Excavatio Pictisf no Viking-agd han e farmsteadt a s Buckquoy, Orkney by Anna Ritchie

INTRODUCTION The promontory known as the Point of Buckquoy forms the N side of the Bay of Birsay, one of only three sheltered bays on the W coast of the mainland of Orkney (fig 1). The Point was named after a farm, Biggaquoy, which no longer exists, and the name was probably derived from Norsd Ol e bygg-kvt, meaning bere quo r barleyo y enclosure (Marwick 1970, 60-1) lony B . g tradition soile th f Birsa, so consideree yar mose dth t fertil f mainlano e d Orkney,e aideth y db sand from the Bay of Birsay which, as it consists mainly of shell, is rich in calcium carbonate and is still usefertilisea fieldse s da th n .o r Oat bered san speciea , barlef so y which callee useb o dt d 'big', wer mose eth t important crops grown locall medievan yi post-medievad an l l times (Stat Acct 1795, xiv, 320). Until recent times, a fleet of fishing boats operated from the Bay of Birsay, Poine sidth S Buckquof f o e to th n Castro d stil o s yan maii e aGe lth n landing-plac boatsr efo , althoug fishine longeo hth n commerciaa s gi n o r r eveo l nsubsistenca e level.N Skippi(O o eGe skipa-gja) on the N side of the Point may indicate another, formerly useful, landing-place. flae tTh promontor f Buckquoo y y rises some 4-6 m abov foreshoree th e t consistI . f o s Lower Stromness flags of the Middle Old Red Sandstone, which are threaded with igneous dykes (mainly camptonite) and capped with boulder clay (Wilson et al 1935, 57, 63). The bedrock splits naturally into flat rectilinear slabs and consequently the beach below the Point of Buckquoy provides an almost inexhaustible source of superb building stone. Until 1970-1, when an excavation was carried out over a total period of 10 weeks, a long low mound, truncated by the cliff-edge, existed on the S side of the Point of Buckquoy (NGR HY 243282). It had suffered considerable damage from coastal erosion and from traffic along a track leading acros boae sitth e t o s eth causewanaustt e th o shelterr t s o head e y Castrf th an d o t s o a aGe at the tip of the promontory which gives access to the tidal island known as the Brough of Birsay. e continuinTh g threa f thio t s destruction prompte e Departmenth d e Environmenth f o t o t t sponso rescua r e excavatio mounde th f o n . Locally moune th ,knows dwa Sinclair's na s Brae, but this a , s name operated solel oran yi l tradition proposes i t i , calarchaeologicao e dt th l l site Buckquoy, afte promontore rth whicn yo t standshi . Before excavation moune th , d survivea o dt maximua o t d an mm lengt heigh0 2 f approximatelf ho o t y 0-. Usin5m g initiall grie yth d method and finally total area excavation, the mound was excavated down to the level of the natural subsoil wit consequene hth t destructio l archaeologicaal f no l features. Fro layoue e mth th f o t structures which survived until then, it may be estimated that at least half of the original site had previously been destroyed by coastal erosion; of the remaining part, in most cases only the basal wall-courses survived, owing robbinpartle th o yt earlief go r building orden si construco t r w ne t ones and partly to damage caused by modern ploughing. Little will be written in the description of RITCHIE: PICTIS VIKING-AGD HAN E FARMSTEAD BUCKQUOYT SA , ORKNEY 175 the excavation about soil stratigraphy because very little survived; the stratigraphy was essentially one of stonework, and the task was not so much one of distinguishing soil differences but one of disentangling building debris and tumbled stones from the remains of walls and paving, often only the basal wall-course. The crest of the mound, 0-5 m high, covered the superimposed remains of eight stone structures. Sections were recorded but they are not considered useful enough to be worthy of publication. Nevertheless, a process of successive rebuilding on the same spot allowed recognitioe th archaeologican a f no l sequence spannin 10te gperioa hth o centuriet dh fro7t e mth s AD.

OJVKNL"Y-| ? ISLANDS

_ 7 E N T LA N D ,4, FI R.TH

0 2 kw5 1 0 0 1, , ,$

FI G1 Buckquoy : locatiop nma

In summary, the site consisted of a series of farmsteads, one on top of the other, of which werI I phase d e Pictisan sI phased han s III-V were Norse sixt ;a fina d han l phas activitf eo s ywa represente Viking-aga y db e burial inserted lasint e ruine th ot th f farmsteadso .

EXCAVATION Phase : earlyI Pictish farmstead (fig 2) ; 9 pi , s8 Phase la, house 6. The earliest surviving structure on the site consisted of part of a house characterised by rectilinear cells opening off a central chamber in which the hearth was situated. The walls showed the use of two building techniques: one employed upright slabs as the internal walle base th facth f , e t o wit ea h loose boulders behin basae th s dla wal lothee fillingth d r usean , d horizontal drystone masonr e inneth s rya wall-face from ground level upwards (pe i 8a)Th . externa walle parbotth n i lf o th case likels si havo yt e been buil turff o t , becaus outeo en r stone facsurvivedd eha ; where externa lhousee pavinth f o , gther N simplexiste s p e ewa th ga yo a dt between the edge of the paving and the back of the inner wall-face of the house. The latter survived heigha o t onlf courseso o t ytw . 176 I PROCEEDINGS OF THE SOCIETY, 1976-7

1970-71 -phases t-JL

• • i•:•/•... . • . - 'outline cf '" .-..' '-; \ lat&r'wtdlinq •'•'• -1'' '-•

FIG 2 Buckquoy: Pictish phases RITCHIE: PICTISH AND VIKING-AGE FARMSTEAD BUCKQUOYT SA , 7 ORKNE17 | Y

/ ^' T •' psi-wka

1, Id'

Tntrial-

^^••••^-•••^^^••^K^•v^r^-vvV^iSi^ I

-Irurial ^5£"'#,^;V^A - i^^O;^/-:'---^.'/ ! '.'•• ^>

am-;3;fe.SftSvv v•••• 1 slS^i.AWWV.; •//;v'

•:/.'v

.-.-'/ -'

