Proc SocAntiq Scot, (1988)8 11 , 267-276

The Scottish medieval towerhouse as lordly residenc lighe th recenf o tn ei t excavation Christopher J Tabraham*

SUMMARY Forpastthe century, architectural historians have taken lead examiningthe in castellatedour heritage and produced the models which form the basis of our present understanding. Only recently have archaeologists begun to broaden that appreciation. One area in particular where the picture may be changing medievalthe is towerhouse, which generallywe perceive free-standingas self-and contained. This short paper takes a select group of towerhouses in southern Scotland and seeks to show archaeologicalhow investigation substantiallycan alterperceptionour themof residencesas of lordship.

INTRODUCTION yeae Th r 1987 marke centenare dth publicatioe th f yo firse th tf nvolumo Davif eo d MacGibbon Thomad an s Ross's five-volume epic Castellatede Th , Domesticd an Architecture of Scotland (1887- 92). No class of monument surviving from our medieval past has been better served or more intensely studied than the castle. Our fascination for ecclesiastical ruins has a longer history, but the Scottish castle has received the greater attention down the years. We owe an enormous debt to professional architects like MacGibbon and Ross, and to architectural historians of the calibre of Mackay MacKenzie (1927), Stewart Cruden (1960; reprinted 1981) and John Dunbar (1966), for picking over castellater bonee ou th f o s d heritag publishind ean majoe gth r seminal works. Until quite recently, excavation scarcely played any part at all in this learning process. Where it was undertaken t tooi ,for e kth f followin mo g walls clearind an , g away wit abandoy e hga th l nal obstructive rubble and soil to reveal yet more stone walls for architectural historians to peer at and puzzle over. But the advent of more sophisticated archaeological techniques after the Second World War, coupled with the emergence of an interest in medieval archaeology in its own right (see Hinton 1983) openeopportunities w ,ha ne p du castln si e studies increasingls i t i s a , y doin othen gi r medieval areas, particularl r burghsou n yi . e amounTh f archaeologicao t l work n medievacarrieo t ou d l site n Scotlani s s stilr i d fa l outweighed by that on monuments of a greater antiquity, but what has been done has already begun to contribute enormously to our present understanding. Archaeological excavation clearly plays the lead role at those timber castles where no masonry survives; witness Peter Yeoman's work at the Giffards' mott Castlehilt ea Strachaf lo Grampian i n (Yeoman 1984) vitaa alst s I . lo ha par plao t t ya complex castle sites where a masonry phase, or phases, may be but one development in the evolution

* Historic Building Monumentsd san Brando0 2 , n Street, Edinburgh 268 SOCIETY OF ANTIQUARIES OF SCOTLAND, 1988

of the site as a whole; Gordon Ewart's work on the peninsula site of Cruggleton in the Machars of Galloway (Ewart 1985) has been the most significant such case to date. But archaeology also has an important part to play at those castles where all the principal phases are masonry and where the lead, dateo t bees ha , n architecturae taketh y nb l historian. Elsewhere in this volume is a report on an excavation which I carried out in partnership with George Goo t Smailholda m Tower prominena , t landmar Bordee th n ki r country. Prio excavao rt - tion, it was widely assumed that the residence of the Pringle laird was confined within the towerhouse, with service office skitchen- , bakehouse, brewhouse, stable groupe- c ,et d barmki e arounth n i t dni enclosure. That perceptio somewhaw no s ni t altere resulexcavatione a s th d a f to , whic broughs hha t ligho t seconta d residential unit, comprising probabl outen ya r halchamberd lan , standing alongside e towerhousth verd ean y likely contemporar previousld ha ye witW . yhit excavate Threavn do e Island Gallowae th n i , y Dee unearthed an , substantiao dtw l buildings, again contemporary witha towerhous agaid ean n interprete havins da g serve maie th residentiann a di (Gooe lus Tabrahad& m 1981). These discoveries have prompted me to look again at the Scottish towerhouse, particularly our usual understanding of it as a residence of lordship.

