Proc Soc Antiq Scot, (1995)5 12 , 813-879

Towards an architectural history of J Philip McAleer*

ABSTRACT The Tironensian abbey Kilwinning,of compared sisterits to abbey Kelso,of beenhas relatively neglected, despite the fact that the extent of the remaining ruins at both sites is about the same - and equally dramatic. At Kilwinning, those remains consist of parts of the church - the south arm and gable of the transept, the wall of the south aisle, the west front - and parts of the east and west rangescloister,the of including chapter-housethe slype.a and Especially conspicuous well-are preserved ensembles such as the chapter-house entrance, the gable of the transept and the arch to its eastern chapel, the cloister portal, and the arch at the west end of the nave aisle. A rapid survey reveals that each of these elements clearly represents a different building phase, yet the chronology abbey'softhe construction, particularly church, the that neverof has been worked out. Thisis what this paper attempts to do. The task is not without its difficulties since, owing to the lack of documents, dated be parts the can only approximately processthe by stylisticof analysisand comparison. Even then, some buildingspartsthe of stubbornly refuse yieldto their up histories, and continue to tantalize the prying investigator.

INTRODUCTION ruine Th f Kilwinninso g Abbe Ayrshirn y i three th 303432S ef o majo e(N e those on ar )rf eo house s e Tironensiaoth f n Orde .n i r 1 The e lesar ys extensive tha remaine nth t Kelsa s o (Roxburghshire), founded in 1128, and at Arbroath (Angus), founded in 1178,2 both of which were royal foundations Davio t e dI (1124-53du , Williad an ) Lioe mth n (1165-1214), respectivelyn I . contrast, the patrons of Kilwinning were the de Morevilles. The probable founder has been identifie Richards a d Moreville d e (ob. probable 1189)th d an , e foundation dat 1187,s ea 3 although dates of 1140, 1157, and 1191, have been cited in the literature;4 Richard's father, Hugh (ob. 1162), has also been credited.5 The possibility that a church existed at Kilwinning as early as 1184 adds to confusion.e th registee th s A 6 r books have been lost sinc earle eth y 18th century, exace 7th t datf eo Kilwinning's foundation cannot now be determined until and unless fresh documentary evidence turns up.8 The abbey was certainly established by 1202/7 when an of Kilwinning appeared as a witness to an agreement between the of St Andrews and Glasgow at a council held in Perth.9 The site (illus 1) is now partly occupied by the church, erected c 1775 (at which time abbee nave th th f yeo churc pulles hwa d down dominates i d 10) an ,bell-towe a y db r constructen di

•• Technical Universit f Novyo a x 1000ScotiaBo ,O HalifaxP , , 4 Canad2X J aB3 814 SOCIET ANTIQUARIEF YO SCOTLANDF SO , 1995

ILLUS 1 Aerial view (1984) of the abbey ruins and the parish church from the south-east (Crown Copyright: Royal Commission on the Ancient and Historical Monuments of Scotland)

the early 19th century. The latter replaced the surviving north-west tower of the fa?ade structure (apparently a twin-towered western transept), which had been damaged by lightning in 1809 and

whic s subsequentlhwa y demolishe 1814.e meagrn di Th e remain abbee th f yo s church which 1

stretch between the two later structure1 s are dominated by the south (gable) wall of the transept which survive nearlo t s s fulyit l original height s connectei t I wese . ruine th th f t o so fa9addt e structure by the south wall of the nave which is preserved to a relatively uniform height of about 3.65 m. Neither they nor the fragments of the attached monastic buildings to the east and west have been subject to a close analysis, either in terms of their stylistic elements or the possible building phases. Generally, the ruins have been dated as a unit to the late 12th and early 13th centuries.12 Little archaeological investigatio takes nha n placechurce finas plae th it f Th .no n l hi phase was published, along with some detail f standino s g arches 1878n i , , although befors thiwa se eth s clearesitewa f latedo r domestic structures. lase 1th 3t Jusf centuryo td beforen briea e , th e f McALEER: TOWARD ARCHITECTURAN SA L HISTOR KILWINNINF YO G ABBEY 815

descriptio remaine th f no s appeared t withoubu , indicatioy an t e sequencth f no r periodeo f o s

construction. Nothing has been published on the physical fabric of the abbey buildings since 4

then. 1 Any attempt to interpret the remains encounters two major problems, at times difficult to surmount, even if they do not constitute absolute obstacles. The first problem has two aspects. One is the general character of the development of the later medieval (Gothic) styles of architecture in Scotland. The revival of early forms in the latest period (after c 1380) introduces an elemen f uncertainto t y regardin datine gth f undocumentego d buildings whic compoundes hi d by the nature of the second aspect, the survival of or retention of certain early or 'Romanesque' forms well after other 'newer' motifs have been accepted. The latter tendency is demonstrated by semicirculacontinuee e th th f o e dus r arch long afte introductioe th r pointee th f no dy formb d an , e retentioth f decorativno e motifs conventionally describe s Romanesquda e well aftee th r

succeeding Early Gothic motifs have appeared and with which they are mingled. The former 5

can be recognized by the appearance of the semicircular arch in buildings o1 f the period

subsequent to the Wars of Independence, as well as the revived use of the barrel vault. The 6

occasional tendency for the later builders to favour some of the simpler early details help1 s to further possible confusion.17 The second problem - which perhaps constitutes a greater obstacle - is the existence at Kilwinnin f numerougo s undocumented restoration physicae th f so l fabric. Man f thesyo e repairs may perhap bettee b s r describe s extensivda e patching; however the e labelledyar , they often confuse or obscure the reading of the fabric, for they invariably seem to be present at important junctures and, indeed creaty illusioe ma , eth junctura f no e where none existed r thehavy o , yma e eliminate whice don h did. In the analysis which follows, the consequence of these uncertainties will at times be very evident, either in the decipherment of the relationship of one building phase to another, or in the identificatio e correcth f o n t perio phasea r dfo . Therefore, where appropriate, alternative interpretations wil considerede b l t musI . e admitteb t e whic dseey on thae hmma th t most mora t seeo t eno mo convincin perspicacious y ma e m o gt s observer readine remainse th th n I .f go , I generally have found myself drawn to the solution of an individual problem that seemed to best fit int o'logicala ' sequenc histor e abbee th th r f eyfo o building wholea s sa . Other disagrey sma e I hop , andso e f thei , y will follo inte wm o print with their correction thao s accuratn a t e fabric histor Kilwinninf yo g Abbe finally y ma composede yb . Non e lessth e , despit foregoine th e g reservation r cautionso s archaeologican a , d an l stylistic analysis of the existing above-ground remains suggests they are the result of at least three major building phases (illus 2). The first, which consisted of two campaigns, may be identified as probably comprising a small transeptal church and the eastern range of the associated monastic buildings, althoug t mushi admittee b t d thasuggestee th t dearle forth yf mo challengee churc b chapter-house y date th f th hma eo d dan e entrance disputed. d Phasean o stw thre considerable ear y less problematic. The separatee yar d fro firse f vermo th t e yphason y eb limited extenconstructioe cloiste portae representh d w th - t an o ne t l r- a f nto successivt ebu separate campaigns which can be interpreted as the enlargement of the early church: first at the east end, possibly repeatin e initiagth l plan (chancel, transept, eastern chapels) largea n o , r scale; then toward weste th s , perhaps introducin aislen ga d navmosd an e t certainl ytwina - towered western transept. It is the remains of this work at east and west which now dominate whice sitd th ean h constitut ruinse bule th eth fourthf A k.o , minor, phase, representee th y db existing west wall, marks the contraction of the church due to the abandonment of the western structure. 816 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1995

ILLUS 2 Diagrammatic plan of the abbey church and monastic buildings, showing the consolidated masonry and the reconstructed outline of the church; compiled from plans held by Historic Scotland (222/187/40, 58-9), with a suggested dating scheme. Numbers in the key denote: (first church) 1, 1160s; 2, 1180s; (second church) 3, 1190s; 4, 1230s over 1160s; 5, 1230s; 6, 1250s. The early 19th-century bell tower and the current church (c 1775) are shown in outline. (Drawn by Margaret Finch)

THE FIRST CHURCH

Description and interpretation Certain feature existine th f o s g standing fabric suggest that earliepartn a f so r smaller structur havy ema e been incorporated into a later rebuilding. Of this hypothetical first church, it is possible to identify only the lower soute partwesth d f han so t s wallsout it soute f th ho s hd transeps nave walan it e limit f m o Th l. f thiar tso s easte indicatee b th wor t ,y a visibl : kma dby e insid (latere eth , enlarged) transept slightla , y irregular vertical joint eight course ssoute higth hn hi wall, 4.8 fro 2m wese mth t wall (illuverticaa ; 3) s l joint betweee nth exterior south-west angle of the transept arm and the west wall of the east range of the cloister (illus 4); and, at the west, a vertical joint at the end of the nave, 1.60 m west of what may have been the original westernmost cloister doorway (illu 43)e onl& Th 5 s.y architectural feature three ar s e doorways f whico , hs i onle yon unaltered broaa d dan , buttres stronf so g projectio south-wese nth t (0.5a ) 6m t cornesoute th f hro transept arm.

81 I 8SOCIET ANTIQUARIEF YO SCOTLANDF SO , 1995

ILLU S5 Interior fac f souteo h aisle wall, looking east

suggestee b n Ica t d thaplae f thith tn o s first building consiste rectangulaa f do r chancel (chapel-less?a , ) transept aislelesn a d an , s nave l withoual , t vaults, although other interpretation possiblee sar . visiblThero n e eear remains of the early chancel, nor any firm evidence of transept chapels. The remains of the rubble core of a wall (maximum extent soutvisible th fooe f hth transepe o tet walth a 2.2 f , o l 1m) t (illus 3), continuin ease f th o t o gt the vertical joint described above, could, however, be evidence of the south wall of a chapel: the west end is a flat surface (maximum length 0.15 m) which suggests the possibility of an interior re-entrant angle between the inner e soutplanth f htransepe eo soute walth th f hd o l hypotheticawalan a t f o l l chapel. Furthermore t cannoi , e b t assumed that the first transept lacked chapels due to the construction of chapels at a later date. A wooden-roofed, aisleless navdivisiony ba forsuggestes e i y th eithef m laco e an n s(i f th k o r y dexteriob r buttresses or interior responds) on the surviving south wall; both its internal and external faces are completely uninterrupted flat mural plane late e scale s th (illurf Th eo rebuildin. s5) t Kilwinninga g also strongly suggests that firse th t churc smals aislelessd hwa an l untit ,bu l further excavation this hypothesis must remain simply that. The architectural detailing is slight in the extreme. There are no base mouldings to the walls, inside or oute masonr e transepTh . th f yo t walls consist f courseo s f variablo s e height constructed with blockf o s variable size, ranging from nearly square to enlongated oblongs; the majority avoid either extreme. The exterior casing of the south nave wall has been mostly robbed out except for a short stretch at the west end. The variable shape of its stones and height of the coursing is comparable to that found in the transept. By contrast casine interioe th sout,e th f th g o f h ro wall (illu , whics5) intacts hi , consist rough-surfacef so d stones of rather irregular squarish oblongs lai waverinn di g course variablf so e width coarsenese Th . masonre th f so y reflectiomaa e yintentiob e th f no plasteo nt interioe walle rth th f .ro Of the three portals, the best preserved is that at the west end of the south nave wall giving access to the cloister (illus 6). The exterior (cloister side) of its, jambs and the semicircular arch of its single order are chamfered (chamfe f abouo r mm)6 7 t ; therimposo n s ei t moulding highee th ; r inner arcjambd han plaine sar . portae Th jambrebatee s i l th doo splayea e d r sar dran fo d toward interiore sth .

820 SOCIETY OF ANTIQUARIES OF SCOTLAND, 1995

ILLUS 8 Interior face of west wall of south arm of transept with blocked portal to cloister

A portal, wit similaha r semicircular arch transept'e t (hth a , ef o nort d sh en wes t wall, formerly also gave access to the cloister (illus 7). The portal, 1.47 m wide internally and 1.22 m externally, was of one order internall externallyd yan lacked an , imposn da t mouldin decorativy an r go e carvings externas It . l jambs (and arch?) were chamfered originally portae lates Th . wa rl partially timblockee grandeth e w t a tha ne d - a t r portal was inserted into the east end of the south nave wall - and altered towards the exterior in order to creat closeea t wit narroha w rectangular opening rebate doora originae r dfo Th . l for mfulls i y visible th n eo interio transepe th f ro blockin e t (illuTh . s8) g- considerabl y thinner tha thicknese n walth fluse s i th - lh f so wit interioe hth r transepe planth f eo t wall creatioe closee Th . th f no t involve nortinsertioe w dth ne h a f no jamb and flat lintel towards the cloister walk. The new jamb and lintel are chamfered, as is the original south jamb. That the new work was chamfered to match the existing jamb seems to be established by the fact that the inserted threshold block abuts the lowest block of the south jamb which is chamfered. The third portal, located on the interior, at the west end of the south wall of the transept arm, probably led to a spiral stair in the south-west angle and, possibly, to an adjoining sacristy (illus 8 & 9).18 It also has been altered to a rectangular form. But the east jamb, which was rebated, and first three courses of its west jamb (as well as the rebuilt courses of that jamb and lintel stone) are chamfered.

Affiliations The type of plan proposed for the first church at Kilwinning is similar to that familiar from the earliest Cistercian churche Englandn si : Waverley (Surrey) 1129/32c , ; Tintern (Monmouthshire), after 1130 Fountaind an ; s (Yorkshire) 1135.c , Cisterciae 1Th 9 n churches certainl transepd yha t chapels - one at Waverley, and two at both Tintern and Fountains - in each arm.20 In Scotland, a similar plaappeared nha d just slightly e Cisterciaprioth o t r n example firse th tt a schurc f ho McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF 821

Holyrood Augustinian a , n foundation establishe y Kindb g Davi c 1128. s bee, dI ha nt 2I 1 reconstructe possessins da gprojectina g straight-ended chancel, transept arms, each witha rectangular east chapelaislelesn a d an , s nave.22 If the first church at Kilwinning was built on a plan similar to that at Holyrood or the early Cistercian churches furthen ca e r on ,speculat e whether mannee th n i ,f contemporar o r y Cistercian churches, solid walls may have separated the transeptal chapels from the chancel.23 That the transept arms wer enave oper theith efo o nt r fulpossibls l wa widt s a y h - alread case t yth ea Tintern - is more likely, for there is no evidence of responds at the junction of the south arm's west wall and the south nave wall (which is actually a chamfered angle), nor of any extension towards east.e th 24 It is unlikely that chancel, transept arms or nave were vaulted; there is no evidence of rib vaults and there is no articulation of the wall in anticipation of them. Nor do barrel or groin vaults seem very likely. The absence of external buttresses to the south wall of the nave could be accounted for by presence th cloistee th f eo r walk nav e vaulteds andth ef wa i , absence th , f internaeo l respondy sma be due to the use of corbels in their place. However, the use of corbels in the original construction would imply (aisle) vaulting. It might also be observed that, in conjunction with the later rebuilding of the nave, old corbels were not reused but, rather, new corbels were inserted into the wall. At the same time it may be noted that buttresses are not always omitted from an aisle wall adjoining a cloister. Buttresses on the exterior of a south aisle wall against which cloister walks were intended and built are found at such major buildings (with 12th-century naves) in as the at Durham and at Ely (Cambridgeshire) and Peterborough Abbey (Cambridgeshire). In Scotland, the west wall of the south arm of the transept and the south aisle wall of the nave of St Andrews (Fife) have very wide external buttresses with angle shafts contrastn I . t Arbroatha , soute th , h aisle wall, whic contaiy hma earliese nth t masonr abbee th n yi church flas i , t both externall internallyd yan s a , soute th s trui f heo aisle wal t Augustiniaa l n Abbey (Roxburghshire). Premonstratensian Abbey (Roxburghshire) also lacks external buttresses to its south aisle wall, despite large pier-like responds on the interior. The evidence is not unequivocal; none the less, it would seem that a wooden-roofed aisleless nave is a distinct possibility. The limitation of the architectural detailing to chamfering provides little evidence for any precise dates for the construction of the first church and is, indeed, deceptive in its simplicity and austerity, because the doorways consequently have an early 'primitive' appearance. In fact, early Romanesque form Scotlandn si , durin firse g th 12te t halth hf o fcentury , would probably have been more decorativ would ean d have include imposn ordero a dd tw san t moulding. Similar unmoulded arches appear at Dunfermline Abbey (Fife) in the nave gallery arches, which may date from about the mid-12th century. It is of considerable interest that Dunfermline's gallery arches succeed the richly moulded arche mais it f nso arcade.25 Similarly t MagnuS t a , s , Kirkwall (Orkney), choie inth r elevation 1140c , gallere th , y consiste f arcadedo s with three unmoulded ordern a d san impost formed by a chamfered stringcourse, but no capitals, following above an ornamented main arcade. progressioe 26Th n from decorate plaio dt austere) r n(o paralleles i nave th en d i elevatio f no (Cumberland) where unsubdivided gallery opening unmouldeo tw f o s d orders succeede arcadn da ordero whicn ei tw e s hth eac h receive dholloa w chamfer.27 Severae th f o l arches appearing in the east range of the cloister, dated c 1170, at Dryburgh, founded 1150/2, by Hug e Morevillehd e similaar , naturen i r t slightlbu , y more 'elaborate'o than tw i , f t o the e yar orders or with a label (illus 10).28 While it seems there can be little doubt that the portals in the south and west wall of the transep t Kilwinnina t earle onlf gi ar y- y becaus theif eo r subsequent alteratio dat a wil s t ne a e(a b l earle seenth yn )i 13th beecenturs ha n t i suggeste y- d thawestere th t n cloisteportae e b th y o t l rma

McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY 823

ILLUS 11 St Mary, Auchindoir: interior face of north portal

much later (illus 6). earln 29A y dat suggestes ei transep o similarits it tw y e db th t portalo y t b d san likelihooe th portaa f do thin i l s position, considerin wese th ge tn i thawalth e f to l on ther s ewa transept. obviouo Thern e ear s insertio e sign th jambe f th so f r archnso o , especiall jambe th s ysa consist inside and out of neatly alternating headers and stretchers. A very similar but smaller scaled (north) portal ruineexiste th t t Mary'sa dS s Churc t Auchindoiha r (Aberdeenshire), which alsjambs oha f header/stretchero s t intse s o wala f rougheo l r irregular masonry (illus II).30 Discontinuitie coursinn i s ease th tt porta e gjama th t f Kilwinnina blo parallelee b y e th gma n di portaparloue th o t lDryburgt ra h (illus 10) additionn I . sligha , t disagreement betwee extradoe nth s stoneandcurvthe the eof s placed ove findit r parallea s Kilwinningat l westhe t in ,the fac of e west transepportathe of l t (illu Thuss7). , there stronseebe mto g argument considesto portathe r l latea originat rno insertiond an l . 824 SOCIETY OF ANTIQUARIES OF SCOTLAND, 1995

One may find a similar contrast between the masonry of the inner and outer surfaces of a wal t Jedburga l remainine th n hi g three original eas te sout bayth f ho s aisle wall e (westh f o t processional portal) wala , l which also lacked either interior respond r exterioo s r buttressesA . more extrem t stilebu l only partial paralle founs i l t Arbroata d h wher soute eth h nave aisle wall, which similarly lacks internal responds or external buttresses, is built of rough and variable masonry belo internan wa l stringcourse, above f morwhico s i e t hi regula 'improvedd an r ' quality. The section of wall distinguished by the coarse masonry ends on the interior (or begins?) 1.01 m west of the west jamb of the eastern processional doorway. On the exterior, the quality of the masonry is less of a contrast with the interior than it is at Kilwinning, but in general it is more random than the masonry in the other standing walls of the east end.31 Arbroath, a sister Tironensian abbey, also offers a parallel for the position and form of the west processional portal at Kilwinning, in a portal at the west end of the south aisle wall. It is, however reverse e s rebatei interiore th ,t th i r n splae do fo , th , y openin t towardgou cloistere sth :

thus, the interior jamb (the arch is not preserved) is chamfered while the exterior is plain. A 2

portal in this position, of course, is not in itself unusual. 3

Tentative conclusions Althoug e meagreneshth materiae th f o s l provide sligha t finabu y st basian lr judgementfo s I , would like to propose tentatively a date for the construction of the first church during the third quarter of the 12th century. modeste Th , indeed, small-scale Cistercian-type plan, already know Scotlann ni t Holyrood)d(a , would be appropriate during this period for a non-royal foundation, a reflection of its modest character. Although the plan is of a type associated with the earliest Cistercian foundations, it is a type that

continue appeao dt Cisterciat ra n abbeys founded between 114 severit1160.e d 0an Th austeritd yan f yo 3

the minimal architectural detailing, retardataire in appearance3 , would fit well into this period, but it has to be admitted that such 'bare-bones' architecture continued int 1170e oth laterr so witness a , cloistee sth r buildings at Dryburgh, unless that date is much too conservative for the east range there. If a church was on the site of the abbey church at Kilwinning before 1184, its fabric certainly is not part of the conspicuous remains visible today - except as tentatively identified above - which, at the earliest (as will be discussed below), date to c 1190 (the new portal to the cloister), or otherwise from well after 1200. If this proposed dating of a 'first' church constructed sometime between 1140/50 and 1160/75 indeed has any validity, it in turn implies that the foundation date was more likely 1157 than 1187 foundere th ; , consequently, would therefore have been Hug Morevillee hd , rather than his son Richard.