^ ^^-SV*^^s \ ! »l¥ / 1*/ / xf -]mnh> ,,,,,,,,,,f V ? f 5 ^ 7- f 9 .i,<

FI G3 Buckquoy : late Pictis Norsd han e phases

P 178 | PROCEEDINGS OF THE SOCIETY, 1976-7 One complete slab-lined cell measured 2 m by 1-5 m to 2 m in area; only one upright slab remaine formee d th wallE alon t N rbu s , gexistencit f othereo indicates presencswa e th y db f eo their chocking stones. The total surviving internal length of the house was approximately 8 m. By analogy with a similar cellular house at Gurness in Orkney (the two house-plans are compared in Ritchi likels i t i e y , 1974 tha1) g t,fi housroughls wa e6 y symmetrica eitheo lt centrare sidth f eo l chamber and therefore that its maximum internal width may have been about 9 m. The main hearth was rectilinear with a stone kerb but no formal paving, 0-85 m wide and an incomplete length of 1-25 m at the cliff edge; it contained peat ash and a little charcoal. A subsidiary open hearth lay partly within one of the cells towards the back of the house. onle Th y area withi house nth e wher originae eth l floor deposit remained uncontaminated by Period Ib activities was within the slab-lined cell; this deposit yielded a perforated bone pin or needl ecatalogue (nose , 1 . finds)f eo . halE Phasesubsequentlhousf s S fo e , housewa eIb 6 Th . 5 y dismantle orden di builo t r da smaller versio same th f eno plan-form , hous (pe5 i 9a) curvin.A g portio wall-facef no courseo tw , s high, built across the central chamber of house 6 also belonged to this phase, but it was too fragmentar alloo t y w further interpretation structuree Th . f phaso s , likeIb e thos f laeo , were built directly on the natural clay subsoil, and the stratigraphy of the two periods of building there- fore depends upon the relationship of the stone walls rather than on soil levels. As a result, the artefacts from phases la and Ib must be considered inseparable, with the exception of the per- forate (non mentione) dpi 1 . d above lace soif k.Th o l stratigraphy should also imply tha greao tn t period of time lapsed between phases la and Ib. House 5 was a compact and almost complete building of trefoil form, with three rectilinear cells surrounding a central hearth on the fourth (SE) side of which lay the entrance. Except at the entrance, wher uprigho etw t stone slabs were used survivine th , g walls were buil f horizontao t l drystone masonr f whicyo r hoccasionallo onle yon courseo ytw s remained outen A . r wall-face housee th sidN survivef e o , th but n wits do a earliee , hth r hous wall-fillina , e6 f eartgo r ho turf above the basal course should be inferred. Although the S wall of the house had been robbed entirel latey yb r building activity innes linit e f th ,ero indicatefacs ewa vera y dyb shallow groove rilled with 'split peas'; this is interpreted as the effect of rainwater percolating down the inner face of the wall, part of which, judging by three surviving chocking-stones, may have been lined with upright slabs. Internally house th , e measure desige three th th d f endo cellan 3-6y 2-7 , b 0sm 5 m showe d considerable symmetry with each containing a similar floor-area averaging 1-30 m by 1-60 m. The cells were divided fro centrae mth l heart stonhw lo are ey akerbb s projectin gcenti w onlfe y-a metres above the level of the floor. The rectilinear hearth, which occupied most of the central area kerbes ,wa thren do e side paved san d with stone slab measured san d approximatel y 0-4b 0m y hearte th seale s f removabla o opehe wa y d d0-7b Th nen . 0m e horizontal slab, whic notched hha s carved to fit the side slabs (pi 9b). The hearth was filled with peat ash and some carbonised twigs of heather (calluna). Pavin frod hearte gle m th h throug entrance h th hous e th extended o ean t d some 1-30 m beyond. The entrance was 0-65 m wide at its outer end and 0-80 m wide at the inner end, where it was very close to the hearth; access into the house must have been past the NE side hearthe oth f . Phase I artefact assemblage. artefactw Fe s were foun association di n wit structuree hth f so this phase. Apart fro perforatee mth (non belongin) dpi 1 . houso gt , there6 twe ear o simple bone skewer pins (nos 2, 3), an incomplete pin-shank (no. 27), a bone point (no. 36) and a perforated limpet shell (no. 93). The most distinctive artefacts were found amongst the wall-filling of the circular chambe f houso r (phase4 ) whereII t overlaei y phas eI structures bone ; thesth ee ear RITCHIE: PICTISH AND VIKING-AGE FARMSTEAD BUCKQUOYT SA , ORKNE9 17 | Y spoon (no. 43), a fragment of bone mount (no. 45) and a bone comb (no. 50). In assigning these three finds to phase I, it has been assumed that their provenance implies derivation from the early occupation rather than that they were lost during the construction of house 4; no other artefacts were found elsewhere in the walls of house 4. None of these objects is closely datable but none would be inconsistent with a date in the centurh 7t y AD perioe th , d which wil arguee b l phasr dstratigraphicafo n o eI l evidence. Bond ean horn spoons with sloping shoulders characterisin bowe shape g th th hav 3 f l4 e o lik . e ebeeno n foun contextn di s datinearle mid-firsd th yan o g t t millennium AD, e.g. Keis Skirzd san a Head broch-complexes in Caithness (Anderson 1901, 122-7, 144-5; NMAS GA 463, 784), and Jarlshof, Shetland, from wheelhouse phase III (Hamilton 1956, 64, fig 29, 8). The closest parallels for comb no. 50 are those from the Broch of Burrian, N Ronaldsay (MacGregor 1974, no. 153, fig 11) Bustod an n Crannog, Ayrshire (Munro 1882, 219 218)g unusuafi ,n .A l featur thif eo s come th s bi suspension hole at either end; one hole at one end is normal, designed to allow the comb to hang fro ma belt , perhaps wit iron ha nintermediatn a rin s ga e fittin s founa g t Ihrea d , Gotland (Stenberge Buckquoe holeo rth 1961tw n s o 50)e g fi ,yTh . com suggesy bma t tha t huni t g hori- zontally, rather than vertically, possibly to avoid too much movement while hanging from the belt. The Broch of Burrian has also yielded two bone mounts (MacGregor 1974, nos 130, 131) closely comparable to mount no. 45 and to later mounts no. 46 (phase II) and no. 124 (phase VI); these mount likele sar havo yt e strengthene mouthe dth leathef so r knife-sheath mountf s(c s still attache sheatho dt s from Lund, Sweden, Blomqvis t ale t 1963, figs 218-20). Phase II: late Pictish farmstead, house 4 (figs 2, 3; pis 10, 11 a) structuree Th phasf so e Ib were dismantled virtuall flooo yt r level, leavino tw r go onle yon course f wallinso positionn gi time th et a ,whe n builts houswa ;e4 dons thiswa e probablo yt obtain building stone as well as a measure to level the site because house 5 was dismantled even beyond the area required for house 4. The NW wall of house 4 was built immediately on top of the basal course of the E wall of house 5, utilising the upright slab on the E side of the house 5 entrance as part of the outer face of the new wall (fig 2; pi 9a). The circular chamber at the NW end of house 4 was semi-subterranean and had been dug into and below the earlier building-level; at the back of the chamber, the floor-level was some 0-36 m below the phase I ground-level. A natura sub-soie levee th l th risf o ln ei l beneath hous meane6 t that orden i , achievo t r leveea l floor ove entire th r e lengt circulae f housflooe th ho th f o , re 4 r chambeintg odu s naturawa r l boulder clay walls It . , which survive maximua o dt m revetteheighs wa 0-f , o t8 m d against both overlyine subsoie th th d an l g remain phasf so ebuildingI largy sb e upright slabs which were then backed at ground level by horizontal stonework. Above the upright slabs, the wall consisted of horizontal course stonef so s which presente dsmoota h inner facwhict ebu h lacke dformaa l outer face; up to six courses survived. Elsewhere the house wall consisted of a basal course of stones with botouten a innen ha d r facan roccasionad ean l fouremaino t p r u course f so s abovee Th . S Ehouse par th suffere d f eto ha d considerable damage from ploughin deed gan p plough-scratches were visibl mann eo y stones. The plan-form of house 4 consisted of four interconnecting rooms set in a line SE-NW (pi house anotherth d f eo an 10)d .en , Ther serveE S pave a e y dwerth b entrances dt o a et tw se e on , pathwa maie walyth W nf leadinS o lliving-hall e th E gn S i int t house e ose th Th . , e SE fro e mth entrance was 0-75 m wide and was furnished with a low sill-stone and, just inside the house, a horizontal pillow-shaped slab maintaining the height of the sill-stone. This entrance led into a small vestibule, 1-65 m by 1-25 m, which was unpaved. From the vestibule, entry was gained into a larger room, measuring 3-3 m by 2-2 m, which was again unpaved; a whalebone socket (no. 62) 18 | 0PROCEEDING THF SO E SOCIETY, 1976-7 found againswalE thif S o le s tth roo havy m ma e(fi ) functioneg2 dooe pars th d a f r o tfurnitur e but, if so, it had probably been displaced from an original position rather closer to the doorway. This room was otherwise featureless and its function is therefore unknown. The next room was the main living-hall, for it contained the hearth and possible traces of flanking benches wal W thiseconf e S o l sTh e . living-hallth d n i entranc t house se th s t o i ;et wa was 0-9 m wide and, although there was no sill-stone, the paving of the entrance extended some 0- int 6 hallm e heart e oth kerbe s .Th s hwa thre n wa do d uprighew en side lo E y S s b t s slabsit d an , left open to facilitate the removal of ashes. When found, the hearth was filled with fine