Proc Soc Antiq Scot, 118 (1989), 219-229

Medieval corn-drying kilns at Capo, Kincardineshir Abercairnyd ean , Annemarie Gibson*

SUMMARY Radiocarbon dating of grain from circular corn-drying kilns Abercairnyat Capoand indicates that this type of kiln, traditionally regarded as being 16th- to 19th-century in date, was in use at least 500 years earlier.

INTRODUCTION In 1984, three corn-drying kilns were excavate Scottise th y db h Central Excavation t Unita e ,on Abercairny, Perthshire (now Tayside Region Capot a o ,tw 224 ) 4 Kincardineshir d (NG90 ) an N RN e (now Grampian Region) (NO 626 675) (illus 1). The three kilns were of the same basic design and varied onl detailyn i threl .Al e were isolated structures round ,ha d kiln-bowl shord san t flue werd san e sunk into hillsides (illus 2,3,4). Although corn-drying kiln thif so s typ knowe ear n from many partf so , verhavw yfe e bee subjece nth f archaeologicao t l investigation. Similar kilns have been traditionally regarde 16ts 17td a han h century A datDsomn d i reportee an ear havo dt e been stiln li use in the 19th-century (Martin Brown, pers comm; Close-Brooks 1980, 340). A letter written in the 17th century by James, Lord Drummond, mentions foundations being dug for round kilns (Fraser 1825, 130). Such documentation and the absence of earlier dating evidence have prompted the conclusion that these structure post-medievae sar daten li . However result e excavatione th , th f so t sa Capo and Abercairny, coupled with the recent excavations of similar corn-driers at Chapelton, Tayside (Pollock 1985, 367), indicate that these round kiln 13te th h s n werei centur e factn us i , n yi , and before. Grain-dryin necessara s gi y ste cropn i p productio temperatn i e climes where summer cooe sar l and moist. There are three main reasons for grain-drying: as part of the malting process; the seed for nexe th t year's harvest mus driee b tstorager dfo ; and, most importantly preluda s a , grindino et g (Fenton 1978,375) grain-dryine Th . g process attache theso dt e kiln wels i l documented (Firth 1974, 24-25; Feachem 1957,49-50). A fire was lit at the mouth of a covered flue and the heat produced was drawn alon passage gth enteo et kil e rth n bowl below raisea d floor, usuall woodef yo iror no n struts. graie dries Th nthiwa n do s floobeddina n ro straf go sucr wo h like. In the Northern Isles up to the beginning of this century, and on mainland Scotland before the introduction of agricultural improvements over the course of the 18th century, almost every farm had its own kiln (SAS, V18, Kilsyth, 309). Much of the previous research on corn-drying kilns has been based on extant examples in Orkney and Shetland where two basic types survive: four-sided kilns and

"Central Excavation Unit, HBM, 9 Melville St, Edinburgh 3 220 SOCIET ANTIQUARIEF YO SCOTLANDF SO , 1988

rounded kilns, both being integral parts of longhouses or barns. The Scottish mainland and Hebri- dean examples, however generalle ,ar y sited away fro homesteade mth circulae ;th r bow flud e elan ar sunk into the hillside and they often had no accompanying shelter. Fenton (1978,379) does point out that the apparently distinct regional variations in corn-driers (at Freswick, Caithness) may partly be e coincidencresule th th f o t f survivaleo , citin existence gth four-sidea f eo d kiln, found withina rectangular Viking structure.

ABERCAIRNE TH 224Y4 KIL90 ) (illuN N(N ) s2 During exploratory excavations in March 1984 in advance of the proposed levelling of a natural gravel mound at Abercairny, Perth (illus 1), a corn-drying kiln was discovered. Previous to this, in December 1983, a cist had been discovered on the top of the mound and excavated. situatekils e south-south-ease nwa Th th n do knole t poinsida th t f a leo t wher slope eth e rose sharply flubeed d t almos ean ha kile ncu t nTh .pi t level int hillsidee oth , wit flue hth e entering from south-easte th e circulaTh . r chambe 1-c ( r5 diametermn i stone-flaggea d )ha d floor gape th ,n si which were sealed by a pink clay. The lower portion of the chamber wall was of vertical, drystone construction, founde flaggee th n do d floor, which uppe e ros heigha Th 0-o e. c t r 7f m portio to e th f no