THE MEDIEVAL TOWERHOUSE AS RESIDENCE: THE PRESENT VIEW The Scottish medieval towerhouse has for a long while now been viewed as a free-standing, largely self-contained entity, lofty and forbidding, 'closed-up and inward-looking' as Cruden so tellingly puts it (1960, 108-9). He continues: '. . . a simple rectangular tower achieving height but not extent is an inevitable result of conditions and requirements of impoverishment, the need for security and necessity of obtaining it cheaply. To these demands of security, economy and uncertain future a plain rectangular tower with exceedingly thick walls constructe obvioue stonf th o s l deal wa s answer als s reductioe owa th t courtyarI .e th f no d castl famila o et y dwellin littld gan e more.' veie Thith ns s i struc MacGibboy kb Rosd nan centursa y earlier (1887,1,144): mansione Th nobilitye . . th f . ' so , being constantly liablattackee b o e t burnt d dan , were necessarily built of stone; while from the improverished conditions of the barons of the time, they were of the simplest form. All that was required was a stronghold sufficient to accommodate the owner's family and personal retainers proteco t d an , t them from sudden attack. These castles consis squara f to r eo oblong tower, with thick wall' . . s. Not that the towerhouse is seen in complete isolation. MacGibbon and Ross again (ibid, 145-6): '. . . there can scarcely be a doubt that all these keeps had a courtyard connected with them, enclosed with a good wall. This court or barmkin was essential to contain the stables and other offices.' Mackay MacKenzie coine phrase dth e 'farmyard attachments' (1927,195 describo )t e these subsidi- elementsy ar servicr o , e offices, thereby emphasizing their inferior status. 'self-containede Th ' qualit towerhouse,'e th f yo incorporatin. . . essentiae th l gal l ingredients of the normal medieval house in a remarkably compact form . . .' (Dunbar 1966, 37), is stressed throughout the published record. If there are residential units present, integral with or adj acent to the towerhouse, these are, almost without exception, seen as a secondary development, ranges erected at some later dat responsn ei upturn family'e a o eth t n i s fortunes arrangemene Th . Crichtot ta n Castle, in Lothian s frequentli , y give o illustratnt e this evolution from simple towerhous o compleet x courtyard castle (Simpson 1957) rare e.Th exceptions, lik towerhouse eth greatd ean hall unit buily tb Duke th f Albaneo t Dounya e Castle about 1400 (Pringle 1988) ,cautio o seet t mno n against this empirical evolutionary view; and the complex late 14th- and early 15th-century establishments of the TABRAHAM: THE SCOTTISH MEDIEVAL TOWERHOUSE AS LORDLY RESIDENCE 269

likes of the Douglases at Bothwell Castle, on the Clyde, and Tantallon Castle, beside the Forth, where hall-blocks are central elements in the ensemble of buildings, are seen as lordly residences of a quite different nature. But the information coming out of the ground from both Threave Castle and Smailholm Tower strongly suggests tha traditionae th t lmedievae vieth f wo l towerhous fule th l s ea extent of a lord's residential needs may be seriously flawed.

TOWERHOUSE DOUGLASEE TH F SO S In orde demonstrato rt e this last point tako t wisI , verew a hno y select grou towerhousef po s belonging to one noble Scots family. Firstly, each will be examined solely in terms of its upstanding architectur havy ema t assesfunctioneo i et w sho residenca s da lordshipf eo . Thereafter wilI , l bring into focus evidence from excavation to see how this might alter our perception. towerhousee Th s tha havI t e selected wer earle builth Douglasf sr o tfo , undeniabl mose yth t significant aristocratic famil southern i y n Scotlan e latth e earld n di 14tan yh 15th centuriesn I . choosing residences pertaining to one family, I hope that I will be able to compare and contrast them more tellingly. I am conscious that my analyses can only be regarded as general and that there are factors which perforce compromise conclusions somy m f eo particularn .I awarm a ,I e tha castlee tth s have changed over the years and that it is impossible from this distance in time to analyse each castle completes init d state firs s whetwa builtt ni . Some element longeo sn r survive; others have beeo ns altered that they defy untangling. Nevertheless accepte on f i , s what follow broan si d terms, thenI think that the comparisons are, broadly-speaking, valid. three Th e towerhouse southern i l al e nsar Scotland powerbase th , Douglasese th f eo . Threave Lordshie th 622)9 n i 73 , Gallowayf p(NGX o N R Neward ,an 421294) T k (N Forese th n Ettrickf ,i to , were residences of the senior line stemming from Black Archibald, or Archibald the Grim, who died at Threav 1400n ei . Hermitag 960)6 Liddesdale49 n ,i Y residencee (N th s ,wa Archibald'f eo s cousin, William Douglas, the first earl of that House, whose estate passed in time to the junior line, or Red Douglas earl f Angusso . Now we can reasonably assume that each of these Douglas residences catered for a broadly similar accommodation need, for each was the chief seat in its respective lordship or estate. Each would have been expecte provido dt e suitabl householdes lodginghi eare d th an lr sfo ,s thahi s i t immediate family, officials and domestic staff. It is not possible to determine precisely how many folk knoe w talkint we e ar ye ver s f hera gy o r littlefo e abou detailmedievae y th t an f so l nobleman's househol alont dle e Douglasethose th f eo s (Wormald 1985,94-5). What little researc bees hha n done -for example, Keith Stringer's analysis (1985,149-76 Earf )o l Davi Huntingdon'f do s household dan followin late th e n g12ti h century Margared an , t Sanderson's review (1974, 31-48 Archibishof )o p Beaton's 'kin, freindis and servandis' in the mid 16th century - points to substantial numbers being involved. Not all would require to be accommodated within their lord's castle(s) of course, but the figures serve to illustrate the size of a magnate's following and the potential accommodation requirement Douglases e likee th th f thi o sf o t s o s d permanenAd . t provisio t asidneee se n th eo dt space for visiting guests with their travelling households (in 1591, for instance, earls attending the Justice court of King James VI were permitted a travelling household of 24 (Mitchison 1983, 8)), and clearlwe ear y dealing wit requiremene hth provido t e substantial living space. These figures should borne b individuae looe minth n ei w t ka s da l residence earle th Douglasf f so s o .