peat-ash, stone-pavea s wa t beneat h bu as de hfloorth stonA . e spindle-whorl (no. 83), affecte heae th ty db of the fire, was found beneath the paving of the hearth. The hearth measured 0-7 m by 1-4 m, and the remains of a stone-lined post-hole were found on either side, probably representing the upright posts for a spit across the hearth. A fragment of bone mount (no. 46) was found in the post-holeS identican A . l hearth with flanking post-holes belonge primare th o dt y e phasth f eo post-wheelhouse chamber 1 at Calf of Eday, Orkney (Calder 1939, 175, fig 1, pi LXVII, 2). At the NW end of the hearth was a shallow pit containing ash from the fire and part of an iron knife- blad lates e (nowa r paving-slabs t o seale. pi 97)tw e y db th ; . The remains of low stone-kerbing on either long side of the hearth suggested that there may originally have been wooden benches or platforms lining the walls, but this can be no more than a very tentative interpretation: the kerbing may have functioned simply as minor sub-divisions of the floor-area as in house 5. An iron rivet (no. 111) was found within the kerbed area on the NE side of the hall. At the NW end of the living-hall lay the entrance into a circular chamber, which measured diametern i 3- 4m . Initially entrance th , onls eywa 0-6 wide5m , between substantial upright slabs three th sided S smale an th ln uprigho sidN tee slabwhicth n so h wer l thaeal t survive thaf do t e entrancth sid f o e e slabs (show t thi A solin sni . stage2) d g blacfi , n neitheko e circulath r r chamber nor the living-hall was paved. After an interval of unknown duration, the circular chambe paves wa r d with massive slab e entrancth s (pd s wideneian ewa 1) 1a y o b 1-7dt 5m dismantling its original N side to floor level, swinging back the long slab which had formerly been chambee parth f o t rpositioe walth o t l foun s whicn ni wa dt hi when excavated, restine th n go basal stonee originath f o s l wall-core pavind an , g ove aree th r a thus free pavine dTh . (fig3) g extended over the small compartment which formerly lay to the S of the entrance into the circular chambe intd oan rliving-hale parth f o t l (pi . 1Whil1a) e ther obvioue ear s structural problems involve thun di s widenin doorwaye gth , this interpretation seem explaio st survivine nth g evidence best, particularly the free-standing condition of the long wall-slab. Some subsidence occurred in the central area of the floor of the circular chamber, and an extra laye slabf lais o r dswa dow orden i maintaio rt nlevea l floor pavine separateTh s . gwa d overall from the original floor-level by a thin layer of brown sand, 0-04 m thick, and the paving thus provided a relative sequence of artefacts within the occupation of the house: the painted pebble (no. 87), pottery sherds (nos 64-7) bond an , e point (nower) 38 . e primare founth n di y occupation deposit, whil bono etw e pin s20)d (noan , bons7 e poin iro n pard a nf tan to (no ) 39 . knife-blade (nower) 99 . e foun dsecondare lyinth n go y paving. foue Th r room f housso weree4 divided froe on m anothe shory b r t cross-walls buila n i t particula characteristid ran c way. Onl basae yth l course survived thid s,an consiste substantiaf do l stone slabs set upright to form box-shaped wall-ends, with the tops of the slabs approximately level with one another. Upright slabs have been found as the footings of radial piers in wheel- houses and later related structures (e.g. at Jarlshof, Hamilton 1956, 67, pis Xb, Xlld), usually in the form of a single slab incorporated into the inner face of the pier, and it is likely that the box- RITCHIE: PICTIS VIKING-AGD HAN E FARMSTEAD BUCKQUOYT SA , ORKNE1 18 | Y ends in house 4 represent a development of the same building tradition. It is suggested that they acte bases da r piersfo horizontallf so y coursed drystone masonry. After the abandonment of house 4, the circular chamber became filled with tumbled stones and earth up to a height of 0-6 m above floor level at the back of the chamber. The remains of a roughly constructed stone hearth, measuring approximately 0-9 m square, were found on top of tumblee th d filling othe;o n ther s r ewa structura l evidence associated with this heartht i d an , appeare havo dt e been nothing more tha ntemporara y camp-fire. immediate th o T housf o eS weree4 remaine structureo th tw f so s which stratigraphin o , c grounds, shoul broadle db y contemporary with hous Bot. e4 h overla earliee yth r houswerd an ee5 same builth t ea t leve direc y houst neithes a lan bu t, d e4 associatio rha ns wit lattere wa he th On . incompleteln a y surviving are carefullf ao y laid paving, whic originalld hha y been boundes it n do N and E sides by a wall or fence, because the small stones that had acted as fillers between the paving and its boundary were still in position. At its NW extremity on the cliff edge, the paving abutte bace shora d th f ko t lengt curvinf ho g wall. This wall whicf o , threho t onlo e ytw course s survived, was built of horizontal drystone masonry with a smooth face to its inner curve and a rough outer face, and it had been truncated to the S by coastal erosion and to the N by the con- structio housf no Norsn i e3 e times functioe .Th thesf no incompleto etw e structure unknowns si . Phase I artefactI assemblage. Apart from those artefacts already mentione precedine th n di g alss sectionowa foun4 2 . ,d bonno corneinsidmaie E n S th epi e f neo r th hall hous e n i Th ., e4 other artefacts associated with phase II were found outside and to the SW of the house, including ogam-inscribee th d spindle-whorl (no whic) .84 immediatel y hla y entrancoutsidW S e eth e into the main hall. The bone pins include no. 7 with a flat ovoid head and straight shank; looking at the pins from the site as a whole on figs 4 and 5, it is noticeable that hipped shanks tend to be confined to the smaller pins where the risk of loss from clothing was perhaps higher. Pin no. 28 is an un- providet finishei d an cleares e , d7 s th exampl. samno e ts th indicatioa f een o typ pi f eo n that some pin-makin t leascarriegs t a Buckquoy a t wa t dou animal-headee Th . (non ) appeardpi .24 o st represen catta thi d ,san possibilit supportes yi presence th t boney dca b f eso amongs animae tth l bone sample from this phase (Appendix 1). The fragment of bone mount (no. 46) has a slightly convex upper surface, unlik flae eth t fragment from phas e(noI bono . 45)therd tw ee an ,peg ear s stil positionn i l . Relatively little potter recoveres ysherdwa x si t ds bu wer fro l wholea site al s m eeth a d an , associated with phase II, mostly found on the floor of house 4. The fabric of the pottery is fine and well-fired with little grit, 6 mm-12 mm thick, and its outer surface is buff to red in colour, while the inner surface is black. None of the sherds provides enough information to allow the vesse e reconstructede forb th f mo lt o ; only wall-sherd flat-baseo tw d san d sherds were recovered, apparently from large jars, and none is decorated. On the basis of the fabric, this pottery seems comparable with the wheelhouse wares, Hamilton's classes II and III, at Jarlshof in Shetland (Hamilton 1956, 64-6, 81-2) witd late an ,h th e wheelhouse pottery from Clickhimi Shetlann i d (Hamilton 1968, likel159-60)s i t i y d vessele tha fore an , th als s th tf m o oswa similar. Both of the spindle whorls belonging to phase II (nos 83 and 84) are made of chalk, probably derived as pebbles from local glacial deposits (Appendix 7) and presumably chosen for its creamy colour and tractable nature. Several other chalk whorls have been found in Orkney (MacGregor 1974, 92), includin froe Broce gon m th Lingrof ho w (NMA 11E S)G whic mads hi e fro mvera y bees ha n4 carefull8 . facno e f similao eyOn . incise84 r . chaldno wito k t oga n ha m inscription encircling the perforation (pi 13a); Professor Kenneth Jackson reports that the inscription is unintelligibl tha d fore e an tth letterin f m o g suggest earle datscentura h th y8t n e i y (Appendi. x6) 18 | 2PROCEEDING S OF THE SOCIETY, 1976-7 otheo Tw r ogam inscription portabln so e artefact knowe sar n from Scotland, both knife handles, from pre-Norse context t Gurnessa Mhi c Orknen si Ba cd Connaiyan Nortn no h Uist (Wainwright 1962 , 98) neithed 96 , ,an r contain late letter-forms (Padel 1972, 12). Whateve meanine rth e th f go Buckquoy inscription, its presence on a spindle whorl would seem to imply that the whorl was specia owners it o t l . The ogam-inscribed spindle whorl is undoubtedly a Pictish artefact, and the same identifica- tiobees nha n argued elsewher paintee th r efo d pebble(Ritchi7 8 . no , e 1972, 298) 9 tota.1 A f o l painted pebbles have been founddomestin o l al , cNorthere siteth n si n CaithnessIslen i r so d an , 15 of the pebbles can be dated to the period from about 200 AD to the 8th century, thus spanning the historical Pictish era. The painted designs are simple, mostly solid dots and wavy lines; the closest parallel to the Buckquoy example is one from Jarlshof (NMAS HSA 4108; Hamilton 1956, 77, fig 39) which has also been painted overall with small circles. The purpose of painted pebble s unknowsi n but analogy ,b y with later tradition s charm-stonea e us , r curinsfo g sick peopl animaleand s seems strengthene likelis yand contemporarthe dby y story about StColumba's holy pebble recorded by Adomnan (Anderson and Anderson 1961, 399-405; Ritchie 1972, 299). Four fragment irof o s n knife-blades were foun phasn di I contexteI s (nos 97-100)t a e ,on least belongin single-edgea o gt d blade wit hcurvea d back (no. 97).