EAS E WESD TH TAN T RANGE CLOISTER^E TH F SO "

Description and interpretation ease Th t cloisterange th consistf w eo remaine rno th f slypea so f chapter-housese o th , undercrofe th d an , f o t dormitorye th verticaA . lsouth-wese jointh t a t t angle buttres transepe th f o s t arm wal,e betweeth ld an t ni north of the slype's archway, seems to suggest this range belongs to a separate building phase from the (first) church (illus 4 & 12). Furthermore, the ashlar masonry of the east range is characterized by the use of a conglomerat whicn e (i pebble e hth vere sar y conspicuous churc e wallfoune t th th )no f sn o hdi - earl later yo . slypee Th , formerly barrel-vaulted, finas wasit n li , phase passagewaa , y wit portaha eac n walli ld hen , initially buma t y have bee nclosea d room formin sacristyga library/vestrya r o , . Therbarrel-vaultea s ewa d chamber, presumabl f similayo r dimensions, abov. eit archwae Th y fro cloistee mth r int slypee oth , whic neithes hi r rebate r splaye dno doora r similas dfo i , r McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY 825

ILLUS 12 East range, portal to slype and entrance to chapter house

to the two cloister portals identified as belonging to the first church, but is slightly more elaborate (illus 13). Both the inner and outer jambs are chamfered (the chamfer of about 57 mm), and there is a chamfered label surrounding the arch towards the cloister. On the interior of the slype, at the west end of the north wall, there is evidence of a blocked-up opening passagewar o m2 y1. wide infile th : l doe angle t bonth sno t eda wit wese hth t waltherd cleaa an l s ei r vertical joint 1.2 m east of that (north-west) angle (illus 14 & 15). Along the base of this (north) wall there is the remains of a bench, composed of two courses, one wide, one narrow, which is part of the original construction. After the blocking of the putative portal at the west end of the north wall, the bench was extended unde blockagee rth othen I . r respects barrel-vaultee th , originas it n di t slypl no form s wale i t Th .a l slype th clearlf s ei o ease d insertioth n yen ta n (illus rebated16)s portai e t i Th . n i l , wit internan ha l splad yan rounded jambs toward exterioe sth r (illus 14).35 ease wald portad possiblTh y en tan l ma l y even post-dat perioe enlargemene eth th f do rebuildind an t g churche th f o t completeleasI e .d oth fen t y replace originae dth l east wall f whahavy o ma te bee ncloseda room, which thus could have functione sacristya s da , vestr libraryr yo t thisBu . argumen weakenes i t e th y db fact that evidencthero n s earlieen i a f eo r wall existedd ; ha and t i f i ,, woul portaa t dno l simply have been inserted in it? On the other hand, if it had always been a slype, why was it necessary to insert the east wall with its portal? Other evidence to support the supposition that there may have been a portal connecting the slype (or closed roominterioe transepe th th o t )f o r s founi t d insid south-wese eth t transepte anglth f eo . Therea t a , location corresponding to the obverse of the blocked portal, there is a small vestibule and the lowest steps of newea l stair vestibule Th . coveres e i remaine th smala y db f so l ribbed vault stouo tw : t unmoulded ribs spring from either side of the portal opening into the transept; however, their keystone, plus the southern half of the diagonal ribs, is missing (illus 17). The wall opposite the doorway from the transept, ie the south wall

828 I SOCIETY OF ANTIQUARIES OF SCOTLAND, 1995

ILLU 7 RemainS1 f ribbeso d vaul f vestibulo t staio et r vice

of the vestibule; does not bond in with the east wall, although the latter does bond with the east jamb of the doorway (illus r doe18) soute No .s th h wall bon vic e witn step e di mase th th ehf th f spropeo s o wests it o rt . e soutTh h wall therefore could represen infile th t l (and more passaga f o ) e (perhaps angled diagonally?) through the thickness of the wall leading to the 'slype'. The erection of the south wall certainly necessitated the removal of the two southern ribs of the vestibule vault: only the removal of the masonry forming the south wall would allow for the completion of the ribbed vault.36 e chapteTh r house broaa , d rectangle, projecte beyonm 4 ease 5. d dth e undercrof t th fac f o e d an t dormitory above. No details of its internal elevation or mode of covering are preserved for, although the south wall stands for a good height, its ashlar facing has been robbed out above the bottom four courses. Unlike the slype chapter-house th , e show evidenco sn latef eo r alteration. The facade of the chapter-house comprises the conventional three arches, but the details are rather unconventional 19).(illu & semicircula e 2 31 s7Th r main archwa flankes yi smalley db r side arches with sills four courses higher tha threshole nth centrae th f do l doorway. These side lights each hav singlea e large continuous jambe rolth t la , wit hchamferea d label single .Th e remaining base (north jamb, north light badls )i y weathered, but the profile seems to have been composed of a torus/scotia/torus sequence. Midway through the thickness of walle th tympanua , m witsemicirculao htw r arches survives, althoug centre hth e shafcapitad an t l have vanished in each case. The subarches are unmoulded, and the rear-arches and jambs of these lights are simply chamfered. centrae Th l archwa richls yi y moulde distinguishes i d dan face th t y thad profilb e mouldinge th tth f eo s jambe o th fidenticae ar s l with arce thosth h f eo (illu s 19) exterio.o Thertw e ear innee r e orderon on rd an s separated by the moulded forms of the doorway proper. The outermost order and that to the interior of the chapter-house were both composed of an angle roll with a thin fillet (best seen on the interior order) flanked b pairya f thio n hollow rollsd jamtriplea e san d Th .f shaftbha o t s engaged agains faces it t , whil othee eth r exterior paia orde f detaches o r rwa d shafts (botmissing)w hno . Onl detachee yth d shaft based ha s s (which are now much weathered); they seem to have consisted of a single, rather pronounced torus. The capital was

830 SOCIETY OF ANTIQUARIES OF SCOTLAND, 1995

forme continuoue parth y db f o t s impost moulding tha acrosorderse n th ra t l st al compriseI . dpaia f toro r i separated by a flat-faced band, articulated by narrow fillets. To emphasize the uniqueness of the shaft, an additional narrow torus appeared belo continuoue wth impose th s f loweo t e mouldingon r additionn ;i e th , upper part of the shaft, which was formed from a block bonded into the wall, received a shaft-ring. The undercroft to the south of the chapter-house has been much rebuilt (illus 20): of the early structure there are only the lowest courses of the east wall with the remains of three narrow windows with jambs and sills chamfered externall internan a d yan l splay. windowNone th f eo s preserve openings heae it s th f do . Only three courses of the north jamb and part of the sill block of the southernmost window is preserved; of the middle window, ther onls ei sils yit l block northere th f ;o n window erodee th , d sill bloc thred kan e course botf so jambs hit s remain. A fragmen wese th f t o cloisterangidentifiee e t b th latere y ruine f eo th th ma rf , n s o di possibl y post- , domestic structures whic occupw hno y tha site.e t parth ease f 3Th 8o tt wal f thio l s range abuts churce th h wall just 0.6 wese wes4th m f tcloistee o jamt th f bo r portal (illus 21); l.llm sout f thiho s juncture , e remaintherth s doorwaa ei f so y with chamfere rebated dan d exterior jambs (the chamfef o r 76 mm) and an internal splay, 1.06 m wide, increasing to 1.57 m. The jambs of this portal are preserved for only three course r thre northe southe lates fourta pard fo th th f ewa d an o n tr n t ho s;o i an ,converte d intoa narrower window, the sill of which is at the top of the preserved height of the door jambs. At the juncture wit churce hth h wall, ther ebona seem e f seveb do o st n courses (the stumwale th lf po therefor e stepping nortdowe th ho nt door jamb). Five courses above this point (countin churce th n go h wall), ther thia s eni slab corbelled acros angle th s e (purpose?) above whic ease hth twese walth f t o lrang e once again bondn i s regularly with the church wall for the remainder of the preserved height (five courses). These courses extend ove insertee th r d window around which there a 'mandorlaseem e b o st f disturbanceo ' f o soute th d t hen .A this range, there is the remains of another portal with chamfered exterior jambs, rebate and internal splay, of which only the threshold blocks and one course of the jambs are preserved (illus 22).

ILLU 0 SExterio2 r fac f easeo t wal f easo l t range McALEER: TOWARD ARCHITECTURAN SA L HISTOR KILWINNINF YO G ABBEY 831

'ffHi^'fll'—v^i'fSilSffp^ '

ILLUS 21 Junction of east wall of west range with south wall of church

Affiliations chamferee Th d archwa slypee th f yo , wit chamferes hit d label similas i , severao rt l archwaye th n si east possible rangth f eo y contemporary foundatio f Hugno Morevillee hd : Premonstratensian Dryburgh. The archways to the library/vestry there (of two orders, the outer with plain jambs and a chamfered arch, the inner with chamfered jambs and arch and chamfered label; illus 10), to the slype/parlour (chamfered inner arch with taller plain superordinate arch and chamfered label), and to the day stairs to the dormitory, have all been dated to c 1170.39 difficulIs i t finexaco y t t dan t paralle chapter-house th r fo l e ensemble.40 Indeed Scotlandn i , , no close parallels seem to exist. The general composition of the openings to the chapter-house at Dryburgh is similar to that at Kilwinning: a larger semicircular archway flanked by smaller

openings each subdivided into two by a recessed tympanum. These flanking lights have a plain

1 4 I SOCIET 2 83 ANTIQUARIEF YO SCOTLANDF SO , 1995

ILLU 2 SThreshol2 f portado f wes t souto a l d t hrangen e

outer arch (and inner pointed arches contrasd an ) t wit e centrahth l doorway whic s richlhi y moulded and shafted. However, aside from the semicircular arch, there is no resemblance or duplication of details with the chapter-house ensemble at Kilwinning: there are capitals instead of a continuous impost moulding, and there are bases as well as a discontinuity between the profile of the shafts and arch mouldings; the inner jamb is continuous, although with large rolls separated by a giant dogtooth. Altogether it is therefore quite different from the portal at Kilwinning. The heavier, more conventional, forms at Dryburgh might suggest, if anything, that it is earlier in date; however, that judgement presuppose t onl sno lineay a r evolutio strictea t thae nbu likels on rni o yt have happene Scotlandn di . A large continuous roll is found on the inner order of the south portal of St Mary's Church at Auchindoir (Aberdeenshire) contexe outeth n a n i f , ro t order with monolithic shafts, crocket capitals, a moulded arch with a filleted angle roll, and a dogtooth decorated label.42 It provides, nevertheless paralle a ,aspece side th on e r f Kilwinninlighte o t fo l th f so g ensemble. unusuae Th lshaft-ringe forth f mo coursea pars a ,f o t d block rather than between detached (en delit) shafts finy paralleda ma , jamb outee e th th f n r i lso wes e ordeth f t ro porta f Jedburgho l , possibly from the 1180s.43 At Jedburgh, a shaft-ring was carved in the middle of each block of the coursed shafts. soute lace parallelee th f base b kTh f o h o cloisteportan e e ease th th d sca th o y en t ldt b a r aisle at Furness Abbey (, formerly Lancashire) which also has continuously moulded inner and oute rdiffereno jambtw f (o st heights), without capital r imposso t (illus 23). mouldinge 4Th 4 s McALEER: TOWARDS AN ARCHITECTURAL HISTOR KILWINNINF YO G3 ABBE83 | Y

are composed of an angle roll flanked by hollows, and there is a continuously moulded plinth under both orders, although the outer one stops short of it. Another exampl e absencth f o ef base eo e chapter-houss founi sth t a d e doorway, possibly from the late 13th century, at Augustinian Inchcolm Abbey (Fife), founded in 1123/53.45 However, here agai e othenth r t similardetailno e ar s, even though they includ semicirculaa e r arch, continuous roll mouldings detached an , d shafts. Continuous impost moulding alse sar o lackind gan mouldine th g profile quite sar e different. Less elaborate doorways t possiblbu , y close datn i r Kilwinnino et g than Inchcolm, were found at the destroyed parish church of Lasswade (Midlothian).46 The south portal with unmoulded arcjambd himpostan d sha s forme extensiony db stringcoursee th f so towee Th . rnave arcth eo ht possessed continuously moulded arcjambd han s with simple moulded imposts neithen I . r case were bases evident. The former north (now east) portal at Kirkliston Church (West Lothian) has an inner order with the same profile, a roll set against a hollow, for arch and jamb; there is no evidence of there having been any bases.47 The south portal of St Cuthbert's Church, Monkton (Ayrshire), also had an inner keeled order continuously moulded - jambs and (semicircular) arch interrupted only by a thin impost - in contrast to an outer order with shafts (illus 24).48 The inner order also lacked base mouldings. Non f theseo e churches, however e dateb n d ca ,wit y han precision. A closer stylistic parallel, but one more distantly located, may be found in the lavabo at Mellifont firse th , t Cistercian abbe Irelann yi d (Louth). e semicirculaTh r archelavaboe th f o s , 49 which has been dated to 1200-10, are characterized by the use of the same profile of mouldings for both the jambs and arches, a continuous impost serving as a capital for the two outer and single inner orders, the appearance of both keels and fillets, and a lack of bases to the individual orders. On the other hand, the lack of close parallels among monuments on either side of c 1200, makes it necessary to consider if the chapter-house portal ensemble might not be the product of the late phase of Scottish Gothic when semicircular arches were once again in frequent use, forming a kind of Romanesque revival.50 Of these late portals, perhaps onloffere yon sligha s t similarit chapter-house th o yt e entrt ya Kilwinning soute : thath s hi t porta f Tullibardino l e Church (Perthshire), attribute late th e o d15tt h century (illus 25).51 Others employing a semicircular arch do so either in a context of 13th-century moulding capitad san nort e case th l th formshf e o n inne i s a : r porta t John'S t a l s Churc Perth,n hi 52 wher late eth e dat reveales ei tale th l basey db s with their downward flaring silhouette r whero ; e the details of capitals, bases, and foliage are clearly of a later period, as at the west portal of the towe t Dundea r e Church (Angus). t Tullibardine5A 3 e mouldingth , f arcjamd o s e han bar continuous, interrupted only by an impost; bases are lacking. The impost with its two tori separate unarticulaten a y db d scotia bears onl superficiaya l resemblanc weatheree th o et d formt sa Kilwinning which were once sharply defined , too hollowe So . ,th filleted san d rol f Tullibardino l e are conspicuously shallow compared to the deep hollows and especially vigorous rolls at Kilwinning. Amon late gth e portals, ther I believe , eis evidenco n ,revivae th r f shaft-ringsefo o l , curioue th s i r s basno e desig t Kilwinninna g found later.54

Conclusions The east range of the monastic buildings appears most likely to have been constructed shortly after the first church, as the comparisons for its features seem to suggest a date after 1170/5. The earliest date, based on the similarity of the slype portal with the cloister portals at Dryburgh might be 1170. On the other hand, the closest parallels for the richer architectural detailing of the chapter-

McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY | 835

house arches implie latesa r date durine on ,las e gth t quarte 12te th hf ro century appearance Th . f eo the shaft-ring perhaps indicates a date in the late 1180s. Once again, a foundation date for the abbey earlier than 118 s implied7i , although constructio ease th t f cloistee rangno th f eo n i r particular could be placed after 1187.

THE SECOND CHURCH

PHASE ONE CLOISTEE TH : R PORTAL

Description portae Th soute lth leadinf h churce o cloistee walth ease th d f th o o en glht t a rmark dramatia s c contrast, with respec bot o t scale richneshd th e an decorations it f so earliee th o t , r work, eve chapter-housne thath f to e archway t doe I .t bon sno wit n di wese hth transepe t walth f o l t and, indeed, completely overlap norte th s h jamb of the earlier portal in that wall - which as a consequence was converted into a closet with a rectangular doorway utilizing the original south jamb as part of the narrowed opening (illus 7). Toward cloistere sth portae f threth o , s i el outer ordercontinuousla d san y moulded jamb (illus 26).55 The six missing jamb shafts of the outer orders were monolithic (there is no sign of shaft-rings). The bases (tw missine wese oar th tn jambgo ) consiste toru a scoti d f do an s a followe broada y db , flat x torus.si e 5Th 6 capitals encompasse chalica f o o e dtw shape possible on , y figurated wite on ,h water-leaf wito tw h d stiffan , - leafthreThe . e orders comprise (thdone e outer) finely moulded, includin appearancgthe arrithe an son e of angle-roll (the on e, middle) wit thiha n roll bande shafts-ringy db designs) o (the tw on ef innerd (o s an , ) which formerl deepla d yha y undercut chevro r lozengno e pattern meeting ove thia r n rollcontinuoue Th . s inner jamb utilized three thin rolls (the angl keelede eon ) flanke hollowa y db . Finally labe e decorates th , wa l y db dogtooth of a modest size. interiore Onth jambe th , s- whic h were rebate slightld dan y tall e splayeth , f unmouldeo d- d rear-arch continue upwards three (west) or four (east) courses above the springing of the exterior arch. The level of the springing of the exterior arch is marked internally by a very narrow course. The broad rear-arch itself is only slighly pointed (illus 27)seae Th m. wit earliee hth representee b r waly ma lverticaa y db l joint whics hi packed with small stonejame th ease bf th s o t (illu 0.9o t 8m s 28)significance Th . f thio e s vertical seams i rendered ambiguous, however, becaus masonre eth y t obviousleasi f o t y represent patchinsa walle e th du f go to a (later) respond having been ripped out. One might have expected the seam to abut the west side of the vanished respond but, rather, it is located farther west.57 A seam to the west of the portal is not evident: the jamheighe th bf o tstone havy sma e been determine heighcoursee e existine th th th f y o db tn i s g wall (the t constructejambno e ar s d wit regulaha r alternatio f headerno stretchers)d an s . Abov voussoire eth e th f o s rear-arch, the upper (five) courses of the wall are the result of restoration, the masonry here characterized by distinctiva e vermiculatio surfacese th f no .

Affiliations Amon mane gth y decorate beginnine th d 12te d 13te portalth hth an f h o f go centuryd s en fro e mth , there are again no very precise parallels. These portals would include: at Kelso, the north and west portals t Dryburgha ; cloistee th , chapter-housd an r e portals t Jedburgha ; cloistee th , wesd an r t portals; at St Magnus Cathedral, Kirkwall (Orkney), the west portals; at Paisley Abbey (Renfrewshire), the east and west portals of the cloister, the north porch, and the west portal; at Elgin Cathedral (Moray), the west portal; and at Inchcolm, the chapter-house portal once again. Others which should als mentionee ob e preservedar parise th t da h churche f Lamingtoo s n (Lanarkshire Kirklistod )an n (West Lothian). Wit water-holdine hth f regarprofile basese o th t th s notabl i f o ye dt t o e i , t gno e s thai t i t

McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY I 837

ILLUS 27 Interior of east processional portal to cloister

type, becaus t becom scotie no eth s aha e horizontal ,emphasi e although th broade o th t e n o s, du , rather flattish lower torus, it is moving in that direction. They are somewhat similar to the bases of the portal at the east end of the south aisle at Dryburgh (late 1190s?) or to those of the former north portal (now locate easte th t Kirklisto t )a d a perhapf no s1200.c 58 Water-leaf type foliage appears in Scotland throughout the west end of Kelso, in conjunction with multi-scallop type capitals, but Kelso is not precisely dated.59 It also appears early on in the work at St Andrews Cathedral which was started 1160/2.60 Water-leaf also appeared at Jedburgh

(along with crocket) in the 1180s. The two right-hand capitals (corresponding to the outer orders) 1

formee oth f r north porta t Kirklistoa l alse nar o water-leaf6 ; thercontexe th en i f fouthe o te ryar other capitals which see havmo t ecrockee beeth f no t variety t preciselno latt a , ebu y datable appearanc e typeth r ,e suggestefo eeveth f i n d dat f shortlo e y o beforto e eb y 120ma 0 conservative.62 singlee Th water-leaf capita carves i l d fro sam e stiff-leao mth etw bloce th f s kexamplesa l al , together formin capitale portal'e gth th f o s s east jamb.63 Stiff-lea primitiva f o f e form appearn si

the early work of the crypt of Glasgow Cathedral, before 1199, and what may be described as 4

proto-stiff-leaf in some capitals of the north aisle arcading at Holyrood6 , perhaps c 1195. But 5

alread ydevelopea d 'wind-blown 'capitale forease th appeared th f m tha o f t leas o sa e n on tdo 6 blind arcade of the presbytery of Arbroath, sometime after 1178, certainly before 1200.66 About same th e time, what coul consideree db earln da y for f stiff-leamo capitals e usee s th r th wa dff fo o west porta f Jedburgh.o l e 1190s67Th , therefore decade th , seee whicb n e i mo t h stiff-leaf replaced

water-leaf for capital decoration in Scotland. Once established in Scotland, stiff-leaf had a long

