6 3 PROCEEDING E SOCIETYTH F O S , DECEMBE , 192712 R .

III. A NEW SURVEY OP . BY W. DOUGLAS SIMPSON, M.A., D.LiTT., F.S.A.ScoT. In my book upon The Castle of Kildrummy: its Place in Scottish History and Architecture, published in 1923,1 have given a full historical accoun architecturad an t l descriptio f thino s noble building. Since that date, however, further excavatio e clearin froth y ruine d iv m th f nan go s have reveale d importand an man w yne t features conn ; whilow - y m e tinued intensive study of the castle, both on the structural and on the documentar increasey ym sided an , d knowledg experiencd ean e th n ei general subjec f mediaevao t l military engineering, have suggestea t dno few corrections, additions d improvementan , e accounth n o s t already published. Accordingly I welcome the opportunity, afforded by this paper,1 of submitting a brief new survey of the ruins, along with an up-to-date and more accurate plan.

I. HISTORICAL INTRODUCTION. The only specific account which we possess with regard to the foundation of is a note by Sir Robert Gordon in his Genealogical History e Earldomth f o f Sutherland,o "written about 1630. After describing the life and actions of that great ecclesiastical states- man, Gilbert de Moravia, Bishop of Caithness from 1223-45—the founder of Dornoch Cathedral, and the last Scotsman to be canonised—Sir Robert proceeds to tell us that "this Sanct Gilbert wes appoynted be King Alexande e Secon th thesaurere b s majestio dt hi e nort r th f fo ro n h i c ; and dureing the space he had this office he built the castle and f ortresse of Kildrume in , with seaven tours within the precinct of the said castle."2 Such a notice, in a chronicler writing four centuries afte evente th r , mus f courso t treatee eb cautionde witdu l hal . Fortun- e presenatelth n yi t case collateral circumstances exist which tend very strongly to support the accuracy of Sir Robert Gordon's account. In the first place, we have to recollect that he had a peculiar interest in Kildrummy, and also special opportunities for research into its earlier history. Sir Robert was tutor to the son of John, twelfth Earl of Sutherland, and Earl John was married to a daughter of Alexander, t shoul1I explainee db d that this paper appear furthea s sa r instalmen systematia f to c survey of the early in Mar, which I am making under a research grant from the Carnegie Trust for the Universities of Scotland. 2 Genealogical History Earldom e oth f of Sutherland, . . 181332 ed . ,p SURVEW NE A F KILDRUMMYO Y7 3 CASTLE. fourth Lord Elphinstone, the owner of Kild rummy. The connection thus formed betweefamilieo closa tw time s e e eon esnwa th th one t A . Sutherland writs were removed from Dunrobi Kildrummo nt r proyfo - tection r RoberSi . t Gordo knows ni havo nt e visite Donside dth e castle, o hav t d access charteit d ha e o an t s r chest. independenn A 1 t linf o e evidence is supplied by the Breviary (1509), which states that Bishop Gilber employes wa t Kiny db g Alexander II., garrisonin"n i d gan building n repairincastlei d an s g other e e benefiedificeth th r f o fo ts monarch and the state."2 We have also the entry in the Martyrology of Aberdeen, dating from about the same period, in which we are informed that Bishop Gilbert " built castles agains wile furth e d f tth yo and barbarous tribes."3 The language of these two entries places the erection of what Cosmo Innes justly called " the noblest of northern castles,"4 against its proper political background as a detail in the measure outstandine th f o se leadinon go t achievementp gu Alexandef so r e Second'th s reign reductioe th , greae th f tno provinc f Moraviaeo e th , district between the Spey and the Dornoch Firth; an old Celtic palatinate which—stron s allegiancit n e Housi g th f o Macbetht eo e bitterld an , y resentfu Normanisine th f o l g tendencie rivae th f ls o Hous f Canmoreeo — had for two centuries fiercely resisted the process of compulsory in- corporation withi expandine nth g real f Scotlandmo . Kildrummy Castle is located just half-way between the two important ancient centres of Brechi chaia d Elgin f nan f o formd strongholds n o e an , on s , royad an l baronial, which controlle e greadth t route northwards from Strathmore ove Monadhe rth r o Strathbogid througo s an d r an , hMa e into e disaffecteth dp (fig ascribe) area. 1 ma .e secon th A 5 o dt d quartef o r

e GenealogicalSe 1 Historye Earldomth f o f Sutherland,o . 296p ; als r WoSi . Frasere Th , Sutherland Book, vol. i. p. xl; vol. iii. p. 194; vol. ii. pp. 147, 349; and his The Lords Elphinstone of Elphinstone, vol. i. pp. 141-4. * In temporalibus et spiritualibus ipsum pro singulis regum in boreali Scocie parte agendis et castrorum custodiendis edificandis et aliis edificiis pro utilitate regis et rei publice reparandis prefecerunt, see Bishop A. P. Forbes, Kalendars of Scottish Saints, p. 355. 3 Contra rabiem indomitarum silvestriumt e gentium edificans—ibid., . 130p . Sketches of Early Scotch History, footnote, p. 79.

4 6 See my Castle of Kildrummy, pp. 3, 49-51; also my Castle, 2nd ed., pp. 3-5 ; and my papeRoyae "n Th ro l Castle of Kindrochi Proceedings, "n r i Ma n i t vol. Ivii 82-5. .pp . perfectlIs i t y clear that the significance of such very powerful castles as Kindrochit and Kildrummy, near the heads of river basins amid the mountains of Western , is to be found not in the east-and-wes r blind-alleo t y strateg thesf yo e narrowing valleys t rathe north-and-soute ,bu th n ri h or transversal strategy of the trunk roads leading across the Mounth towards Moravia. On an east-to-west strategy the position of these castles is meaningless; they are at " the back end of r to-da thino t Ma e moder termt n f yi th ko ap f o so nto nowhere. rairoad e ar an ld e comW " - munications ascending the Dee and the Don; whereas the mediasval mind would regard the whole problem transversely ancienn I . t time cross-countre th s y routes, over the Mountd han northwards, greater werfa f eo r importance tha roade nth s ascendin valleyse gth . Mr W. Maokay Mackenzie (The Mediaeval Castle in Scotland, pp. 20-1), albeit sceptical about what he calls the assumption "that in the planting of these fortified structures a strategic

SURVEW ANE F KILDRUMMYO Y9 3 CASTLE. the fourteenth century brings the castle into intimate relationship with the important Mounth passes. It is depicted as a great building of hewn stone, towered and battlemented, and south of it is marked the Capel Mounth Pass, with the significant words hie unum passagium.1 n accordancI e with normal mediaeval procedure e castlth , e thus founded -by a royal minister, with a national end in view, would be entruste e hereditarth o dt y wardenshi e locath f l po feuda l landholders, Celtie th c Mormseor throwd ha f Marw o so n ne wit wh ,n thei i e t hth lo r regime, unded an r their novel titl f Earleo s were activ n organisini e g parishes, as at and Migvie, founding and endowing monastic houses, suc s Monymuskha throwind an , earthworp gu k castles like eth Peel of Fichlie and the Doune of Invernochty—all these being the out- ward and visible signs of the Norman penetration. A phenomenon very

familia studento t r f thio s s process, alik n Scotlani e Englandn i d dan , 2 constane isth t juxtapositio f parisno h churc castled han s representin,a g respectively the ecclesiastical and civil nuclei of the early parochial organisation whicn i , parise hofteth s hwa n co-extensive wit manore hth , and the priest was a younger son of the lord. Examples of this associa- tion—so fule f earlieimporo th l r rfo t topographical r historou f o y country—are frequen n Aberdeenshirei t t Lumphanana s a , , Midmar, , Auchindoir, Migvie, Coull, and elsewhere. Accordingly it will at the first sight seem highly remarkable that the important case of Kildrummy appears directly to violate this so constant rule. Here (see sketch-map e ruine, th fig ) 2 . d mediaeval parish church still picturesquely crowns its burial mount, and to the southward stretched the ancient principl observed,s ewa forces "i admio dt significance th t line f "f positioneo th e o greae th n tso mediaeval route throug royae case th th f l eho n castle i Mar. t Bu "s with their associated burghs, he seeks to find the explanation not so much in considerations of strategy as in the idea that there were older centre populatiof so t thesna e points tha d castlee an ,th t s were raised " witn ha ey dominatino et locae gth l24-5. highway. e populatiohimsele pp th h , t t a t no fd Ye point"nan s t tha t burghexise ou tth no td sbefordi castlese th e t werbu , e founded contemporaneously with them. Surely this suggests that the military position of the castle was the first consideration. afted l KildrummyAn al r chaie greateste th f th castle,l no al f o s under review t merelno , s ywa never associated with a royal burgh, but also (as we shall see) was deliberately planted on a site apart fro earliee mth r local centr populationf eo . privilegem a 1I reproduco dt relevane eth . GoughtS . portio(M , p thif Genno sma . Top. 16) through the courtesy of Bodley's Librarian, who also kindly had the photograph taken at my request datThe e .assigne tex the tha is t din t favoure Dby dr Craster, Keepe Westerof r n Manuscript e Bodleiath n si n Library therd topographicae an ; ear whicp l ma featureh e th n i s appear to me clearly to point to such a date being approximately correct. It is interesting to note that Kildrummy is the only castle shown in this graphic manner and on so large a scale, all the other castles being indicated merely by conventional tourelles of small size. This fact, together wit precise hth e Mountindicatioo tw e hth f Passesno , suggests strongly thacartoe th t - grapher was familiar with the castle, and wished to accord it special distinction on his map : possibly as a favourite residence of David II. and his aunt, the Lady Christian, sister of King Robert Bruce. reproduces wholi e p Th ema Nationaln di MSS. of Scotland, Par. ii t. iiiNo . Castlee Th e oSe f Kildrummy, . 53-67pp . 1 40 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. burgh, of which ample documentary evidence exists from the fourteenth century onwards desertew , whilno e deth fields still bear such significant names as Boroughmuir, Milltown, Malt Croft, Tolbooth, Schoolhill, and Grallowshillock. t froBu m1 this ancient centr f burghaeo l lifcastle th e e stands strangely aloof gooa , de south-westward th mil o t ea n o t Ye . closer examination what at first sight seems to be an exception to KH'TJPUMMV • S'T" CQNNttTCD W|TH M L U K U I II I^Y THC MED1EVAL 6Utt

mons castri in old charters is a regula4 r technical term for the motte or earthen moun abandonen a f o t d early Norma e sitn f th e castleo d An . this early castl t Kildrummea stily identifiede lb yma . About hal mila f e south of the church, and on the farm of Milltown of Kildrummy, close by the riverside, is a well-marked, flat-topped eminence named Gallows-

See The Castle of Kildrummy, pp. 279-86. • •: '1 Antiquities of the Shires of Aberdeen and Banff, vol. iv. pp. 734-5. . . "Integram mllam t burgume n baroniai de Kildrymmie, firmism cu burgalibus, annuls redditibus,3 terris rudiset ejusdem, rudiscum burgalibus Surrowmyir,lie et Surrow-lie cum hauch, Auchinvany et Gallowhillok, cum fortalicio, manerie loco," etc.—"Registrum .Mctgni