PICTISH HOUSE-TYPES Althoug t confineno s i thit ho i dt horizonta f so periode us e th ,l masonr uprighd yan t slabs in combinatio wall-buildina s na g techniqu characteristia s ei c featur post-brocf eo h architecture in the Northern Isles and Caithness. It is a natural response to the type of building stone available, and can be found in structures of all periods from neolithic chambered tombs to the vernacular building f relativelo s y recent times. Building technique alon s noti e , therefore a diagnosti, c facto r datinfo r usefue b g conjunction n purposesi l ca t i t bu , n with specific house-typess i t I . possible to isolate plan-forms of houses which are confined to the middle centuries of the first millennium Awhichd Dan , more specifically, represent Pictish building traditio ScotlandN n i . t Buckquoa 5 House d an ys6 beloncellulare th o gt type f houseo . Unlik amorphoue eth s housing complexe earliee th f so r iro suce nHowmas ag h a e Brae, Orkney (RCAMS 1946 . 195,no , 48-50p p 93)g cellulae ,fi ,th r hous a detache s ei d structur whicn ei hcentraa l chambe surs i r - rounde roundey db rectilinear do r cells whole ,th e entere singla y db e entrance bese .Th t surviving example elsewhere is the 'late dwelling' at Gurness, Orkney (RCAMS 1946, no. 263, p 78, fig 132), but others have been found in post-broch settlements such as Berwick, Orkney (Watt 1882; RCAMS 1946 . 679 252-3)no p , p , direco .N t evidenc bees foreha e n roo th f mfoun o o t f s useda d on these houses, but an advantage of the cellular design is that it reduces the span of the roof, wit protrudine hth g walling betwee celle nth s actin effecgn i internas a t l roof-support combinaA . - tio f partiano l corbelling ove celle timber-framind th r an s g ove centrae th r house l areth f eao woul mose dth seee t b likel mo t y solutio roofingo nt . Using this method maximue th , m lengtf ho ridge-timber necessar rooo yt f hous woul5 e d have been just unde, spanninm 2 r e aregth a between the entrance and the walling on either side of the opposite cell, to form a hipped roof. Hous represente4 mose sth t sophisticated example know figure-of-eighte dato nth t f eo type of house. Simpler house f thiso s type have been Udale founth t UisdN a , t (Crawford 1973; ,9 1974, 9-11) and in the post-broch settlement at Yarrows, Caithness (Anderson 1873, 137, fig 1, F and G), and comparative plans of these structures drawn to the same scale may be seen in Ritchi tere e 1m1974g Th . fi 'figure-of-eight, ,26 t idealt servei t no s distinguisi bu ,o 's t h this type of house, where the cells are set in a linear plan, from the cellular type where the cells are arranged RITCHIE: PICTISH AND VIKING-AGE FARMSTEADS AT BUCKQUOY, 3 ORKNE18 | Y roughl circla n yi e roun centrae dth housele are th structuree f ao .Th t Yarrows a t closelno e ysar datable but they are likely to be late in the post-broch sequence. The Udal houses belong to Crawford's phas earlo t centurie h h ey9t 7t XI (Crawfor e , dateD .1 th A s o dt t a d s 1974A . 9) , Buckquoy figure-of-eighe th , t houseUdae th t la s were replace earle centurh th y9t n a d i y yb Norse settlement. Nothin know s gdetailei e th f no d internal arrangement Yarrowe th f so s houses, but the stone-built hearth and flanking platforms of Buckquoy house 4 are repeated at the Udal. The circular cell at the NW end of house 4 was filled with large stones and it is likely that these represented a collapsed stone-corbelled roof; the diameter of the cell was well within the maximum figur m-4-4 f o e 5suitablm r corbellinefo g (Dunbar 1966, 233) timber-frameA . d hipped roo envisages i f d housee ove th rese slab-buile f rth o t Th . t box-end internae th o st l sub- dividing walls seem best interpreted as the bases for piers of drystone masonry, probably sharing a common ancestry in building tradition with the piers of wheelhouses; these piers would have risen to a greater height than the external house-wall and thus provided internal support for the central part of the roof, while the ends of the couples rested on the house-wall itself. The longest timbers required would have been the ridge-beams over the main living-hall which had to span e externaTh distanca . m l 4 roof-coverin f eo l threal f ego Pictish house moss si t e likelyth n o , analog locae th f l yo vernacula r traditio f lateno r times havo t , e been tur r thatcho f , perhaps combined with flagstones at the wall-head (Dunbar 1966, 232). Althoug floor-aree hth f housao clearl s i e4 r largeyfa r than tha housf perhapd o t an e5 s comparable with tha f housreao o t n , l e6 implication drawe b n nsca abou respective th t e social units involved becaus possibilite th f eo y that other house r outbuildingo s s belongin eaco gt h phase of occupation had already vanished over the cliff. e deliberat vien th I f wo e symmetr possibld yan e anthropomorphic qualitie f houso s , e4 together with the Pictish love of symbols, a plan of the house was submitted for comments to a social anthropologist, Dr Anthony Jackson. He pointed out the similarity between the design of the N part of the house (circular chamber plus living-hall) and the symbol known as the 'circular disd rectanglcan e with square indentation' e distributioth , f whicno confines hi Orkneyo t d , Caithnes Sutherland d scepticas san wa e makinh f t lo ,bu g this similarity int ofira m identification. On the anthropomorphic appearance of the house, Dr Jackson drew attention to the ethno- graphic parallel provide Africae th y db n Dogon house (Forde 1954 explainedd ,an 97-8) 5 g ,fi : 'The structur housef eo ovel worlse al rth d often reflect men's religiou dividese viewar d dsan into sacre profand dan e areas parallelee b . t Thesno y de witdivisionma r o h y theisma r cosmological views. In the case of the Dogon, the house structure mirrors their cosmology, as does their field syste rituald mDogoe an .Th n house represent man.sa Jacksor D ' n suggesweno t n o t t aboue tth Picts: 'Their cosmology could be as intricate as the Dogon who, after all, are an isolated and primitive people. I do not think we should underestimate the possible complexity of their belief.' 'circulae Th r disrectangld can e with square indentation bees ha 'n interprete representatioe th s da n of a human figure (Thomas 1963, 60), and it is conceivable that this symbol and the design of hous derive4 e fro mcommoa n idea. Long cist burial (fig 3; pi lib) c , sitee th ,f juso tN beyone domestith e aree o th dT f th a o c settlement stone-buila y la , t long cist containin inhumation ga n burial side tope th e f Th .slabo s s were immediately beloe wth modern turf-line, and ploughing had chipped some of the slabs. The cist-pit had been dug through contemporare th y topsoi int d naturae olan th l clay subsoilexcavatee th d an , d clabeed yha n used cise filo t th lt afte buriale th rbeed ha n made beecise d tha Th . n constructe lininy db slightle gth y slopin t witpi g he sideninth f seo uprigh foothree d tth an tstont eheada e e th e t on slabs,a o tw , 184 | PROCEEDINGS OF THE SOCIETY, 1976-7 slabs along each long side. The slabs measured 0-02-0-06 m in thickness and from 0-3 m by 0-3 m to 0-66 m by 0-47 m in area. The average depth of the cist from the tops of the side slabs was 0-44 