HERMITAGE CASTLE Let us first consider Hermitage. This castle has a complicated development (RCAMS 1956,1, thernumbed a e an 75-85e) ar puzzlef 63 r o o ,n s still awaiting explanation. What seems clear enough 270 SOCIETY OF ANTIQUARIES OF SCOTLAND, 1988

is that William Douglas firse ,th t earl, whe came nh e into possessio lordshie th late f no th e n p1360si , took occupation of an existing and recently completed masonry building, more like a fortified manorhouse Dacre , builth y eb t lor f Cumberlanddo hel d Englise lande ha th dth f o so hwh , king. William Dougla t intspu o effec majoa t r rebuilding programm thaf o td centuryeen whiche th y ,b , resulted in the creation of a mighty towerhouse. This comprised a large rectangular block, most probably four storeys high, to each corner of which was added a smaller tower of similar height. arrangemene Th accommodatioe th f o t t entirelno s ni y cleart appeari t bu , havo st e bees na follows: in the main block, at ground level three storage cellars; at the first floor level what may have bee nlarga e hall wit smalleha r chamber perhap ease th secon te t endsa th n ;do floor another hald lan chamber each wit hhiga h ceiling, suggesting tha principae tth this tope swa th l , t suitedirectla d ;an y beneat rooe hth f wit wall-walks hit battlementsd san , another hall whic havy hema been intendeo dt serve as temporary quarters for the defending garrison in times of siege. In the adj acent corner towers we hav mixturea chamberf eo servicd san e offices lattee th , r confine groune th firsd o dt dan t floors, the former occupying the upper storeys. The total floor area, assuming a four-storey main block, amounts to some 16566 ft2 (1539 m2), which can broadly be broken down into 11916 ft2 (1106 m2) of residential accommodation (halls and

chambers 465d )an 0 ft (431 mservicf )o e (kitchen, bakehouse, prison, storag forth)o s d ean . This 2

spac expresses ei tabln di 2 e for illumn i s 1.

o o R A R A I N 1000 S Q u A R E f E

T HERMITAGE THREAVE ILLU S1 Tabl e comparin e flooth g r aren ft(i 2a) available within the towerhouses of Hermitage, Newark and Threave

NEWARK CASTLE Now how does Newark compare? This towerhouse is considered to have been built for Archibald, fourth earl of Douglas and son of Archibald the Grim, very early in the 15th century (RCAMS 1957, mor61-5a s i 44)o en t , I .conventiona l towerhouse than Hermitage simpla , e rectangular block probably five storeys high beneath the battlements. Despite the subsequent alterations (there is good reason to suppose that the existing attic storey is a later addition), the original arrangement seems clear enough - storage at ground level; a kitchen area on the floor above; thire halth e dn th lstoreo y and, above that furtheo ,tw r floors therl al som e n ear I . e 5726 ft2 (531 m2) of floor area, roughl favoun i 2 f residentiay3: o r l accommodation (3564 ft2 (331 m 216o 2t ) 2 ft2 (201 m2) (illu. s1) TABRAHAM: THE SCOTTISH MEDIEVAL TOWERHOUSE AS LORDLY RESIDENCE | 271