8 6 8 83 SOCIET ANTIQUARIEF YO SCOTLANDF SO , 1995

ILLUS 28 East end of south wall of church (east of east processional portal)

life. A lustily rich form occurred on the six capitals of the inner portal of the north porch at Paisley,69 whil largese eth t displays appea nave th en i rarcad t Holyrooea d Abbey 1230.c , 70 Fully formed stiff-leaf also appeared in the somewhat later eastern chapels of the Glasgow crypt, after 1233, choie thed th 71n an 1242.rc i 72 Due to the extensive damage, the exact form of the chevron or lozenge motif is difficult to reconstruct diagonae Th . chevroe l barth f so n spannin holloe gth w betwee outee nth e r th roll d san middle one are widely spaced and seem to have faced in opposite directions, so a large scale lozenge spanning two blocks is the most likely result. This may be contrasted with the outer order wese oth f t porta t Brechia l n Cathedral (Angus),73 also badly damaged, wher diagonale eth a s- pair to each voussoir - all pointed upwards so that a series of arrow-like motifs would have resulted, or with the middle order of the portal at Lamington Church,74 where the 'arrows' point McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY 839

away fro e e centrarc mth th e s outethe a hf th o eyn ri alst Mary'sordeo S d o f o y r El , (Cambridgeshire). Similar motifs more usually occu smallea n o r portale rth scal n t i Kelso sa s ea , Jedburgh Arbroathd an seriea : f interlockinso g roll lozenges ove rola r l occuroutermose th n o s t major wese ordeth f t o rporta t Kelsoa l similaa ; r interlocking lozeng soute founs ei th hn d o jamb , betwee detachee nth d shaftswese th f t o porta, f Jedburgho l similaa d an , r form again appearn so

outee th r order, whil variantea , involving only arrow-like chevrons middl e founs i ,th n do e order. 5

t anotheYe r variation occuroutee th n r so archivol wese th tf o tporta t Arbroata l h wherei thina n 7 angle roll is abutted from either side by a series of zig-zags executed as a roll, thereby producing the effect of bisected lozenges. A very similar motif is also found in two orders of the south portal at Kirklisto contexe th semicirculaa n i f o t r arcothed han r orders with multiple roll hollowd san s

which are more Early Gothic than Late Romanesque in feeling. The south-west fa?ade portal at 6

Kirkwall, c 1220?, also displays a large roll-profiled chevro7 n of two different patterns, one

formin seriega hexagonsf so othee th , r 'arrows' simila Brechin.o rt 7

middle e motith Th f o fe order thia , n roll decorated with disk-lik7 e band e r ringso son s i , which, in varied form, occurs on a number of the same portals. It is found at the Kelso west portal, miniature in scale, on two of the (innermost) minor orders. A ringed roll also was found on the

north(?) aisle portal at Jedburgh. The outer order of the west portal of Cistercian Holm Cultram 8

(Cumberland anotheis ) r example, where7 , however for not muc m ringas thicthe ,do hof sare kbut

a contrast in diameter with the roll as at Kilwinning and Kelso. They are also present on one 9

wese ordeth f t o rporta t Arbroatha l outee th d rporta e ordean ,th t Lamingtof a lo r n Churche On . 7

voussoi arcn a hf o rremain t Paisleysa , wit largha e roll bande rola y ld b flanke nutmeg.y db f O 0

these buildings, Kels undoubtedls oi earliese yth t (1160 r 1170sso Arbroatd an ) latese hth 8 t (surely, at the earliest, the early 13th century).81 The outer order, consisting of thin rolls and hollows, (several of the rolls with a distinct fillet), can again be paralleled in a general way by many of the same portals. Finely moulded order e archivoltfoune th ar snorte n o dth f ho s porta t Kelsoa l innee th , r Jedburge ordeth f o r h

south (cloister) portal, and the inner orders of the west portal at the same building. To this list 2

ma addee yb d oute third ran d ordersoute th f hso porta8 Kirklistont a l . Dogtooth, however, is found on few of the above portals. It does, however, occur on the interior and exterior labels of the west portal at Jedburgh, and again on the exterior framing the

portatympanum e t Marylabee th S th f f o lo n ,o Auchindoir.d an , gianA t dogtooth also occurt sa 3

Dryburgh, where it 8 was used on both the inner jamb of the chapter-house portal and the south aisle cloister portal, thus spannin perioe gth d fro 11701190se e mth th otheo o st N . r decorative motift sa Dryburgh overlap with Kilwinning smallea n O . r scale dogtoote th , h motife alsuses th n owa do interio exteriod an r r windowe label th norte f th o s f ho s transept' exterioe s th chapels n o rd an , jambs of the north lancets, dating from the 1220s. Dogtooth also appears at Paisley on the jamb betwee south-wese shafte nth th f o s t cloister portal similaa n i , r positio norte th hn n o portal n o d an , outee th r wese ordeth f t ro porta l (an flankins dit g blind bays). archivolte 84Th l (foural f so ) portals at Paisle mouldede yar , wit otheo hn r decorative motifs appearing additionn I . , dogtoot s alshi o found as a major motif on the Kirkwall north and south aisle west portals, c 1215-20, in the crypt of Glasgow (after 1233 Elgi1242)y e b ;th n o , wes rear-archee t th portal, n chapteo e d th 8f 5an so r house vestibule at St Andrews (mid-13th century).86 Uniquely, a similar combination of a ringed roll, finely moulded orders, and a lozenge-like patter wes e founs ni th tn portado t Arbroata l h which, however circulas ha , r abaci, splayed rather than stepped jambs, and a semicircular arch. The last three features, as suggested above, probably indicate a date in the early 13th century.87 Formally cloistee th , r porta t Kilwinnina l g seem t intfi oseriea o st f portalso s characterized 840 SOCIET ANTIQUARIEF YO SCOTLANDF SO , 1995

by a distinction between the moulded jamb of the actual opening and its surrounding orders which, in contrast, have detached shafts simpleste Th t .necessaril no f i , earliese yth f thio t s formate th s i , present east porta Kirklistot a l n ('shortly before inne1200'?)n a s r jamha t I . b orde notes f a o r - d - continuous profile interrupted by the capital, and without a base, in contrast with the two outer orders which have plain unmoulded archivolt monolithid san c shafts with torus/scotia/torus base profiles. Although the jamb capitals form a continuous band with the outer orders, they are of a different design e easterTh . n portal fro soute mth hcloistee aislth o et t Paisle a r s basicallyi y

similar, although less austere than that at Kirkliston. The mouldings of the inner jamb and arch 8

have the same profile consisting of rather coarse shaft8 s or rolls and hollows. There is but a single outer order foliag e capitale Th th . f eo s form mora s lesr eo s continuous frieze. Instea shafte th f dso at Paisley norte th , h transept north porta 1165-75c ( l t Furnes)a hollows sinneha e th t rsa jamd ban arch, interrupted only by an abacus and, of course, is without bases; there are rather squat water- leaf capital grandes it o st r displa f threyo e inneorderse th middle n r th jamO . d eban orderse th ,

emphasis is on the hollows at the expense of the rolls. The west portal (c 1175) of Cistercian 9

Calder Abbey (Cumberland) 'advances' beyond Furness'8 s north transept eliminatioe portath y b l n abacue oth f imposs innee sa th r r jamfo t b whic modestls hi y accente solitara y db y angle roll.t 90A Holm Cultram, the west portal (ell 80) of the nave follows a similar format to Furness. It has no capitals at the inner jamb (which is like that at Paisley, but with thinner shafts), the abacus only continuing from abov water-leae eth f capital foue th rf senframino g orders whic more har e finely moulded south-wese Th . t porta 1185c ( l t Cistercia)a n Jervaulx (Yorkshire), more modes scaln i t e order(reflecteo tw s s it rathe n di r than three) richet bu , detailn ri , also include contrassa t between e dooth r openins enframinit d an g g orders throug e former'th h s continuous mouldings, conspicuously baseless, and the latter's moulded bell capitals, detached shafts, and bases raised on plinths.91 (Dogtooth of medium size is found on the jamb between the shafts and again, smaller, on the inner order.) Owing to the elimination of the impost/abacus, as well as of a capital for the inner jamb, and the presence of stiff-leaf capitals and dogtooth, Kilwinning seems the latest of this series.92 This format necessarila doe t seee b sno mo t y Cistercian typet contrasti s a , s wit wese hth t portal at Fountains Abbey (Yorkshire), c 1160, where there are major and minor orders (the major ones detached shafts) with tall capitals forming a continuous frieze linking them with the door jamb which is slightly set apart by its greater width and the complex of mouldings on its return lattee faceth rn i ,respec somo t t e extent anticipatin latee gth r developmen innee th f r o tjam a s ba

distinctly separate element. numerouse Th variedd an , , fine hollow archivole rolld th f an sso t 3

mouldings, derived from easter9 n French sources, do, however, probably account for similar sets of moulding laten si r doorways. Portals at Arbroath (west), Brechin (west), Arbroath (north), Paisley (south, north and west), Holyrood (west), perhaps climaxin Elgit ga n (west), are contrasn i , forma e th o t t described above, characterized by more widely splayed jambs with - except at Arbroath - minor orders between the major, and with curved abaci, so the main jamb reads continuously with the sequence of the outer orders thesf o l e Al . portal datee b n d sca betwee mid-13te th earle d nth yan h century. Typologically, the portals to the chapter-house and nave south aisle at Dryburgh (c 1170 and c 1195) seem midway between thes formatso etw distinctioA . s maintaineni d betwee doore nth - opening surroundin e propeth d an r g orderchapter-house th n si e porta virtuy continuous b lit f eo s mouldings whic e furthehar r accente large th ey db dogtoot h between e soutthemth hn I . aisle portal, capital abacd san i have been reinstituted, althoug shafte hth coursee ar s distinction di o nt the detached monolithic shafts of the three outer orders (as also in the chapter house portal). In both cases, however, the jambs are splayed despite the stepped profile of base plinths and abaci. McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY | 841

The archaic flavousemicirculae th f o r rstrongl e archeth d san y stepped arch order s- wherei e nth flat planes (although moulded) at right angles of the voussoirs are emphasized - is belied by the lush stiff-leaf foliag wese th tf ejambo processionae th f so l doorwa contrasn y(i crockeo t t capitals e easonth t side) e soutTh . h porta t Kirklistoa l n alss elementoha n a f bot o s hha t typei r fo s extended inner jamb with slender (baseless?) angle shaft and three orders rising from five monolithic shaft major/minoa n i s r arrangement place diagonala n innee do th s r A .orde s i r decorated like the others, and the capitals form a continuous band, the portal is more conventional forman i t than Dryburgh's.

Conclusions t firs a impossibls a y t cloistee seew b ma t ne t mdati o t e no s rea th y portama t I l with precision, tha , withiis t rangna f leseo s thadecadeso n tw generae portaa th s f A . o l l 'Transitional' typet i , could be (and has been) conventionally broadly dated c 1180. Comparison with numerous other portals which have similar features, those also only approximately dated, suggest possibla s e construction date betwee n120118c c d slightlr 0o an y later appearanc e stiff-leao Th . tw e th f f eo capitals, however, makes it most unlikely that the portal is any earlier than the 1190s. Although the chevron motif has a life into the 1220s (Kirkwall), the appearance of water-leaf does not seem to have extended beyon 1190se dth . Even suc datingha , however, would see mleno t d suppore th o t t argument advanced earlier that the abbey was founded well before 1187, which date indeed might b terminusea post portal e portaquerne th th r f I fo .l date 1190-1200sc , the fabrie nth c into whict hi insertes i d mus oldee tb somy rb e decade orden si alloo rt e wconstructioth e f timth o r e efo us d nan church, and then the need, or a feeling of the need, for changes to develop. One could also entertain the possibility that the doorway was the result of a gift of Richard de Moreville, either made shortly befor s deatehi 118n hi r stipulate9o s willhi .n d i Thihav y sma e bee origie nth f no the later confusion about the founder of the abbey.

PHASE TWO REBUILE :TH T EASD TEN

Description and interpretation constructioe Th portacloistere w th ne o t la f ,n o givin navee ghavy th accesf ma ,eo t fro i ease d o mst th en t bee necessarna y preliminar remodelline th o yt , rather gor muca rebuildine n th o ,h eas e d largeth en t f go r scale. The new rather grand portal allowed the closure of the smaller unpretentious original portal at the north end of the west wall of the transept's south arm, and could have facilitated access to the monks' choir when it was removed to the nave while rebuilding proceeded. The reconstruction of the east end, possibly on an enlarged scale, produced a plan with an east aisle to the transept's arms, providing two bays in each that could function as chapels, and an aisleless chancel. e verticaTh l seam betwee enlargee nth originae d ease th form th d transepe t n an th sidlo s f m eo ar t structur reae b dn moseca t clearlexteriore th n yo , wher broae eth d buttress immediatele th f o d en ye easth f o t slype marks the beginning of the new work (illus 14 & 16). Its plinth design, of a roll one course above a sloping setback (which consist course)e on f so , continue course son e lower alon survivine gth g south facf eo the east aisle setbace th , k expande courseo tw o dt s ove narroa r w chamfered course. Ther a basa r s ei efo detached nook-shaf narroe th n wo t eas buttresse t facth t seemi f et o bu , shafs a neves wa t r place position di n for there is no sign of a capital. Inside, it will be recalled that the eastern limit of the gable wall of the first church has been identified with a simple vertical joint 4.82 m from the west wall of the transept (illus 3 & 29). It must be admitted that this juncture is problematic, for the masonry east of this seam - between it and the respond of the south arch easappearance w th t ne s aisle oha th f e - bein f eo g older than that otherwise identified wit firse hth t churcn hi 842 I SOCIETY OF ANTIQUARIES OF SCOTLAND, 1995

ILLU 9 SLowe2 f interioro easd en t r south wal f transepo l t

the transept. This area of masonry, eight courses high (to the bottom of the three later memorial plaques whic hid w sectioe wale heno th th lf nuppeo e equa th o respond)e rt l th hal t itself o fno s fi uniforme forth t ,a east end, belo eastere wth n plaque more th , lesr eo s regular coursing replace s i stop d san ashlay d b wida f ro e range of sizes and shapes (illus 30). In its lowest seven courses, the respond does not course in with the transept wall, but just enough is visible between it and the eastern memorial plaque to reveal that the upper eight courses do. horizontae Th worlw junctur transepne d k e ol wit th e f hth eves eo i t n less clea probabls ri (illut I . s3) y markeremnane th stringcoursa y db f o t soute th hn eo interna l face, 12 full courses abov e existineth g dirt wese th tn internao floor d an ; l face, abou same th t e leve - ljus t abovblockee arcth e f eth h o d west portao t l the cloister formee Th . r stringcourse die towardt sou easte sth : therevidenco n e s ei th f ebees of that ha nt cu i t narrow course of blocks which continue its line to the east (before it disappears behind the memorial plaques). The masonry on the south wall courses evenly across the entire width between the level of this string and the sill of the three lancets above (a full 19 courses). However, it does not course across to the west face between thi wese s th stringcours tn walo e l on wesundee d th sil e ef tan o rth llancet , eight courses higher; the masonry abov highee eth r west string relate (newe th o st ) windo whicf o w - h onl soute yth h jamb remains - in the west wall, the sill of which was lower than that of the south lancets (illus 31). Somewhat confusingly, the masonry does not course regularly across the south-west angle between the two strings which mark the levelwindoe th f so w sills sout west d soute h an additionth n i :f h o wes e walld th fiv e t en t a ,th ,e upper courses suddenly become seven. On the exterior face, the line between the new and old is clearest on the west face, due to a change in the shape of the stones, from squarish oblongs to a greater use of elongated oblongs (illus 4). The horizontal junctio clearle soutw th no hn t yn o fac no evident s i e t seemI . s most likel (gabled yen thae )th t wals wa l completely rebuilt at least above the level of the upper barrel-vaulted chamber (illus 32). If this, indeed, was the case, it might explain why its plane diverges (to the north-west) with respect to the walls of either barrel-

844 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1995

vaulted chamber vaulte th : s would have prevente divergence dth e betwee planee nth s from being apparent during construction. As to the new south-east aisle, even though only part of the southern of the two bays remains, it is enoug reconstruco ht desige th t n (illus 33)piee rTh . betwee chapelse nth , which doe t lin witp sno e u hth nave south wall, is unexpectedly complex, as it was originally composed of 12 shafts of which two were detached (the axial one on each of the east and west faces) and of which four - the diagonal pairs - were keeled (illus 34). The axial shaft, north and south, also keeled, was larger and a shaft of the same size is found soute onth h respond responde Th . , however t otherwisno , e shafted s formei , broaa y db d buttress-like form with widely chamfered angles (illus 30)formee Th . r pie responr ro d betwee norte nth h chance e chapeth d an ll (wall) may be indicated by the remains of one round shaft, thinner than those of the pier, built into the south wall of the 18th-century parish church - if it is in situ: this half-shaft lines up with the minor keeled shaft of aisle th e pier, rathe (latere linth e rf th e o tha) axian e navi nt th i les i one arcadr no ; e (illus 35). The pier and the remaining arch of the east aisle of the transept have a number of distinctive features which emphatically separate them both stylistically and chronologically from the cloister portal. The bases consist of a narrow roll and hollow above a flaring ovoid foot. The capitals (illlus 36), of the general moulded bell type, are notable for their compression: a roll between two astragals form their upper edge succeeded by an abacus consisting of a hollow flaring out to a fillet - all elements following the section of the supports below innee clearlt Th order.e r arcno e ordeth e th yh - sar f definecomposeo rs wa tripled- a f do largf o t e rolls e middleth , , larges e wit on thfilleta e otherTh . , finer mouldings, consistin f rollghollowso d an s , included a line of dogtooth facing the transept space. The label was flanked by rows of nutmeg and had foliage stops.94 Of the upper elevation little remains. A wall shaft of thin triplets, the middle one with a fillet, rose fro impose mth t abov detachee eth d piere secone shafth Th .f o td storey seem havo st e bee talna l gallery-like arcade of which the three bases (on curved plinths) for shafts at its south jamb remain. The fragmentary south spandre gallere th f o ly opening retains portion blinda f so , rather plate-like, encircled quatrefoil (illus 37)t I . mouldea s ha d frame circulad an , r knobs marke junctioe dth e lobes th f no . Immediately above th s ei chamfered south southernmose jamth f bo t clerestory v/indow. These remains suggest that the second-storey openings consisted of an arch of two orders, with detached shafts, subdivided by a central shaft. Above, the clerestory may have consisted of a triplet of narrow opening f equas(o l height?) with simple chamfered jambs possibilite Th . f threyo e opening clerestore th n si y s suggestei face th t y thadb survivine th t g south jam locates bi d ove outee th r r gallere ordeth f o ry opening, indicating that the clerestory composition was wider than the arch below. The aisle was rib-vaulted (illus 38). In the remaining south bay, the diagonal rib was formed by a triplet of largee rollth s- r middle roll wit filletha . Both the diagona wald an ll ribs sprang from small bell-shaped corbel-capitals. The inner jamb of the aisle's south window is plain, except for a simple label. The jambs of the three south lancets of the gable wall were moulded - a thin continuous roll - with an orde f detacheo r d shafts witseto f shaft-ringhtw o s moulded an s d capitals formin outen a g r surroundA . passageway ran through the wall at their sill level. Above the lancets, at the base of the gable, there was a small oculus, with a thin roll as a frame, flanked by pairs of short lancets with chamfered jambs: a label skipped continuously over all the arches. A second passageway also ran through the wall at this level, connecting wit e south-eas vicha th t ea t angle staiA ( . r vic eroof-space begalevee th th f t o lna e ovee th r chapel vaults lowes it : r exi completels i t y invisibl lateo t e r erestorativdu e conservation uppee th ; e r exith o t t gable passagewa s stilyi l visible. transept'e lancee th Th ) n i t s west wall f whico , he sout th par f ho t jamb remains detachea d ha , d jamb-shaft wit rola hrina d l mouldegan d jamb toward interiore sth . exterioOe transeptne th th f o r ,heav a ther s yewa solid buttres south-ease th t a s t corneaislee th f .o r The south lancet of the chapel bay had a chamfered inner jamb with an outer order formed by detached shafts, moulded bell capital mouldea d an s abutmene dlevee th arch formeth e f t o th l A .f o tr aisle roof againse th t eas transeptte walth f o l angln a , e turret with coursed angle shafts began. (Althoug south-ease hth t cornes rha been rebuilt, the weathered base for a shaft remains.) In addition, there is also a large capital with sword-leaf foliage corbelled into the wall at this point: its purpose is not at all clear, especially as it supports the springing of an arch. e soutTh h lancet gable th f eso were more austerexterioe th n eo r than the yinteriore werth n eo , their

McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY 847

<»tt ILLUS 36 Detail of south respond of east aisle of south arm of transept

jambs being simply chamfered; a thin continuous label linked them. The level of their outer sills stepped up to the west, apparently in order to avoid the pitch of the roof over the adjacent east range of the cloister, yet the raggle of a steeply pitched roof crosses the sill of the middle lancet to reach an apex between the central and west lancets (illus 32).95 The eave of the roof would have formed a broad overhang above the east end of (newle th y created?) slype. (Thersigno n f beae o s ear m ends. therd An )e must have been some kina f do sloping setback where the new east wall of the slype was reduced in thickness at the end of the upper chamber.96 Abov oculus e lancete th e th s ,swa slightly vesica-shape contrasn d(i interioe th o t r circle), witha heavily moulded frame, and above it a single plain lancet. (The vesica is off-axis towards the east, both with respect to the middle lancet below it and the single lancet above.) e staiTh r e south-westurreth t a t t cornetransepe th f s beeo r ha t n much damage repairedd dan e th , repairs obscuring some of the original connections to the dormitory range. It could have been at the time of rebuildine th eas e thad passage th en tth tf go e betwee putative nth e former sacrist r library/vestryy(o e th d an ) transep blockes wa t , possibldup orden yi construco t r existine th t g newel stairlowese Th . t surviving steps res solin o t d masonr evidenc e y- whicb y htheif ma eo r insertion into earlier walls largA . e roughly patched area on the south most probably has replaced the original and/or later entrance to the upper barrel-vaulted chamber. From abou barre e levee th th tf lo l vault ove uppee rth r chamber insidturree e th ,th f eto (now stair-less) fivd certainls i eha buile exitd on dan f yso (illus 39) firste :th , angle south-easte th blockeo dt w no s i ,d (and invisibl exterioe lateo th t n e ero patchinrdu refacing)d gan secone th ; d faces north thire th , d east fourte th , h fifte northth h d oncan , e again eastlowese Th . t (blocked south-ease ) exith o t t presumably gave accese th o st level over the upper barrel-vaulted chamber and probably allowed the vice to serve as a night stair from the dormitory level over the chapter-house and undercroft. The four upper exits establish that there were two

McALEER: TOWARD ARCHITECTURAN SA L HISTOR KILWINNINF YO G ABBEY 849

level f superimposeso d passageways across bot soute weshd th han t transepte wall th soute f th o s hn I . wall, they were at the sill of the lancets, probably ascending to the roof space over the aisle vaults, and at clerestory level, where no doubt they connected with the east vice; at the latter level, the passageway separated the inner oculus from the outer vesica. In the west wall, the levels would be the equivalents to the gallery and the clerestory.