Sigilli Segum Scotorum," : \ t 1593-1608i' , . NoalsCf .o 2125 No , . 51 . 4 " Cum, turre ruinosa"—ibid., 1634-51- . 484. No , •> • ' .'•.•:•.':' A NEW SURVEY OF KILDRUMMY CASTLE. 41 hillock on the Ordnance Survey Map. This name, however, is quite unauthenticated: the real Gallowshillock is north of the church, on the far f Templetonmo s correctli d an , o namee mape sitys th e Th n . do at Milltown is invariably known to natives of the parish as the Castlehill, and is without doubt the mons castri of the old writs. Long- continued ploughing has worn down its outlines and smoothed away all traces of a bank or ; but as viewed from the farm of Westside, across the Don, it still has every appearance of a well-preserved motte. From this poin f vantageo t , also t i wil , e realiseb l completelw ho d y Norman was the lay-out of the ancient burgh. The thatched cabins with their tofts straggled along the summit of a hogsback ridge, having the Don on the east side and a marshy bottom to the westward. At e burgth f h o e sout rosth d e timbereth en eh d e earlmounth f yo t Normaparise th s e opposithth n wa church t e castled a th een d n o an ,, nort hs circula it sid f o e r mounded burial-ground, perched upoe th n summit of one of the "kaims" of fluviatile detritus which are so marke locada l feature. l thi s thoroughlAl i s1 y mediaeva n arrangei l - ment ; for example, it is exactly paralleled at Coull,2 where also we find the church on the north side of a churchyard which itself is at the north end of a level area stretching along the Tarland Burn, and terminate e castlee soutth th e gaty th f b hwhic,o t o de h opena n i s northerly direction e ancienTh . t road from passe e castlth s e e easd oformth n an e teaster th s n boundar e churchyardth f o y o N . doub aree th t a between churc castld occupies han mediaevae wa th y db l village. t woulI d thus appea re earliethatth n i ,r infeudatioe stageth f o s n of Mar—probably in the twelfth century—one of the Mormeeors had thrown up a timbered earthen castle to serve as the civil nucleus of a parish e churcth , f whicho t woulh(i d seem) occupie a mucs h older site of Christian worship, and indeed of human habitation.8 Between the e castlchurcth d e han grevillaga p wu e community, foure whicth -n hi teenth centur s organiseywa a full s a dy articulated burg f baronyo h . departurw Ane s takewa e n wite adventh h f Bishoo t p Gilbere d t Moravi royae th s la lieutenant longeo N . o servt r e local administra- n instrumentiva es a ends t f largebu o , t r national policies a ston, e castle on a great scale was planned, for which a more suitable site was found, endowed with plenty of elbow-room and strong in its natural defences, on the well-defined promontory thrust out from the hills into the river valley a mile to the west. The older castle -was

r thes . BremnerFo eA r kaimD e s,se Physical Geography Basin,n . Do 58 e . op th f 1 sketch-mae Se Proceedings,n pi vol.. Iviii47 . p .

2 * The Castle of Kildrummy, p. 279. The church is under the invocation of St Bride. 42 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. then abandone feld dan l into e parisdecayth t hbu ;churc h kepe th t village beside it on the ancient stance. And thus we have worked out a very pretty and interesting study in the evolution of a parochial topography in which local and national purposes have divergently played their part. The Castl f Kildrummyo e s vouchea , y contemporarb d y record, makes its d£but upon the stage of history on Tuesday, 31st July 1296, on which day it received within its gates the English monarch, Edwar s returhi d n I.no , march from Elgin, aftee downfalth r f o l Balliol. Edward remaine e castlf Augustth o t t ea dd 1s ove an ,e th r thereafter continue s progreshi d Kincardiny b s ee Cairn O'Neith d - an l na-Mounth Pas Brechin.o st n 1303I , durin s seconhi g d great invasion

e northoth f , Edward again reache1 s returhi d n Elginno journed an , y paused for a few days (4th-9th October) at Kildrummy. In 1305, 2 Donald, , was a minor in ward under the care of his uncle, Robert Bruce, Ear f Carricko l thuo wh s, exercised control oves hi r estates and the Castle of Kildrummy. Herein lies the explanation of significana t paragrap famoue th n hi s Ordinanc e Settlementh r fo e f o t the Kingdom of Scotland, promulgated by Edward I. in September 1305, whereb e Ear th yf Carric o l s directeki d "to place the Castlf eo Kyndromyn in the keeping of a man for whom he shall answer."3 The sequel is well known. After his defeat at Methven (26th June 1306), Bruce sent his wife and sister, under the charge of his younger brother Nigel, to Kildrummy for safety. On the near approach of an English army, led by Prince Edward of Carnarvon, the Queen and her ladies fled north, while Nigel gallantly and successfully defended the castle until, sometime before 13th September, s forcewa o surt e d h '- render through the treachery of the blacksmith Osbarn, who set fire to e corth n supply e greastoreth n ti d hall. Thereafte e Englisth r h

" tumlit doun equartea " l "castlee al th rf "o 4 ; tha the, is t y threw down on es sidit f eo orden i rendeo t r t uselesi r a militar n si y sense. Wevidencw e ho shal e f thio se el s partial demolition e subth -d an , sequent rebuilding, is still clearly visible on the west side of the enclosure. The restoration would doubtless take place when the Earl of Mar, who had been captured at Methven, returned to Scotland in the general exchange of prisoners after Bannockburn (1314). In 1336,

1 See Bagman's Boll (Bannatyne Club), pp. 110-1, 179, 183; J. Bain, Calendar of Documents relating to Scotland, vol. ii. Nos. 800, 822-3; Hume Brown, Early Travellers in Scotland, p. 5. • 'Botuli Scotice, . vol53p . ;i . Baiu's Calendar, vol . Nosii . . 1397, 1399, 1400; Calendar of Patent Soils, 1301-7 161. ,p . 3 Botuli Parliamentarism, vol. i. p. 268. whole Th 4 e stor f thiyo s famous sieg finels ei y describe Barbouy db Bniss hi . n ri M (ed. W . Mackenzie, pp. 58-61). See my Castle of Jfildruinmy, pp. 187-92. SURVEW NE F KILDRUMMA YO Y3 4 CASTLE. the restored castle stood another , this time an unsuccessful one, when Bruce's heroic sister e Ladth , y Christian , a determinebea f of t d assault by the Earl of Atholl, acting on behalf of Edward Balliol and the English interest. The later fortunes of the castle do not concern us here, except in so far as they shed a fitful and scanty light upon its structural history.1 It stood again in 1361, in 1404, in 1442, and in 1530, on which last occasio s recordei t ni o havdt e been burnt. 26tn O h2 February 1654, e castlth e opene s unwillinit d g gate o Colonet s l Morgan, actinn o g behalf of Cromwell. In 1689 or 1690 it was burned by the Highlanders of Dundee' than o s ts armyoccasiowa d ,an n describe "s greaa d a t castle in the mouth of the Highlands, called the castle of Killdrumie, sur- rounded with great walls wherein their was much building, and being for the most part totally burnt and destroyed, the reparatione of it cannot be under nyn hundred pounds sterline."3 It was sufficiently patched up, however, to form the headquarters from which the Earl of Mar launche ill-starree dth d " Fifteen,"4 which involve anciens dhi t line in ruin, and sealed the fate of his ancestral castle. By the victorious Hanoverians it was plundered and dismantled,5 and has since remained a roofless ruin, though some parts are doubtfully said to have been occupied as late as 1733.6 Throughout this long and stormy history, documentary evidence bearing on the structural history of the fabric is neither bountiful nor precise. Between 1431508d 5an , however hande th f n castlo i se s th , ewa e Crownth accordingld an , fine yw d entriee Exchequerth n i s Rollsf o payment e garrisoth o d expenditurt s an n e upo e buildingsth n e Th . first of these entries appears under the years 1437-8, and is sufficiently detailed to warrant our identifying the work then executed with the added in front of the main entrance.7 At the same time the chapel was re-roofed with tiles and a stone fireplace was repaired.

1 A sufficient historical sketch, which might easily be extended, will be found in my Castle of Kildmmmy. 2 R. Pitcairn, Criminal Trials in Scotland, vol. i. part i. p. 246. 3 Sir W. Fraser, The Melvilles, Earls of Melville, and the Leslies, Earls of Leven, vol. ii. p. 168. 1 See p. 2 of An Impartial Account of the Sebellion in the Year 1715, contained in a Letter from a Gentleman in Scotland to his Friend in New , and published as an appendix to The History f Scotland,o . W.J y ,b M.D r Jame(D . s Wallace), Dublin, 1724. (Macbean Jacobite Collection, Aberdeen University Library. Thi apparentls i s thire yth d edition previoue th , o stw having appeared at Dublin in 1720 and 1722. For the author see Diet, Nat. Biog.) 6 "The Earl of Mar has got a scar These forty years he'll be the waur, They've broken his ha's wi open force And ta'en five hundred highland horse," —A. LAING, Donean Tourist, . 451p . • Castle of Kildrummy, pp. 253-7. 7 Ibid., pp. 116, 214-6. See Exchequer Rolls, vol. v. pp. 57-9. 44 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. The total expense seems to be £31, 13s., but includes sundry small item t chargeablno s e agains e fabricth t n 1451I . n expenditur,a f o e s incurrewa . £138d d. upo6s , e castle.nth n 1464 I have 1 w e roofin, eth g towerso otw f e " burgeth , s"e nialditouth d r "an s tour. 146n I d " 28 an 1469, £10 s spen0i t each year upon e constructio"th repaie d th nan f o r Castl f Kildrummeo e serieth y f payment d ";o s 3an s close 147n si 1 with £80 charged agains t e fabri "r e th d castle." repaith M an c f s o rA 4 Mackenzie rightly comments, these sums in the aggregate " must repre- sent a fair amount of construction." At the same time it should be

observed that comparisons with othe5 r similar wor e extremelar k y unreliable: if the building of the tower at Rothesay Castle cost £191, 7s., that of David's Tower at Castle, no greater undertaking, cost over £400difficulte Th . s thai y t ther s littlei e work t Kildrumma y Castle to-day thaconfidentln ca t e assigneb ye th o t d fifteenth centurye musw t t remembebu ; r thae " mucth t h building" withi courtyarnthe d recorde very now 168 dlargelin 9has y perished.6 The lofty structure with a corbie-stepped gable, abutting internally e nortonth h halle curtai e wes th s assigne i th , f writer d o to nt ol y db s Alexandero t , first Lord Elphinstone acquireo wh , d Kildrumm 1508n yi , and fell at Flodden five years later. It has every appearance of

belongin thio gt s date t embodiebu , remnante sth olden a f ro s structure.7 Doubtless by including this building, which is not a tower in the strict sense of the word, Sir Robert Gordon was able to count his " seven withi precinct.e nth " History records three occasions (1306, 1530, and 1689) on which Kildrummy Castle suffered by fire. It is not surprising, therefore, to find tha e masonr th te existin th f yo g ruins shows very marked traces of burning. Such trace e particularlsar y eviden e gatehousth n i t d ean the adjoining curtain walls. That these signs of scorching are not wholly due to the latest conflagration (in 1689) is shown by the fact tha tmos e somth f t eo seriousl y affected stone e thosar s e forming part of the original internal wall-face of the west gatehouse tower, now expose e fallinth y gdb awa af yno inserted vault, shale whicw s l hsee(a ) dates probably from the fifteenth century. If this dating be accepted,

1 Exchequer Soils, p. 463. * Ibid., vol. vii. p. 277. • Ibid., . 559pp , 650. Ibid., vol. . viii79 . p . 4 6 The Mediceval Castle in Scotland, p. 60. ' Cf. W. D. Peokham on " The Architectural History of Amberley Castle " in Sussex Archceo- logical Collections, vol. Ixii. 34-5pp . e pointH . t thaou s t Bishop Sherburne (1508-36) spent much money in building, but there is little evidence of his operations at Amberley Castle. s wors "ratheHi wa k r tha f decoratino t adaptind e gan standar th s tim hi o t g ef o dtha n rebuilding, work which may make a vast difference to the comfort of the house but very little to its historical ground plan." 7 Castle of Kildrummy, pp. 99, 228. SURVEW NE F KILDRUMMA YO Y5 4 CASTLE. we have here probably tracee firsth tf conflagrationso , whe e castlnth e was set ablaze by Osbarn the Smith in 1306. It may be recalled that Barbour describes the fire as having been particularly intense at this part of the castle, so that although the gate was destroyed the assailants were unabl forco et entryn ea .

