Hill p15. Annis ‘Carlisle Castle – A survey and docu- 70. Sieges of Antioch and Jerusalem: Rogers mentary history’ p126. p25-39, 47-63. 89. HKW p858; John R Kenyon ‘Medieval For- 71. GS p23 (ch16), 25 (ch18). tifications’ p46; M Biddle ‘Excavations at Winchester, 1969, eighth interim report’ in 72. HH p83; ‘William of Newburgh Book 1’ Antiquaries Journal 50 (1970) p277-326 translated by P G Walsh and M J Kennedy especially p291-2. p73. 90. ‘The History of the Albigensian Crusade - 73. ‘De Expugnatione Lyxbonensi – The Con Peter of les Vaux-de-Cernay’ translated by quest of Lisbon’ translated by C W David W A Sibly and M D Sibly, Castelnaudry p143-5, 165. p133, 52 n45, dates of writing p xxv, Tou 74. ‘An Arab-Syrian Gentleman & Warrior in the louse p276, 277. Period of the – Memoirs of Usamah 91. Armitage p91 n2. Ibn- Munqidh’ translated by Philip K Hitti p102-4; long quote in Kennedy p103-4. 92. D J Cathcart King ‘The Trebuchet and other Siege-Engines’ in Château-Gaillard IX-X 75. Ibn al-Qalanisi in ‘Arab Historians of the Cru p457-469, quote from William the Breton sades’ edited by Francesco Gabrielli p49-50;

The Castle Studies Group Journal No 21: 2007-8 227 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence

and possibility that Chadabula was a trebu 108. Liddiard on the ‘battle of Bodiam’ p7-11 chet p461. (and its implications), chart p72 showing much reduced levels of sieges in the four 93. Jim Bradbury ‘Philip Augustus’ p147-50; Sir teenth and fifteenth centuries; Matthew John Maurice Powicke ‘The Loss of Normandy’ son ‘Behind the Castle Gate’. p255-6; Sean McGlynn ‘Useless Mouths’ in History Today, June 1998 p41-6; John S 109. HKW p910, 1009. Matilda fortified Wodstock Moore ‘Anglo-Norman Garrisons’ in Anglo- in 1142, GS p91. Norman Studies 22 p205-259, p242 and 110. J G Coad and A D F Streeten ‘Excavations at n312 for garrison numbers and casualties. Castle Acre, Norfolk, 1972-77’ in Archaeologi- 94. Medieval Archaeology 36 (1992) p195. cal Journal 139 for 1982 p138-301, particularly 178-94 for the country house and its transition 95. HKW p76 (plan of tower), p844; R Allen to a tower. Brown ‘Castles from the Air’ p214-6. 111. King’s data for attacks on castles for the 75 96. ‘ Its History & Buildings’ towers listed by Thompson p64-6 revealed: edited by Ron Shoesmith & Andy Johnson 45, 60%, were attacked. 30 of the 75 castles p145-54 (by Glyn Coppack). were in royal possession at 1189 (from Allen 97. Stephen Friar ‘The Sutton Companion to Brown ‘A List of Castles, 1154-1216’ reprint Churches’ p451. ed in Charters) of which 24, 80%, were at tacked. While some of King’s incidents were 98. Derek Renn ‘The Keep at New Buckenham’ minor he excluded occasions when attackers in Norfolk Archaeology 32 (1958-61) p232- declined a siege because of the strength of 5. fortifications e.g. Bristol in 1138 and Newcas 99. Bill Woodburn and Neil Guy ‘The Castle tle in 1173. Studies Group Conference – April 2005’ in 112. Quotes from R A Higham’s favourable review Castles Studies Group Journal 19, Arundel of Liddiard’s ‘Castles in Context’ in Castle p9-24, consider D’Albini a possible builder Studies Group Journal 19 p256-7. Another c1138 (p16), without ruling out Henry I. positive reviewer, Amanda Richardson in Me- King (King ii p469) thought the shell late dieval Archaeology 50, 2006 p405-6, would, eleventh or early twelfth. I favour Henry however, have liked more explanation of why because of Arundel’s importance in south military interpretations are avoided. coast defence after Robert Curthose’s 1101 invasion. Caen stone, used in the shell, 113. e.g. Peter Purton ‘Donjons: some heretical might have been difficult to obtain in 1138. thoughts’ in Castle Studies Group Newsletter 16p86-88, questioned the pre-eminence of 100. Philip Dixon and Pamela Marshall ‘Norwich ideas of symbolism over multi-purpose and castle and its analogues’ in SRWE p235-243, practical interpretations based on military at p238; HKW p38-9. history, views advanced with some 101. Robert Liddiard ‘Castle Rising, Norfolk: a ‘trepidation’ as he lacks a background in ar ‘Landscape of Lordship’?’ in Anglo-Norman chaeology. Studies 22 p169-186. 114. e.g. Liddiard Ch 4 p70-95, especially p78-80 102. Jean Mesqui ‘Châteaux et Encientes de la and 84-87; Sarah Speight ‘Castle warfare in France Médiévale – Vol 1 Les Organes de la the Gesta Stephani’ in Château-Gaillard XIX Défense’ photo and plan p119, 123. p269-74. 103. T A Heslop ‘Orford Castle, Nostalgia and 115. Ravaging in the ‘minor’ rebellion of 1123-4, Sophisticated Living’ in A-N Castles p272- Hulme ‘Henry I and Norman Castles’ in Cas- 96. tles Studies Group Journal 20 p216-223, espe- cially p217-8. All chroniclers of Stephen’s 104. John Rhodes ‘Orford Castle’ English Heri- reign emphasise violent disorder, examples tage guide for plans. from GS include p41-2 (ch29) robbery, 75 105. Barbara Harbottle ‘The Castle of Newcastle (ch55) slaughter and plundering in Lincoln, 97 upon Tyne’ (guidebook) contains a section (ch74) pillaging Wilton nunnery, 101-4 (ch78), and plans; abandoned siege, JF p45. 112 (ch86), 127-8 (ch100) general lamenta- tions, 145 (ch114) firing crops to cause starva- 106. A point made by Philip Dixon ‘The Myth of tion. the Keep’ in SRWE p11. 116. John Gillingham ‘Richard I and the Science of 107. OV vi p467-9. War in the Middle Ages’ in Anglo-Norman Warfare edited by Matthew Strickland p194-