ILLUS 1 Locations of Abercairny and Capo GIBSON: MEDIEVAL CORN-DRYING1 22 KILN I S

turf pink grey claasd hyan dark grey ash

.B

top soil

turf ash charcoal 1 disturbance

e AbercairnILLUTh S2 y kiln

wall, constructe turvesf do , adde dfurthea thio t 0- c rs 7m height . These t continu turveno d sdi e eth vertica lowee l facth f eo r stone wal t splayelbu d outwards. Turves found withi kile n th nhav y filma el bee represeny n rooma para r f o o ft t collapsed turf walling lowee Th . r three course stone th f eso element were more smoke-blackened than blackenine thosth f eo abovep to ge probabl,th y marking positioe th raisee th f ndo floor. The flue was c 2 m long and had two construction phases. First, the flue had been cut into the hillside to a depth of c 0-9 m and a width of c 1 -6 m; clay was found adhering to its south side. Its width was subsequently reduce erectioe th y lin a b 1- o dverticaf f t ne 2o o m l stone slabbing which formeda 222 SOCIETY OF ANTIQUARIES OF SCOTLAND, 1988

new wall. Turves found within the flue may have been part of the flue cover. The floor of the passage was covered with laminated, charcoal-rich deposits containing carbonized grain, some of which ran unde secondare rth y stone walling described above thoughs i t I . t that these deposits represent debris from successive firings of the kiln. Slight scorching at the entrance to the flue probably marks the positio intt subsoifire e e poinsoute cu e o steth th s A th th . f nf po ho twa o t l wher flue eth e enteree dth chamber. The bottom of the step was almost level with the supposed position of the raised floor and it could be that this step marks the point at which the contents of the kiln were loaded and unloaded. ture fTh wallinkil e bace th n th f bowkt o g beea d ha ln extensively burn somd an t e burnt turved sha collapsed on to the floor. It is reported that this type of kiln 'not only required frequent reparations but was extremely liable to accidents by fire' (SAS, V18, Kippen, 349). It appears that the Abercairny kiln was abandoned after one such fire. The flooring spars were removed and a massive boulder (1-5 mxl-5 rolles mwa ) d int kile oth n bowl. No finds other than charcoal or carbonized grain were recovered from the kiln. Abercairnn A y estate improvemen 177f o 5p (SROt ma 1006P RH ,) show proximite sth e th f yo excavated kil Millntono nt far,a m tow whicf no surfaco n h therw eno tracs ei e als(illup oma se 3).Th show milsa l witkeyhole-shapeo htw d barn kilns nearby (author's interpretation) likels i t I . y that these large barn kilns represen typw f kilne eo na t introduced durin e periogth f agriculturado l improvement 18te th hn si century 1796n I . Statisticale th , Account Kilsytr fo h reported: '50 years ago . . . every farmer had his own kiln for drying his corns. A common kiln is now erected near each miln, where every farmer get grais sr boil'hi npe dried . 6 (SAS, t da V18, Kilsyth, 309).

Abercairnn ILLUA S3 y estate improvemen f 177o p 5 ma t GIBSON: MEDIEVAL CORN-DRYING3 KILN22 | S

The smaller and earlier Abercairny kiln is of the type owned by an individual farmer which was situated away fro othee mth r farm buildings becaus frequenf eo t accidents (Feachem 1957, 60)s i t I . possible that the Abercairny kiln was associated with an earlier, medieval, settlement on the site of Millnton.

THE CAPO KILNS (NO 626 675) Quarrying during June 1984 at the Tilcon Quarry, Kincardine and Deeside, revealed what appeared to be two stone-lined pits. The HBMD was notifed by Mr I Shepherd, the Grampian Regional Archaeologis smala d ltan excavatio undertakes nwa Centrae th y nb l Excavation Unit. Both features proved to be corn-drying kilns which had been destroyed by fire. The kilns lay 7-0 m apart on edg e gravea th f eo l outwash terrace abov Rivee eth north/north-wesa r n Norto k hEs t facing slope.