II. DESCRIPTION OP THE REMAINS. The site of the castle1 is a bold promontory projecting upon two sides, nort westd han , int oa dee p ravine, know s Bacna k e DenTh . e ravinflankth f o es hav ea slop f nowhereo e less s thait t na 45°d an ; bottom, abou fee0 6 t t belo e castlewth , briskly flow smala s l burne Th . extreme promontory e pointh f o t whicn o , e donjohth s placedni , forms e highesth e tcastl th par a f roc eo s t i k area d outcroan , p benched to form a platforai for the tower. Here the slopes are very steep and perhapregulary ma d san , have been scarpede otheth n r O e side. th f so castle, easd south an te groun th , d a broafall gentld n i s dan e descent toward e picturesquth s f Kildrummyo n De e , abou yard0 20 t s awayn O . these sides, therefore ,a ditc h (fig ) abou.3 fee5 breadth8 n ti t stild an ,l fee0 2 n depthi t o s drawfrot 4 wa ,m1 n roun e enclosureth d , meeting eithen o Bac n r middlke De side Th . e par thif o t s soute ditchth n ho , side opposit e gatehouseeth , appear havo t s e been fille eithe, lata up d t ea r e periooccupatioth n e i castled th f r elso no , n ordei e o facilitatt r e carting away material from its ruins. At its west end the ditch works t upo ou e slopnth f Baco e k Den d hasan , , therefore, never contained t nortseeme i th d waterhavt ho en sa t t e beebu ; n stopped, wit hretaina - moung in r batardeaudo vere th y t e Dena simi betweeth s a d ,- an t ni larly situated castl f Kidwelleo yCarmarthenshirn i e (fig. 29). This part ditche oth f , however s bee,ha n considerably obscure quarryingd ol y db , dumpiny b d an f rubbisgo morn hi e recen ease t timesth t n sidO e . eth ditch averages about 80 feet distant from the castle, but on the south it has been indrawn more closely. This ditch has some puzzling features, and in my former account21 suggested that it may possibly have been a prehistoric work utilise modified mediaevae dan th y db l engineer. While there is no inherent impossibility in such an idea, further consideration s convinceha e tham d t t sufficientherno s i e t evidence l evental t A s. the section now shows the usual boldly scooped-out U-profile of mediaeval ditches s lac It f complet.ko e harmon alignmenn yi t wit e curtaihth n walls, and its great and unequal distance out from them, may perhaps be explaine e suppositioth n do beforng tha du castle ditc e s eth tth hwa e See general plan in my Castle of Kildrummy, p. 75.

1 » Ibid., pp. 139-50. 46 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. was built. Suc a hprocedur e would scarcel e mostth e yb convenient , t mighbu e adopteb t f i dther s an wa e immediate nee o securt d a e defensive enclosure amid a hostile neighbourhood. A similar order of construction, doubtless for the latter reason, was employed at the Welsh Edwardian Castle of Harlech.1 Whil e generath e l dispositio e sit th s welei f o nl enough suiter dfo a large mount-and- castle, evidencthero n s i e e whatsoever that

[Photo W. Norrie. Fig. 3. Kildrummy Castle : Warden's Tower and north end of Ditch. suc y s evean hha r existede donjo th site s certainlf eo Th nha . y never- been mounded, and there is no vestige of a bank within the ditch. All evidence th e accords wit viee hth w tha stone tth e castl remaininw eno g wa earliese sth t fortres erectehave e sitee b W th eo .st n alreaddo y seen that mottea ther s ewa nea parise th r h church thid an s, earlier strong- hold was doubtless abandoned after the stone castle was erected on a different site. Had there been a twelfth-century earthwork castle on groune th occupiew stone dno th y edb building shoule w , d have expected 1 See paper on " Harlech Castle " by C. K. Peers in Transactions of the Honourable Society of Cymmrodorion, session 1921—22 . 64—5pp , . A NEW SURVEY OF KILDRUMMY CASTLE. 47 e paristh he mediaevachurcth s d immediatit an hn i l e burgb eo t h neighbourhood. T oa thirteenth-centur y enginee problee th r f fortifyinmo e sitth ge s abova e described presente obvioun da normad an s l solutione th n O . extreme promontore pointh f o t woule yh d plac donjons ehi , while round the enclosure denned by the slopes of Back Den and the ditch he would carr curtais yhi n walls, flanke projectiny b d g towers enclosind an , e gth domestic buildings. So far as convenience of plan and the configuration groune oth f concernede dar entrance th , e might equally well have been soute contriveth h n e eas o front r th desir e o tn dth o s: includo et e th n ei domestic apartments a chapel, large in size and correctly oriented, led to the gate being placed on the south. The natural place for the hall and principal domestic buildings would be against the long north curtain, with a sunward outlook, at the opposite side from the gate, and in convenient neighbourhood to the donjon or dernier ressort. Thus e castlfath r e conform e normath o t s l disposition a thirteenth f o s - century fortress : but in various points of detail it shows skilful adapta- tion to special circumstances, revealing that the engineer was no mere esprit f visioo pliancyd n nroutinier,an buildine ma Th a . t gbu being situated on fairly level ground, with plenty of space, the engineer was thus unhampered by any special topographical difficulties, and was able therefore to develop his design with entire freedom. Accordingly the plan reveals itself as a very logical and symmetrical one, and illustrates most strikingl ideae y th whic t la thirteenth-centure hth y military engineer was aiming. The castle, then (see plan at end, fig. 31), consists of a great heptagonal enclosure, fou f whoso r e angle e cappesar powerfuly db , round, flanking towers, while a fifth is occupied by a highly developed gatehouse. The remaining two angles, in the south-west and south-east curtains, are of low salient, and t interruptinno , e commangth e gatehousth f do d ean the two adjoining mural towers, are thus not provided with any salient defence courtyare Th . de castl areth f eao measures abou fee2 t18 t from east to westl by 147 feet from north to south.2 The curtain walls are mostly about 8 feet 6 inches thick, and where part of the original parapet remains, on the east side south of the chapel, show a height of abou feet5 3 t . Her e parapeth e a continuou s i t s one, without , and carried up flush from the wall; it has an unfinished aspect, and in time of siege would no doubt have been provided with

1 Measured fro ease mth t curtai soute th t hna re-entran chapee wese th th f to o t lcurtai n opposite. 2 Measured from the middle point of the north curtain to the rear-wall of the gatehouse opposite. 48 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. a timber curtaie Th . n wall, says Barbour n describini , e gth great siege of 1306, that a . t . batallitys " . mwe l tal Within, rycht as it wes with-out." The Snow Towee north-westh t a r t angle forme e donjoe th dth f o n castle t I alon. s completeli e y circular, alike withou withid an te th n enceinte, from which it has a salient of three-fifths towards the west. This has been a noble tower, measuring 49 feet 9 inches in diameter above the battered base, or 53 feet below it, with walls 12 feet thick. n utteUnfortunatela w r ruinno s i , t onli ye basemen th y n pari t t remaining. It forms a circular chamber 26 feet in diameter, in which is a well, 6 feet square, worked roughly in the solid rock. On the south wesd an te towe sidestily th seee f b lma o sr nconsiderabla e portiof no finele th y domed ashlar basemente vaulth n i t . Descriptions still extant of the tower in the eighteenth century show that it was five storeys

high, each storey being dome-vaulted e ape th f eac xo n i h d vaulan , t 1 was lef hoistinr t fo ope e nriney a g r gwateo buckea y windlasd rb an t s from the well to every floor and to the parapet. On the first floor a loop-holed mural gallery ran round the tower. At the re-entrant angle betwee e inne th e towe th e wesnd r th facan r tf o ecurtai n wall there s beeha na rone-pipe—indicate d quite clearle markth y ysb which still remain, and by the oaken pegs, to receive its fastenings, still firmly dowelled int ashlae oth r masonry. In my former account I have already drawn attention to the special characteristic affinitied Snose an th wf o s Tower s greaIt . t development, compared to the other towers, has a very French look, and its internal arrangement, vaulte n eaco d h floor wit n openina he th n gi vault, is distinctively French, as is also the mural gallery. In all these

respect e Snoth s w Tower must have strongly recalle greae dth t donjon 2 at Coucy; and I see no reason to alter my view that the resemblance e marriagth o t ma n 123e i ey du 9wel e betweeb l n Alexanded an . II r Mari e subsequen e th Coucyd e d an , t lond intimatan g e connection between the Coucy family and Scotland.8 1 Some accounts give the tower seven storeys, the latter figure being doubtless due to the inclusio intermediato tw f no e wooden floors tha knowe ar t havo nt e existed. donjone th n i t Coucys sa A 2 , Aigues Mortes (Tou Constancee rd Chateaudund )an . Thern a s ei English paralle t Hawardena l Mackenzir M . e (The Mediceval Castle Scotland,n i 56). p , objecting to this suggested relationship between Coucy and Kildrummy, says that such vaulting on all storeys, with oubliettes in the vaults, is " a late rather than an early feature, "and instances Coxton Tower, 1641. But such a parallel between a seventeenth-century tower-house and the donjon earln oa f y castle like Kildrumm quits yi e besid marke e th poine th :regar vaultinn e i t th o dt g and the oubliettes in the Snow Tower is that these features do not occur elsewhere in Scotland