Affiliations enlargee Th d plan t hav, whicno ey beehma n very different typologically kina fro s olde dmth wa , of reduced versio t Andrew f thaS n o f o t s (begun 1160 r Arbroat)o h (founded 1178), bot f whicho h had projecting presbyteries and aisled choirs. At St Andrews, solid walls separated the chapels at the end of the aisles from the presbytery, and again those of the transept arms from the west bay of the choir aisles t ArbroathA . , whicshortea d hha r choir (two bays t Andrews S fouo t ,t a r ) there wer solio en d walls evidence Th .t clea no t Kilwinning ra s e i solit bu , d walls between chanced an l chapels shoul consideree db possibilityda . This would mea nplaa n simila thao t rt Dundrenna a t n Abbey (Kirkcudbrightshire . 1142fd ; ; existing building begu n1160/remodellinc g1180)c , which,

however, had three transept chapels, or almost exactly like Abbey (Wigtown; fd. 1192), 7

closer in date. The same plan was use9 d again, almost a century later, at the last of the Cistercian 8

houses in Scotland9 , that of (Kirkcudbrightshire), founded in 1273." The new work in the south aisle is stylistically distinct from the cloister portal with regard to every detail of the latter, and since it formed a complete elevation there are features for which there wer earlieo en r parallels. Twelve-shaft piers are found in Scotland before 1200 at Dundrennan, in its south transept and nave arcades (the westernmost pier of the south arcade), where they were all coursed, and the

major (axial) shafts keeled. Twelve-shaft piers of a somewhat different design were found in the 0

early 13th-century nave at Holyrood10 . Placed on a lozenge-shaped base, rather than a round one as at Dundrennan, their core was stepped and the shafts were neither keeled nor filleted.im Twelve- shaft piers relating to Holyrood appeared sometime after 1230 at Valliscaulian Pluscarden (Moray)

in the south arm of its transept. The canonical type occurs again, much later, in the nave at 2

Sweetheart, where the major axia10 l shafts bear a wide fillet; although exceeded by the 16-shaft piers at Exeter Cathedral (Devon) 1280-1340c , latese ,th thef to exampleye mustype.on e e th b tf so 103 e SweethearTh t pier l lac peculiae al s kth r featur f thoso e paia f Kilwinning eo f f o ro e us e th : detache delif)n d(e shafts.104 Outside Scotland, coursed 12-shaft pier appeared sha d earliee th n i r nav t Augustiniaea n (Yorkshire), nort e chevee th th h d 10 n i 5transepan t t a t Cistercian Byland (Yorkshire; soute 1170s),th hn i transepd 106an fivd ean t easter nnave bayth ef so

at Benedictine (Lincolnshire). In the south, 12-shaft cluster piers appeared in the 7

choir of St Albans Abbey (Hertfordshire).108 10 Responds that do not correspond in their design to the piers forming an east arcade in a transepfoune ar t Dryburghda m ar t , wher norte souted th han h transept responds differ from each othe s wela r s fro a lpiere mth , mos f tripleto t e obviousl f thius o s e n th shaft y yb s facin largga e half-shaft on the pier. At Glenluce (after 1191/2), the south transept respond, in the form of a double-layered buttress with an axial half-shaft, addresses an eight-shaft cluster pier. Also, the soute th f ho transepd responen e th t t aislda t Arbroatea h (afte rfore 1200?th a f halm o f n i o f s i ) large polygonal pier with four thin shaft against se s t alternate facets chapee th ; l composepies rwa d of eight keeled shafts.109 The triple wall shafts rising from the pier abacus relates Kilwinning to some of these same buildings, Bridlington, Bardney,110 and Byland,111 as well as perhaps more significantly to the nave

elevation of Holyrood (1220s and 1230s?) and the choir of Glasgow (early 1240s onwards). In

2 11 850 SOCIET ANTIQUARIEF YO SCOTLANDF SO , 1995

the last two, three slim wall shafts rise from the impost of the piers, relating visually to the axial shaft of the pier below.113 The arrangement at Kilwinning seems closest to these two where even their axial shafts have a narrow fillet, as also at Kilwinning. As a motif, encircled quatrefoils appeared in Scotland as early as the time of the construction uppee oth f r stagewese th tf so crossin g towe t Kelso,a r wese lats a thath s ea f d t 11 o tporta 4an t a l Elgin, c 1250.115 Large encircled sunk blind quatrefoils appear at Dryburgh, c 1225, at the juncture between the presbytery and choir: there is a rolled frame, but the ends of the lobes curl back rather than terminatin ball-likn gi e knobs. Foiled shapes are, however, found most frequently decorating tympana rather than spandrels. They appear variously unencircle r framed o mouldinga y b d , circula r vesica-shapedo r s a ; completely piercin wale gth l plane, especiall case f th tympana eo n yi r suno , k t int a wale s oth (a l Kilwinning) s reliea r o f , e walformth ln planeo s s appropriati s a , r walefo l surfaces between arches. To the south, pierced tympana with varying designs are found in the choir and transept at Ripon Minster (Yorkshire) e 1170th n d e i 1180schoi,s th an f Hexhao n ri ; m Abbey () c 1180-1200?, ; throughou f Lincolo l al t n Cathedral c 1192-, c e 1240th n i ; spandrels of the subarcade of the choir triforium of (Yorkshire); and in the north transept of , c 1230-50. In other locations they are found in the Lady Chapel of Hereford Cathedral, c 1220, and on the west front of Binham Abbey (Norfolk), c 1240. Small unencircled quatrefoils pierce the tympana of the second stage arches in the Ripon choi transept.d an r gallere 11Th 6 y level archechoie th f Hexhan o rsi m displa yvarieta f pierceyo d forms in the tympana of the south side; in the north elevation, the tympana piercings are all in the

form of quatrefoils, although subject to some variation. The tympana of the north transept arm 7

e alsar o pierced (four different shape usede ar s : trefoil, quatrefoil11 , encircled cinquefoild an , vesica); the tympana of the south arm are not pierced. Pierced tympana become a major theme at Lincoln,118 carried through from one phase of building to the next, generally in the form of trefoils and quatrefoils, beginning wit minoe hth r east transep choid an t r (after 1192-1210)s . Theiwa e rus continued in the major east transept, c 1210-20, which probably influenced the use of encircled quatrelobes and cinquelobes in the gallery tympana of the great transept of York, c 1220-50.119 At Beverley, the quatrefoil was favoured, sunk into the spanadrels. In the Hereford Lady Chapel, encircled quatrefoils, circula vesica-shapedd an r , help wale filth l l betwee windoe nth w arched san the vault. The nave of Lincoln, finished by 1240 at the latest, remained faithful to the unencircled form, but with a more liberal use of punched forms. The motif is used in groups of three (generally quatrefoils) in the tympana above the pairs of openings to the false-gallery. Deeply sunk pointed trefoils and both round and vesica-shaped encircled cinquefoils (the cusps not plate-like, but of ) tympane fil spandrele th l th arcadine d th aan f so g flankin wese gth t porta f Binhao l a mn i vigorous display. In the presbytery of , c 1240, unencircled quatrefoils are punched into both the tympana and spandrels of the gallery arcade, and, two decades later, unencircled trefoils appear in relief on the spandrels of the main arcade of Lincoln's Angel Choir. Of particular interes appearance th s i t f similaeo r forms, additionally, spandrele th n i s between the major gallery arches of the Lincoln nave. Lincoln, however, may have been anticipated by the use of similar forms in the choirs of the great Yorkshire abbeys of Rievaulx and Whitby formere th t A .1215-25c , , encircled quatrefoil bote ar sh sunk int spandrele oth s between the paired openings of the gallery, and pierce the tympana of the subarches, in each bay of the choi transept;d an r 12t Whitbya 0 e 1220s,th n i , 121 trefoil quatrefoild an s s pierce e subordinatdth e spandrels of the subarches (as at Rievaulx) and in a few cases - at least - also decorated the major spandrels, either reliefn i 'sunkn i r o , 't Kilwinning) a for s m(a t theiYe . r appearanc t Rievaulea x may have much earlier precedents in Scotland, for the dado arcading of the surviving north and McALEER: TOWARD ARCHITECTURAN SA L HISTOR KILWINNINF YO G ABBE j Y85 1

east walls of the aisleless choir of Coldingham Priory (), possibly dating from the

1190s, has a quatrefoil, trefoil, or vesica in each spandrel. A little later in Scotland, small trefoils 2

and quatrefoil12 s pierce the tympana of the south gallery of Holyrood's nave.123 The tympana of the false-galler f Glasgow'yo s choi likewise ar r e pierce encircley db d quatrefoils wit plate-traceryha - like effect, again similar to that at Kilwinning, but are perhaps later in date. The general character of the two bays of the transept elevation would have been similar to the gallery level of the Holyrood nave (but without the further subdivisions), and the choir clerestory of Glasgow, including the latter's wooden roof. That means the elevation would have been of the same type as the choir of Hexham, except with vertical articulation. e presbyterTh f Arbroatyo imposinn a d hha g base plint rola f lho over three chamfered setbacks. The east end of Dryburgh likewise featured a prominent base course (as it remains on the north side of the chancel and north transept) composed of a roll above several sloping set backs arrange pairso tw .n di The design of Kilwinning's transept fa9ade, dominated by tall lancets, recalls Arbroath's south transep 1190-1214/15)(c t , wher lancete two ther seare wit oculugablhan the ein s (over three interior registers of arcading),124 and Rievaulx, c 1215-25, where there are three lancets.125

Lancets also dominate at Cistercian Croxden (Staffordshire), c 1220-5 to c 1253/4, in the south 6

transept (and west facade).127 12

Conclusions The likely plan of the new east end of Kilwinning - aisleless chancel, crossing, and transept with an eastern aisle servin generia f o chapels ga s c wa typ s- e which provide particulao sn r evidence undertakinge date foth th f r eo . motifMane th f yo s characterizin transepe th f o soute t ga tth m har Kilwinnin lona d gg ha currency r instanceFo . , 12-shaft piers occu earls a r Dundrennas ya n (north transept, c 1185) and as late as Sweetheart (after 1273). Even so the pier form is unique and does not fit into any developmental tendency, and the respond of the aisle arch is without an exact parallel for its particular form. Triple wall shafts similarly begin in the late 12th century and continu mid-13the t leaseth a o t t . Decorative foil f variouo s s shape slona alsd g oha histor t ybu they see havmo t e been more usually associated with tympana than spandrels unti e 1220sth l . Decorated spandrels make an early appearance in the choir of Rievaulx, then in the naves of Holyrood and Lincoln, suggesting a currency of c 1220-40. The resemblance of the encircled quatrefoi t Kilwinnina l thoso gt etympane use th nave n dth i choie f Holyroo ef o th ao t n a ri d dan Glasgow, as well as the appearance at both of individual pier types - involving 12-shafts in one cas e- combine d wit unusuae Scotlann hth i e us l thref do e shaft verticar fo s l articulation l pulal , l Kilwinning into the orbit of those two buildings - or perhaps, more precisely, between these two buildings e wor. Thath f enlargin , ko is t g Kilwinning would fall betwee constructioe nth e th f no nave of Holyrood and that of the choir of Glasgow. Considering the work as a whole, a date of beginnine c 123th r 0fo f constructiogo n woul more db e comfortable. Other aspectse sucth s ha gable design - following on after Arbroath and Rievaulx - would also fit neatly into the period after 1230. probable Th ecloiste w existencne e rth portaf eo arouny b l d 1200 automatically pushee sth earliesthe rebuildinat - t intearl - easothe y the nexend t yeargthe of t centurysof therif eBut . r threo o e tw decade f o p wasga betweesa full-scale portae th nth d an l e rebuilding suggestioe th , n thacloistee th t r portapreluda s schema wa l o et rebuildinr efo g loses strength, unless ambitiod nha been frustrated by a lack of means for that period. It is tempting to associate this campaign of rebuilding with the continued beneficence of the de

McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY 853

Moreville family: either of Richard's daughter Ela (Eva) de Moreville, wife of Roland of Galloway, Constable of Scotland, or of their son Alan (d. 1234).128 According to Timothy Pont, it was another(?) daughte f Richaro r d (anwifes hi d , Avici Lancaster)e ad , Dorothea, wif f Philippueo e sd Horssey 'accomplisheo wh , fabric e saie dth th d f Monastery'.ko 129

PHASE THREE NAVE TH : E

Description and interpretation The only evidence for the rebuilding of the nave is a series of corbels inserted into the south wall of the (hypothetically aisleless) origina arce pield th nave han rd openinan , g intwese oth t transept (illu 40)& .5 s They provide evidence of an aisle, a bay system, and a vault. Four corbels remain in situ, two others are gone, nava thero s s f seveeo ewa n bays nave shape th Th .ef eo pier soute unknown.s si th f ho ease d th en tt 13 A 0 wall, there is a semicircular plinth which must have been for the respond of the arch opening from transept to aisle. The west respond to the arcade is preserved as part of the south-west tower pier, along with the west second-storee jamth f bo y arche lean-te fronn th si f o to roof ove aisle rth e vaults. wale angle th Th lf eo betwee n aisl broatransepa d s ei an dm chamfear t widem 2 t appeari :0. r s original to the first church. If the suggestion that the first church was aisleless is correct, then it follows that the respond indicate semicirculae th y db r plint insertes hwa d int wale oth l (illus 28). Althougo n h w therno s ei indicatio responshape e th th f f eno o d which rose fro plinthe mth , fro extension ma wess it n tn o sid cleas i t ei r that it was accompanied by a single shaft or re-entrant angle (from which the vault rib may possibly have sprung). The corbels are somewhat unusual in shape, although they resemble a moulded bell capital to some extent. Their exact profile is difficult to determine due to erosion; none the less it is clear that there was a series of roll mouldings at the top, middle and bottom of the bell. respone nave Th th eo d t arcadeas e th (south-west e t n sidth eo f eo ) tower pie distinguishes i r e th y db appearance of chamfered orders in place of some shafts: there are two broad chamfered orders towards the towarde on nav d aislee esan th , flankin largega , broadly 0.1 fillete0m d axial shaft wese Th .t respone th f do former (false) gallery arcade also displays an axial shaft with a broad fillet, but is flanked asymmetrically by major and minor shafts (a pair [a,b] to the gallery, with at least four towards the nave in an aab?b rhythm). e arcTh h openin soutw g ne froh e aislmth e intwese oth t transep s completi t e (illu s desige 40)Th . s ni characterize basw ene whic n i typea scotie e y db hth on , essentialls ai y eliminated rollw o placee lo tw :sar a n do plinth wit hsmala l right angle between them (illu scapitals e 41)Th . , much weathered sharo d ,rathee eth r 'compressed' nature of those in the east transept; however, their profile is quite different (illus 42). It consists of a necking splia rind t gan rol l separate narroa y db w bellabacue Th alss . oswa forme splia y tdb roll, witha small astragal on the vertical face below it.131 The use of wide fillets (57 mm) on the axial shafts is also another characteristic feature: keeling does not appear. The broad fillets continue down onto the base mouldings. five eTh orderarce eace th h ar f ho s larg scaln ei distinctd co-ordinatee ean ar d an , sizn di e wite hth responde shaftth f o s s 42)(illu& identica t . 0 Thes4 no e shapeyar n i l largese th : t half-shaf - axia e e th on l- t is filleted, while the soffit order is polygonal; the pair of outer shafts and their corresponding orders are round; very thin roll foune sar d betwee maie nth n orders. Immediately abov arce eth h fro fla e aisle mth t th s ei (north) jamb of an arch that formerly opened from the roof space over the aisle vaults into the west transept. conversioe Th aislelese th f no s nav ehav y int aislen ma eoa e beedon n accompanie lengthenings it y db . A vertical seam, 1.6 m west of the west processional cloister portal, affecting the five courses visible below a memorial wall plaque, suggests that the south wall was extended two bays (illus 43). The extension is well built, with regular large exterioe ashlath n o r r face (illus e coursin 44)contrasn th I ., it ashlad o gan t e t ar r much less regular and more rough-faced on the interior. As the wall does not bond with the respond of the arch at the end of the aisle, it could be argued that this section is part of the early wall of the first church. Accordingly, the fa9ade wall of the first church would have been either just west of the east wall of the west rang extensionn wale a r wel th eo s i lf I l . wesrouge it th ,f o t h characte innes it f ro surface could possible yb

McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY 855

ILLUS 43 Interior face of two west bays of south wall of church, looking west

explaine intentioe resule th th f s o dta plasteo nt keepinn i t i r g wit remaindee hth easte walle n th th O .o t f o r the exterior, it may at first appear that the seam between the old south nave wall and the new aisle wall is covere walla y db belongin post-Reformatioe th o gt n remodellin wese th f t go rang e (illus 45). Howevert i , could be argued that the new masonry abuts the west face of this wall which, it has been suggested above, not only stands on the site of the east wall of the original west range but preserves some early masonry in bond with the church's south wall. Work included a new portal to the cloister placed in the eastern of the two 'new' bays. The jambs and pointed arch of the new portal, now blocked, are simply broadly chamfered towards the cloister, and are provided with a moulded label. The jambs are splayed towards the interior, and the rear-arch is a segmental curve; there are no mouldings. same Ath t evaulte timew ne a ,d wing parallelin westere gth n cloiste range least a th buil r s o to et rwa t- intendes wa t cleastarted t i no replacf o s rdi t i t I . r supplemeneo originae th t l range. latter' e trac13o th 2N f eo s west wall is preserved, and, as described above, the new work extends along the south wall of the church as far as the west face of the east wall of the early range. As mentioned, a new portal was inserted in the east bay which suggests thaintentioe th tteao t s r downwa wese nth tinitiae bloc o rangt th d p klu e an wes t processional doorway stume originae . Yetth th f s po a , l east wall still remains, this schem nevey ema r have been fully carried out additionn I . voussoire th , wala f so l arc s hbaysw it remai t westernmose ne A .th o n ntw i e th f o t east endcourseo springine tw , th f s o vaul a preservef e go ar t d (illus 46). Ther alss ei oseriea f ashlaso r blocks bonded intchurch'e oth s south wall betweee th vaule d nth an t new cloister porta (intendeda r fo l ) north/south wall (illus thicknese 46)Th . thif so s wall suggest havy ma e t si been meant to support an upper level; at ground level, it may have been a solid wall pierced by openings, rather tha open na n arcadecasey an , n evidenci o ther: n s actualls ei wa t ei y built. e Thib o st wals wa l weste th whicparallelef o o ,t onle e hth on y y visibldb e evidenc extension a s ei n (spur 'middlee wallth n )i f o ' the west transept's south wall (5.18m west of the south-east angle of the transept), with the base of an angle- 856 SOCIETY OF ANTIQUARIES OF SCOTLAND, 1995

ILLUS 44 Exterior face of two west bays of south wall of church

shaft (illus 47). Thus south-ease th , t exterior cornetransepe f thiw th o ne f s overlapped o r en wa t e th y db west claustral range which, possibly vaultea hav o t d s eha dwa , undercrof aisleso tw n .i t A bench-like plinth, originally composed of two courses - the upper about twice as high as the lower (0.23 m) - each chamfered and now not all complete, wrapped around the exterior south-east angle of the transept (illus 48). It was 0.35 m. wide on the east face. On the south face, there was a shallow projection (75 mm) which apparently formed a respond for the inner divison of the contemporary west range. The plinth increase heighgrounn e si th s a td level drops slightl wese th to y t alon soute gth h face. Ther e traco th n f s eo i chamfer around the south-east corner where the high plinth-bench is found.