Immediately, the contrast between the provision at Newark and Hermitage is striking - Newark is a third of the size of Hermitage in terms of available floor space. Yet both were the residences of wealthy and important earls who were closely related. Both were the chief seats in their respective estates and it is not immediately apparent why there should be such a wide discrepancy.

THREAVE CASTLE e momenth r fo , t thenr thirus t ou d,Le o t pasDougla n o s sthiw towerhoussho e se d ean compares mose th . f tThreav o evocativ medievar e ou on l s eal i f eo l castles, buil Archibaly b t e dth Grim as his base in Galloway following his elevation to that lordship in 1369. At the time of his death ther 1400n ei mights hi , y rectangular towerhouse consiste fivf do e storey grount a s- d leve storagla e provision, includin g prisona flooe th rn o ;abov kitcheea n thire areath dn o ;store e yhallth a n o ; fourth storey two private chambers and at the top, beneath the roof, an upper hall which, as at Hermitage havy ema , been intende servo dt temporars ea y siege quarter earl'e th r s sfo househol d (Tabraham 1988). Adding up Threave's available floor space, we find that it almost exactly corresponds with that at Newar k- 551 0 ft2 (511 m 2)- wit remarkablha y similar rati residentiaf oo servico t l e (illu. 1) s Again, we puzzle at this apparent discrepancy at Hermitage and Threave in the amount of accom- modation provided. Both residences were built by men of equal rank and wealth in areas of comparable insecurit instabilityd yan Archibaly Wh . Grie dth m shoul contente db , seemingly, witha towerhouse residence a third of the size of his cousin's is not immediately explained.

EXCAVATIO THREAVT NA E CASTLE Up to this point we have been considering these castles solely from their surviving architectural remains. As regards Hermitage and Newark, we have no alternative for neither has been archaeolo- gically investigated t Threav;bu (Goos eha Tabrahad& m 1981, 90-140). freelI y admit that whe begae nw n excavatin Threavn go e Islan 1974dn i regardee ,w d Archibald the Grim's castle much as our predecessors had done. 'This lofty grey stronghold', MacGibbon and Ross described it (1887,1,157), reinforcing our perception of it as an isolated and fully self-contained residence surrounde dfast-flowine onlth y yb flae th t gmarshd water Rivean e e th yf sr o De haughs n .I our ignorance, we assumed that the foundations of two buildings encountered by the Office of Works in 1923, whilst the enormous programme of consolidating the upstanding masonry was drawing to a close, probably belonged to an earlier castle of the Celtic lords of Galloway which is reputed to have stood on the island and been burnt by The Bruce in 1308 (Skene 1871, 345). Excavation showed this assumption to be mistaken. buildingo tw e botd Th ha sh been erecte e d14t th lat hn i ecentury , probabl r neao n ryo completio adjacene th f no t towerhouse. Clearly then towerhouse th , neves ewa r intende stano dt d alone as the full extent of Archibald's Gallovidian residence. Quite how these buildings functioned it was impossible to determine because of the paucity of the surviving remains - mere wall-footings and no superstructure to speak of - for both were dismantled about 1450 to make room for the new artillery work. From the surviving evidence, however, it would seem that both structures were at least two-storeyed, apparently with service accommodatio groune th n no d floo plausibld an r y residential space over. Both structures may be compared to a building surviving within another of Archibald's castles, Bothwell, on the Clyde. Here Archibald and his son, also named Archibald, the fourth earl, remodelled the castle which had been drastically damaged during the Wars of Independence in the firs e 14tt th hal hf o fcentur y (Simpson 1925, 165-93). Adjacen now-vanishee th o t t t oncdbu e enormous square tower at the north-east corner, which most probably served as the earl's principal 272 SOCIETY OF ANTIQUARIES OF SCOTLAND, 1988 lodging fine w two-storeye,da d building, simila Threave'o rt s Buildin floon i g1 r area almosd an , t identical to Threave's Building 2 in ground-floor layout, though substantially bigger. At Bothwell the functio perfectls ni y cleaouten a grear r- ro firse t th halt n flooo l r abov undercrofn ea storager fo t . Perhaps Buildin Threavt a g1 more th s ei closely comparable, considerin flooe gth r are a- 217 8 ft (202 mlin2) at Bothwell compared with 180n0 ft2 (167 m2) at Threave (illu2 s 2). Bothwell Threave Threave Tantallon Hermitage Smailholm (D (2)