thirteente inth h century. 3 See The Castle of Kildrummy, pp. 104-10, 152-3. SURVEW ANE F KILDRUMMO Y Y CASTLE. 49 Next in importance to the donjon conies the Warden's Tower at e north-easth t corner e bes,th t preserved castlee toweth n i r. This fine specimen of mediaeval military construc- tion (fig ) measure4 . 7 fee 3 sinche 6 t n i s diameter above the battered plinth, with wall fee8 s inche7 t s thick stild - an re ,l mains to a height of nearly 60 feet. It three-quartea s ha r salien botn o t h fronts. The tower contains four storeys. Its basement was a prison, and is entered by passaga th n eei wall, well secured inney b outed an r r doors closing against interior e prisoe th th garderobea s n ni I . . The manner in which this basement was t quitceileno es di clear. Roun wale dth l (fig. 6) runs a continuous corbel table which at first sight seems to have carried a timber floor. But a prison not vaulted over in stone would be an extremely unusual thing; and it is noteworthy that abov the corbee l [Photo B.C. Marshall. table ther bana es i d mS- *• Kildrummy Castle: view of War- orn roug,h -heartin, ,. g den's Tower fro south-easte mth . all round the tower, which rather looks as if ther beed eha low-pitchea n d " mushroom" vault restin centra n go e uppee pierTh .r three storeys were reached by a newel stair in the gorge wall, but opening separately from the prison e firsth entry t n flooO e fou.ar r r large mural chambers narrowing outwardly to loopholes 7 feet long 4d J an inche s broad, slightly fantailed below, t devoibu f plungeo d . This roo s probablmwa y a store. The upper two floors provided living [Photo Marshall.B. C. rooms, and each was furnished with a fireplace. Fig . Kildrumm5 . y Castle: Originally these rooms were probably equipped Window in Warden's Tower, with mural chambers similar to those below, narrowing outwardly to loopholes, or at best, to small windows; but these would seem to have subsequently been replaced by large windows of pronounced Edwardian type (fig. 5), consisting of two lancet rakea n o sd basframed ean d wit cusp-corbelleha d lintel.1 These questioe Th 1whetheo t s na r these window originae sar insertionr o l doubtfua s si l one, there being features in the bonding which tell both for and against this view. I have long regarded VOL. LXII. 4 50 PROCEEDING E SOCIETYTH F O S , DECEMBE , 1927R12 . window t onca s e recal e typ th f lopenin o e o constantls g y founn i d Edward the First's Castles in North Wales. The rough masses of hearting (tig. 6) which project from the wall-face on the second-floor level have very muc e dispositioth h d appearancan n e corea falleth f f o so ne groined vault. A small portion of the parapet walk, reached by two steps up from the roof of the tower, may be seen on the east side, and three of the plain dished and plunged runnels still remain; but the

[Photo W. Norrie. Fig. 6. Kildrummy Castle: Interior of Warden's Tower. (The arches over the mural chamber moderne sar : originally there were lintels loophole Th . ee seeth n no left, with the buttress-like wall adjoining, is a restoration.) parapet itself, which rose flush from the wall-head, has almost entirely perished. The special security enjoyee Warden'th y b d s Towe s showi r y b n the fact that o directhern s i et communicatioe n th betwee d an t i n castle courtyard. It must have been reached only from the hall by the passage between the chapel and the kitchen. On the other hand, the the insertionss ma inducem a t bu d, her stato e t matte e open eth a s na r questio n afte carefua r l consideration on the spot with Mr J. S. Richardson, who favours the view that they are contem- porary with the tower. A iNEW SURVEY OF KILDRUMMY CASTLE. 51 occupant of the tower possessed at all times an easy and private means of e posteregresth y sb n adjoinin norte th n hgi curtain. The Brud Maulan x e Towere south-easth t a sd south-wes an t t angles are twins in their general arrangements. Each is 30 feet in diameter, with semicircular salient to the field, but projecting as a square building toward e courtyardth s o thae s interior, th t s form oblong rooms with trilateral outer ends. Only the lower storey of each towe s unvaulted s intactwa i r d e uppean , th ; r storeys havd eha mural recesses with exactlloopn a f o s y similar patter e thoso nt th n ei Warden's Tower. Owing to their lack of plunge, these loopholes can- not command the base of the wall—a fact which suggests that the summits of all the towers were provided with hoardings for this purpose n eacI . h towe e basementh r s enterei ta door t y b no d, specially defended e squarth n i e, gorge wall; the upper floors were reached by mural stairs in the curtain adjoining, convenien servo t t e bot e towehth d an r e ramparth t •walk . direco n Ther s t wa e communication between the basement and the first floor. All four tower s thusa s described have bases of identical profile (figs. 7 and 20, e walth No, l 5) .risin g almost vertically through some six or seven courses, then [photo G- p- H- Watson- setting back sharply from a bevelled Fig. 7. Kiidrumiuy Castle: south side of mouldin *r >thre fo g eP coursesI , afte• r whicf h Pl'nthj Maule Tower. e toweth s continuei r d with unbroken verticalit summite th o yt . Since my previous account was published, further excavation has disclose e completth d e ground-plas i t i e th a f r o nfa o s , preserved. This has been a large and notable structure, consisting of an oblong block, 68 feet 6 inches in breadth, passing out frontally into greao tw t round half-engaged towers fee 3 basan 3 ,i t l diameter, between which lies the trance, 8 feet in width. The total over-all depth of the gatehous towee feet1 6 Th s .e i r feewall9 inchee 6 t sar s thick shod ,an w differena t profile fro e othercastle mth th n sei (fig. 12), steeply battering directly from the ground to a height of six courses. As originally built, the basement of each tower contained an oblong, unvaulted chamber, with a semicircular bow towards the field. Behind each, in the rearward portion of the gatehouse, is a lodge of greater breadth, obtained by thinnin gwes e trancee wallofe th th fth t f n sI o lodg .remarkabl a s ei e fireplace (figfee7 , t. 8) 8£ inches wide, with heavily chamfered jambs 4: fee inche9 t s high, curve t belod ou d alswan o abov carro et e hoodyth , 52 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. whic s perishedhha segmeiitae Th . l backin fireplace th f go e appearo st be an insertion. This fireplace is of unmistakably Edwardian type, and can be assigned without hesitation to about the year 1300.1 The gate- house bears evidence of successive alterations. At first the partition walls betwee e rearwarth n de towe lodgeth d r an sbasement s seemo t have bee woodnof , wit hcentraa l stone pier (havin gsplayea d plinth) carro t e flooyth r latea above t a rt datbu ; e stone gorge wall doord san s wer ee gorg th built n ee easI .walth t n i towel a stonr e with fluted piscina bow bees ha l n re-used basemene Th . thif o t s tower, owine th o gt generally eastward slope of the ground, is at a lower level than that of

[Photo R.C. Marshall. Fig. 8. Kildrummy Castle : Fireplace in West Lodge of Gatehouse. (The uppe stone o lefe rtw th t f jamso restored.e bar )

the west tower. Hence it occurred that when the gorge walls were wese builtth tn ,i towe r the(fig) agains10 p y. u wer incorpord n eru an t - The same type of fireplace is found, among; other buildings, at Carnarvon Castle (fig. 9), in wor1k dating from 1285-9 ; at Conway Castle, begun in 1285; in the solar of Ludlow Castle, circa Byware th 1283-92 n i dd ToweTowee an ; th t f Londona ro , alse o th buil o Edwart y s b t A . dI remoter provenance f thio s typ remarkee f fireplaceb e o y ma t di , tha exampln a t e occure th n si Tour de Sel at Aigues Mortes, in work assigned to post 1289; see C. H. Bothamley on " The Walled Town of Aigues Mortes," in Archceological Journal, vol. Ixxiii. (1916), p. 258, and Plate xiii. No. 4. The occurrenc Edwaro type t th Britain e i e f e First'ethudu o y dth e nsb ma s connection with Aquitaine. The only other Scottish fireplace at all resembling this one at Kildrummy with which I am acquainted occurs at St Andrews Castle, in a part of the building assignable to the sixteenth century; but the condition of the jambs of the fireplace suggests that it was rebuilt into its present position from some older building s doubtleswa t I . s originally made durin Englise gth h occupatio castlee th f no .

A NEW SURVEY OF KILDRUMMY CASTLE. 53 ate oldee ease dth th tr n towei pier t pietakebu s e , r th (fig rwa ) n .11 dow n angorge dth e wall carried right acros lowee towee th sth t ra r levele th , material e pieth rf beinso g roughly reinserted, wit e spla hth t awa ycu y soute o nth obtaio t hs sida nflusa o es h facingstila t lA late. r period, vaults were inserte botn di h towers theid an , r basements, thus covered in, see havmo t e been filled wit ha soli d packin rubbley dr f go , doubtless

Fig. 9. Fireplace in Eagle Tower, Carnarvon Castle.1

to render them proof against artillery. The floor of the east tower (fig. 11) is paved with herring-bone ashlar, older than the inserted vaulting, under which the pavement runs. A large forework was also built, prolonging the trance, and containing an outer gate 27 feet in advance of the original one, and a middle gate between the two. Beyond this outer gate the forewore wallth f s o continuee kar abutments da e drawbridgeth r sfo , e axl f th whico e h turned, doubtless withi a nlea d jacket n socketi , s

1 This illustration is reproduced, by permission of Sir E. Vincent Evans, LL.D., from the Transactions of the Honourable Society of Cymmr odor ion, 1915-6. SURVEW ANE F KILDRUMMYO Y5 5 CASTLE. provided for the purpose; while the bridge when down rested on each side upon a carefully constructed bench, and spanned a masonry-lined pit. Behind the outer gate on the east side was a side gate (figs. 12 and 13), afterwards buil e oppositth t n up.O 1e sid squarea e containe sth newel stair up to the room whence the bridge was worked. Probably

[Photo H.P. Watson.G. Fig . Kildrumm12 . y Castle: Barbica Sidd nan e Gate.'

whe forewore nth addeds kwa r subsequentlyo , originae th , l outer portal of the castle, between the two towers, was refashioned: its , which at first it must have possessed, was taken out, and a double gate was substituted. The disturbance of masonry here is still very evident. measure) 15 d t (figs an fee pi 1 4 s2 inchee 1 2 t. Th lengtfeen 8 si d t han

1 This infilling has now been taken out, as shown in fig. 12. * This illustratio reproduces ni d fro r WmM . M. Mackenzie's recent boo Mediaevale kTh Castle in Scotland, by kind permission of the publishers, Messrs Methuen & Co., Ltd. 56 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. 10 inche breadthfeen s8 i s sectionsti o buils deepd i tw t an ,n I i t . , clearly marked verticaa ofy fb le masonr jointh n i t eithen yo r innee sideTh r. portion fee,8 inchee eas6 t th fee te inche6 wes0 n sid1 o sth d t n tean o s side ,n extension appeara e b o d durint s an , g peaceful conditions must have been boarded over, otherwise acces sgaine e sidee coulb th t -o dno t e neweth gatd l s noticeablan estairi t I . e thae masonre th t th f o y barbican wall s continuoui st abovpi e eth s throughou s lengthit t , sug- gestin as ng wa alteratio that extensioe pi th t e th n f no mad e while eth

[Photo W. Norrie. Fig . Kildrumm13 . y Castle: Vie Sidf wo e Gat excavateds ea , showing later infilling.

barbican was in course of construction. On the west side the wall of the pit beyond the joint is recessed back 5 inches, so that two rows of con- tinuous corbellin requiree gar carro dt bridge e bencth yth thir n eho fo s side; whereas, on. the other side, the set-back does not exceed 3 inches, and only one corbel-course is necessary. The side walls of the pit are formed in good coursed rubble; the masonry of the two ends is in beautiful ashla f close-jointero d block coursen si s averagin incheg9 n si height: the inner end wall is doubly battered against the slope of the ground. The pit drains by a central stone-built gutter emerging at the outer end by a sluice which is carefully checked for an iron grating. In soute th h wal anothes i t thila d sen r sluice, whic bees hha n provided with