KIL (illuNA ) s4 rouna d Kildha kiln-chambenA diametern i 2-c r( 8 m ) wit hshora t flue (1- long5m ) leading into it from the north-west. The back of the kiln had been clipped during gravel extraction and a later machine cut (measuring 3-2x 1-2 m) had been sunk to floor level on its west side. The chamber had a stone-flagged floor bedded in clean pink clay. A turf wall was built on this clay and stood to a height of 1-2 m, reaching the top of the chamber pit. shalloflus e eTh wa w (0-1 wit, 5m) h clay turf-lined -an d sides smalo .Tw l pestholes, which were found straddling the outer part of the flue, probably indicate the position of a wooden structure. Vertical stone slabs propped agains ture t th f entrance walth chambee t la th havo ey t erma serveo dt strengthe poine fluth e t whict nth a tea loaded s kile hth nwa . collapseA d networ carbonizef ko d branche floo chambee e founsides d th th san rf wa n so do r and flue. The majority were round-sectioned (c 0-10 m in diameter) but some had been squared off. obviouo Thern s ewa s patter positioe th o nt squaref no opposes da round-sectioneo dt d pieces. This network of branches may represent elements of the collapsed remains of the raised floor of the kiln. dense Th e soot deposit adherin chambee wale th f th o l o gt r suggests tha raisee tth d floohighes wa r r than the present day ground surface. However, it seems more likely that this network of rounded branche collapses si d roofing material. A spread of bright orange peat- or turf-ash covering the carbonized wood was overlain by a thick layer of regular, smoke-blackened turves. The upper fill also contained turves and clean sand. Unti therw bees lno e ha n suc evidencw o littln h ho r ekilno o t s esa were covered fro elementse mth . Cape Th o excavatio provides nha firse dth t clear indication roofina f so g structure indebtem a I ( . o dt Dr A Fenton for his interpretation of the Capo evidence.)

Finds e rectangulaOn r bloc f stonko e (0-45xO-30xO-2 ) wit5m a circularh , central, indentation diameter)n (0-i 5m , forme dchambee parth f o t r floor otheo N . r artefacts were recovered.

KILN B (illus 5) rouna d Kilha d nkiln-chambeB diametern i 3- c ( r5 m ) wit shorha t flue (2- long0m ) leading int t froo i north-west e mth kile bee d e bacth nha n f Th .k o clippe d during gravel extractione Th . survivin gchambee parth stone-flaggef a to d rha d floor wit gapy han s seale cleay db n pink claye Th . turf walling of the chamber was built on this floor and reached a uniform height within the chamber. 224 SOCIETY OF ANTIQUARIES OF SCOTLAND, 1988

KEY

turf soot

burnt ash wood

clay burnt wood 2m

ILLU S4 Capo, kilnA GIBSON: MEDIEVAL CORN-DRYING KILNS 225

Quarry face (vertical)

soot

burnt wood

ash

ILLU S5 Capo, kilnB | SOCIET 6 22 ANTIQUARIEF YO SCOTLANDF SO , 1988

Dense soot-blackening on the turf walling indicated that the raised floor was positioned above this level. fluee Th t ,lined whicshallono s s wa , hwa w (0-15 smal o m)Tw . l circular (0-10 diametermn i ) depressions were foun indicateithen y do ma flurpositioe d e sideth th e an origina f n eo a f no l wooden structure. The chamber floor was covered in deposits of carbonized grain, sometimes several centimetres thick. Mixed with this were burnt straw and lengths of burnt wood, often flat in section. The flue contained similar piece f carbonizeso d grain o woon t .dbu These remains most likely result from accidental firing wheuse n burni e kile s ,th nth n wa t wood being collapsed strutraisee th f sdo floon ro which the straw and then the grain had been placed. A spread of bright orange peat- or turf-ash covere carbonizee dth d remains. Wedge f graveso l uppee founth chambee n di rth filf o l r suggest deliberate t tur backfillincu f e materiath d gan thi n li s contexcollapsee b y ma t d upper walling.