ILLUS 2 Block plans of great halls at Bothwell, Threave (1 & 2), Tantallon, Hermitage and Smailholm

The arrangement of the accommodation at Bothwell during the Black Douglases' period of occupation, insofar as we can make it out from the somewhat patchy remains, is perhaps a good deal close thao rt Threavt ta e than migh caste appea firsth t ea e b t mightsigho e rt th te (illuyse n sca 3a)e .W north-east tower servin earl'e th s sga principal lodging, containin gmodestly-sizea firse d th hal t n o l floor, where he could entertain more intimately, sandwiched between basement storage and upper private chambers. Handily placed adjacen greats hi s thio t thall swa mora , e capacious banqueting and reception room, complementin somewhae gth t cramped provision withi towere n th kitche e .Th n block servin greae gth t halllargelw no , y disappeared, stood immediatel north-wese th o yt e th f o t

a) BOTHWELL

b)THREAVE

c) NEWARK

T= TOWER H-HALL L - LODGING K- KITCHEN C - CHAPEL physicae ILLUTh S3 l relationship between tow- erhouse, great hall and other accom- modatio t Bothwella n , Threavd an e Newark TABRAHAM SCOTTISE TH : H MEDIEVAL TOWERHOUS LORDLS EA Y RESIDENCE 273

great hall. Beyon greae dchapeth e soute tth hals th lho lwa t wit h ancillary residential accommodation withi roune nth d towesouth-ease th t ra t corne elsewherd ran enclosuree othery th b e n ei n si us r ,fo househole th visitind dan g guests. No looe ww k agai t Threavena . Upon first inspection entireln a , y differene t th castlen i t bu , excavatee lighth f o t d materia dissimilaro s t lno loft e continue th yW .e towerhousse o et e servins ga the earl's principal lodging, incorporating a reasonably-sized hall in the middle, with private cham- bers abov kitched e an storag d nan e beneat unlikt no h - north-ease e thath n ti t towe Bothwellt ra t .Bu grouped aroun maie dth n eas ttowerhouse s fronfrone hi th f f o to t o door) tw m e 5 ,ar , withi(1 t f 0 n5 large and doubtless imposing structures (illus 3b). (There are other building foundations to the south awaiting further excavation.) The nearer one, Building 1, quite probably housed the outer or great firss it hal tn flooro l ; Buildin conceivablg2 y housed ancillary residential quarter firss it tn sflooro , perhaps with a chapel in the east-west orientated projection at the south-east corner. In essence, the arrangemen thal al tt differenno s ti t from Bothwell without ;bu t Bothwell's great curtain wall, which was in any case substantially a legacy from the 13th century. At Threave, in its naturally defensive island setting, it could be argued that a stone curtain was not so essential. The requirement for a great hall, independent of but complementary to the provision of a smaller, more intimate hall within the lord's private lodging, is demonstrated not only at Bothwell but elsewhere in Scotland. Tantallon Castle, for example, was built in the 1350s most probably by William Douglas, the first earl and creator of Hermitage (Tabraham 1986, 5). Here, adjacent to his own lodgin colossae th gn i l Douglas Tower placee ,h dtwo-storeyea d hall block lesr fa s ,a substantia l building than the tower but providing a further 2200 ft2 (204 m2) of accommodation (illus 2). And ther gooo n s edi reaso supposno t e tha earl'e tth s residential needs wer differenty ean , whethes wa e rh in residence at Tantallon or Hermitage.