58 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. an iron grating moving up and down in a well-wrought chase cut in projectino tw g stones. Fro innee mth r portae gatehouseth f o l , abutments abou feet4 t thick extend back on either side some 7 feet into the courtyard. These abut- buile on d mentf wito e gatehousee hth sar probabld an , y carrie arcn da h with a meurtriere over the portal, just as in the rearward extension of the gatehouse at Caerlaverock Castle. As there is no evidence of any comnumication betwee basemene ne uppeth th d r e gatean floort th f -so house, it is possible that this was supplied by external wooden stairs rising on either side along the inner face to a platform carried by these abutments. Fro firse mth t floor upwards, acces havy ma se been gained

consolidatioe b yvicth a n ei 1 n whic re-entrane h th fillp u s t angle between gatehouse th wese th td curtaiean n; suc staiha r would also conveniently have serve e garderobeth d s arrangemenheree Th . e otheth n ro t side of the gatehouse may have been similar. The earlies f theso t e successive modifications that have broughe th t gatehouse into its present state, namely the building of the stone gorge walls in the two towers, was probably done about the end of the four- teenth century; the masonry is of a kind found nowhere else in the castle, very much joggled, and resembles masonry of this type in the Church of Bothwell, founded in 1398, and in the contemporary work at . Perhap t withou no e fac s th i st t significance that

Kildrummy abou2 t this time (1374-88) was in the hands of the Douglases, who also owne e barondth f Bothwellyo previouy pointes m A n i . t sdou account, the forework is evidently the building whose cost is detailed

in the Exchequer8 Rolls for 1436-8, and the alteration of the original outer gate, between the towers, is probably contemporary, or shortly after- ward ste a l events pal s t followin wa :a t si g upo constructioe nth e th f no new advanced entrance. The vaulting of the two towers, and packing their bases against cannon fire, Avas doubtless don periodt e somea th f eo s e fifteentth n i h century whe knoe nw w fro e Exchequermth Rolls that building was in progress;4 and the blocking of the side gate illustrates the same tendency towards consolidation, which indeed is a marked featur f fifteenth-centuro e y developmen t othea t r gatehouses, sucs ha those of Tantallon and Caerlaverock. Probably in the seventeenth century e finath , l stag s reacheewa d whe e drawbridgnth e passet dou of use, the pit being filled in, and a stepped and cobbled causeway carried right throug trance hth e int courtyarde oth . imaginee On uppen sa r platform somewha pattere th f o t n of that whic throws hi n across the1 rear portal of the great gatehouse at Pembroke Castle. 2 See my paper on "The Architectural History of Bothwell Castle," in Proceedings, vol. lix. p. 179, note 2. . 218 3cit.,pp . , Castle 225Op .« of Kildrummy, . 2U-6pp , SURVEW ANE F KILDRUMMYO Y9 5 CASTLE. In its final form the total length of the entrance passage, from the outer end of the pit to the inner end of the rearward abutments, was abou feet8 8 t . Thu l attackal ss effectivelwa y canalised e lonth gd an , narrow trance was defended by a and by three pairs of folding gates, with probably a fourth pair closing against the courtyard. Doubtless also there would be the usual meurtrieres opening in the flooring above. Along the inside of the north curtain is set the great hall, with the kitche s loweit t r easterna o r nsolae th endr grea o rd ,an t chambet a r s uppeit r end, nex e Snoth t contemporare w on Tower f hale o s i Th l . y build wit curtain.e hth t measureI 1 fee1 s7 inchefee3 t0 4 inches 6 t y sb , •within wall fees6 t thick . threy b Toward t eli larg s norte wa th se t hi windows, wit hsmallea easlightine rd looth en t t screenspa e gth e Th . dais s windosecure n o a wa lights tw y d n thib w do f an ,so s sidwa e inner barred e shuttersout th lesseo n htw O sidr.e windowsar e , fur- nished with side benches alsd an ,o another dais window fee3 , t 10| inches daylighn i t width innee Th .r jam f thibo s window show semicirculasa r hollow sunk, in a broad splay, and terminated in a broach stop below, and a grotesque corbel mask above (see fig. 20, No. 7). At Maryculter Church, built in 1287, and at Cowie Church, which was consecrated in 1276, the same hollow jamb-moulding is found (see fig. 20, No. 9).2 The external reveal of this windo s beewha n much s apparentlalteredwa d an , y converted int odoora t presenA . s externait t l moulding show wide sth e casement e fifteentoth f looo N hp . lightcentur8) . e screenth sNo , y n 20 (figo s . this recesa side d an ,s here doubtless indicate positioe doore th s th f .no On the inside wall at this point is a socket, perhaps for an iron rail apparently connected wit screene heasth e hallTh te .wal th ,f whiclo s hi now reduce mera o dt e foundation, must have possesse doorso dtw e on , leading int othekitchee e oth th d r ndescendinan woodey gb n steps into the passage between the kitchen and the chapel, which is the only means of access to the Warden's Tower and the north . The floor of the f woodo hals ,wa l wit ha shallo w basement belowfee 0 e spa4 tf Th no . is too great for single timbers, so that there must have been dwarf walls, or a row of posts, as in the late thirteenth-century hall at Ludlow Castle. A hole as if for a handrail, beside the north-east stair door, and a bench below, may indicate a wooden ladder descending into the basement, which was apparently unlighted. The ledge on which the

also 1S t Carnarvo oa n Castl halle eth , although never completed buile s oon i f, d wite th h curtain against whic t ihi s setthereford ,an s designewa e d fro outsete mth , beinge parth f o t first work at the castle, 1285-91—see C. R. Peers on "Carnarvon Castle" in Transactions of the Honourable Society f Cymmrodorion,o 1915-6. 12 . p , e DrSe W. Kell Transactionsn yi Aberdeene th f o Ecclesiological Society, vol . (1886-9.i ) par2t i. p. 28. 60 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. hall floor rested is formed by an ordinary scarcement along the south wall t alon bu norte ; gth h wall this ledge appear projectina s sa g bench, now much damaged. The scarcement on the south side appears to be resule alterationn th a f o t ; probabl e halyth l floos beeha rn heightened, and the bench on the north side was made available to suit the new e halth levell n tracthero I a fireplace n .e north-eas f s th o ei e t A . t corner a newel stair led up to the walk of the north curtain, and doubtless als o at o gallery ovee screensth r . From this gallery access woul e obtaine b de diagonalle chapel th th t o A t d . y opposite

[Photo W. Norrie. Fig. 16. Kildrummy Castle : South Wall of Hall; Warden's Tower in rear.

corner of the hall another newel stair, well secured, served the lord's private apartments. On the exterior of the south wall was originally a plinth consisting of a bowtell above a plain chamfer. To this wall has been applied a comparatively modern building, probably of the eighteenth century, wit wally fragmentsd hdr ol o s f madtw o d p eu an , partition f similao s r unsubstantial materials flooe f thiTh o r. s building is paved with slabs of the handsome Correen stone (andalusite mica- schist), so commonly found in old Aberdeenshire farm buildings. After this late buildin s added gwa plinte th , h alon e exteriogth e r th fac f o e cloures halwa l d awa o gaiyt n extra space t remainbu , s e intacth t a t points where the partitions butt against it (fig. 16). A NEW SURVEY OF KILDRUMMY CASTLE. 61 The kitchen, about 24 feet square, still retains traces of its fireplace, with aumbry norte , th sink draind hn i an ,l curtain al , solae s Th ha r. been almost entirely rebuilt, apparentl e firsth t y Loryb d Elphinstone (1508-13) structure th , occupyinw eno s placgit e (fig bein) .17 ga tower - house of the L-plan, with vaulted basement and crow-stepped gable curtaid ol erectee nth walln e earlied o th reli A f . o cr arrangemens i t a high window overlooking the hall from the first floor of the old solar

[Photo Norrie.W. Fig. 17. Kildrummy Castle : Elphinstone Tower and Hall, seen from Courtyard.

apartments thao e lors , th l tthas goind al coulwa tn go e d ey kee n pa on.1 At the north-west corner of the hall, remains of a straight mural stair, ascending westward e visibl higa ar , t ha e enort th leve hn i l curtain. On the east side of the courtyard is the chapel, measuring some 4 feet7s 9 position 1 It fee .y b t , wit hale hth lprivate betweeth d ean t ni apartments, is paralleled at the Welsh thirteenth-century castle of

As at , , and the (now destroyed) House of the Knights Hospitaller1 s at Linlithgow. There are English examples at Ludlow Castle, Dudley Castle, Penshurst Place Gread an , t Chalfield Manor. 62 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. Caerphilly orden I . securo t r e correct orientatio t obliquelse s i t ni n yi the curtain, which its east end overrides. The floor of the chapel .was

[Photo W. Norrie. Fig. 18. Kildrummy Castle: View of East Front, showing Chapel Gable. of wood, and the undercroft was entered by a door on the south side agains e curtainth te easTh t . window (fig ) consist.18 f threo s e tall lancets, 14 feet 6 inches in daylight height, rising from a moulded string d havinan , g above them a fourt h smaller lancet s whicit n i h SURVEW ANE F KILDRUMMYO Y CASTLE. 63 present form appears to be a rebuild of the fifteenth century. Exter- nall e threyth e window wroughe ar s t wit hplaia n double chamfert bu , inside (fig. 19) the widely splayed scoinsons are closed above with moulded bonnet-headed rear-arches springing from filleted wall-shafts with capbasesd san . Roun outermose dth t arch-mould has bee traina l of dogtooth, now greatly wasted. All this detail (fig. 20) is of very good

[Photo W. Norrie. Fig. 19. Kildrummy Castle: Chapel Window, interior view. (This view shows the peel- e wall-facinth f e chape o e basemene leve th g th th f in o n lo gi l t floorp u t ; alse oth roof-raggle in the north wall, the wall-shaft cut to receive a water-table, and the inserted joist-holes, all belonging to the later lean-to erection.)

mid-thirteenth centur e north-easy th type t A . t cornee chanceth f o r l is an aumbry, and beside it a door opens into a small vestry with a garderobe, constructed in a special abutment against the shoulder of the gable. Probably there was a gallery at the west end of the chapel, entered fro e correspondinmth g gallery ove e screene hallth rth ,n i s whics turit nn hi woul norte reache e staie db th th n h i r y curtaindb . Ther cleas ei r evidence tha chapee afterthoughtn th ta s i l , thrust ou t throug ease detaie th curtainth e t f hs th windoo l a t distinctlf bu o ; ws i y thirteenth-centur lesy no type s is eviden ,it t tha alteratiothe t an nwas 64 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. early one, probably while the castle was in course of erection. The proof thatjjthe out-thrustin e chapeth f go l gabl s secondari e e b y yma

FT.