m. The floor of the cist was unpaved and measured 1-7 m by 0-45 m at the head and 0-3 m at the foot; the corresponding measurements at the top of the cist were 1-85 m by 0-6 m by 0-45 m. grave alignes Th ewa d SW-NE. SW , wite heae th hth o dt bode beed Th y ha bacn s it lai kn d o directl clae th y ciste n flooyo th ,f wito r heae hth d facing SE, arms at its side and both feet turned inwards so that the right foot overlay the left. The condition of the bones was poor, and parts of the hands and right foot were missing as a result of post-burial decay skeletoe .Th bees nha n identifie thamals a d a f to e age yeard0 betwee3 d s an 5 n2 probabls wa o ywh suffering fro mtuberculoua s conditio spine th f eno (Appendi . Therx5) e were grave-goodso n bode beed Th .y ha n covered cise th t half-filled an , d with natural clay subsoild an , two stone slab thed sha n been placed ove centrae rth cistle parth .f Wheto n found, each slad bha onembedded een cist-filline th othee n dth i d r restingan g agains innee th t side r facth e f slabseo , likels i t i y d thaan t this alternately sloping arrangemen originas wa t neithes a l r slalargs bwa e enough ciste adequatelth .f o p to spao y t e nth evidenco Thern s e wa associato et e this long-cist burial wit particulay han r phas occupaf eo - fels t wa tha t i siteradiocarboe ta d tioth ,an n no n date from bone collagen sufficientl e coulb t dno y precise to answer the question of whether the burial was Pictish or Norse. Inhumation burial in long cists was practised at both periods in N Scotland, the closest examples in this case being the graves in the early Christian and Norse cemetery on the Brough of Birsay (Radford 1962, 168, 178). It is possible that the long cist at Buckquoy did not belong to the occupation of the site at all and that its presence there was fortuitous. Its location on the Point of Buckquoy rather than in the cemetery on the Brough of Birsay suggests that it ought to pre-date the latter. Phase HI: early Norse farmstead, house 3 (fig 3; pi 12a) Norse farms are a considerably better known settlement type than are Pictish domestic sites, and the degree of erosion loss becomes very obvious from the plan of the surviving structural remains of the Viking-age settlement (fig 3). Each of the three buildings represents a stratigraphi- cally and chronologically separate phase of the farmstead, and each must originally have been accompanie t leasothea e y on tdb r building whic vanishes hha d int seae onle oth .Th y building which survives of the initial phase of Norse settlement is a byre (house 3), and the original dwelling hous bees eha n lostsame r Th anothe. eo r dwelling house must hav econtemporare beeus n ni y with the barn (house 2, phase IV), while the dwelling house (house 1, phase V) lacks any surviving outbuildings. The byre (house 3) was built partially overlapping the NW end of house 4; this section of the byre walbees ha l n destroye ploughiny db recenn gi t timesdoubo n t ther s bu , t ewa abou e th t stratigraphical relationship of the SE wall of house 3 to the underlying walling of house 4. Internally, the byre measured 4 m across, and it survived to a length of 8 m. The entrance must have been in a missing portion of the wall. The wall itself was built with a substantial inner face of drystone slabs, backe turfy lattede b represente s th ; rwa house sida dth W f y aloneo b N e gth basal compacted sandy layer, bordered by a line of small stones which indicated that the original walbeed ha ln about 1- thic5m k (pi 12a). Ther danges ei assuminn ri gtura f backing when only basae th l course wala f so l survive thin i st cas inne,bu e eth r face survive maximua o dt m heighf to five courses outereason o a n therd existedd y s ,ran ha n0-facee , wh wa t i 6 m f ,i , shoul t havdno e survived. byrdividee s Th ewa d intarea o shora otw y sb t cross-wall projecting t righa , t anglee th o st NE wall, about 0-7 m into the interior of the building. No features survived in the area to the N RITCHIE: PICTISH AND VIKING-AGE[FARMSTEADS AT BUCKQUOY, ORKNEY | 185 of this partitio indicato nt functions cross-walle eit th f o S , e stone-linebuta , th o ,t d drain nra down the central axis of the building. A sample of the silty material filling this drain was kindly examined by Dr S E Durno; phosphate analysis gave a figure of 660 mg (P2Os) per 100 g of the sample, a very high value which indicates animal life in the near vicinity and which supports interpretatio thif no buildinse parth f byreto a sample s ga .Th e also included plant cells resembling those from leave f grasseso s , including Anthoxanthum odoratum (Sweet Vernal-grass), Cyanosurus cristatus (Crested Dog's Tail), Holcus lanatus (Yorkshire Fog) and Molinia caerulea (Purple Moor-grass) stonee th f o slabe sOn . linin draie alss gth e onkere wa th th parf bo f o t underlying hearth of the phase I house 6, uncovered presumably during levelling operations when housbuil s utilised wa tan conveniene a 3 s da byre t th sla er drainbfo . Remain pavinf so g survived eithen o draine r sidth f eo . At a later stage in the use of this building, the cross-wall was extended by 0-4 m by rough extra walling, which showed a straight joint with the original wall-end for the basal four courses and only above that level was roughly bonded in to the earlier wall. A lightly built arc of low wallin attaches extendee gwa th o dt d crosswall kerb-stonew wherlo N ,o e opeth etw o nt s com- pleted an oval area measuring 1 m by c 1-2 m. This is likely to have acted either as a storage compartment for fodder or as a setting to hold a water barrel or milk churn. Circular settings of upright slabs abou tdiameten i 1- 2m r were foun t Jarlshoda fearl e botth yn hi wheelhouse phase and in the Viking-age levels (Hamilton 1956, 61, pi IXb, 110, fig 53, 1A), and sunken settings or impressions for wooden barrels are a common feature of Icelandic Viking-age farms (Stenberger 1943 ; Eldjar, 72 fig, s35 n . 19487) g fi , House 3 appears to have functioned as a combined byre and barn. Apart from the partial paving of the byre, the floor was earthen and showed patches of heavily burnt soil throughout the building, perhaps indicating tha rooe caughd th t fha t collapsedd firean . Afte abandonments it r , the byre end was used as a convenient dumping place for a midden, which will be described under phas structuresV eI dens.A e scatte enigmatif ro c tubular objects immediately beyonwalE N le dth proved to be the result of poor drainage; thin sections were examined by Mr J Romans, who kindly reported tha objecte th t 'irose appeath ne b staine o rt d pipes' which form around vertical root channels in very poorly drained soils. The site as a whole is extremely well drained, and the localised formation of these pipes is inexplicable, unless it occurred at a later date, during phase V, when this particular area was covered by several layers of paving. Phase III artefact assemblage. This early Norse assemblage is very sparse owing to the difficult distinguishinf yo components git s with absolute certainty houss beed :a ha ne3 usea s da midden, and there was no definite floor deposit, it was impossible to attribute finds from its interio perioe occupations th it f o rt do . Find attributee sb than tca d certainl