THE DOUGLASES' TOWERHOUSES RECONSIDERED IN THE LIGHT OF EXCAVATION Let us now return to the three towerhouses and see how excavation has altered the picture. In terms of available floor area, Threave can now be demonstrated to be a good deal larger than at first supposed. In fact, it has slightly more than doubled in size and there are still more structures awaiting

F O O R A R E A I N

S Q u A R E F E T HERMITAGE THREAVE

ILLUS 4 Table comparing the floor area (in ft2) available at Hermitage, Newar Threavd e an lighk th f o tn i e recent excavation 274 SOCIETY OF ANTIQUARIES OF SCOTLAND, 1988

excavation (illus 4). Perhaps when the full extent of Threave's accommodation around the year 1400 is known, it will turn out to be little different to Hermitage. Ther cousinseo stiltw le remainth , y Willia puzzlwh e Archibaldso th d t ms an e a , should have buil quito ttw e different towerhouse meeo st t their similar residential needs t ThreaveA . , Archibald appear havo st e been developin traditioe gth n establishe precedine th n di g centuries, maithaa f to n keep, or donjon, with a high degree of defensive capability, surrounded by lesser residential and service units - precisely what he acquired at Bothwell in the 1360s following his marriage to Joanna Moray t HermitageA . , cousin William contrive produco dt morea e fully self-contained residence. These quite separate approaches resulte a differenc n di e manne th n i ewhicn i r e variouth h s component lordle th f syo residence were arranged. Hermitag more th s e radicaegreawa e th r t lhalfo l was place ensemblheare e thirth th e f t dtth o a dn store maiee (o th nf y o block) wit lord'e halhn th lsow sited elsewhere, most probably in the largest of the attached towers and at the same level. But these are differences at the margin; they do not affect the central fact to emerge from excavation - that, generally speaking, a similar amount of accommodation, and in roughly the same proportion of residence to service, was being provided at both. So where does this leave Newark? A towerhouse whose total floor area is so similar to Threave's that one might justifiably raise more than an eyebrow at its present dating, 50 years after Threave. Unlike Threave, Newark is not built upon a naturally defensive island but on a bluff overlookin Yarroe gth w Water surroundes i t I . stonea y db barmkin wall tha generalls ti y considered to be of 16th-century date, largely on account of the oval gunholes that puncture it at intervals. Be mayt thai s ta , wha cleas ti r even fro mcursora y examination throug jungle hth e envelopethaw tno t si is that the wall is certainly not of one build. Royae Th l Commission' sInventorye plath n i hint t thisa s (RCAMS 1957). Straight joints, particularly in the projection along the east side, strongly suggest a work of more than one period, whils latrine tth e chutnorth-ease th n ei t corner point existence th o st residentiaf eo l accommodation her t somea eevolutionare hole dateth w o dt f I . y theory laid predecessorsr dowou y nb , thee nw would view these outbuilding mucs sa h later additions. 'Acountre sth y improved, buildings providing enlarged accommodation were extended around the inside of the walls of the courtyard' (MacGibbon & Ross, 1887,1,146). I have my doubts. Taking the average ground area of the hall-blocks that we hav eothee seeth t nra Douglas castles thef o me woul,on perfectlt dfi y int north-ease oth te parth f to barmkin wall, interestingly in a similar position to those at Threave, a short distance from the entrance int towerhouse oth e (illus 3c). Excavatio t Newarna k Castle might well radically alter rou perceptio eare Douglas'f th o lf no s chie wealthfs seahi n i t y holdin Forese th f Ettrickgf o o t .

EXCAVATION AT SMAILHOLM TOWER accepw no viee e th t Iww f that e standar,th d for f towerhousmo itsely incapabls eb wa f f eo meetin residentiae th l gal l demands laid upon individual member highee th f so r nobility, with their large households and their onerous courtly, ceremonial and hospitable obligations to maintain and sustaie fulfilw n nca , this arguemen move w s eta dow sociae nth l lesse e scalth o ert nobilit landed yan d gentry? Smailholm Tower, near Kelso, was built in the 15th century by a head of the Pringle surname - t cleano s whicri (se e precedine ehth on g pape thin ri s volume) Pringlee Th . s happe havo nt e been closely associated with the earls of Douglas for they had served as squires of Black Douglas, were tenant Douglasese th n si ' holdin Forese th f Ettricf go o t held kan d important administrative posts Warde th f withif thao so e t nForeston . Pringl f Smailholeo alss possessiomn owa i sizeabla f no e landholding scattered throughout the Border country and, although he was clearly nowhere near the TABRAHAM: THE SCOTTISH MEDIEVAL TOWERHOUSE AS LORDLY RESIDENCE | 275 league and stature of his master, the earl of Douglas, he was a well-to-do member of the landed gentry t withouno , t standin Bordee th n gi r community. The archaeological investigation at his Smailholm residence shows that the accommodation within the simple towerhouse (1394 ft2 (130 m2) of residential accommodation; 850 ft2 (79 m2) of