Fig. Kildrumm20 . y Castle: Moulded detail—(1), Section through Chapel Window; (2), Plan of Chapel Window; (3 and 4), Enlarged Plans of Chapel Window; (5), Profile of Base of Warden's Tower; (6), Pla Wesf no t Window, south wal f Hallo l ; (7), Mouldin t interioga f ro reveal, east jam f Wesbo t Window; (8), Mouldin t exterioga r reveal; (9), Mouldin Doon go r of Maryculter Church. briefl t forthyse . Externall soute th n hyo sidjointe eth , whic packes hi d in with small chips, is clearly visible where the curtain wall meets the ERRATUM. Fig. 20, No. 6, is a plan of the west window in the north wall e hallo th fe correspondin refe;th 8 Noso d t r an .7 g window in the south wall. A NEW SURVEY OF KILDRUMMY CASTLE. 65 norte chapelth n h O sid.breace eth h curtaie mad largeth a s n ei nwa r one joine th : t occur distanca t a s f 1e0o feet chapele nortth f d ho an , e chapeth e rebuill th gabl d tan e portioe curtaith f o nn for mstruca - tural unit. For some reason not apparent, the rebuilt part of the

curtai t bacse inche7 ks ni s1 behin e olde th de inside th r facen ,O . the footing of the old curtain remains in situ, and is carried right through the interior of the chapel; and the splicing of the wall, where the breach took place, is quite evident at a low level on the north side belo e vestrywth . e Lastlchaith f evidenceo n yi e th , interior side walls of the chapel butt without bond against the inner e curtainth fac f o e . At a late date the chapel was desecrated, and a two-storeyed lean-to of confined dimensions was built against the inside of the east gable, ashlae th r facing bein ggaio t tor t n nou space rougd an , h n holei g sdu e heartin e joiststh th r ,g fo whil e windowe shaftth eth f o s s wert ecu away to receive the wall-plate.2 It is curious that no traces exist to reveal the structure of the hall roof where it abutted upon the lofty solar wall. The chapel has not been vaulted, but there is otherwise no evidence as to how its roof was managed. Probably it was of the simple trussed-rafter type, without tie-beams d eithean , r ope r boardedo n , whic s usua wa hn smallei l r churches of the thirteenth century. The Exchequer Rolls show that renewes rooe th wa f 1437-8n di , bein t thaga t date covered with tiles.

t CarnarvoA n Castle s originalla , y built granare th , y onls slatedywa , 3 othee th r buildings, includin spacioue gth s hall, being thatched.4 Along the west curtain were lean-to constructions, indicated by the foundations show plan no ncorbele (figth weather-tably d .b san 31)d ,an e for a wall-plate (fig. 21). At the south-east corner of the courtyard is a bakehouse with three ovens, probably of the fifteenth century; a fourth ove s beeha n n inserte e adjoininth n di g Brux Tower e courtyarTh . d s cobbledi , sloping toward e southth s . e nortth hn I curtain, close agains Warden'e th t s Tower porta s i , - cullised postern, leading out to a small triangular area of ground from which, -withou y apparenan t t direct contact wit e castleth h a vaulte, d underground passag bure dowd th eBacle n n i o n t k vaulDene s Th .tha 1 See Mr C. S. T. Calder's Note appended to this paper. 2 The chapel is said to have been "occupied as a place of worship so late as the year 1733" (A. Laing, The Caledonian Itinerary, vol. i. p. 253). If this is true, the subsequent lean-to is probabl linkee b wit o p yt evidencee u d th h f eighteenth-centuro s y occupation suggestee th y db rang f buildingeo s alon soute gth hhalle sidth I thin .f eo k little therb n e eca doub t thae th t castle area was occupied by agricultural buildings after its final dismantling. The lean-to in the chape hardln lca y have been anything more tha nshea d suitabl sucr efo hpurposea . 3 Exchequer Rolls of Scotland,. 59 . volp . .v . BC .4 Peer Trans.n si Cymmrodorion Soc., 1915—6. 9 . p , 5 VOL. LXII. 66 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. fallen, but the deep trench of the passage remains throughout its length ashlae th p r to wal e exposes (figth i l e sidt . a th e 22)d n nexdo an e , th t castle. It has already been observed (p. 50) that there is no direct com- munication between eithe Warden'e rth s Towee postere th th r d o r nan castle courtyard, from which they coul e reacheb d d only througe th h hall by the passage between the chapel and the kitchen. The portcullis chamber of the postern was probably reached from the newel stair in e Warden'th s e easTowerth t n I curtai. ne sout th walf thio ho t ls

[Photo Norrie.W. Fig. 21. Kildrumniy Castle: Interior view of West Curtain. (The two rows of putlog-holes are modern.)

towe s anothewa r r postern broae th e lices ,n dth o leadin areo t t a gou betwee ditche castle th nth d , whicean h areencloses awa .a y db This postern is now built up, but the portcullis chamber overhead still partly remains, and was reached from the Warden's Tower. As usua thirteenth-centurn li y castles fine t Kildrumni,w da mose yth t careful regard for sanitation. From the number and size of the garde- robes we may argue the presence of a fairly large permanent household. The garderobe in the chapel, and that in the prison of the Warden's Tower, have already been mentioned. The latter is the only garderobe within the compass of a tower. In connection with all the other towers the garderobes are arranged in the curtain walls adjoining, and muat A NEW SURVEY OF KILDRUMMY CASTLE. 67 have been reached by mural passages. The shaft of each garderobe descend) (fig23 . s verticall e thicknesth e walln yi th discharged f an ,so s at its base in a buttress through a vent divided by a central post, in order to prevent ingress or egress by so unsavoury a means.1 One garderobe south-wese th t a , t angle, vents throug singlha e opening.n I 2 connection wit wese hth t gatehouse tower, what appear gardea e b o -st robe of a different type is found, consisting of vertical shaft, at the base of whic ha culver s carriei t d right throug e curtaihth n from baco kt

[Photo C. R. Marshall. Fig. 22. Kildrummy Castle : General view from north-west.

front. The garderobe at the south-west angle also has a culvert carried through the curtain. It would seem that the drainage of this part of courtyare th collectes dwa d int orunnela carried an , d alon innee gth r

1 Suc precautioha t extravagantno s nwa casA f prisoner. o e s escaping fro e keemth t pa Newcastle by a latrine is instanced in J. H. Parker and Turner, Domestic Architecture in England in the Fourteenth Century, p. 19. Alexander I., surprised by the men of theMearns and Moray while building a hall (palatium) at Lyff, near Dundee, in 1107, escaped per latrinam, Joannis Forduni Scotichronicon, bk. v. chap, xxxvi. (ed. W. Goodall, vol. i. p. 285). The hole down which, accordin e well-knowth o gt n story, Montrose refuse climo t d orden bi escapo t r e from Pitcaple Castle (see New Statistical Account, vol. xii. pp. 561-5) was perhaps a garde- robe shaft. 1 The corresponding garderobe in the south-east curtain has probably been of similar design. 68 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. e curtainth sid f o e , discharging throug e culvertshth o tw tha o e s , th t garderobe shafts were constantly flushed e shafTh t. nex e gateth t - house is placed in a special interior construction of the same build as the gatehouse t separatebu , d fro curtaie westware mth th o njoint a y db t distinctly e interiorvisiblth n o e , indicating thae gatehousth t d an e curtain are not of contemporary date. Externally this joint is not in evidence, as the wall here seems to have been scabbled. Althoug e castlth h e througs it l al h vicissitude s preserveha s s thirteenthdit - century plan and its original unity of design, it is apparent, even on the most superficial examination, that the build- ing t varioua s s stage n theii s r stormy history have undergone very considerable reconstruction and repair. The north curtain, with the hall set against it, the remnant n anciena f o s t solar embodied e Elphinstonth in e Tower e chapeth , l gable, and the six round towers, are all faced with most excellent freestone ashlar (fig. 24). The plinth on the towers is continued also along the north [Photo C. B. Marshall. curtain, and is stepped up at either end Fig. 23. Kildrummy Castle: Garderobe to accommodate with the levels of the in south-west Curtain next Maule ground thin I .s curtain Snoe ,th w Tower,1 Tower. and the great hall is found what appears to be the oldest masonry in the castle (see fig. 22), consisting of more or less cubical, wide-jointed ashlar of quasi-Norman aspect. The curtain e easth t n o sids e (fig .e south-west 18)th d tha n ,an o t , have been much patche t varioua d s periods t werbu , e clearl f inferioo y r finish fro e e south-wesoutset mbasth th e f Th eo . t curtai recentls nha y been a roughlexposed s i d yan , constructed plint f smalo h l stones. In spite of their inferior character, however, I see ho reason for disbelieving that these curtains are not substantially original work; they bon witn d i garderobe towerse th hth d an , e shafts everywhere ear of the same pattern. It is quite a common thing in medieval castles for the towers to be carried out in masonry superior to that of the curtain

I have now ascertained that the piece of cubical, wide-.iointed ashlar facing on the north e Snosidth f w1eo Tower Castley m f f Kildrummy,,o o discusse 7 12 . p repait a da s i r carriet dou with old stones about the beginning of this century. . A NEW SURVEY OF KILDRUMMY CASTLE. 69 walls. The finer masonry and plinth in the north curtain are perhaps

to be 1 explained on the assumption that the castle was begun in a more ambitious styl efouns thawa dt ni possibl sustaino et . The west curtain shows masonry of totally different type (fig. 25), extremely rude, with l stone sized slopeal an f s o s s thrown together,

Photo W, Norrie. Fig. 24. Kildrummy Castle: View of Chapel Window and Warden's Tower. but containing a good deal of original ashlar blocks re-used. It rises from a base consisting of two courses of larger oblong blocks \vithout batter. This wall is clearly an early reconstruction ; the joints between it e oldeth d r Snoan d Maulwan e Tower e verar s y eviden 26)td (figsan , 5 2 . and on the inside of the curtain near its north end a fragment of the r examplFo t Dirletoea Bothwelld outene an th rn I war. t Corfda e Castle, whic docus hi - mentee worth 1 f Edwars ko a d d e towerI.th , e beautifullar s y finishe ashlarn i d , while th e intervening curtains are of rougher work. At Harlech, where the masonry throughout is of rubble tower e wore better th th ,fa n ki s s i r curtainse thath n i .