Affiliations Once again there is a new style: none of the details of the west arch and pier is found in the east transept, and vice versa. The peculiar form of the corbels for the aisle vaults may find a parallel in two in the crypt of Glasgow, which are themselves 'unique' forms there. They are found on the sout hsoute facth f heo octagona norte lth piehnortd e facth ran f heo octagonal e pier belth e f th :o l corbel belo mouldes wit d abacu crosses si threy db e widely spaced rolls.133 Many, although not all, of the choir and nave arcade pier bases at Arbroath are characterized eliminatioe byth scotiae th f no ; instead smalo tw , l roll placee sar d ove largera r onet leasa e n .on I t cas pieee (north l basetal s survive easternmose th )- t south arcade pie r- ther e were rollsonlo ytw . filleThe t doe seenot s mhavto e been used. At Rievaulx vicinit, piersomthe the the sin e of yof north transept chapels have two rolls separated by a right angle and, as at Kilwinning, the broad majoe filletth f ro se 16-shaft-bundl shaftth f (o s e piers s expressedi ) , althoug huppee onlth n yi r torus and angle. Double roll bases also are found in the undercroft, and on one pier base (and on a 'demonstration pier' from the claustral buildings on display in the Commendator's House), at (Roxburgh).134 Double-roll base crossin e e founth ar s n o d g pier t Dornoca s h

McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY | 859

Cathedral (Sutherland) 1222/3-45c , , wher ease eth t respond nave th eo st arcad e (whose basee sar

not visible) have a wide fillet on the large axial shaft. In addition, the much-restored bases found 5

in the two-bay antechambers leading down to the cryp13 t at Glasgow Cathedral include a few which eliminate the scotia to replace it by a single or double angle between the two tori. Wide fillets occur on the north-east aisle pier at Dryburgh, and even wider fillets were apparently used in the nave piers, as can be seen from the remains of the pier at the pulpitum and the north-west respond. (The nave pier bases also eliminated the scotia in favour of a rather flat band.alreads ha t )I y been noted that broad fillets occurre Rievault da x where they were expressed in the upper mouldings of the bases. Broad fillets are also found in the crypt (see the axial pier of the east chapels, for instance) and in the choir of Glasgow Cathedral where once again base and capital profile quite ar s e different interestf o s i t I . , though, thafillete th expressete ar s base th en di

moulding plinth.d t Paislean sA y Abbey responde th ,soute th f hso aisle wal composee ar l f do 6

triplet of shafts13 , with fillets which are also expressed in the bases. The responds and piers for the now vanished west towers continu wide us eo large t fillet th en o saxia lbas e shaftsd th ean t bu , capital profiles are completely different from those of Kilwinning and some of the minor shafts are keeled. implicatioe Th chamferee th f no d orderease th towee t n sidth so forf e ero th f piem o r fo r the arcades is not certain. Chamfered orders do occasionally occur in Scotland as at Kelso, Kirkwall, Dryburgh, Paisley, and Sweetheart. Only at Dryburgh and Sweetheart - two buildings widely separated in date - do they form a major motif. At Kelso, chamfered orders are found in the two remaining bays of the nave (two in the west bay, one in the east bay); all the orders were chamfered towards the aisle. The inner order of the nave arcade was chamfered at Kirkwall e ordere galler s welth th a ,s f a navo n slyi retrochoird ean t DryburghA . , thee yar norte founth hn di transept, whil t mucea h later Sweetheart onl middle yth e chapee ordeth f o r l and nave arcade chamfereds swa chamferee Th . d orders were more discretely place t Paisleyda , being restricte archee side th f th e o s o dt facin g int towerse oth ' basesnorte th f Englandhn o I . , f chamfereo e us e th d order founs si t Bridlingtonnortda e th f h o wes e arcadet d a th t en td a , an , Hexham, where the arcades of the south transept arm have three chamfered orders. The wide date rangf thesth eo f eo e buidings spanning more tha centurna y- Kels Sweethearo ot - t proves only the longevity of the form, but reveals little about the date of the aisle arch at Kilwinning.

Conclusions apparene Th t complete discontinuit styln yi ewese betweethad th t tan a t ease word e th nth en t t ka may be exaggerated by the loss of the nave. Yet the uniformity of the corbels in the south wall does provide some evidence or argument for the extent of the break, and does not suggest any evolutioy kin ba f gradua do y b styley n n i breae ba l Th . k betwee wes e worthae d th nt th an f ta t k o the east seems as sharp as that between the east and the east cloister portal and, in a similar fashion imply ma , y bot certaiha nmastew lapse arrivane tead th a f tim an f rm eo d f o o lean masons. Even les chapter-house scase th thath f eo n ni e entranc t beei s ne ha possibl produco et e convincing parallel varioue th r fo ss feature wese th tf baso e wors th e k - moulding shafd an s t forms, the use of fillets and chamfered orders, or the capital type - which would provide a relatively precise date and context for the new work. None the less, without being able to produce a set of narrowly dated parallels, it is possible to suggest a date in the 1240s for the reconstruction or remodelling of the nave leading to the completion of the west transept in the 1250s. 860 SOCIETY OF ANTIQUARIES OF SCOTLAND, 1995

PHASE THREE WESE TH : T TRANSEPT"7

Description rebuilte Th , aisled nave terminate transept-lika n di e space which extended unde laterao rtw l towers whicf o , h the north one survived until the early 19th century.138 As the face of the pier towards the nave is devoid of shafts, being merel flaya t surface s unlikeli t i , y navee thae th t crossinf a Th .ther o s d ewa en ge arcth t ha west face of the pier bears a group of five shafts which rise up above nave gallery level. The axial shaft is the broaa s largesdha filletd pairan e t th ;f flankin so g shafts, widely spaced thinnee ar , simpld an r y round (illus 49). Of the corresponding respond against the interior of the west wall, only a rectangular plinth, 1.098 m by 1.948 m - in contrast to the shaped plinth of the pier respond - remains, without any trace of shafts on it. exterioe Th r appearanc transepe partle th b f en o y ca treconstructe d fro viewe mth s abbee madth f yeo ruins befor collapse eth north-wese th f eo t tower.139 They show tha towere tth s wer thren ei e stages diminishinf o , g height, articulated by stringcourses. There were stepped angle buttresses as well as a mid-wall buttress on the west and side faces which ros middle higsecons e eth a th s ha f eo d storey thirde .Th , uppermost, store paia narrof d ro yha w lancets on easwes d towardd an t an t navee side snont th th e bu ,n faces eo additionn I . , there appear havo st e bee nsmala l traceried window on the east fagade of the second stage, at about nave clerestory level, and a larger traceried window on the west fafade of the lower storey placed on the nave side of the mid-wall buttress. The towers orginally may have been cappe plaia y dnb continuous parapet, with stepped gables rising nort sout d saddlebaca hr an hfo k roof. Only one view shows the towers' relationship to the nave (illus 50).140 It depicts a tall wide arch, at least as high as the window on the later west wall, in the south wall of the tower. The face of the wall immediately above the arch is plain. There is the suggestion of a stringcourse below the level of the first setback of the buttresses, above which the sketcher appears to show a line of broken masonry running across tower'e th s south face (fro highese mth t surviving wese pointh f t o twall ) which coul interpretee db e th s da remains of a barrel vault. It is at this point that one might have expected some evidence of the former roof structure t nonbu , e appears. Above this, ther slendea s ei r rectangular apertur e soutth n hei facethend an , , abov uppermose eth t stringcourse narroo tw , w lights with louvre hood san d moulds. Now all that is visible, in addition to the arch opening from the south aisle into the transept, is the lowest course easte th f , so south weswhicd m an ,ar t he wall th woul f so d have forme soute basth e r dhth efo tower (illus 51).141 broaThera s dewa buttress externally nave th line e f th internae eo arcad n th i , d ean l respond, and another at the south-west angle where a stair vice was located. The west wall has been denuded l masonroal f y abov base eth e course. architecturao Thern e ear l details preserve interioe th n do r surfacef o s othee poorlo th tw r y preserved walls double-chamfereA . d tracee plintb n dhca intermittently fro wese mth t side of the buttress to the junction of the west wall of the west claustral range against the transept's south face.142 It steps up thrice along the west face: in the middle of the west face of the buttress in the line of the nave arcade (now robbed out); twice betwee buttrese angle nth th d e turrean s t (onl e southeryth s i g njo actually preserved) oncd an ; esoute morth hn e o face , just befor junctioe eth n wit wese hth t range narroe Th . w chamfer from the base plinth turns on to the wall projecting to the south (forming the west wall of the new west behinrange)n ra d mouldeda an , d base which probably serve angle dshafa th n ei t create reductioa y db n thicknese inth wale th t thi a lf so s point, wher portaea havy lma e been located.143

Affiliations wese Th t transep t Kilwinnina t f onlo ye fivon s gei know havo nt e been constructe Britain di d nan e groupe lates th th e other f s o i tTh . foune ar s t Benedictina d e Bur t EdmundyS s Abben y(i construction betwee e 1140s)nc 1120 th d ?an , Benedictin Cathedray eEl l (after 1155/60?)e th , Tironensian (after c 1175?), and Benedictine Peterborough Abbey (as it was to

Dissolutione th ) (between 117 d 1194).7an Lackin gwestera n arm surmounted an , twiy db n 4

towers wese th , t transep f Kilwinnino t Englise th f closess gi lateso e he th exampleso on t14 t , that a t Peterborough differencee Th . obviouse ar s : Kilwinning's transep smalles i t scaln ri lacke d ean e dth

giant loggia-like porch adde Peterborough'o dt course construction.s th it n sf i e o

5 14

862 SOCIET ANTIQUARIEF YO SCOTLANDF SO , 1995

ILLUS 51 Vie f wes o f soutw o m t transephar t from north-west

Althoug westerha n transep majofouns e ti th t da r Tironensian hous Kelsof eo , thereo n seem e b o st formal relationship between the two transept structures despite belonging to the same Order.146 The earlier transept at Relso was distinguished by having three arms of almost equal dimensions as well as a single axial lantern tower interestins i t I . noto gt e thatowee th t r othee pierth t rsa major Tironensian housn ei Scotland, Arbroath, were flat surfaced towards the nave, and also that the towers were without any internal

subdivisions. In a sense, only the presence of the second-stage openings and the mock clerestory above 7

separate twin-toweree sth 14 d western structur Arbroatf eo h fro transepe mth Kilwinningt ta .

Conclusions The reasons for the appearance of the west transept at Kilwinning are not at all evident, as indeed is also the case with Kelso, where it was, however, part of the original concept of the building. This is all the more puzzling as none of the Tironensian churches outside Scotland had such forms at their west ends. As there is no remaining evidence for the location of altars against the east e transep e endth wall th f o sn i st arms, althoug e lengte walth hth lf ho (2. 1 m) could have accommodated one, a simple liturgical explanation is not forthcoming. There is a recess with a piscina and, at each side, an aumbry located at the east end of the south wall in the west transept at Kelso. correspondina s 148A g arrangemen ease lackingth s i tnorte s walla th m d har n botf i tso an , h transept arms are decorated with intersecting arcading, it seems unlikely that there were altars in transepe th t arms piscine Th .aumbrie d aan s probabl nave y th servet ef a o wes e altad n dth a en t t ra Kelso which, like tha t Kilwinninga t s parochial.wa , 149 Thu wese reasoe th th s r t transepnfo t a t Kelso, as in the case of Kilwinning, is unknown. McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY 863

PHASE FOUR: THE INSERTED WEST WALL

Description and interpretation A west wall closing off the nave from the transept space, was eventually inserted between the tower piers (illus 49).150 It was not aligned between their flat faces but was, instead, placed farther to the east. As it survives nave th t overlapwhicy i ef , o arcad wa d ha responde en suggestn th s ei e th t a shavy s thama e t i t continued acros aislee th s s well a sthereford an , e completely engulfe wese th d t responds. Ther vera s ei y narrow west porta windowd an l formee th , r witchamfereo htw d orders externally lattee th , r with onee ,th inner jamb being simply a plain right angle. The portal jambs were rebated and splayed (1.37 m wide at the inner rebate; 1.77 m wide at the outer edge of the splay). The heavy chamfered tracery of the window was of the simple, uncusped Y-type with a transom. There is a chamfered setback on the exterior at the level of the window sill which is matched on the interior by a right-angled setback just below the window. It has been suggested that the west wall was inserted due to damage to the west end, especially to the south-west tower, caused by the Wars of Independence.151 Equally, the wall might well have been built in the immediate post-Reformation period collapse , th e soutpossibl th o t f he o toweresula ydu s f earlie(a ro t r damage?), in order to allow the nave to function as a parish church.152 However, the author of the report of excavatione th s carrien 198i t 5dou suggeste d thahave waly th t ma le been inserte inadequato t e ddu e foundations unde transepe rth t towers.153

Affiliations Despit face eth t tha'newe th t ' western processional cloisteportae th o t l r employe chamferee dth d order, unaccompanied by any mouldings, the window and portal of the west wall could be much chamfee th f lata o t ee latea r perio us paralledatea n r i e pos s e dth Th .ha t n i l15th-centur y window inserted in the north wall of the sacristy (c 1411-49) at Arbroath.154 Chamfered jambs and Y tracery appear in the windows of the upper storey of the early tower at Dunblane Cathedral (Perthshire) whic bees hha n 16tdatee th h o dt century. 155

Conclusions collapse datth e e south-wes th f th eo e reasonf d eo Th an r fo st fagade tower remain ambiguous. Until a thorough excavation is conducted, questions about the adequacy of the foundations, the completion of the building and the date of its destruction must remain matters of speculation. Since the surviving north-west tower, as it appeared in the late 18th- and early 19th-century views, seems perfectly vertical and intact, the inadequacy of the foundations as a cause for the constructio wese th e doubted b f t no waly ma l spir A s constructe. ewa s lat da s 178 ea 9 without causing any apparent damage to the tower due to its added weight. In 1814, when the tower was demolished resule th f damags o ta , e cause lightniny db 1809n gi foundatione th , s were blowt nou with the use of gunpowder.156 Stylistic comparisons, however suggeso d , t thawese th t t wall could jus s wel a tposte b l - Reformation in date as later 14th or 15th century.

GENERAL CONCLUSIONS t beeAlthougno ns possiblha t hi dato et severae eth l building phasee abbeth f yo s wity han precision, their general sequence is, I believe, clear. One can summarize the sequence, and check briefly b , it y examinin reversn i t gi e order. Thu wese th s t wall, wit s windohit doord wan s i , clearly, whatever its date, the latest work: there is no doubt that it was inserted between the piers at 864 SOCIET ANTIQUARIEF YO SCOTLANDF SO , 1995

navee westhe th ef Th .o west d transepen t t latese itselth s ti f elemen churce th f o th propes i d an r stylistically placed towards the mid 13th century as the last stage of a complete rebuilding of the original church t followI . fron navee o s mth , althoug conservee hth dremaine t statth no f eo o sd permi judgo t s u te with complete certaint f thyi e nav transepd ean t were conceived togethes a r o r separate and sequential constructions. The associated reconstruction of the west range of the monastic buildings remains problematic, both in terms of the builders' intention and the extent of its execution. Again rebuildine th , ease cand th en tstylistin f ,go o c grounds placee b , d aroun fourte dth h decade 13tth hf eo century s exacIt . t relationshi earliey fore an th f m ro o p t churcsite s th ei n ho once more problematic s rebuildinscale it th f t eo bu , g suggests tha earlien a t r structur f muceo h more modest dimensions was likely. The east processional portal, once again on the basis of style, is clearly earlier than the rebuilding of the east end and so is a separate phase, however short in duration or limited in ambition. hypotheticatracee e th Th f so l first churcidentifiee b n hca d mainlbasie th f theisn o y o r later alteration. The construction of the east processional portal caused the blocking up and converson of the portal in the west wall of the south arm of the transept into a closet. Thus the latter preceded e easth t processional porta daten i l reconstructioA . s initiait f no l modest forminimad man l decoration allows two other early remnants to be identified: the portal to the stair vice in the south- west cornetransepte wese th th f ro td processionaan , l doorway. There are four aspects which remain particularly problematic, most notably the ensemble of chapter-house th e entrance r whicfo , h therexaco n s ei t stylistic parallel preponderance th ; e th f eo evidence favour earln a s y date, especiall thers ya cleao e n seem e rb supportino st g evidencs it r efo rebuilding at later date. Two other problematic areas are related: they are the vestibule and vice in the south-west angle of the east transept and the adjoining slype. While their place in the general sequence seems clear, their origina sequence l formth d an s f event eo s that resulte thein di r present appearance remain enigmatic. The fourth problem is the nature and character of the western range. Its original form was theoretically very simple; exactly when it was torn down is more ambiguous: was it mostly destroyed in the enlarged scheme initiated at the time of the introduction of the west transept, a scheme which was never completed?

APPENDIX 1

AREE TH A BETWEE SOUTE NTH HCHAPTER-HOUSE GABLTH D EAN E The effect of the enlargement of the transept upon the adjoining north end of the east range of the claustral buildings, and the functional connections between them, is not immediately apparent, for the constructional histor f thiyo smose areth s tai difficul deciphero t t . Althoug severae hth l component evidente sar exace th , t sequence of the changes affecting their original form remains uncertain. Part of the problem of interpretation is due to the fact that most of the lower portion of the south face of the south-west angle-buttress (illus 14 & 32), frospringine levee th m th lowee f o lth f gro barre formee l th vaul f (o rt possible sacristy levea o t lp u ) equal to the sill of the south-east lancet, has been patched and rebuilt (this is especially evident on the south side of the interior of the stair vice). e exteriorOmoso nth tw te prominenth , t factors are e lowe:th r level barrel-vaulted corridor (the sacristy/slype): with possible evidence of a blocked doorway at the west end of its north wall and, at its east f rebateo t splayed endse dan a , d uppee jambsth d r levean ; l barrel-vaulted space (illu , 15)norte 14 s Th . h wall slightlx e latterosi th f f o , y irregular course f coarso s e grey stone, extends betwee e south-wesnth t angle buttress and a broad buttress at the east which marks the beginning of the new construction and the extension of the transept (illus 14 & 16). The vault supported by this stretch of wall was of the same length as the lower McALEER: TOWARD ARCHITECTURAN SA L HISTOR KILWINNINF YO G ABBEY 865 eass it onet endA . uppee ,th r barrel-vaulted spac closes thia e wa y ndb wall, less than halthicknese th f f so the one below (illus 52). That wall, containing a doorway with rebated jambs, as mentioned above, is not the original end-wall of the corridor. The south jamb is clearly inserted into the wall between the chapter house and the slype; its insertion involved a little refacing of the wall to east and west. The chapter-house wall increase thicknesn i s s 1.0o 1.1t 49m minsertee th eas f o t d jamb t thabu ,t thickness appears originat i s a l continue ease th chapte te o swalt th f o l r house. Furthermore, therbono n s dei betwee norte nth e h jamth d ban nort hslype wallth f eo (sout htransepte walth f o l : illus 16). It would appea putative wald th rf en o lthabarrel-vaulte e e eth th tsacrist f o d yan d chamber abov, eit which may have been bonded into the south-east angle of the original transept, was completely removed whe transepe nth enlargeds wa t sacriste Th .the s ynwa converted int oslypa buildiny eb end-walw ne ga l with a rebated doorframe, although the rounded outer jambs suggest that this portal wall could be a century later than the enlargement of the transept. The original length of the sacristry may be indicated by the point where the chapter house wall increases in thickness. (Does this increase in the thickness of the wall imply that the chapter hous vaulteds ewa westere ?Th lengtne halth f thif ho fo s wall coul thinnee db r becauslevelo tw e s eth of barrel vaults provide dresponsa e thrust th vaulta o e f easet o s e wallTh .t th hal f ,o fwhic s freehwa - standing, neede increasee b o dt thicknesn di orden si withstano rt thruse dth f vaultso t entire Th . e south wall of the chapter house was abutted by the undercroft.) Immediately outside this new end wall, there is the broad flat buttress marking the beginning of the eastward extension which included the new east aisle (illus 52). The new buttress, of rather shallow projection with a nook-shaft at its east side (presumably, it was constructed against the original east end wall of the putative sacristy), does not bond with the end wall of the slype - another reason for considering that latetha e b rlatte insertiony rma r somFo . e reaso t apparent easnw no uppete ne walth e f th ,r o l chambes rwa made much thinner than the new lower wall, so that a gap was created between its outer face and the west buttressese anglth f ewhico p nicels ga a h,wa y face well-cuy db t block f ashlarso t abouA . levee th t l where

• :st*m~-^ * r!^***k*

ILLUS 52 Lower part of south gable wall of transept: barrel-vaulted chambers and buttress of new work | SOCIET 6 86 ANTIQUARIEF YO SCOTLANDF SO , 1995

the extrados of the barrel vault abutted the wall, the broad buttress was reduced to about half its width, apparently in order to accommodate the pitch of a roof over the east range. Part of the raggle sloping up to the middle lancet consists of a lip projecting from the surfaces of the blocks it crosses. The narrowed buttress was then continued upward to end with a setback below the lower east corner of the eastern lancet. Several of these features allow or encourage the speculation that both barrel-vaulted spaces were closed at the east by a thin wall at the time of the enlargement of the transept, with a narrow strip of wall between their outer (east) face and the projection of the broad buttress. At a later date, the end wall of the lower space was replaced by a much thicker wall (of which the east face abutted the buttress), with a portal. The reason for the broad overhang of the roof over the end-wall of the barrel-vaulted spaces is not apparent; it must have produced a rather curious effect. The rebuilt lower south face of the south-west angle buttress is now flush with the plane of the south wall of the upper vaulted space (illus 14). At the level of the springing uppee oth f r barrel vault ease th , t south-wese angleth f o t buttres s fouha s r courses wit angln ha e rollo tw ; more are found four courses above (three courses above the reduction in width of the buttress: (illus 52). Thes e mosear t likely insertion f reuseo s d blockrepairo t e angl th e s edu sa s rol s accompaniei l V-a -y db shaped groove to one side: on the lowest two courses, these grooves do not line up but, rather, are placed on opposite buttress e faceth f so thire th ; d course lack paralleline sth g groove altogether does a ,bloce e sth th n ki ninth course fragmenA ( . t wit same hth e mouldin s visiblturrete inside gi th th larga f n n eo i eo ,e aref ao rough masonry which is the reverse of the area of patching on the south face above the lower barrel vault.) The interior faces of the north walls of the lower and upper barrel-vaulted chambers are not in the same vertical plane, and the face of the transept above the upper chamber is yet again set back - however, on a plane diverging toward e north-westh s t (illus 14). e masonrSom th e sout f th eo f he transepyo th fac f o e t above the upper barrel-vault is similar in character to that of the upper vaulted chamber: the stones are more square than e 'newthosth f eo ' work whic lone narrowd har gan jointe th ;quite sar e coarse heighe e th ;th f o t coursing varies and is rather irregular (it now slopes down towards the west); same kind of greyish coarse ston uses ei d (illus 52). Although ther some ear e irregularitie coursingn si t clearl, no the o yd define either vertical, diagonal r horizontao , l breaks between wha t workd mighol r describee .o b t Changew e ne th s da n si height of the courses, which involve one block angling over or into another, might suggest a juncture between earl lated yan r wor slightla n ki y diagonal line betwee uppee th eastere f ro nth barred nen ease l th vaul t d an t cornemiddle th f o r e face lanceth t r thafo t t twer similano t ei r disjunctions undoubtedle occuth n i r y later masonry to the east of the east lancet. Such discontinuities or irregularities as these joggs in coursing level appear to represent merely building breaks of presumably short duration. In the end, it might be possible to sugges masonre t th tha l al t y abov uppee levee th eth f ro l barrel vault belong enlargemene th o st t witbu th some masonry from the original gable wall reused. None the less, one has to admit the possibility that some of it is in situ from the earlier building. barrel-vaulteo tw e Th d spaces must have been equa heighn chaptee i l th o t t r house e slope th Th f . eo roof suggests the monks moved through this roof space to gain access to the dormitory (illus 32). This arrangemen havy ma te been simila thao t rt Arbroat a t h wher entrance eth dormitore th o et througs ywa ha roof space. The pitch of the roof is completely evident on the exterior face of the south transept gable at Arbroath: its base line corresponds to the floor level of the wall passage reached by the spiral stair entered from a door at ground level just west of the south respond of the arcade of the transept's east aisle.