Finds uppee Th r fill produced several sherd f 13th-centurso y pottery uppe.e Halth f o rf stona f eo rotary quern was found incorporated into the stone floor of the chamber.

DATING The recent excavations at Abercairny and Capo have produced the first group of radiocarbon date sucr sfo h sunken corn-drying kilns. Sealed deposit graif so n froAbercairn e flooe mth th f ro y flue produced two dates, both of AD 1075±50 (GU 1927, GU 1928), and a single deposit of carbonized grain from the chamber floor of Kiln B at Capo produced dates of AD 1330±50 (GU 1929) and AD 1275±50 (GU 1930). A 13th-century date for the Capo kiln is supported by the recovery of a small assemblag f 13th-centureo y pottery fro uppes mit r fill (George Haggarty, pers com). Similaritien si both the construction and the siting of Kiln A and Kiln B suggest that they were broadly contemporary. 11th-centure Th Abercairnye datth r efo yearliese kilth s ni recordetr datfa o es thir dfo s typf eo kiln. The earliest date postulated for the north Scottish rectangular kilns is before the 14th century, on the evidence of an excavated example at Freswick, Caithness. The earliest known example of a north Scottish round kiln was excavated at Jarlshof, Shetland and is thought to date from the 14th or 15th century (Fenton 1978, 384, 379). As yet there is not a sufficient number of well researched and securely dated sites to devise any convincing chronological typology. llthe 13th-centurd Th -an y dates fro Abercairne mth Capd yan o roun medievae d th kiln d san l date suggested for the Chapelton examples indicate that this type of kiln enjoys a far longer history than previously suspected. Both radiocarbon dates are substantially earlier than the 16th- to 19th- century dating bracket traditionally attached to these kilns, and it is quite possible that the 11th- century date of the Abercairny kiln does not reflect the earliest chronological limit for the use of such kilns. Corn-drying kilns of one form or another have been in use since prehistoric times. Agrarian communities devise dvarieta methodf yo grain-dryinr sfo mand g an thes f yo e will have leftraco n t n ei archaeologicae th l record (Fenton 1978, 375). Rudimentary corn-drying kilns have been recordet da Unival, North Uist and Tigh Talamhanta, Barra and possibly Skara Brae, Orkney (Feachem 1957, 45). Given the existence of these early kilns, the llth- and 13th-century dates for the Abercairny and Cap opaucite kilnsth d reliablf yan ,o e dating evidenc corn-drierr efo s generally, automatic allocation of other corn-drier post-medievae th o st l perio questionede db musw no t . GIBSON: MEDIEVAL CORN-DRYING KILNS I 227

APPENDIX: CARBONIZED GRAIN

Ala FairweathenD r

ABERCAIRNY The sample was taken from a charcoal-rich layer on the floor of the flue-passage. The layer was homogeneou overlaid san laminatey nb d charcoal-rich deposits. Nearly halsample th f e consist cereaf so l grains (table 1), most of the remainder being unidentifiable carbon fragments mainly derived from cereal part wood an s d twig charcoal substantiaA . l proportio fractioe th f no n smalle rconsist m tham n1 f o s mineral grains, sand, soil peds, etc littl.A e less tha materiae n th hal f o f l largewoos wa dr m tham n2 charcoal, the remainder cereal remains. Cereal cereale samplse th comprisf th o abouf d so consise% an % 60 t e 52 hullef o t x d la barle e th f yo six-row type Hordeum vulgare (syn H polystichutri). Most of the grains show signs of having begun to germinate, the embryo being prominent and protruding. Separated embryos were also found. Many grains e rathear r flattened givin e appearancgth f havineo g been squashed whe sofa n tni condition. This i s consistent with the observed signs of germination. Oats make up about 40% of the cereal fraction and were of a small grain type consistent with A vena strigosa and/or A fatua (bristle or grey small oat and wild oats respectively). This is confirmed by the frequent finds of flower bases with fracture marks diagnostic of both species (ie both types found). There is no evidence for the presence of A saliva and all broken and entire are interpreted as A strigosa and/or A fatua. A fatua was present in significant quantities as estimated by the frequency of basal fracture scars, but the exact proportion of grains attributable to A fatua is hard to determine because the naked grains are not certainly distinguishable fro mstrigosa.A There are a very few culm nodes or parts of cereals other than grains and light chaff (glumes, lemmas, paleas) preservatioe Th . thesf no e part theid san r quantity suggest tha sample th t e consist graif so n which beed ha n threshed fro strawe mth , entire chafe .Th f would have become brittl broked ean n of situn fi during or after carbonization. This tends to be confirmed by the presence of several entire flowers with caryopses (grains) still within lemmas and paleas.