SURVEW ANE F KILDRUMMO Y Y1 7 CASTLE.

older foundation has recently been exposed (see plan, fig. 31). Very prob- abl e curtai ys th presen it n ni t fordaty mema fro reconstructioe mth n castle oth f e afte dismantlins it r Englise th y gb 1306n hi . Barbour tells Princw uho s e Edwar than do t occasion cause quartera d l "e al th f "o castl"e tumlib o et tWr doun.M Macka d ean "; y Mackenzi s pointeeha d out that this phraseology signifies a casting down of one side of the fortified enclosure. evidence fabrie Th 1th f ceo here stand completn si e harmony with the literary record. The architectural detail of the buildings throughout has been of an exceptionally high order. A large number of carved fragments have been recovered in the course of the excavations. These include richly moulded mullions and transoms, voussoirs, caps, rybats, and other pieces ranging fro thirteente mth fifteente th o ht h centuries, whil enumbea f ro Renaissance pediments and finials indicate that the tradition of ornate buildin wels gwa l maintained unde Elphinstonee rth late th e n ssixteenti h and early seventeenth centuries. One fragment of an inscribed stone shows the letters M.B. in raised late Gothic characters; another has the letter A on a panel within an ornamental frame of late pattern. With these evidences of fine craftsmanship must also be associated the beautiful fourteenth-century sepulchral slab wit eight-pointen ha d cross of calvary, now lying forlorn in two pieces outside the north curtain near the postern gate. The general appearance of the castle before the Snow Tower fell in 1805 is shown by a water-colour painting formerly at Clova House (fig. 27). III. THE ARCHITECTURAL PROBLEM. In my former account I drew attention to the different profile of gatewao tw e th y towers suggested an , d that these tower "e distinctlsar y later" tha e others.nth Since the e completenth d excavatioe th f o n

gatehouse has 2 thrown new and searching light upon this important question. Despite its fragmentary condition, it is now clear that this gatehouse belong a typ o t es otherwise unknow e thirteenthth n i n - century castles of Scotland. Among the latter the entry is usually, as at Inverlochy and Lochindorb, merely a portal in the curtain wall, which localle b y y ma thickene orden di admio t r portcullia t s chamber above. In more developed cases, as at Rothesay, the entry is through a square gate-tower t KirkcudbrightA . , Tibbers, Coull d Bothwelan , fine w la d still more elaborate entrance, consistin passaga f o g e betweeo tw n towers, with porters' lodges in the rear. But even at Bothwell, it is lUedicevale 1Th Castle Scotland,n i . 59 . p 1 Castle of Kildrummy, pp. Ill, 129.

STIRVEW NE A F KILDRUMMO Y Y3 7 CASTLE. clear fro ma glance plan th t ea tha e havw t e something very differ-

r lesfa s d importanan t en t tha Kildrumme nth J y gatehouse lattee Th .r has formed a great self-contained unit, a " gemel tower" in niediseval parlance—that is, a large isolated rectangular block with twin towers e frontth o ,t tranc portersd an e ' lodges between them o below(n d an , doubt) apartments of considerable size and distinction 011 the upper floors. It will at once be recognised that this is the type of gatehouse found e greaith n t Edwardian castle f Wales o t sHarlecha s a , , Beaumaris, Llanstephan, Aberystwyth, Kidwelly Caerphilly.d an , s amii t I d these castles that the inspiration of the Kildrummy gatehouse must surely be sought. With this consideratio Edwardiae nth n fireplac entirn i s i e e mentionee b y harmonyma t i d alsd an o; tha e speciath t l profile found in the gate towers, where a sloping plinth rises straight from the founds thas i , t normally adopte Welse th n di h castles. Nor rests the matter here. In studying the Welsh gatehouses, which vary greatl detailn yi I hav, e been strongly impressed wit e closth h e resemblance between tha f Harleco t he gatehous th Castl d an et a e Kildrummy (fig. 28). e majoExcepon r r fo t point—the omissiot a n Kildrummy of the rearward stair-towers—the two plans are practically identical. In each we have the oblong self-contained gatehouse block with twin tower frontn i s , whose outlines pass smoothly int e walloth s e trancoth f e betwee e otheth n n rthem o side t bu ,s for ma shar - pre entrant with the curtains. In each, the towers contain an apsidal chamber, behind whichrearware th n i , dgatehousee parth f o t largea , r room is formed by encroaching upon the trance wall. At Harlech, as at Kildrummy, one of these rearward rooms has a fireplace. In neither gatehouse has the basement any vaulting. In both, the garderobes are contrive re-entrane th t da t angles betwee curtaine towere th nth d san . But this is not all. From the plans and table of measurements submitted herewith vere ,th y remarkable fact emerge gatehouseo stw tha e th t e sar not merely closely similar in their arrangements, but they are almost identica dimensionsn i l . e Excepabsencth r t fo Kildrummtea e th f yo rear towers, and for minor variations in detail, the same measured drawings may almost have been used by the builder of each. It is diffi- cult to believe that such a coincidence can be merely an accident. In this connection it becomes important to ascertain whether there are any known links between Edward the First's building operations in Wales and in Scotland. Fortunately the documentary evidence is not altogether silent upon this point. Edward's master of work in North Wales, between e buildinth d f ha Rhuddlan127g1299d o o an 9 ,wh , Harlech, Conwayd an , Beaumaris Castle s chargehi n frod i s an ,m Jul 129, y3 Decembeo 0t , 28 r Proceedings,e 1Se vol. lix . 173p . . 74 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. s himsel1293wa , f Constabl f Harleco e h Castle s Jamewa , e Sanctd s o Georgio. frow mlettea No 1 f Edwardro , dated June 30,1302 t appeari , s

Fig. 28. Gatehouses at Harlech and Kildrummy Castles, drawn to same scale. Measurements. Harlech. Kildrummy. Total over-all breadth (excluding rear towers) 69 feet 68 feet 6 inches. Total over-all depth (excluding rear portal) 62 feet 61 feet. Diameter of towers ...... 34 feet 33 feet. Thickness of tower walls ..... 9 fee inche6 t s fee9 inches6 t . Length of room towesin . r 20 fee inche6 t s 20 feet 6 inches. Breadth of rooms in tower . . . . . 14 feet 15 feet. Length of rear room ...... 22 feet 6 inches 22 fee inches9 t . Breadth of rear room ...... 21 feet 20 feet. Thickness of back walls of rear room 5 feet 5 feet. Thicknes trancf so e wall reaf so r room 4 feet 3 feet 6 inches. Widt trancf ho e ...... 8 feet 8 feet. Thicknes reaf so r wal roof lo towermn i s. 4 feet fee3 inches6 t .

that Master Jame Sancte sd o Georgi mastes owa workf ro t Linlithgowsa , wher e famouth e s " Peel s the n courswa "i n f erection.n o eo d An 2 1 For Jame Sancte sd o. Morris GeorgiE . e WelshJ e Th , ose Wars of Edward . 145I.,,pp 219, 268-9; also Calendar of Chancery Rolls, 1277-1326, pp. 178, 182, 275, 326, 353; Calendar of Close Rolls, 1288-96, p. 423; ibid., 1296-1302, p. 239. Morris describes him as the architect of the castles with which he was connected ; and in this view he has been followed by Harold Hughes (Archceo- logia Cambrensis, vol. xiii., 6th Series, 1913, pp. 276-7), and C. K. Peers (Trans. Cymmrodorion Soc., 1915-16, pdescribeo . 29)wh , s Jame Sancte sd o Georgi s "oa master-mason. entriee th l al s t Bu " refer to him as '' master of work," i.e., the man who had the business oversight of the castle-building. H neves ei r referre master-masona s a o dt clea s i t i r d that,an , although charged wit adminise hth - trativ financiad ean l responsibilit t himselno s architect e worke th fwa th e r h ,y fo . Accordino gt the Ancient Monuments Commission, Report on Flintshire, pp. 26 and 28, the architect of Flint and Bhuddlan Castles (both buil 1277n i t ) Kichardwae son knows i o , "havo nt wh e constructed severa greae th f to l military workreige th Edwarf f no so " dHarlecI. h Castl ascribe. s ei T . G y db Clark (Mediceval Military Architecture, vol. ii. p. 79) to Henry de Elreton, but I do not know on what authority. Harlech is akin to Rhuddlan and Beaurnaris rather than to Carnarvon. z J. Bain, Calendar of Documents Relating to Scotland,, vol. ii. No. 1308. A NEW SURVEY OF KILDRUMMY CASTLE. 75 October 8, 1299, the roll of a court held at Linlithgow shows that James de Sancto Georgio was at that time deputy-governor of the town, his chief being William de Felton, who, it is significant to note, had been constabl f Beaumarieo s Castl 1296n ei , with Jame t GeorgS f so e under hi mastes ma latteworks.e f ro engages th s rwa A 1 d bot t Beaumariha s and Linlithgow in 1299, it is clear that his transference to Scotland took plac than ei t year. Again e master-masoth , n employe buildinn di e gth magnificent castl f Carnarvoeo s Waltenwa f Herefordo r e appearH . s chargn i e worth f keo there from 1288 onwards 131y b beed 5d ha n an , succeeded by his deputy, Henry of Elreton.2 Now in March, 1304, we find that Walte f Herefordo r , master-mason •workins wa , Edinburgn go h Castle.3 From the above facts it is clear that Edward I. employed the same administrative and technical staffs on his in Wales an Scotlandn s accordingldi i t i d an ; y most suggestive th o fin t en i d gatehouse at Kildrummy, and the windows inserted in the Warden's Tower, such unmistakable links with the Welsh fortresses. It is indeed abundantly clear that the " Hammer of the Scots " has left his mark most legibly upon our " noblest of northern castles." If then we are agreed that the gatehouse at Kildrummy is Edwardian, the question at once emerges: what preceded it ? That the entry was always in this quarter is self-evident. There may have been originally two towers here, with the portal between them, as at Bothwell. Or the two curtain havy ma se come together upo nsquara e gate-towert a s a , e sai b s thaRothesay i d w t no therevidenc l o than n Al s ca tei . r efo either of these views. A third possibility must not be left out of account, namely, thatpresente th gatehouse y predecessor; an never d ha thae th t design of the castle, in fact, was not completed till the Edwardian gate- house was built. In my former account I drew attention to the distinct masonry evidence that the erection of the castle occupied an appreciable perio f time do tha d e engineersan ,th t s thea , y naturally would, began their operations on the back or north side, and worked round towards e Welsth f o h e castles e on front th t A , . Kidwelly (fig. 29), there occurs a remarkable parallel instance of a gatehouse forming an integral part of the original design, but not completed until a later date. It is clear from the most cursory glance at the plan of this castle that its great . StevensonJ , Documents Illustrative Historye oth f of Scotland, r Willia . 393Fo p vol. . ii .m de 1Felton, see Morris, op. cit., p. 263. His connection with Scotland appears to date from the Falkirk campaign (1298), in which he was summoned to serve with thirteen valets. He was in

command of the infantry from Anglesey, ibid., p. 287. 2 See C. B. Peers on Carnarvon Castle in Trans. Cymmrodorion Soc., 1915-16, pp. 7, 10, 15. 3 Bain's Calendar, garrisoe th lisa f . 1536o tn I f StirlinNo n o . vol . ii . g Castle "probablt ya its surrender to the Scots in 1299," occurs the name of " Richard the mason " and two companions. In the total list of about ninety people, only one man, William of Lanark, bears a distinctively Scottish name (ibid., 1119). this I No .s Richar architece dth Flinf o Rhuddlantd an t ? 76 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. gatehous s conceivewa e d standsfro w outsee mno th t i ; ts whera d an e the whole plan is hinged upon it, and the entry to the inner enclosure is left almost undefended in reliance on the formidable outer gate- house. Yet e castlwhil th e res s substantiallf th i eo t e lateth f ryo thirteenth century e Perpendiculath , r e detailgatehousth f o s e make it equally evident that this part was not built until a full century n onlca y later e concludW . e that durin e intervath g a temporarl y defence, of palisaded earthwork possibly, must have supplied its place. Somethin e samth f e o g kin y welma dl have beee cas t th na Kile - drummy I conceiv. e tha gatehouse th t ee tim herbuils th f eo n ewa i t

Fig . Pla 29 .Kif no d welly Castle, Carmarthenshire.