APPENDIX 2

THE SOUTH-WEST STAIR TURRET OF THE TRANSEPT As described in the text, the doorway giving access to the stair vice in the south-west angle of the transept has been remodelle leadw dno s t (illuI int . osmala s9) l vestiblbeginnine th vice d th ean e f gwhicho , above eth first four masa step d f restoreso san d masonry occupyin site f fougth eo r more holloa s i , w shaft (illus 39). This small area is one of the most puzzling and difficult to interpret in all of the abbey ruins. In its initial form, the doorway was arched, with a chamfered and rebated eastern jamb (which still remains unaltered) and a western jamb which was also chamfered externally. The west jamb was rebuilt McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY I 867

abov bottoe eth m three courses (which bond wit adjoinine hth g wes transeptte walth f o l ) whee arce th nth f ho doorway was replaced by a lintel; only one block at the east with its eastern edge cut to a curve to conform to formee linth e f th eo r arch remain evidencs sa . it f eo The east jamb of the portal bonds in with the east wall 1.09 m wide of the vestibule. The lower eight courses are of narrow oblong blocks; a very high course (of two blocks) follows, then there is a setback which mark springin e levee th sth ribe f o lth s f abovgo e which ther five ear e more courses. This (east) wall doet sno bond with the narrow south wall (0.89 m) of the vestibule (illus 17). The south wall appears uniform in construction; however, it is not all in one plane (illus 18). The lowest three plana course n i ee whicsar h form acutn a s eeaste anglth t ) e(a wit remaindee h th tha f wale o t th lf o r above. In addition, in each of these courses, the west end of the block is angled (or cut?) back to the south, forming a vertical joint with the masonry of the vice proper which is to the west of that of the wall above. Above these three lowest courses nexthe , t five mee masonr victhe t the e - consistin yof fougof r courseat s- an obtuse angle sixtA . h course carries across this anglquartea n i ed whilr en curv nexe o eth tw te (the stones reused from the arch of the doorway?) against the mass of rubble which now constitutes the masonry of the vice at this level; six more courses follow before the underside of the ribless half of the vault is reached. The south-west wall, which is the ashlar face of the mass of rubble on which the lower steps of the vice placede ar , consist sevef so n courses befor t giveei s awa rubbleo yt firse .Th t three correspon lowee th o drt three of the south wall, which (the latter) appear to overlap them (the former), and the heights of the courses are not the same. As noted above, the next four courses correspond with the next five courses of the south wall. At the west end, the first four courses abut the newel of the stair vice (illus 53); the suceeding three courses are stepped back towards the east, the area between them and the stair vice wall above the first four steps in situ being filled in irregularly by ashlar blocks in step-like form which are clearly the product of an effort to conserve the ruins. The four steps in situ are integral with the newel (illus 53). The north half of the lowest step has, it would seem, been cut back, so it is now a triangular wedge pointing outwards (to the east). The three upper steps appear to be the same height as the coursing of the curved inner wall of the turret and one can make out the ghos f fouo t r more wale stepth ln si abov e (wher stepe eth s have been replace ashlae th y drb blocka n si step-like form). Above this point - that is, above the level of the seventh course of the turret wall - the south side of the turret is blocked with rubble up to a level approximately equal to the ribbed vault of the vestibule. Above this level turrete th , , which survives nearl fuls it l o yt origina l height bees ha , ns it robbe l al f do steps, except for a short section of five complete steps and fragments of two others. In the turret wall there is firs larga t e patche soutde th are hn a o sid e which probabl locatioe uppee th th exin a s o rt tf ywa no barrel - vaulted chamber. Above this area, there is a blocked exit, facing south-east, which formerly led to the level ove e barrel-vaulteth r d chamber abov sacriste eth y and, presumably dormitorye th o t n o , ; then quartea , r revolution abov , therexin eit a ts e i facin g lowee nortth o hrt passagewa wese th transeptn e y i wal th s f i o lt (i t in the area between the lintels of these two exits that the few remaining in situ upper steps are found). There then follows a third exit, facing east, leading to the lower passageway across the south gable; a fourth exit, facing north, connecting with the clerestory-level passage of the transept's west wall; and, finally, a fifth exit, facing east, and corresponding with the clerestory level of the south gable. Ove squarise rth h area formedoorwae eassoute d th th y an td b d hremaine ywallan th ribbea e f sar s o d vault (illu ribo situn ssi tw 17)e sprin.Th g from positions flankin insertee gth d doorwa e linteth f o l y (just a t the setback formed at the top of the high course of masonry in the east wall). They are massive and unmoulded keystone th ; missings ei e soutTh . h less i par t s (i ttha nvaule halfth constructes f i t o ) f rubbledo , some of which seems original, which meets the south wall in an irregular horizontal line. The south edge of the vault over the vestibule was formerly bounded by an arch, of which only the unmoulded springer remains, which separate frot d i vic space e mth th e f epropero . e nortTh hvicee wallth sectioe f , th o tha , is tn betwee curvee nth d (weste th turree ) th d wal f an to l doorway same th , appearf e renovateo e buil e th b s do a st d wes e partth t f doorwayse jamo th f bo whicf o , s h(a noted above) onl firse yth t three courses belon initias it o gt l phase irregulan A . rwese jointh t t a taffect e th s four courses immediately below the springer of the west arch of the vault. vestibule th f I e initially formed bot passagewaha adjacene th o yt t sacristy (later altere slype)a o dt d an , gave acces stairo st s descending fro dormitorye mth , theunderstandabls i t ni therbony o n es wh dei betweet ni ansoute dth h wall existine Th . g south wall shoul dproduce blockine theth th e passagewae nb f th o t f go t ya 868 SOCIETY OF ANTIQUARIES OF SCOTLAND, 1995

ILLUS 53 Remaining steps of newel stair in south-east vice of transept

time th e thasacriste th t convertes ywa d int oslypea . Whe passagewae nth blockeds ywa soute th , he th hal f o f ribbed vaul removes wa t d becaus blockage eth mads ewa e thicker tha doorwae nth y which must have existed between the vestibule and the sacristy. (The north face of the blockage is to the north of the northern jambs of that doorwa whicf y o actua o n h w therl visiblno s ei e trace.) readilt Whano s souti te y th evideny h walwh s i tl cleathero y n plans e wh eri d continuouon ean n i l al t sno verticas i o tw lweste joine reasoe th th t Th .a tr nfo courses (10 & 11) which together end in a quarter curve at the west, is equally obscure, unless it is simply the consequenc re-usinf eo g stones doorwayfroe arce th m f th h o . difficuls i t existinI e th decidparo y t tf o an t gf e i steps belonge originae th o dt l stairs leadine th o gt dormitory. As the upper three of the four steps in situ correspond in height to the courses of the interior wall of the turret, and as there is no sign of disturbance or variation in the coursing of the wall above them, and above where the succeeding four steps above them would have been, it seems most probable that the entire stair vicrebuils time enlargeds ewa th et wa a t tha ease d th ten t . McALEER: TOWARD ARCHITECTURAN SA L HISTOR KILWINNINF YO G ABBEY 869

ACKNOWLEDGEMENTS Research for this study was made possible by part of a grant from the Social Sciences and Humanities Research Council of Canada. I very much wish to thank Richard Fawcett for his comments, encouragement and support, all of which have been greatly appreciated. Illus 2 was draw r publicationfo Margarey nb t Finch; s fundethi grana wa s y b dt fro e Society'mth s Angus Graham Bequest.

NOTES 1 There were three other smaller houses: Lesmahagow (Lanarkshire), founded in 1144 by David I; Fogo (Berwickshire), a chapel granted to Kelso by Patrick Corbet, 1253/97; and Fyvie (Aberdeenshire), founded by Reginald le Chen in 1285. A fourth abbey was established at Lindores (Fife) by David, Earl of Huntingdon, in 1191. The Order, founded c 1105 by Bernard of Tiron (ob. 1116/17), settled at Tiron, near Chartres, in 1109. See Cowan & Easson 1976, 66-71. 2 Cowa Easson& n 1976, 68- 66-7d 9an , respectively. 3 Cowan 1986, 267-8 Macleo. ;cf d 1878, 128-9 (foundation date). 4 1140: Cowan 1986, 268, cite (incorrectn da ) referenc J Spottiswood' o t e , Historye Churchth f o f o Scotland [sic] (London, 1677), 407': there is no reference to Kilwinning on Spottiswoode 1677, 407 (nor are either de Moreville or Kilwinning mentioned in the index); in the Appendix (by T Middleton) editioh 4t e nth (1677o t dato n ) s citedei . 114 gives 0wa Chalmery nb s 1807-24, vo , 5483 l , without citing his authority. 1157: lost register (Eglinton no 279), formerly in London. 1191: according to Timothy Pont; see Dobie 1876, 254-5. 5 Dempster 1622, 69-70; Spottiswoode 1677, Appendix (by T Middleton), 15. (Richard: Eglinton no 279, and Pont, Dobie 1876, 254.) Cottineau 1935 & 1937, vol 1, 1519: founded 'avant 1062, ou 1107, par Hughes de Morville', citing Barrett 1913, 79, and Montrond 1856, col. 426. For Hugh (briefly), see Edwards (1924), 332, who favoured a foundation date of c 1150, and Cruden 1986, 83-4. Gestae Th Regis 6 Henrici, Stubbs , 1867312-131 l ,vo , give accounn sa fountaia f o t n havin witn gru h blood for eight days at 'ecclesiam Santi Vinini in occidentalibus partibus terrae regis Scotiae, infra Cuninham, no longe a castello de Hirun', in June 1184. (Anderson 1908, 286, 327-8.) 7 At that time, they were in the possession of the Earls of Eglinton: Cowan 1986, 267. establishmene th r Fo Tironensiae th f 8 o t n Orde Scotlandn ri Barroe se , w 1960, 29-38 acceptee ;h d Hugh foundere a sAndersoth d di s a , n 1954. 97 , 9 Registrum . Kilwinnin98 o 1843n , l 184 , ,vo g appear lis a f monasterie o tn s i Scotlann si Gervasy db e of Canterbury Mappas hi n i , Mundi, 1210c : Stubbs 187 , 441-2 2 1880 9& l vo , . 10 Grose 178 91791 & (1791)2 l ,vo , 214, pi between 212-13 e alsSe .o Dobie 1876, pi facin8 26 g ('Kilwinning Abbey' viea : w fro south-ease mth t wit parise hth h churc right)e th t ha . 11 Anonymous 1814, 723^t, pi facing 721 (contents page). For another view of the original tower, see Denholm ('J.D.') 1802, 243-4, pi facin (content1 g24 s page) (draw DenholmJ y nb , engrave Scott)R y db . Crudeg E 12 n 1986 ,mose 61Th .t precise identificatio buildinf no g phase limites si briee th fo d t sentenc e or two in Discovery Excav Scot 1961, 54, and idem 1963, 60 (repeated in Stevenson 1985, no 56, 94-5). 13 Galloway 1878, 89-114. 14 MacGibbo Rosn& s 1896-7 (1896)2 l ,vo , 73-86. notablA 15 e exampl f boteo f thesho e tendencie wese th s ti s porta f Arbroato l h Abbey, which wile b l discussed further below. 16 Eg of the retention of the semicircular arch, the portal of the west tower of Dundee Church (Angus; MacGibbon & Ross 1896-7, vol 3, 126, fig. 1049), of the mid 15th century; of the reappearance of the barrel vault, Seton Collegiate Church (East Lothian) 1470-150c , 8 (Cruden 1986, 184-6, pi 65): both structures have otherwise clearly identifiable Late Gothic forms. 17 Eg the rather 13th-century mouldings and capitals of the late 15th-century inner portal of the so-called Halkerston Towe John't S t ra s Kirk, Perth: MacGibbo Rosn& s 1896-7 , 1203 1045l g fi ,vo , . 870 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1995

18 Ther s somei norte e th evidenc f hadjoinine o wes e sidth d th f eo en t t ea s now)gi t slypi s , e(a tha a t doorway formerly opene de transept betweeth d an .t ni These areas, difficul interpreto t t , wile b l discussed further below. For detailed descriptions see Appendices 1 and 2. 19 Waverley: Fergusson 1984, 26-7, fig 3; other later examples cited by Fergusson are Sawley (Yorkshire) 115c , 0 and Norwayn i , , Lysa, after 1146. Tintern: Brakspear 1929; Clapham 1934g fi , ,77 25. Fountains: Gilyard-Beer & Coppack 1986, 178-83, figs 14 & 15. 0 2 This typ plaf eo n doe t sees no m havo t e been typica earliese th f o l t French Cistercian churches: Citeaux I (cons. 1106), Pontigny (fd. 1114), and Clairvaux I (1115). The earliest example of a long aisleless nave, transept arms with chapels and (apsidal) presbytery is Ourscamp, fd. 1129: Fergusson 1984, 27, and bibliography cite 19-22n earliese d n pla e th f nTh o . t Tironensian foundatio Scotlandn ni t Selkira , k (Selkirk) unknows i , n (the sitneves eha r bee nsuccessos it located) f o ease t d t Kelsno a rTh en t .s oha been excavated: according to a 16th-century description, it had an eastern transept and crossing tower, chapelse th f bunature choio e their th td th o , f ran ero number t knownno s i , . 21 Holyrood: RCAHMS 1951, 130, fig 290; width of nave, 7.01 m. (Widths of other naves: Waverley, 7.3 Tintern; 1m 8.5c , Fountains; 3m als: lengte m o transepth a 2 f 5. h,o t arm). 22 A very similar plan seems to have appeared in a non-Cistercian context in Wales, at the Benedictine prior f Ewennyo y (Glamorgan) whic s beehha n recently redated fro mc 114 1116-26o 1t : Thurlby 1988, 281-94. Ewenny, however, became a priory only in 1141; the reasons for a plan of this type occurring before 1126 in a non-monastic context were not considered by Thurlby. t Fountaina s A 2 3I (1147-52) I s , Buildwas (1150-60), Roch c 1170)e( Abbed an , y Dor eI (1170-80) : Fergusson 1984, 42-3. 95 , , 9263 , 24 The evidence for the relationship of the transept arms to the nave at Holyrood was not actually uncovered (RCAHMS 1951, fig 289). At Waverley and Fountains (also Lysa), the (lower?) transept arms were separated fro nave mth e spac wallsy eb : Fergusson 1984, 26-7. Fergusson describee dth crossing (termed an 'unsegregated' crossing) at Tintern as similar to those at Waverley and Fountains. Gilyard-Beer & Coppack (1986, 191), however, specifically noted the transept arms at Tintern were not walled off from the nave. Fergusson (1984, 28), pointed out that similar plans were used by the Englandn i g e , , after 1143, Torre (Devon), Cockersand (Lancashire), Egglestone (Yorkshire) Bayhad an , m (Sussex) e Austith y s b welna ,s canonsa l , Kirkha Haughmond man d (Yorkshire). 25 MacGibbon & Ross 1896-7, vol 1, 230-58 (esp. 238) and figs 211-12 (on 246-7); Cruden 1986, pi 8 ('triforum'3 &3 , 1150-80?); Walke Ritchi& r e , 1987120-162 o ,n ; Gifford 1988, 175-80 (esp. 180). The gallery openings at Dunfermline do have a shaft and moulded arch recessed midway through the thicknes walle th f , so whic h make austerite sth more th f eyo visibl more enavth e arcl eth eal o h t peculiar. 6 2 MacGibbo Rosn& s 1896-7 , 259-91 l vo ,2 (esp. 273); RCAHMS 1946 399o ,n , 113-25 (esp. 123); Cruden 1986 ,(choi 8 p3 i r date d1140c 119d )an , 1137-4 Crudef 2(c n 1977, 85-97 [87-8: 1137-42], figs 1 & 5). 27 Althoug Augustinian ha n prior establishes ywa t Carlislda t untino e1123c ls s 113wa wa t , 3i e thase a t created by Henry I: Dickinson 1950, 245-6; Knowles & Hadcock 1971, 139, 152. If begun 1123/33, e navth e most probabl t beeno nd reacheyha y 113db 6 whe cite nth y reverte Scottiso dt h control (annexation by David I) until shortly after the death of the first , Adelulf, in 1156. Its construction in the years 1136-56 would therefore place it in the ambit of the architecture of the northern kingdom. Also see Gem 1989, 24-5, 28. 28 MacGibbo Rosn& s 1896-7 , 448-641 l vo , ; RCAHMS 1915 258o n , , 132-48 (esp. 138-44); Baldwin 1985, no 68, 118-19; Cruden 1986, 83-7. Richardson & Tabraham (rev) 1937/87 offered no further thoughts on the dating of the cloister range. 29 R Fawcett has suggested to me that this portal, or at least its arch, might be a much later insertion because of the fineness of the cutting and of the joints. It was probably blocked up when the nave was rebuilt and (possibly) lengthened two bays prior to the construction of the west transept, because a new larger portal to the cloister (now blocked - the blocking post-Reformation) was included in the easternmost of the two new bays. McALEER: TOWARD ARCHITECTURAN SA L HISTOR KILWINNINF YO G ABBEY 871