228 The Castle Studies Group Journal No 21: 2007-8 TWELFTH CENTURY GREAT TOWERS: The Case for the Defence

207, p195-6; still an excellent introduction to the Howden. ‘The Annals of Roger of Hovedon’ theory of twelfth century warfare. (Howden) translated by Henry T Riley 1997 117. Itinerarium p213, 214. reprint 118. Howden Vol 2 Part 2 p314-5. Itinerarium ‘The Chronicle of the Third Cru- sade – The Itinerarium Peregrinorum et Ges- Abbreviations and Bibliography ta Regis Ricardi’ translated by Helen J Nicholson A-N Castles, ‘Anglo-Norman Castles’ edited by Robert Liddiard, 2003 (collection of pa- JF. ‘Jordan Fantosme’s Chronicle’ edited pers). and translated by R C Johnston 1981 Armitage. Ella. S. Armitage ‘The Early Nor- JW. ‘The Chronicle of John of Worcester’ man Castles of the British Isles’ ,1912. Vol III edited and translated by P McGurk ASC Anglo-Saxon Chronicle. Kennedy. Hugh Kennedy ‘Crusader Castles’ 1994 Baha al-Din. ‘The Rare and Excellent History of Saladin by Baha al-Din Ibn Shaddad’ trans- King. D J C King ‘Castellarium Anglicanum’ lated by D S Richards 2 vols 1983 Brown. R Allen Brown ‘English Castles’ 3rd Liddiard. Robert Liddiard ‘Castles in Con- edition, 1976 text’ 2005 Brut ‘Brut Y Tywysogyon or The Chronicle of OV ‘The Ecclesiastical History of Orderic the Princes’ Peniarth Ms 20, translated by Vitalis’ edited and translated by Marjorie Thomas Jones. Chibnall 6 vols Charters. ‘Castles, Conquest & Charters – Col- Rogers. R Rogers ‘Latin Siege Warfare in the lected Papers’ by R Allen Brown. Twelfth Century’ 1992 Chatelain. André Chatelain ‘Donjons Romans RT . ‘The Chronicles of Robert de Monte’ des Pays d’Ouest’ 1973 (aka Robert of Torigni) translated by Joseph Stevenson 1991 reprint Chepstow.‘Chepstow Castle Its History & Buildings’ edited by Rick Turner & Andy SRWE. ‘The Seigneurial Residence in West- Johnson, 2006. ern Europe AD c800-1600’ edited by G Meirion-Jones, E Impey and M Jones, 2002. Gruffud. ‘The Life of Gruffud ap Cynan’ trans- lated by D Simon Evans Thompson. M W Thompson ‘The Rise of the Castle’ 1991 GS. ‘Gesta Stephani The Deeds of Stephen’ translated by K R Potter 1955 (chapter numbers WM ‘William of Malmesbury: Historia No- are given because the 1976 edition has radical- vella’ edited by Edmund King, translated by ly different page numbering) K R Potter, 1988 HH ‘Henry of Huntingdon: The History of the WP. ‘The Gesta Guillelmi of William of Poit- English People 1000-1154’ translated by Diana iers’ edited and translated by R H C Davis Greenway. and Marjorie Chibnall. H&B Robert Higham and Philip Barker ‘Timber Castles’, 1992. HKW. R Allen Brown, H M Colvin, and A J Taylor ‘The History of the King’s Works’ 2 vols 1963.

The Castle Studies Group Journal No 21: 2007-8 229