Edward I., as the finishing touch to a long process of construction upon one original design. In his recent Rhind Lectures on The Mediaeval Castle in Scotland,1 Mr W. Mackay Mackenzie, f.S.A.Scot., expressed serious doubto t s a s whether Kildrummy Castle, as we now see it, is the building existing in the thirteenth century, and besieged by Edward of Carnarvon in 1306. After summarising Barbour's accoun e siege th r Mackenzi f M o ,t e com- ments as follows: "It is plain that the idea of Kildrummy in 1306, which to-day.e place se th t thas e mind f ew hi Barbouno " o t s n i i ,Bu d t ha r ther realls ei y nothin Barbour'n gi s description tha incompatibls ti e with the present castle. His language is fairly general. The parts of the castle specifically mentione e "barrath e dar s hald" e "mekilth ; l hall," which was used as a granary during the defence, and through whose "thik burd" the conflagration started by Osbarn appeared "ferst as a 1 Publishe 192Kildrummr n di fo 7 ; . 55-61pp e .yse SURVEW NE F KILDRUMMA YO Y7 7 CASTLE. sterne, syne as a moyne"—a phrase which perhaps suggests that its wooden roof became ignited, but certainly does not justify Mr Mackenzie's assumption tha entire tth e buildin timberf o s curtaie gwa th ; n wall—

'' tha t thata t tym wes batallil tal Within, rycht as it wes with-out" ;

and the "yhet," which was burned, but which the active defenders " muryt up" again. Further it may be noted that Barbour thrice

dwells with special 1 emphasis on the great strength of the castle.

Mr Mackenzi s disposeei mucy la ho dt stress upo e detaith n l which 2 poee th t mentions abou wale th t l that "a t time" being embattle botn do h sides. "This battling, Kildrumme th writesf e o "h s wa , y " 'at that tym,' not of the place when he wrote; there had obviously been some recon- struction. sweeo t t pBu "awa presene yth t Kildrumm strengte th n yo h o sligh os fd incidenta an t a phrasl n altogethea e e seemm - o t sun r warranted procedure. If any significance at all is to be attached to the three words, they may signify (as I have already suggested) merely some temporary defence, a timbesuc s a h r hoarding covered with drenched skins, which afforded the defenders brief and precarious shelter from the flames. At all events, the remark is certainly not sufficient to justify Mr Mackenzie's suggestion that Barbour had in his mind a castle radically differen knee h e wt on whefro e mth n writing about 1380. Commenting on the fact that the Snow Tower was vaulted on every floor r MackenziM , e observes that " wit a hfireproo f tower of this sort in existence, it would have been strange for Barbour to say that the garrison saved themselves from the fire only by taking refuge withi e doublth n e embattlement e wall.t surelth Bu f "o sy thi s strainini s g Barbour's evidenc breakino t e g point. After alle ,h gives us very little detail as to the actual course of the conflagration. Likely enoug e rapith h de flame onseth havy f ma o set preventee dth defenders dispersed along the wall-heads from escaping into the Snow Tower cleas i t I r . fro examination ma e castlth f eno to-day that there has never been a complete passage all round the enceinte at the parapet levelgarrisoe Th . n would thus hav desceno t e e spirath d l staird an s endeavour to make their way across the courtyard through a raging mass of flames. On the general question raised by Mr Mackenzie I would merely offe followine th r g criticism t thaseeme I l argument.m al t o t s s against the thirteenth-century date of the present Kildrummy Castle are bound

Barbour'e 1Se s Bruce,. lineiv . s 59-17bk 5 (ed. W. M. Mackenzie 58-61). pp , . ' Ibid., bk. iii. lines 335-42; bk. iv. lines 65-6, 101-4 (ed. W. M. Mackenzie, pp. 44, 58, 59). 78 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1927. to break down utterly upon the bedrock question of the plan. Studying s aspectsit l becomee al thi on ,n i s s strongly impressed witfactso htw . Th es unit firsit s f designi yto whicn i , l parthal e adjuste- sar co d an d ordinated in harmony as a coherent whole—with the single exception of the chapel, whose abnormality is sufficiently explained, I think, by the desire for correct orientation. In any case the chapel—undoubtedly a thirteenth-century building—is obviously later thae curtaith n n through whic thruss i t hi s orientatio It t out .curtai e th n ni s afteni r all no more abnormal than its presence where it is, and the great prominence s giveeas it enceinte.e to nth t en whole n di Th e character chapele oth f , indeed uniquthins I i , d kquitan a e; e sufficient explanation e obtaineb e remembeo it sw f di r tha ecclesiastican castle a th td eha l founder. The second fact is the completely thirteenth-century nature designe oth f , wit greas hit t donjon, single envelope simpld an , e arrange- e interiomenth f o t r buildings. Whatever sectional reconstructioy nma have taken s lonplacstormd it an gn i e y history above factth o , etw s completely justify, in my opinion, the view that the castle as we know t i to-da s substantialli y e castlth yo gallantls e y defende y Nigeb d l Bruce in 1306. Moreover, the evidence of plan is reinforced by the evidenc e moulding f detailth o el Al othed . an s r detaile chapelth f o s , e styld proportioth an e s lanceit f o nt windows e mouldingth , d an s plinth of the hall, the profile of the bases of the towers, and a great carvee death f o ld fragment assemblew no se Elphinston th n di e Tower, l distinctlal e ar thirteenth-centurf yo overlooe yw charactern ca kr No . the evidence supplied by the insertion of Edwardian windows in an older e presenctowerth n Edwardiaa d f an ,o e n gatehouse whic s clearlhi y secondary. Apart from obvious additions, and from inevitable reconstruction and patchin detailn gi I fee, l strongl impossiblw yho imagino t s i t ei e produce t couli absorptioe th th e e tha f b e dbuildinse o tth t e w ns ga of thirteenth-century remnants in a piecemeal reconstruction of the fourteent d fifteentan h h centuries e morTh e. closel I ystud e th y Kildrummy ruins, the profounder grows my conviction of the fact that the two leading characteristics of the building are the unity and the thirteenth-century natur s designit r Mackenzif eo M . e closes his very stimulating discussion somewhat negativel observiny yb g thas ti "one perhaps justifie suggestinn di g that probably rather more fourtee- nth fifteenth-centurd an y work subsist n Kildrummi s y thas hithertha n o been suggested." With such a verdict I am well enough inclined to agree; but I stipulate that rebuilding, where it has taken place, has followe originae dth lthirteente lineth f so h century. I have to record my special thanks to Mr James S. Richardson, A NEW SURVEY OF KILDRUMMY CASTLE. 79 F.S.A.Scot., H.M. Inspecto f Ancieno r t Monuments s done ha m eo wh , the favour to go most carefully over the ruins with me, so that I have e greath d t ha advantag s exceptionahi f o e l knowledg f mediaevao e l architecture in checking my own views, which, in not a few important details, have undergone modificatio r Richardson'a resulM s f a no t s criticism. It is a pleasure also to acknowledge my indebtedness to my colleague, Professo . MarshallR . C r , M.D., LL.D. f Aberdeeo , n University, s beautifufohi r l serie f photographso Thorntor M o t s ; Taylor . nL o ,wh assiste surveyinn i . FentonWyness J e r dm M ruinse o t g th d ;an , A.R.I.B. A., F.S.A.Scot. prepares ha o measuree ,wh d th d drawings unde directiony rm . Nor would it be seemly to omit due recognition of the courtesy which I have received from the proprietor, Colonel James Ogston of Kildrummy, who has accorded me every facility for examining, measuring, and photo- graphing the ruins at all stages of the long-continued work of excavation and repair. For much information on points of detail in regard to Harlech Castle, e gif sun-printf th o t r fo d an s taken fro large-scals mhi e measured draw- ings of its gatehouse, I am obliged to Mr H. Harold Hughes, F.S.A., R.C.A., A.R.I.B.A. Since the foregoing was in type, I have been favoured with a most interesting communication from Mr C. S. T. Calder, F.S.A.Scot., of the staff of the Scottish Royal Commission on Ancient Monuments. Mr Calder offers an explanation of the anomalies in the lay-out of the chapel whic exactlo hs e seem meem o yt e circumstance o t sth t s thaI t have great pleasur appendinn ei Calder'r gM s communication herewith.

ADDITIONA . CALDERT . LS . NOTEC Y , B F.S.A.SCOT.. , 23rd January 1928. From a study of the plan of Kildrummy Castle it is observed that a breach has been made in the curtain wall between the Warden's and Brux Towere receptioe chapelth s th i e r easter f t o I th fo s .f d no nen also noticed that thidoe d e breact fairlsen th sno t yfi h intende, it r dfo as migh expectede b t t superimposebu , curtaie soute th th n st o h na side and leaves a gap on the north side. (Fig. 30, sketch A.) opiniony m n I , this point represent changa s f desigeo n immediately after operations had begun; and, on the evidence of the now lop-sided breach I woul, d suggese th t t thae firsou th ty t la intentio o t s wa n chapel rectangularly to the existing walls and fill the opening squarely. (Fig sketc, .30 ) hB. To accommodate the widely splayed east windows, a slap in the 80 PROCEEDING E SOCIETYTH F O S , DECEMBE , 192712 R . curtain was necessary, at least as far as the sills, and it required com- paratively little extra effor mako t t extension ea n beyon curtaine dth , and gain the increased internal accommodation within the chapel, which evidentls wa y desired.

BRUX TOWER

SKETCH A bKCTCM B Fig. 30. Kildrummy Castle: Sketch Plans to show alteration in lay-out of Chapel.

A windowed gable on the lines of the first conception, sketch B, presents an element of weakness, as it cannot be effectually covered by the field of fire from the adjacent parapets or towers. To surmount this difficult d enfiladan y e easterth e n wal y swinginb l e chapeth g l southwards is a probable explanation for the departure from the original project, the southward inclination being made in preference to north- ward, to obtain a truer orientation. Proc. Soc. Ant. Scot.] [VOL. LXII.

WHITE LINES i-ltPE INDICATE MODECN iX//M_Lij 6U1LT OM fOJHO F CUPTAlO S K

E LPH I W5TO-N E TOW

OLD EC FOUNDATIONS POSE T A &ASD E

C O u B T y A G D

MAULE TOWtK

E 5 U 6AO H KE

i'5T-" CE.NT. : GIL6LBT

C1CCA I3OO: EDWA12D1AN uT-"ctNTucv-

LATE

i43(b-|455JAMCb I

LOCP LLPHI M5TON E.

DAT E-

IS*-" CENTUC• V SUGGESTED - CHBOMOLOCV KILDRUMMY-CASTLE

CE.NECA!. PLAN T N0\"iA AS - 192.7

AT A A CI&CINA WITH KLITEO tSL)*L IS bull! IN. NOTE. : T-E HALL . THE CII^HEL. .Tut foue LOOPS* IN THE V.'.\ODEM*S TOWLC^^n THE IM THE LAbT CUUIAIH /\ojoiHit-ia AKE OH THE Fl^ST FLOOR •

W. DOUGLAS SIMPSON. NliA W SUttVEY OF KILllBUMMY CASTLE'. FlG . HISTORICA31 . L GENERAL Ft AS.