30 MacGibbon & Ross 1896-7, vol 2, 281-5, esp. 283. r ArbroatFo generan 1 hi 3 l see: Macki Crudee& n 1954/82; Walke Ritchi& r e , 121-2198763 d o n ,an ; now most especially Fawcett 1994a. This section of wall may predate the main construction of the church which Fawcett (1994a, 61-2) places in the 1190s. 32 Could the reverse relationship of the rebate and splay indicate this wall was originally the north wall of a small church or chapel located on the site of the existing north walk of the cloister? Fawcett 1994a, 62-3, does not comment on this feature of the portal. abovee . Se 19 n , 3 3 34 Discovery Excav Scot 1962 , 'Excavatio55 , 196n i 2 abbee soutth f hyo churc reveales hha d e partth f so foundatio cloistee th f no r arcade sectioa footinge d th e 12th-13t an f ,th no f o s h century south range'; Discovery Excav Scot 1963, 60, 'Excavation in the area of the monastic buildings revealed further detail soute th f hso range, largely reduce foundations's it o dt ; Discovery Excav Scot 1983 , 'Limite60 , d excavation was carried out within the slype, and in the area of the E. processional doorway and E. cloister orden i , establiso t r h original occupation surface advancn i s programma f eo f restorationeo ' (also reported by Youngs, Clark, & Barry 1984, 262). 35 For a measured elevation and plan see Galloway 1878, pi 8. 36 The newel stair and its vestbule will be discussed further below, and is fully described in Appendix 2. measurea r Fo d 7 elevation3 , plaprofiled nan Gallowae se s y 1878. 8 , & p 7 i 38 At present, the remains of three barrel vaulted rooms occupy the northern two-thirds of the west range. The two southern portals of these rooms appear to be totally uniform in build, but are certainly later constructio earlienon r foundations. Fawcet suggestehas t possibilitdthe y tha seriethe t latesof r barrel- vaulted rooms which occupy the site of the west range of the cloister might be pre-Reformation, on analogy wit restructurine hth wese th f t go range t Dryburgsa h (Richardso Tabrahan& m 1937/87, 21), Dundrennan (Richardson & Tabraham 1981, 16-17) and even more extensively at St Andrews. At these sites, though, vaulted cellars were created which seem different fro e domestimth c character implied by fireplaces and windows of the range at Kilwinning. 39 See above, n 28, and esp. Cruden 1986, 84 pi 25 and 86. RCAHMS 1915, 138^4, referred to the slype and chapter-house doorways as having been 'modernized', but without offering further explanation. 40 MacGibbon & Ross 1896-7, vol 2, 81, 84, 85, figs 506 & 507. The west wall was considered a 'somewhat late design' becaus capitale eth follot s no jam (sice d wth di )b mouldings. 41 MacGibbon & Ross 1896-7, vol 1, 453, fig 413. 2 4 MacGibbo Rosd nan s(282) 3 1896-7 70 , 2832 g ;l fi Fawcet, vo , ; Shephert51 1985bg fi , d 38 , 1986o n , (date5 lat54e th 11 e, o d13tt h century). 43 See Garton 1987, pi 4 & 74, 81 (for date). 4 4 Hope 1900(nort9 23 , h transept portal) (cloiste2 25 , r portal) Hopr o , e 1902, 19-20 ; Kaines-Smit32 , h 1914, 289, 292; Dickinson 1965, plan, 'mid to late 12th century'; Dickinson 1967, 51-80 (there is no mentio cloistee th f no r portal). 45 Cowan & Easson 1976, 88, 91 (a date of c 1153 is now preferred); initially a priory, it was elevated int abben oa 1235n yi . MacGibbo Rosn& s (332 2 1896-7 75 , 324 ,2 g l datefi , vo , f do towardd en e sth the 13th century); Paterson 1925-6, 232, 245 (chapter-house before 1265?). Paterson & McRoberts 1984, 23, did not mention the portal and merely dated the chapter-house to the 13th century. For Inchcolm, see also Gifford 1988, 242-6, esp. 245 (242, chapter-house after 1235); RCAHMS 1933, no 14, 6-1 Walke; 85 5 g Ritchi(espfi r& , 9) . e, 118-19 198761 o n , . 6 4 MacGibbo Rosn& s 1896-7 471-4, 1 l Wilsovo , ;C McWillian ni m 1978, 276-7. 47 MacGibbon & Ross 1896-7, vol 1, 366-70; C Wilson in McWilliam 1978, 273-4. 48 MacGibbon & Ross 1896-7, vol 2, 286-7 ('mouldings are of thirteenth century forms'), fig 708. 49 Leask 1958, 40-1 ('stands somewhat apart from other Irish work'), pi III; Stalley 1987, 171, fig 71. considereo s s wa t FawcetI y db 0 5 t 1994b, 115-16 3.3g fi ,4 ('... another exampl westere th f eo n Scottish late medieval predelictio revivinr nfo g earlie t offecomparisony no an rd r di forms') e h t r furtheso bu , r analysis. For a discussion of Romanesque revival architecture as part of a 'wider movement intended to express Scottish identity', see Campbell 1995, esp. 302-12; the differences - or similarities - 872 SOCIET ANTIQUARIEF YO SCOTLANDF SO , 199:5

betwee revivalistie nth coriginale formth d an s , however t considereno e ar t discussed), stylee no d(i s i . One is left with the impression that what constituted 'Romanesque' to the later period was primarily the semicircular arch. 51 MacGibbon & Ross 1896-7, vol 3, 330-7; fig 1261. The display inside the church describes the nave and its portal as an extension of the original church of c 1452. e so-calleTh 52 d Bride's Halkerstoe Doorth n i , n Tower, datinlate th e o g15tt h century: Fawcett 1987, 8-9, 27. abovee . Se 16 n , 3 5 54 The 13th-century forms, especially of the bases, of the lateral choir portals of St John's Church, Perth, mose ar t likel resule yth f 19th-centurto y restoration Fawcete :se . 29 t , 198723 , ,22 5 5 Dobie 1876, pi facin ('Doorway6 25 g f Kilwinnino s g Abbey') measurea r Fo . d elevation, pland an , detail capitalsf so , archivolt decoration basd an ,e profile Gallowae se s y 1878. 6 , & p5 i 6 5 Onl merese yth tdecoratio e tracth f eastere o th n no n bases shows a , Galloway nb y 1878 stils i , , pl6 i visible eastere onlth w n yno o ,n inner base. 57 Par f thio t s are patches ai unequao tw n di l vertica t e clealthiy stripsno Th s wh r s i . shoul t i so : e db entire section of wall east of the portal may have been patched more than once; it is also likely that the (missing) respond itself was an insertion at the time of the reconstruction of the nave. r DryburghFo e Crude8 5 se , nr Kirkliston 1986Fo . ,86 e above se (pace7 , 4 n , Wilson McWillian i , m 1978, 274, who describes the bases as 'waterholding'). 9 5 Water-leaf foliage appeare choie th Yorf n rdo i k Minster rebuils a , Archbishoy tb p Roger Pont 1'Eveque (1154-81), fro late mth e 1150 t sRipo1166o t a d n,an Minste e 1160sth n i r, anticipatin 'nearle gth y exclusive relianc waterlean o e f foliag latn i e e twelfth-century northern English churches, both Cistercia non-Cistercian'd nan : Wilson 1986 r Kels, Fo 104-5 . o 63 see n , : Baldwin , 122198570 ;o ,n Cruden 1986, 42-54, date wese dth t transep 1160c t . 60 It appears in one of the few surviving capitals of the dado arcading of the west wall of the south arm of the transept. For St Andrews, see Cant 1974, 78, 84, or 1976, 12, 21; Walker & Ritchie 1987, no 64, 1123-?. On water-leaf at St Andrews, see Thurlby 1994, 53^1. 61 Garton 1987, 78-80, and 81 where she suggested a date 'around the 1180s'; Fawcett (1994a, 64) also place beginnine dth Jedburge th f go h nave 'hardl . earliey.. r tha 1180s'e nth . WilsoC McWillia n i 2 6 m 1978 presenc e basi,e 274th th f sn o , o water-holdinf eo g bases. 63 Stiff-leaf foliage appear havo st e appeare Englann di d1180c Fergussoe Se . citese h ; n ,198440 n , ,64 as early dated examples, St Frideswide, Oxford, 1178-80 (west bays of 'chevet'), , 1182^- (Corona Chapel), Dover Castle, 1181-7 (chapel). This dating seems to be confirmed by Halsey 1988, 132, 159 (133: east bay of nave erected c 1180-5). 64 For the shaft in the south-west corner of the crypt begun by Bishop Jocelin (1174-99), c 1200, see: Radfor Stoned& s 1964, 222-3, pi LXIIIa; Fawcett 1985a, 25-6; Fawcett 1990, 120, 125. 65 Wilson 1984, 130. 66 MacGibbon & Ross 1896-7, vol 2, 30-52; Cruden 1986, 54-60. Fawcett (1994a, 61) considered it likely tha constructioe th t churce th f nt Arbroato h a t begino nd huntidi 1190se th l . 67 RCAHMS 1956, 198, pi 41, fig 239; Garton 1987, 70-5. Garton (ibid, 74, 80) identified the capitals of the west portal as two-tiered crocket, admitting that they differed in that respect from the crocket capitals in the nave and therefore suggested northern French Gothic influence. She may well be correct r dat r Jedburghhe efo s a , , 'aroun e 1180s'dth , would sugges appearance th t f stiff-leaeo f there woul rathee db r precocious considerin s advengit Englann i t same dates di th eo dt period (see above, n63) 68 Fawcett (1994a, 66, n 16) also suggested that the choir of Coldingham Priory (Berwickshire) should more likely be placed in the last years of the 12th century than after 1216, therefore an(other) example of fully developed stiff-lea 1190se th n i f. 69 MacGibbon & Ross 1896-7, vol 3, 7-28. Howell 1929, 68-9, dated the inner doorway (pi facing 68) to the 14th century. 70 Wilson 1984, 130-1, 134-5; rudimentary stiff-leaf had appeared in the capitals of the north aisle blind arcade 1195c , . McALEER: TOWARD ARCHITECTURAN SA L HISTOR KILWINNINF YO G3 ABBE87 Y buildine extendee Th th f go 1 7 d east limb under Bishop William Bondington (1233-58 certainls wa ) y 1242y Fawcete undeb se y : rwa t 1985a, 27-8 r 1990,o , 109, 125-6. FawcetR 2 t7 inform tha e foliage sm th t partl s ewa y renewe plaste n di 19te th hn ri century . 73 MacGibbon & Ross 1896-7, vol 2, 203-15. 74 MacGibbon & Ross 1896-7, vol 1, 376, fig 339; Stevenson 1985, no 59, 98. The inner continuous jamb/order is carved with spaced zigs and flanking bar-linked diamonds. (The outer order is ringed with bobbins.) 75 Garton 1987, 73. 76 None the less, C Wilson in Me William 1978, 274, 'late 12C. 77 RCAHMS 1946, 116 (north-west portal, c 1215; central and south-west portals, c 1220), pi 20, fig 167; Ritchie 1985, no 49, 97-9. 78 Garton 1987, 76-7. standine Th g9 remain7 s have been date d1175-80c : Fergusson . 1984, p59 i 62 , 80 It is stored in the cloister walk. 81 These dates seem to parallel the appearance of the motif in England: see Garton 1987, 74 and n 51 for examples. 82 Garton 1987, 75-6. 83 See above n 42. 84 Howell 1929: south-west cloister portal, 59, 70 (13th century); north porch, 68-9, pi facing 68; west doorway, 70, pi facing 68 (13th century). 85 MacGibbo Rosn& s 1896-7 , 121^2 l vo ,5 (esp (132)9 . 54 133) g ; fi Richardso, Mackintosn& h 1934, 7-8 (1950, 7-8); Shepherd 1986, no 47, 108-9. 86 Gifford 1988, 366. Fawcett disagrees with Cant 1974, 21-2 or 84-5, who suggested that the conventual buildings were built under Prior John White (1236-58); Cruden 1950, 20 (chapter-house entrance ensemble, mid-13th century) thinke h ; s they must have been started earlier. 87 Mackie & Cruden 1954/82, 27-8, pi 3: c 1190. The north portal at Arbroath (at the west end of the north aisle, immediately proceeding the tower), however, has a strongly pointed arch and finely moulded orders: idem, 26, c 1200. Surely a date of 1190 for the west portal is far too early; the retentio semicirculae th f no r arch givearchain a t i sr whic cai thinnese s beliehi th motife y th d b f so s splayee th d dan structur jambse th f eo . 88 Howel , pil 59 facin1929 , : Transitional'22 ,g 29 . e outeTh r orde 9 8 r herpeculiaa s ei r motif forme fivy db e thin rolls bandebilletsx si y almosn db a : t identical patter middle s founni th n o de e doorwaordeth f o r f Bishoyo p Pudsey's Hal Durhan i l m Castle, date d1160-73c Furnese th r Fo s. porta Fergussoe se l n 1984, 60-1 , date, p56 i d1160/5-75c . 90 The Calder west portal has been dated c 1175 (cf the crossing consoles at Furness): Fergusson 1984, 61-2, . p58 i 91 The Jervaulx nave is dated c 1185: Fergusson 1984, 84, pi 87 (cf 83, c 1190). 92 The west portal at Abbey (Yorkshire), with three of its four orders and the double orders of its inner jamb decorated with vigorous layered chevro thren ni e different pattern fourte th - sh ordea r large roll covere latticewory db k itself execute rola n - ldshouli mentionee db d capitalheree Th . f so the detached shafts are uniformly of the water-leaf type, providing a standard dating for it of c 1180, and a description of it as Transitional' in style. Fergussoe Se 93 n (511984 3 , p1 i ,44 ,Kirkstal l west portal, 1165 , 82-384 ; , Byland west portals 1190)c , . 94 For an early engraving of the arch to the chapel bay of the aisle, see Dobie 1876, pi facing 256 ('Doorway f Kilwinninso g Abbey'). FawcetR 9 suggeste5s ha t d- wit h reservation s- rootha e th tf line might represen latea t r alteratiod nan that the dormitory could originally have stopped short of the transept with just a passage connecting the night stairs to the dormitory. I think this is unlikely because the raggle is a projection from the wall face, rather than a cutting, so the roof would seem to have been planned with the gable wall. Appendie Se fullea r fo r 6 x] 9 descriptio thif no s problematic area. 7 9 MacGibbo Rosn& s 1896-7 , 388-91 l vo , 8 351)g (planfi 9 ; 1899b38 : , 57-97, plan ; RCAHMpII i S 1914 398o n ,, 217-22; Richardson 1949; Fergusson 1973 (remodellin g1180-5)c ; 1984, 66-7 Stel; ,83 l 1986, no 73, 151-2; Cruden 1986, 76-9 (c 1160). 874 SOCIETY OF ANTIQUARIES OF SCOTLAND, 1995

98 MacGibbon & Ross 1896-7, vol 3, 132-9; 1899c, 199-233, plan pi I; Stell 1986, no 72, 150-1. 99 MacGibbon & Ross 1896-7, vol 2, 334^4 (plan: 335 fig 756); 1899a, 3-55, plan pi III; RCAHMS 1914, no 380, 200-8, plan fig 151 (202); Cruden 1986, 137^3; Stell 1986, no 71, 148-9. 0 Fergusso10 : nort17 n hg 1984 fi transep , 83 , t remodelled 1185; 12-shaft piers, influenc f Bylando e e Se . also Fergusson 1973, 232-43, esp. 237-9. e Holyrooth f 10O 1 d piers, Hoe ) observedy67 (1987n 5 ,25 , 'the coursed compound piers bear little relatio usuae th o lnt norther nHoe w cluster'no y e 1994Se . , 89-92. 102 MacGibbon & Ross, 1896-7, vol 2, 146-60. See also specifically Hoey 1994, 94. 103 Well before Exeter, 16 shafts had appeared in the in the north in the coursed piers of Cistercian Rievaulx 1215-25c , . 4 Hoe10 referre) yclusteree 67 (1987th n o d5 t 25 ,d pie t Kilwinninra 'interestinn a s ga forlorf gi n exception inviolabilite th o t clusteree th f yo d pier'. 105 Prickett 1835, 39-50 (architectural description), esp. 48; Dykes 1854, 40-54, esp. 48-9, pi facing 49; Pevsnerl972, 196-9. 106 Fergusson 1984, 81, and n 46; 82, 1170s; the principal shafts were keeled (an eight-shaft pier was used firsd an soute t i nfouth m har r nave bays). 7 Brakspea10 r 1922, 1-59, esp. 22-9; Pevsne Harri& r s 1989, 112-14. Wilson (1986) 80 , n 108-9 , 9 g fi , derive piee dth r form from Kirkstea km)0 d(1 , whic arguee hh begus e 1150sth dwa n nf i Fergusso c : n (1973, 238 n 2) who thought it was influenced by Byland. 12-shaf8 A Hoe 10 . 64 y t n 1987hav piey 5 soute ma re25 , th occurre f h o wes e arcadd d th Ol t en t d a t ea Malton Priory (fd. 1150): Hoey, 254 n 62; Pevsner 1966, 232-3. For pier design in Early in east-central Scotland, see Hoey 1994. 9 Fawcet10 t (1994a ) suggeste67 , d that these aspect Arbroatf so h derived from Hexham begu1180se th n ni . A similar discordance between (choir) respon arcadd dan e pier occurre t AndrewsS t da wasd an , , apparently common in the north: see Cambridge 1977, 280, and Hoey 1994, 87. 110 Hoey (1987, 251 n 43) argued for the shafts starting at abacus level rather than from the floor as suggeste Wilsoy db n (198681)n 9 .10 , 111 Hoey 1987, 251 n 43. 112 Wilson (1984, 130) dated the nave (including aisle walls) of Holyrood between c 1195 and c 1230, Fawcett (1985, 27, 29) the choir of Glasgow from about 1242 to c 1275. Hoey (1987, 253 n 55), with regarGlasgoe th o dt w choir piers, noted tha t reallt no 'the e y yar cluster s because they lack four-axis symmetry and include responds for some of the shafts'. unusuae th r 11Fo l 3Holyroo Glasgod dan w piers mattee s welth a , s f verticaa lo r l articulation Hoee se , y 1994, 92-3. wese 11Th t 4lanter n towe smald rha l oculi flankin roofe apee th gth f s xo agains , witit t h quatrefoil-shaped inner frames. 115 Encircled quatrefoils flank the gabled niches above the west portal, c 1270: see Fawcett 1993, 21. reviea discussio the 11For wof 6 n surroundin sourcegthe Ripon'sof s desig almosn(in t every respecbut t the pierced tympana) Hoee se , y 1986, 250-5. 117 Pevsner 1957, 173: begu n1180c alse .Se o Fawcett 1994a, 66-7: gallery level, 'a date inte 13toth h century'. 118 Kidson 1977, 1/3/31-3, 59-60, 100-5, 165-6, 173; 1978, 1/5/68-70; Pevsne Harri& r s 1989, 450-5, 459,461. 119 Hoey t commen(1986 no appearance d th )di n o t f thieo s elemen decorative th f o t e vocabularye th n O . relation between Yor Lincolnd kan espe se ,. ibid, 232-5. 120 Coldstream (transep3 4 1986 n 9 ,14 t triforum: 'plate tracery' patter spandrel)e th n i . 121 Clapham 1952, 11-14. abovee ; als12Se 68 o2 n ,Crude n 1986, 136,. p50 i earliese 12Th 3 t pierced tympan buildina f ao Scotlann gi d were probably those inserted unde gallere th r y archechoie th f Jedburghn o rsi , followe same (ory th d b t e a , time as? nave )th e gallery tympanae th : shap vesica s ei a (pointed oval). 124 South transept: Fergusson 1984, 83, c 1190; Cruden 1986, 60, pi 16, complete by 1214/15. Fawcett McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY I 875

(1994a, 65) suggested that the inspiration behind Arbroath's south transept fagade was Jedburgh's west front behind an , d Jedburg wese hth t fron f Kelsoo t . 5 Thre12 e tall lancets wit vesica-shapeha d oculus abov tallese eth t centra (ane on ld short lancets ovee th r flanking lancets) dominat easte eth facad f Tynemoutho e c 1190-1200, : Knowles 1910, 1-50, esp. 12-15, pi V; Fergusson 1984, 83 (c 1200), pi 86. 126 Previously broadly dated 1181-1254 (consecrations): Laurence 1950-1, B1-B50, esp. B30-1, pi II (facing B40); Baillie Reynolds 1946 1970H ;VC , 226-30 t Croxden, long(a espo 9 tw . 22 e lancetth , s appear to have been crossed by the roof of the east range as happened at Kilwinning, see pi facing Hoee se yw 1993229)No . , 38-9 (date). 43 , 127 Hoey 1986, 239, singled out these fa9ades as parallels (if not exactly prototypes) for the domination of maie th n elevatio facade singla th f y no b e e tie f lanceto r s oppose (a straditioe maio th o tw dt n f no tiers, eg Hexham north transept, Whitby east and north fafades, Ripon west front, Southwell east end, Brinkburn west front Beverled an , y main transept addee whicb o t , y d hma Rievaul x east front). e MacleoSe 8 d 12 1878, 132-7, especiall holdinge recore th y gifth a f o f td f o Andre o s f Murrawo e y(d Burr), Bisho Morayf po , 1222^2 Alany ,b 122c , 2 (items 2-6). 9 Macleo12 d 1878, 116. 130 Excavations took place in the early 1960s, but it is not known exactly what was discovered at that time, as the report is not available or has been lost. There are brief and tantalizing references to the excavations in Discovery Excav Scot, 1961, 54: The work of construction was well advanced by the e 12tth hf o centurys suspende d wa en t bu , churce de wes th th befor f r t o ho wese claustra d th e en t l range were completed. Wor recommences kwa mora o dt e ambitiou se 13t plath hn ni century t thibu ,s also was never fully carried out'. Discovery Excav Scot 1963, 60: 'Excavation within the abbey church revealed the sill wall for the nave arcade, and gave further evidence that the nave and west end of the abbey church (for which an ambitious extension was planned in the late 13th) were never completed'. Gallowae 13Se 1 y 1878, p. 2 i remaine barrel-vaultew Th lo f 2 o s 13 d room occupinw sno sitee gth , perhaps formin groune gth d floor of units rising two, three, or even four stories, appear to represent a post-Reformation utilization of the range, perhaps equally drastic as the rearrangement of the west range at St Andrews. See above, n 38. 133 A single battered but somewhat similar capital from a wall shaft survives amongst a collection of miscellaneous fragments stored in the west porch at Holm Cultram: the bell has two rolls below the moulded upper rim. 134 MacGibbo Rossn& , vo , 344-82l2 . 135 Simpson 1923^, 227-38, esp. 237-8 (north-east crossing pier base: fig 5 on 238); Close-Brooks 1986, , 115n isolateon 53 A . d exampl double on f eo e roll bas easternmose founs ei th n di t north responf do the north choir aisle (former Lady Chapel sacristyw no , t Dunblana ) e Cathedral (Perthshire) c 1233, : MacGibbon & Ross 1896-7, vol 2, 86-112, esp. 92 (time of Bishop Clement, 1233-58). 6 Ver13 y wide fillet majoe e founth ar s n r o d maie shaftth f no scrossin g pier t Lincolna s , after 1237/39; narrower ones were to appear in the Angel Choir, c 1260. Wide fillets still make an appearance in the south aisle of Fortrose Cathedral in the 15th century: Close-Brooks 1986, no 54, 116. 137 Further excavations were carried out in 1985 in the area of the west transept. See the report (typescript), Ewart 1985filee Historif th o s n i , c Scotland briefA . notice appeare Youngsn di , Clark& Barry 1986, 194; I am grateful to R Fawcett for bringing this to my attention. 138 See above, n 11. e GrosSe e 9 17813 9 1790, & p 2 i l betwee,vo n 212/13; Denholm, pi facing 241drawingo e Tw . th n i s Hutton Collection, National Librar f Scotlandyo , related e Grose ar , th o et engraving. One, Adv. 30.5.22MS . pencil28in o i n , s ,i , executed with some ruled lines, fro msomewhaa t elevater do aerial viewpoint writtes i t i n inkn o ;i , 'Kilwinning 1789 / communicate. Revde th Pollack. y dW .b 1 ./ Steeple / 2 supposed to be part of the house / Church Yard'. The second, Adv. MS. 30.5.22, no washk in 28hviewpoins d it ; executes an i , n pe tn di correspond s more closel engravine th o yt g except north/soute th tha l al t h plane showe sar paralles n a pictur e th o t l e plane. 140 National Library of Scotland, Adv. MS. 30.5.22, no 28g; it is identified as, 'Old tower of Kilwinning 876 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1995