service), was augmented by a further 1345 ft (125 m) elsewhere within the barmkin. Of the latter, 2

only servicsoms mkitche7 e ft6 (3 th )wa e39 n e(i n sout e blocth n hwese ko sidth f t eo court) e ;th 2 22 remainin outen a ms 8 ftr9 2 (8 2)halchambewa g d 94 lan r placed alon norte gth hwese sidth f t eo cour t (illus 2). The presence of such a building considerably alters our appreciation of Smailholm as a lordly residence. Certainly, the towerhouse still stands at the heart of the complex, the chief residential unit and the most secure in troubled times. But it can no longer be viewed as an isolated dwelling house, surrounded by mere 'farmyard attachments'. As at Threave, the home of his master the earl of Douglas, so at Smailholm the Pringle laird, albeit in a more humble manner, had a residence somewhat more elaborate than we had hitherto imagined.

CONCLUSION Clearly then, there is more to the medieval towerhouse than immediately meets the eye. What were formerly thought of as quite straightforward towerhouse residences at Threave and Smailholm are now demonstrated to be a good deal more complex. It is important, however, at this early stage of archaeological researc overworo t t hno excavatede kth evidence precise .Th e natur additionae th f eo l building t Threava s e canno e determinedb t thougd an ; presence hth ground-flooa f eo r halt a l Smailhol mmors i e easily demonstrated exacs it , t dat constructiof eo n must remai somn ni e doubt. However, what cannot be denied is the fact that the archaeological evidence is changing our overall understanding of the form and function of the medieval castle in general and of the medieval towerhouse in particular.

ACKNOWLEDGEMENT I am grateful to Dr Denys Pringle for his valuable comments during the preparation of this paper.

REFERENCES Cruden, S 1960 The Scottish castle. London. Dunbar, J 1966 The historic architecture of Scotland. Edinburgh. Ewart G , 1985 Cruggleton Castle: report of excavations 1978-81. Dumfries. Good, G & Tabraham, C 1981 'Excavations at Threave Castle, Galloway, 1974-78', Medieval Archa- eol, 25 (1981), 90-140. Hinton, D (ed) 1983 25 years of medieval archaeology. Sheffield. MacGibbon, D & Ross, T 1888 The castellated and domestic architecture of Scotland, 5 vols. Edinburgh. MacKenzie, W M 1927 The mediaeval castle in Scotland. Edinburgh. Mitchison, Rosalind 1983 Lordship to patronage: Scotland 1603-1745. London. Pringle, D P 1988 Doune Castle. Edinburgh. SDD (HBM) Official Guide Book. RCAMS 1956 Royal Commission on the Ancient and Historical Monuments of Scotland, An inventory anciente oth f historicald an monuments countye th f o of Roxburgh, vols2 . Edinburgh. RCAMS 1957 An inventory of Selkirkshire, Edinburgh. Sanderson, Margaret 1974 '"Kin, freindi workeo servandis"d wh san Archbishor n dfo me e :th p David Beaton', Innes Rev, 25 (1974), 31-48. 276 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1988

Simpson D 192 e architecturaW , Th 5 l developmen f Bothwelo t l Castle', Proc Antiqc So Scot,9 5 (1924-5), 165-93. Simpson D 195W , 7 Crichton Castle. Edinburgh. Ministr Workf yo s Official Guide Book. Skene (edF )W , 188 historianse Th 1 Scotland,n i . EdinburghVI . Stringer J K 198, 5 Earl David of Huntingdon. Oxford. Tabraham C , 1986 Tantallon Castle. Edinburgh (HBMD SD . ) Official Guide Book. Tabraham C , 1988 Threave Castle. Edinburgh (HBMD SD . ) Official Guide Book. Wormald, Jennifer 1985 Lords and men in Scotland. Edinburgh. Yeoman ,P 1984 'Excavations at Castlehill of Strachan, 1980-81\ Proc Antiqc So Scot, (1984)4 11 , 315- 64, fiche 4: C1-G3.

This paper is published with the aid of a grant from the Historic Buildings and Monuments Directorate (SDD)