TABLE 1 Carbonized seeds from Abercairny f o s No Seeds found Carex flacca 1 sedge Chenopodaceae (Chenopodium or Atriplex sp) orache, or fat hen 1 Galeopsis tetrahit agg hemp nettle 24 Lapsana communis 2 nipplewort Polygonum aviculare knotgrass 1 Polygonum cfpersicaria redshank 32 Raphanus raphanistrum (portio f siliquano ) 1 wild radish Rumex acetosella shee2 1 p sorrel Rumex obtosifolius/crispus dock 1 Spergula arvensis corn spurrey 14 Stellaria media chickweed 1 f cerealo % portio f sampleno : BARLEY Hordeum vulgare (syn H polystichom) 60% OATS Avena strigosa} Avena fatua } 40% CEREAL volumf o % e excluding large charcoal% lump52 t includingsbu broken grains

REMAINDER Other parts: broken lemmas, paleas, charcoal, soilpeds etc 28% % 20 volume th f (Abouo e % consist20 t f largso e charcoal lumps) mm 2 , 48% 8 22 SOCIET ANTIQUARIEF YO SCOTLANDF SO , 1988

Eleven species other than Avena Hordeumd an were identified e fruitsTh . , seeds, achenesc et , (henceforth referred to as seeds) were mainly those which may be described as weeds of cultivation. Four weed species were frequently represented throughout the sample: Polygonum cf persicaria (redshank), Galeopsis tetrahit agg (hemp nettle), both species with comparatively large seeds, and Rumex acetosella (sheep sorrel) and Spergula arvensis (Corn spurrey) with smaller seeds. Polygonum and Galeopsis are tallish plants whereas R acetosella and Spergula are rather shorter and can be indicators of acid, rather impoverished, land. All the other species identified are to be expected. However, although only one nutlet was found, Carexflacca interestins i this ga s sedge tend favouo st r enriched damp groun flusher do s witha higher pH and mineral content. CAPO totae th f lAbou o sampl % 6 t e volume consiste f charcoado l lumps largee largesTh . r tthamm n2 totae th fraction f l o volume materiaf % o s 31 , ,wa l less tha n, mos 0- whicf 6mm to h consiste brokef do n light chaff components with abou thire on thif t do s fraction madf minerao p eu l grits, soil peds, etc0-6e Th . - 1 mm fraction constituted 10% of the total and was mainly light chaff (lemmas, paleas, awns, and their fragments), with abou minera% 6 t l material. Cereal grains made up 53% of the sample. Oats (Avena spp) made up 88% by volume of the cereal smals t fractiowa fraction oa l graine % size th n80 th remainine f n rangsO i Th . . e mm betwee g2 d an n1 t grainsoa f o , wer, abov% 20 mm e estil2 l withi uppee nth r size range ostrigosafA andAfatua (bristld ean wild oat) .evidenc o Thern s ewa e from flore tpresence partth r sfo e osalivafA (the 'cultivated' oat) which could have accounted for the grains above 2 mm. Floret parts identified were all of A strigosa or Afatua and the size range tends to confirm this. From the floret parts in the sample one may deduce that the crop was A strigosa with A fatua as a weedy component in the sample. Eleve cereae r centh n pe f to l fractio barlef o s ynwa (Hordeum vulgare syn. Hpolystichum)x la e th f o six-row type, hulled. The barley grains were all entire and many showed signs of squashing, but did not appea violentle b o rt y impacted. This would indicate pressure when quite soft, probably wit hmoistura e content of 24% (dry store grain needs to be kept at MC 14% to prevent deterioration). The remaining 47% of the sample was mainly composed of fragments of light chaff (lemmas, paleas, awns othed an , r fragments) which would most probably have been carbonize graie parchina th s na n do g proces facilitato st e 'shellin' ('shellin removae th lemm e s th palef d wa ' lo a an expos o at nakee eth d grair nfo milling). Many grains were found 'unshelled r entiro ' e withi florete nth . Storag oatf normalls eo swa n yi e thresheb flore o t f grai sheaved i s t dan nwa s were often stored whole. Parchin e floretth s f go wa s undertaken before grinding largo N . e straw node fragmentr so s were found. This would indicate thae th t sampl beed eha n processe rakiny db sievinr go perhapd gan s winnowing after harves beford tan e carboniza- sample th tiond e Ha .bee n winnowed, however woule on , d have expecte lighw dfe t weed seed Rumexe s(i acetosella Spergulad an arvensis). This materia f courso y ema l have become mixed-i latea t na r stage, recontaminatin gcleaa n sample. 'Weed' species were represented by 16 different species (including three ignota) identified to species level in 11 cases and to family in two cases. Ignota were damaged and not easily determinable. The four most frequent species were hemp nettle Galeopsis tetrahit ang dag polygonum most closely comparablo et Pperiscaria (redshank). Both have large 'seeds' greater than 1 mm in comparison with the more abundant Rumex acetosella (sheep sorrel comparativeld an ) y frequent Spergula arvensis (corn spurrey). This com- parative frequency of the latter two in the sample would tend to indicate acid-impoverished soil (with the caveat of course that the sample typifies the growing crop). The growing conditions indicated would be expected in Angus at this time from historical references and accounts. Oats of course are quite tolerant of bristlstrigosa)e A ( th t d eoa an H performp w lo a s tolerably wel comparison i l n with other cereals under these conditions.