Abbey as it stood in 1806. / Com'd. by the Rev'd. Mr. Sterling, Min'r. of Port Monteith.', in two different 19th-century hands. 141 This area was completely excavated in 1985; see Ewart 1985, 2 (The South West Tower), for a description of the nature of the foundations encountered. 142 One block of the south jamb of the west portal was formerly in situ; it bore traces of the footings for two circular shafts and, missingpossiblyw otherso no tw s i f t . i o :, Thi t mentiones no bloc s kwa y db Ewart in his report: (1985, 1-2, The West Door). 143 The blocks south of the shaft are chamfered: according to Fawcett threshold blocks can be chamfered. The (blocked) middle portal in the south aisle wall at Dryburgh may be an example of this feature. summara r significance Fo th f problemye o 4 th 14 d ean s presente wese th ty db transept f thesso e buildings see McAleer 1991, 349-56. constructionae th r Fo 5 14 l sequenc severad an e l change f intentiono s s concernine e forth gth f m o Peterborough west fron Peere se t s 1906, 440-1, 444mord An .e recently, Reilly 1991, 208-5; 5 h 7(C thesie th currentls si y being revise publication)r dfo . 146 RCAHMS 1956, 240-2; Cruden 1986, 42-54. 147 For the west front of Arbroath see McAleer 1994, 70-83. At Paisley, the faces of the west tower piers toward navee sth widta , flate f 1.0ho ar ,, 4withoum t shafts. 148 See RCAHMS 1956, 244, fig 303 (on pi 56); Cruden 1986, 46, 50. 149 The naves of both abbey churches were retained for use as parish churches after the Reformation, until abandonned for new buildings in the 18th century. 150 Galloway 1878 Eware Se , . pt 4 i 1985 3 (The2- , Secondary West investigationse Wall)th r fo , primarily along the east face of the wall, which took place in 1985. 151 Galloway p; MacGibbo d i4 1878 an 5 9 , Rosn& s 1896-7 ; Barret, 80 vo , 2 l t 1913. 80 , 152 John Knox recorded that the Earls of Arran, Glencairn, and Argyle pulled down the abbeys of Fail, Kilwinning, and Crossraguel: Laing 1846-64, vol 2, 168. Grose 1789 & 1790, vol 2, 213, attributed the demolition at Kilwinning specifically to Alexander, Earl of Glencairne, in 1560: the destruction at this time, however havy ma e, been directe dwese ease thad morth th ten o t nt o whic et h continueo dt serv pariss ea h church. 153 See Ewart 1985, 2, 3. 154 Mackie & Cruden 1954, 33: Mackie & Cruden (Fawcett) 1982, 11, pi 8. 155 Stevenson 1985, no 51, 88-9; Cruden 1986, pi 56. 6 Gallowa15 y 1878. 91 ,

REFERENCES Anderson, A O 1908 Scottish Annals from English Chroniclers, AD 500 to 1286. London. Anderson, A R 1954 'Augustinian and Benedictine monasteries in Scotland: an enumeration (revised from John Spottiswoode's) of all the abbeys, priories and nunneries in Scotland at the Reformation', Trans Glasgow Archaeol Soc,Serw (1954)3 ,Ne 1 , 91-102. Anonymous 1814 'Description of Kilwinning Steeple with an account of its recent fall', Scots Magazine and Edinburgh Literary Miscellany, 76 (Oct 1814), 723-4. Baillie Reynolds 194K P , 6 Croxden Abbey, Staffordshire Ministr= ( f Workyo s Leaflet Guide). London. Baldwin, J R 1985 Exploring Scotland's Heritage: Lothian and the Borders. Edinburgh. Barrett 191M , Scottishe 3Th Monasteries f Old:o BriefA Account Housese th f o which Existed Scotland,n i before Protestante th Reformation, r Monksfo following Rulee th ofSt Benedict. Edinburgh. Barrow, G W S 1960 'From Queen Margaret to David I: Benedictines and Tironians', Innes Rev, 11 (1960), 22-38. Brakspear 192H , 2 'Bardney Abbey', Archaeol (1922)9 7 , J , 1-59. Brakspear 192H , 9 Tintern Abbey Guide. ednd 2n . London. Cambridge, E 1977 'The early building-history of St Andrews Cathedral, Fife: its context in Northern Transitional architecture', Antiq J, 57 (1977), 277-88. McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY 877

Campbell I ,199 5 'A Romanesque earle revivath yd Renaissancan l Scotlandn ei 1380-1513'c , SacJ , Architectural Historians, 54/3 (1995), 302-25. Cant, R G 1974 'The building of St Andrews Cathedral', Innes Rev, 25 (1974), 77-94 (also in Roberts, D ed 1976 The Medieval Church of St Andrews, Glasgow, 11-32). Chalmers, G 1807-24 Caledonia; or, An Account, Historical and Topographic of Northern Britain, from the Most Ancient Presente th o t Times. vols3 . London. Clapham, A W 1934 English Romanesque Architecture, II. After the Conquest. Oxford. Clapham, A W 1952 , Yorkshire (= Ministry of Works Guide-book). London. Close-Brooks J 198, 6 Exploring Scotland's Heritage: Highlands.e Th Edinburgh. Coldstream, N 1986 'Architecture from Beaulieu to the Dissolution', in Norton, P & Park, D (eds) Cistercian Art and Architecture in the British Isles, Cambridge, 139-59. Cottineau 193H 193& 5L , 7 Repertoire topo-bibliographique abbayess de prieures.t e vols1 . Macon. Cowan 198B I , 6 'Ayrshire Abbeys: Crossrague Kilwinning'd an l , Ayrshire Collections, 14/7 (1986), 267-95. Cowan, I B & Easson, D E 1976 Medieval Religious Houses: Scotland. 2nd edn. London. Cruden 195CathedralS e , Th 0 Andrewst oS f t RegulusS d an Church. Edinburgh. Cruden, S 1977 'The cathedral and relics of St. Magnus, Kirkwall', in Apted, M R, Gilyard-Beer, R, & Saunders, A D (eds) Ancient Monuments and their Interpretation: Essays Presented to A J Taylor. Londo Chichestern& , 85-97. Cruden, S 1986 Scottish Medieval Churches. Edinburgh. Dempster, T 1622 Apparatus ad Historiam Scoticam Lib. II. [sic; Lib.II is a misprint for Lib /] Accesserunt martyrologium Scoticum Sanctorum DCLXXIX. Scriptorum Scotorum MDCIII. Nomenklatura, etc.. 3 pt . Bologna(in1) . Denholm, J ('J.D.') 1802 'Description of Kilwinning Abbey', Edinburgh Magazine, or Literary Miscellany, New Ser(Apri9 1 , l 1802), 243-4. Dickinson Origins195e C J ,Th 0 Austine oth f Canons theird an Introduction into England. London. Dickinson 196C J ,5 Furness Abbey, Lancashire. Ministr= ( Workf yo s Guide). London. Dickinson 196C J , 7 'Furness Abbey archaeologican a ; l reconsideration', Trans Cumberland Westmorland Antiq Archaeol Soc, ser2 (1967) 7 6 , , 51-80. Dobie (edS J , ) 1876 Cuninghame, Topographized Timothyy b Pont, A.M., 1604-1608, with Continuations and Illustrative Notices Latee th y Jamesb Dobie ofCrummock, Scot.A FS Glasgow. Dykes 185H 4W , 'Oprior e . Marnth S f Bridlington't yyo a . Assoc Architectural Socs: Reports Papers,d an 3 (1854), 40-54. Edwards 192,J 4 'Kilwinning Abbey', Trans Glasgow Archaeol Society, Serw (1924)7 , Ne , 330^1. Ewart, G 1985 'Kilwinning Abbey Excavations 1985'. (Unpub TS in files of Historic Scotland.) Fawcett 1985R , a Glasgow Cathedral. Edinburgh. Fawcett, R 1985b Scottish Medieval Churches. Edinburgh. Fawcett 198R , HistoryA 7 of Saint John's Kirk, Perth. Perth. Fawcett, R 1990 'Glasgow Cathedral', in Williamson, E, Riches A, & Higgs, M The Buildings of Scotland: Glasgow. London. Fawcett, R 1993 Elgin Cathedral. Edinburgh. Fawcett 1994R , a '' Higgittn i , J (ed, ) MedievalArchitectured an t Diocesee Ar th t S n i f o Andrews. (= Brit Archaeol Assoc Conference Trans, 14), London, 61-9. Fawcett 1994R , Architecturale bTh History f Scotland:o Scottish Architecture from Accessione th e th f o Stewarts to the Reformation, 1371-1560. Edinburgh. Fergusson, P 1973 The late twelfth-century rebuilding at ', Antiq J, 53 (1973), 232-43. Fergusson, P 1984 Architecture of Solitude: Cistercian Abbeys in Twelfth-Century England. Princeton. Galloway 187W , 8 'Remark existine th n so g buildin t Kilwinninga g Abbey', Archaeol Hist Collectionsr Ay and Wigton, 1 (1878), 89-114. Garton, T 1987 'The transitional sculpture of Jedburgh Abbey', in Stratford, N (ed) Romanesque and Gothic: Essays for George Zarnecki, 2 vols, Woodbridge, 69-81. Gem 198R , 9 'Melbourne, Churc t Michae S t Mary' S f ho d Coopern an i l, (edH Nottinghame N ,) Th Area (= Archaeol , 146J , Supp), 24-30. 878 SOCIETY OF ANTIQUARIES OF SCOTLAND, 1995

Gifford, J 1988 The Buildings of Scotland: Fife. London. Gilyard-Beer, R & Coppack, G 1986 'Excavations at Fountains Abbey, , 1979-80: the early development of the monastery', Archaeologia, 108(1986), 178-83. Grose 178F , 1799 & Antiquities e 1Th of Scotland. vols2 . London. Halsey 1988/199R , e 12th-centur1Th y churc t Frideswide'S f ho s Priory', Oxoniensia, (1988)3 5 , 115-67 (also in Blair, J (ed) 1991 Saint Frideswide's Monastery at Oxford: Archaeological and Architectural Studies, Wolfeboro Falls (N.H.), 115-67). Hoey 198 L ,e 13th-centur6Th y transept Yorf so k Minster', Gesta, 25/2 (1986), 227-44. Hoey 198L , 7 'Piers versus vault shaft Earln i s y English Gothic architecture' Architecturalc So J , Historians, 46/3 (1987), 241-64. Hoey 199L , 3 'Croyden Abbey' Maddisonn i , J (ed, ) Medieval Archaeology Architectured an t Lichfielda (= Brit Archaeol Assoc Conference Trans, 13), London, 36-49. Hoey, L 1994 'Pier design in early Gothic architecture in east-central Scotland, c 1170-1250', in Higgitt, J (ed) MedievalArchitectured an t Diocesee Ar th t Andrewsn S i f o Brit= ( Archaeol Assoc Conference Trans, London), 14 , 84-98. Hope, W H St J 1900 The Abbey of St Mary in Furness, Lancashire'. Trans Cumberland Westmorland Antiq Archaeol Soc, ser 1 (1900) 6 1 , , 221-302. Hope, W H St J 1902 The Abbey ofSt Mary in Furness, Lancashire. Kendal. Howell 192R A ,9 Paisley Abbey: History,s It Architecture Art.& Paisley. Kaines-Smith, S 1914 '[Furness:] Abbey', in VCH, Lancashire, vol 8, 287-98. Kidson (edP , ) 1977 Lincoln: t Hugh'sS Choir Transeptsd an (Courtauld Institute Illustration Archives, Archiv . CathedraleI Monastic& Buildings Britishe th n i . London Isles,3) . Pt . Kidson 197P , 8 Lincoln: Gothic West Front, Nave Chapterd an House (Courtauld Institute Illustration Archives, Archive I. Cathedral & Monastic Buildings in the British Isles, Pt. 5). London. Knowles Hadcock& D , 197N ,R 1 Medieval Religious Houses:England Wales.d ednan d 2n . London. Knowles 191 H e prior 0Th W , y churct Oswin Mart S S d f hyo an , Tynemouth, Northumberland', Archaeol, J 67 (1910), 1-50. Laing (edD , ) 1846-6 Workse 4Th of John Knox. Edinburgh. Laurence, M 1950-1 'Notes on Croxden Abbey', North Staffs Field Club Trans, 85 (1950-1), Bl-50. Leask 195G H ,8 Irish Churches Monasticd an Buildings, . GothicII Architecture Ao t D 1400. Dundalk. McAleer, J P 1991 'Le probleme du transept occidental en Grande-Bretagne', Cahiers de civilisation medievale, 34/3^ (1991), 349-56. McAleer, J P 1994 The west front of Arbroath Abbey: its place in the evolution of the twin-tower fajade', in Higgitt, J (ed) Medieval Art and Architecture in the Diocese of St Andrews (= Brit Archaeol Assoc Conference Trans, 14), London, 70-83. MacGibbon D & Ross, T 1896-7 The Ecclesiastical Architecture of Scotland. 3 vols. Edinburgh. MacGibbon D & Ross, T 1899a The Abbey of the Sweetheart or New Abbey, Kirkcudbrightshire', Ayrshire Galloway Archaeol Assoc Collections, (The0 1 Five Great Churches of Galloway) (1899), 3-55. MacGibbon D & Ross, T 1899b 'Dundrennan Abbey', Ayrshire Galloway Archaeol Assoc Collections, 10 (1899), 57-97. MacGibbon D & Ross, T 1899c '', Ayrshire Galloway Archaeol Assoc Collections, 10(1899), 199-233. Mackie, R L & Cruden, S 1954/82 Arbroath Abbey, Angus (adapted by R Fawcett). Edinburgh. Macleod 187W , 8 'Collections toward histors a monaster e th f yo f Kilwinning'yo , Archaeol Hist Collections Ayr and Wigton, 2 (1878), 115-222. McWilliam 197 C ,Buildings e 8Th of Scotland: Lothian (except Edinburgh). Harmondsworth. [FourcheuxMontrondM ] J M C e [ 185,d ] 6 Dictionnaire s abbayesde t monasterese histoireu o s de etablissements religieux eriges toutn e tempstousn e t Mignee P lieuxJ = ( , Encyclopedic theologique, 16) Paris. Paterson 192 e developmenW J ,6 Th Inchcolf o t m Abbey', Proc Antiqc So Scot, (1925-6)0 6 , 227-53. Paterson McRoberts& W J , 198D , 4 Inchcolm Abbey. ednh 4t . Edinburgh. Peers, C R 1906 'Peterborough Minster', in VCH, , vol 2, London, 431-56. McALEER: TOWARDS AN ARCHITECTURAL HISTORY OF KILWINNING ABBEY 879

Pevsner, N 1957 The Buildings of England: Northumberland. Harmondsworth. Pevsner, N 1966 The Buildings of England: Yorkshire, The North Riding. Harmondsworth. Pevsner 197N , Buildingse 2Th of England: Yorkshire: Easte th Yorkd Riding.an Harmondsworth. Pevsner Harris& N , J 198, Buildingse 9Th of England: Lincolnshire. (re n Antram)N ved d 2n , London. Prickett, M 1835 History of the Priory Church of Bridlington in the East Riding of the County of York. 2nd edn. Cambridge. Stones Radford196& G R L 4A E ,'Th C , e remain cathedrae th f so f Bishoo l p Joceli t Glasgona 1197)'wc ( , Antiq J, 44 (1964), 220-32. Registrum 1843 Registrum Episcopatus Glasguensis: Munimenta ecclesie metropolitan Glasguensis a sede restaurata seculo ineunte Reformatamd a I XI religionem. vol2 Bannatyn= ( s Maitlane& d Clubs, 95). Edinburgh. Reilly 199L , 1 'The Architectural Histor Peterborougf yo h Cathedral'. (Unpu thesisD Yorw bPh kNe . Univ). Richardson 194S ,J 9 Dundrennan Abbey. Ministr= ( Workf yo s Guide). London. Richardson Mackintosh& S J , 193B Cathedrale H , 4Th Kirk of Moray: Elgin. Edinburgh. Richardson Tabraham& S J , (revJ C , ) 1987 . ednh 4t . Edinburgh. Richardson Tabraham& S J , 198J C , 1 Dundrennan Abbey. Edinburgh. Ritchie, A 1985 Exploring Scotland's Heritage: Orkney and Shetland. Edinburgh. RCAHMS Royal Commission on the Ancient and Historical Monuments of Scotland. RCAHMS 1914 Fifth Report Inventoryd an of Monuments Constructionsd an Galloway,n i vo . County2 l f o the Stewarty of Kirkcudbright. Edinburgh. RCAHMS 1915 Sixth Report and Inventory of Monuments and Constructions in the County of Berwick (rev issue). Edinburgh. RCAHMS 1933 Counties of Fife, Kinross, and Clackmannan. Edinburgh. RCAHMS 1946 Twelfth Report with an Inventory of the Ancient Monuments of Orkney and Shetland, vol 2. Inventory of Orkney. Edinburgh. RCAHMS 1951 The City of Edinburgh. Edinburgh. RCAHMS 195Inventoryn A 6 Ancient e oth f Historicald an Monuments of Roxburghshire. Edinburgh. Shepherd 198G 6A I ,Exploring Scotland's Heritage: Grampian. Edinburgh. Simpson, W D 1924 'Dornoch Cathedral: the high church of Caithness'. Proc Soc Antiq Scot, 58 (1923-4), 227-38. Spottiswoode, J 1677 The History of the Church and State of Scotland. 4th edn. London. Stalley, R 1987 The Cistercian Monasteries of Ireland. London & New Haven. Stell 198G , 6 Exploring Scotland's Heritage: Galloway.d an Edinburgh. Stevenson, J B 1985 Exploring Scotland's Heritage: The Clyde Estuary and Central Region. Edinburgh. Stubbs, W (ed) 1867 Gesta Regis Henrici Secundi Benedicti Abbatis (= Rolls Series 49), 2 vols. London. Stubbs 187W , 9188& Historicale 0Th Works ofGervase of Canterbury. Rolls= ( Series 73) vols,2 . London. Thurlby, M 1988 'The Romanesque priory church of St Michael at Ewenny', J Soc Architectural Historians, 47/3(1988), 281-94. Thurlby, M 1994 'St Andrews Cathedral-priory and the beginnings of Gothic architecture in northern Britain' Higgittn i , (edJ , ) Medieval Architectured an n Diocesee A th n i Andrewst oS f Brit= ( Archaeol Assoc Conference Trans, 14), London, 47-60. VCH (Victoria History of the Counties of England) 1970 Staffordshire, vol 3. London. Walker Ritchie& B , 198G , 7 Exploring Scotland's Heritage: Tayside.d Fifean Edinburgh. Wilson 198C , 4 'Holyrood: abbey church' Giffordn i , , McWilliaJ , Walker & Buildingse mC Th D , f o Scotland: Edinburgh, Harmondsworth, 130-41. Wilson, C 1986 'The as "missionaries of Gothic" in Northern England', in Norton C & Park, D (eds) Cistercian An and Architecture in the British Isles, Cambridge, 86-116. Youngs , Clark M Barry & S 198,B J , T , 4 'Medieval Britai Ireland nan 1983'n di , Medieval Archaeol,8 2 (1984), 203-65. Youngs, S M, Clark, J & Barry, T B 1986 'Medieval Britain and Ireland in 1984'. Medieval Archaeol, 30 (1986), 114-98.