thio yt s phase came fro house aree mth th af e o immediatel wherW N ee theth yo yt were seale latey db r structures. They include simple bone pins and a point (nos 4, 5 and 40), two plates from double-sided composite bone combs (nos 51, 52), a single sherd of pottery (no. 76), fragments of single-edged iron knife-blades (nos 101, 102, 103), and part of an iron nail (no. 112). None of this undistinguished materia helpfus i l r datinfo l g purposes. Phase : middleIV Norse farmstead, house (fi2 g 3; pi 12b) This phas representes ewa smala y db l rectangula housf o r buildinN e e assos 3it th ; o g-t ciated external paving overlay the dismantled NW walling of house 3, and it was later overlain itsel housy b f e 1. Onl finae yth l yard-pavin phasf g o indicates i earlie e e3V g ;th fi n rdo levelf so paving were too numerous and fragmentary to include on the published plan. The basal course of stones along the SE and part of the N sides was all that survived of house 2, but it had been 18 | 6PROCEEDING S OF THE SOCIETY, 1976-7 built sufficiently carefully with straight wall-faces, angular internal corner unusualld an s y fine paving that the line of the SW wall, of which nothing remained, was obvious from the edge of the internal paving (pi 12b). The building measured 2 m wide internally, and at least 3-65 m long, withi nwala side0-E s widelS 7 m which wa ,e th . survivin o Thern n o , s ewa g entrancetracn a f eo . The floor was paved with large slabs and small pebbles filling the interstices between the slabs; a gap in the paving in the NE part of the building was caused by the secondary burial of an ox. There was no hearth. An upstanding box-like compartment formed by upright slabs was set into probable wale th th n cornei lbuildinge W th e N f o r , perhap storager sfo . Unlik , whose 1 house d ean axe s3 NE-SW y sla , alignehouss wa e2 d NW-SE, crosswiso et the locally prevailing wind. This fact, together with the size of the building and the carefully paved floor, suggests that havy thisma e bee nthreshina g barn otheo N . r evidenc f arableo e agricultur survivedd unlikels i e ha t i t ybu , that non practiseds ewa alss i t oI . possible that i t functione stablea s da . The building was deliberately dismantled to floor level when house 1 was built, and it was associates it overlaiy b lattee wald th E f dN an r o l n pavee botth y dhb yard. Phase IV artefact assemblage. Finds attributable to this phase came mostly from the midden in house 3; others were stratified betwee layere nth pavinf so g belongind an , V phased go t an V sI one (pin fragment no. 29) came from the floor of house 2 itself. The bone pins (nos 8-11, 14, 16, 18, 19, 21, 29) are all common types (Stevenson 1955), with the exception of no. 9 with its decorated flat oval top, and can be matched from most post- Roman Scottish sites where pins have been numbersy founan n i e dBuckquo Th . y rangf eo pins in general is marked by fine craftsmanship, best seen perhaps on the decorated examples, especially the thistle-headed pin (no. 16). The design of a band of cross-hatching on the stem occurs also on a pin of essentially similar form from Buston Crannog (Munro 1882, fig 212), and on a pin from Freswick (NMAS HR 1002). The bone implement (no. 42) may be an un- finished pin but it would be unusually large in this assemblage of pins, and a function as some sort of pounding implement is more likely. preciso N e parallels have bee ndecoratee founth r dfo d bone handlea (nos i t .i 44)t bu , simple form and bone handles were widely used in post-Roman and Viking-age contexts. One complet efragmene comon d b (noan t) (no belon.) 53 .54 thigo t s phase. Bot doublee har - sided composite bone combs of native rather than Norse form. Incised linear decoration involving various groupings of oblique lines as on no. 53 and on no. 55 from phase V is quite a common alternativ ring-and-doo et t decoratio combn no s from Scottish, IrisMand han xs it sites d an , occurrenc centurh 9t e Buckquot yth a n ei y reinforces Alcock's assessmen predominantla s a t i f to y late motif, based on 10th-century Irish examples (1963, 159). This style of decoration may also be found on Norse comb-types; the three-stroke saltires on no. 53 were executed in exactly the four-stroke th s a sam y eewa saltire single-sidea n so d Norse comb from Sanday, Orkney (Greig 1940, fig 68), so that one set of strokes appears to overlie the other, creating an impression of interlace shapee Th . unusuae d ar end-plate 3 thi5 n . o l s no typ f fragmentar e scombo f eo th t bu , y comb from A Cheardach Mhor, S Uist, with simple incised cross decoration, shows a similar interest in the detail of the end-plate, here decorated with oblique grooves along the edge (Young and Richardson 1960, fig 15, 51). One small base-sher wall-sherdo tw d d an (no ) ) wers 77 .79 (no e, founs78 phasn di V eI levels. The fabric and style of these sherds is identical to those from pre-Norse levels, as is the one sherd (no fro) shere .80 m th phasdd (no an fro) eV . 76 m Phase III; with only five sherds involved, whic havy hema been derived from earlier pre-Norse levels, Viking-age Buckquo havy yma e been aceramic. No pottery was found in the first phase of Norse occupation at Jarlshof, but its place RITCHIE: PICTIS VIKING-AGD HAN E FARMSTEAD BUCKQUOYT SA , ORKNE7 18 ] Y was taken by steatite. Buckquoy lacks steatite in any form, but vessels of wood and leather, of which nothing has survived, may have been an important part of the domestic equipment (cf Addyman 1976, 320-1) organie th ; c material that survive t Sandneda Greenlann si d (Roussell 1936 mord an ) e recentl t Trondheiya mNorwan i y Dubli(Lond an ) gIrelann ni 23 1975 , d21 , 6 Riordai( n 1971) provide77 , a tantalisins g glimps f whahavy o e ma et been lost elsewhere. Highlightin absence gth f nativeo e pottery from mos f earlo t y historic Ireland, Ryan suggested tha alternativee tth makino st g pottery migh more tb e efficient: imported potter containerr yo f so other materials (1973, 632), and this consideration could explain the virtual absence of pottery from Viking-age Buckquoy. One of the gaming-boards, no. 90, was found in the midden in house 3. The other two boards, nos 91 and 92, were discovered on the excavation stone pile after heavy rain had washed them, but it is likely, to judge from the stage of the excavation at that time, that they too originated in the Norse levels. All three boards consist of undressed stone slabs which have been lightly incised wit same hth e desig f intersectinno g lines formin cell6 , g3 mor sor e importantly, 49 intersections, with the central intersection emphasised by a circle. This is essentially the same design as that on the wooden board from Ballinderry in Ireland (Kendrick 1933), which is formed by peg-holes instead of line-intersections; a double incised circle surrounds the central peg-hole incised an , d arcs emphasis ecornee eacth f ho r holes. M. Claude Sterckx (Dept of Celtic Studies, Universite de Haute Bretagne) agreed very kindly to comment on the nature and significance of the three gaming-boards, and his report has been published in full elsewhere (Sterckx 1973a). He recognised that board no. 92 does in fact bea same rth e patter intersection9 4 f no s wit hcircla