ACKNOWLEDGEMENTS authoe Th r would lik tako et e this opportunit thano yt followine kth g people DrummonW r :M d Mora allowinr yfo g acces Abercairne th o st WalkerD yr estatM d ,e Managean Tilcoe th f ro n Quarry, Capo for reporting the existence of the kilns and for allowing access to the site. Thanks are also due to ShepherdI r M , Grampian Regional Archaeologis bringinr fo t Cape gth attentiooe sitth e o et th f no SDD HBMD and to the HBMD in turn for financing both excavations. I would also like to thank the GIBSON: MEDIEVAL CORN-DRYIN9 G 22 KILNS

stafvolunteerd an f l theial r r sworfo k durin course excavationse gth th f eo PollockD : MaverA , T , Webster ReidA , KempM , Cameron ReeN D , d dan Fairweathe;A CronA d contributinr eran fo g specialist reports and P Hill, J Rideout, O Owen and Dr N Fojut for their help and advice on preparing this report. Finally my thanks to Dr A Fenton and Dr G Ritchie for reading and comment- ing on the final text.

REFERENCES Close-Brooks, Joanna 1980 'Excavations in the Dairy Park, Dunrobin, Sutherland, 1977', Proc Soc Antiq Scot, 110 (1978-80), 328-45. Feachem, R W 1957 'Corn-drying kilns, limekilns, stone-lined pits and fox-conduits', Proc Soc Antiq Scot, 90 (1956-7), 45-51. Fenton, A 1978 The Northern Isles: Orkney and Shetland. Edinburgh. Firth J , 1974 Reminiscences Orkneyn a f o Parish. Stromness. Fraser (edW , ) 188 Manuscriptse 5Th of Charles Drummond.H S London Hism tCo y . (TentRo p hRe MSS). Pollock, D 1985 The Lunan Valley Project: medieval rural settlement in Angus', Proc Soc Antiq Scot, 115 (1985), 357-99. Statisticale Th S SA Account of Scotland, Sinclair (ed)J r Si ,, 1791-9. Edinburgh. SRO P 100RH ,6 177 5PlanA f Improvementso Policye oth f of Abercairny, seate th of James Esquire. 1775. Scottish Record Office, Edinburgh. This paper publishedis granta of from aid Historicwith the the Buildings Monumentsand Directorate (SDD)