e roun centrae dth l intersectio othee th s rna boardso tw thad an ,t her pattere eth bees nha n cancelle additioe th y d b extr f no a line circled san s forming no regular or coherent design (fig 10; Sterckx 1973a, figs 1 and 2). e gamTh e represente boarde th battla y s db si e game betwees wa playerso t i ntw d an , common to both Celtic and Scandinavian tradition (Sterckx 1970). This was the hnefatafl of Icelandic sagas (Murray 1952, 55-8) similad an , r games were playe Irelann di brandubhs da d an Walen i tawlbwrdds sa (Sterckx 1973b) Ballinderre Th . y boar consideres di havo dt e been user dfo brandubh (Sterckx 1973b, 741), whereas the Buckquoy boards belong to a Norse cultural assem- blage and are most likely to have been used for hnefatafl; the similarity between the Ballinderry Buckquod an y boards underline commoe sth n tradition behind these board-games. playino N g pieces were foun t Buckquo da likelys i t i ,d giveyan e informalitnth e th f yo boards themselves, that small pebbles or shells were used rather than specially manufactured pieces. Bone playing pieces have been found in Norse contexts in Scotland at the Brough of Birsay, Orkney, Jarlshof, Shetland (Hamilton 1956, pi XXXVII, 8), and Drimore, S Uist , mostl(MacLare2) g fi y , wit37 . hn basano 1974 , 17 l, peg-socket peg-holn o e us r e sfo board f so Ballinderry type; together with the Buckquoy boards and fragments of boards from the Brough of Birsay, and Underhoull in Shetland (Small 1966,244, fig 16), and the possible miniature board from Jarlshof (Hamilton 1956, 145, pi XXXI, 1), these indicate the contemporary popularity of board games. Phase late: V Norse farmstead, house (fi1 g 3; pi 12c) d , The latest surviving building was house 1, an incomplete dwelling house measuring internall 4-yc 5 m- truncate a wid5 m d ean lengthmdn 6 i walls .It s were built wit innen ha r stone face, surviving to a maximum of four courses, and presumably an earth and turf backing of which, no trace survived. The NW wall was represented only by a basal course of stones, but an outer face survive shoo dt w hous e tha walsquaree bees th d f tth e o lha wa n d eln di en mE wideN e .Th 188 | PROCEEDINGS OF THE SOCIETY, 1976-7 end-wale th plan d buils an , wa lt with markedly more substantial stone blocks than wer rese eth t of the walls, suggesting that it was a gable wall. There were two entrances, one 0-75 m wide in the NW wall close to the gable and the other in the SE wall. This latter entrance had been 1 m wide and paved, but secondar a housee bee d t th ,a ha f o n t o i , blocketw e y us b perio e p du th n di massive stones (pi 12d). The floor-area inside this entrance was paved, unlike the rest of the house earthen whica d hha n floor. This sensible precaution agains heave tth y wear suffere entrancn a n di e matchee b are n aca d elsewhere instancr fo , t IsleifsstaSiea Skallakod an r Icelann ti d (Stenberger 1943, figs 97, 104, 23). A shallow but distinct circular socket had been cut into one of the paving stones (fig 3, pi 12d), and it seems likely that this was the base for one of the timber posts providing roof-support poina t a ty one-thirla t houseI widte .e th th f f hdo o , suggesting thaaisleo ttw f so posts were involved post-holeo N . s wer poste eth rese foundf sth o t mus d an , t also have been carrie pad-stonesn do . smalA l hearth represented onlheavila y yb y burnt patcearthee th lon e hn i th gnn flooo y la r housee axith f largso .A e stoncorneE N ee slar th covere bn i disarticulatee dth d skeleto newa f no - born infant (Appendix 5); individual bones of babies were found elsewhere on the site, perhaps betraying a hardline philosophy towards infant mortality, although the infant burial in house 1 could be interpreted as a founding ritual rather than a simple convenience. Faint traces of possible benches along the sides of the house were provided by the align- , howeverif ; foud were an t expectinNW r on ,E stonee no S th stonee o mentn sth o tw g n f so so fino t d trace benchesf so doubtfus i t i , l whether this evidence would eve considerede nb . aree Th a paveds immediatelwa E , S wit d han y evidenc E outsidN e house f th eth o n eo mor eleadin E thaentranc e S renewae th e non o g th t n eo l latethas twa r blocked shorA . t stretch of roughly built walling bordere pavine dth g leadin thigo t s entrance (thi pavine sth wale d gar l an indicated by stippling on fig 3; both overlay the earlier wall of the byre, house 3, as well as the paving associated with the barn, house 2) (pi 12c). Phase artefactV assemblage. horizontae Th l distributio mosf no t finds from this final phase of Norse occupation was misleading in the sense that successive ploughing had scattered them and certainly levelled out any large midden deposits. Among the bone pins, no. 22 has a cylindrical head similar to one found at Sithean a Phiobaire Uis,S t (Lethbridg animal-headeo e 1952tw e 5,8)g th ,fi f . O remarkabla s i d 5 pins2 . ,no y delicate piec f craftsmanshipeo , very mucnative th n hi e manne f animal-headeo r d pins sucs ha fro4 2 m. double-headephase no th , eII fron Broce dpi m th Burriaf ho n (MacGregor 1974, 6 g fi , no. 39) and an unpublished pin from the Brough of Birsay. These small pins with the animal heads set at right angles to the shank are quite unlike Scandinavian zoomorphic pins, which are twice the size, with the animal heads carved along the axis of the pin and a totally different art- style (e.g. Jarlshof, Hamilton 1956, fig 58; York, Waterman 1959, 105, fig 25-7). If no. 33 is accepted as an unfinished pin, it becomes evidence for local pin-production. One almost complete comb (no. 47) and two plates representing two other combs (nos 48, 49) belon single-sidea o gt d typ f compositeo e come bar wit8 4 archen ha d an d7 4 back s No . decorated with incised ring-and-dot motifs. This distinctive type of comb has been found on three other sites in Orkney: the Broch of Burrian, N Ronaldsay (MacGegor 1974, fig 11, nos 149, 150), the Broch of Berwick (Anderson 1883, fig 213) and the Brough of Birsay. It would appear to have bee nlocaa l Orcadian fashion, most probably amon Pictise gth h population doed appeat an ,s no r to have been found outside Orkney (thus militating against MacGregor's suggestion (1974, 80) tha fore havy th tm ma e been derived from one-piece combs foun Scotland)S n di e otheTh . r comb-fragments from this phase belon more th ego t commo n double-sided composite type, again of shonativ6 5 wd elinea an origin5 5 rs incise;no d decoration bese .Th t parallel alternatine th r sfo g PICTISH AND VIKING-AGE FARMSTEADS AT BUCKQUOY, ORKNEY | 189 group froobliquf e so ar m5 5 Cahercommau e. lineno n so n (Hencken 1938 26,10g fi ,Lagor d )an e (Hencken 1950, fig 97, 1283) in Ireland. Stone artefacts are represented only by two spindle-whorls (nos 85, 86) and a pebble whet- stone (no dar. o 88)kTw . onlblue th yee beadartefactar ) s85 , (nos 84 smad glassf e o the d e ,an y ar objectamonw fe e sg th likel havo yt e been imported from outside Orkney. Anothe leae th ds i r weight (no. 96) t 9- , a thi;1g likel s si havo yt e bee nbalance-weigha t representin ertog1 g (one- third of an 0re), for Viking-age weights were rarely precise (Eldjarn 1956, 355-7; Skaare 1976, 50, table 13).

NORSE HOUSE-TYPES The partial remains of three Norse buildings survived at Buckquoy: a byre, a barn and a dwelling-house. All three were rectilinear in plan-form but otherwise were quite different, and their differences appeared to reflect their functions. The byre, house 3, was subdivided internally into twobyre th ,e prope storaga d ran roundee e th area d thad an ,den t survived implies thae tth roof was hipped. No direct evidence of roof-type or material was found in any Norse level, but a basic form involving timber-fram turd ean f covering seems most likely. Bot barne hth , hous, e2 and the dwelling-house, house 1, had straight squared ends and may have been gabled; the small floor-are massive th housf d ao an e eend-wal2 housf o l suppore1 idee gablef th t ao d roofse Th . SE side-wal housf o l showee1 irregulan da ropposit e curveth t bu , e walstraights e th wa l d an , plan-for essentiallms i y thasimpla f o t e hall-house bees ha nt I .argue d elsewhere tha housee th t s claime boat-shapee b o dt t Jarlshoda Underhould an f Shetlann i l facn i te dsimilarlar y irregular hall-houses, and that the true boat-shaped house as seen in Iceland and the Faroes does not exist in Scotland (Ritchie 1974, 33). Ther currentle ear y nine excavated Norse settlement Scotlann si d databl perioe th o et d from 12te th h o t centurie h 9t amazingln e s(a th y small bod f evidencyo r sucefo hmajoa r phasf eo Scotland's history): Birsay (Cruden 1965), Westness (Kaland 1973), Skail d Buckquoan l n yi Orkney, Jarlshof (Hamilton 1956 d Underhoul)an l (Small 1966 )n Shetlandi , Freswicn i k Caithness (Curie 1939) and the Udal (Crawford 1974) and Drimore (MacLaren 1974) in the Outer Hebrides. Ther insufficiens ei t evidenc includo et structuree eth t Gurnessa Orphid san n i r Orkney (RCAMS 1946 . 26378-9p p no , , . 485 ,175p no , ) among this lis knowf o t n Norse settle- ments. All but three of these Norse sites (Westness, Freswick and Drimore) were established earlief o literall p to r settlementsn yo , som whicf eo beed hha n occupiecenturh 8t e latth ys d a s ea AD. Beneath the houses at Freswick was found tantalising evidence for a wattle and daub structure builfoundatioa n o t n laye claf o r y (Curie 1939, 94) plas datd it ;unknown e nan e ar t seemi t bu ,s most likel havo yt e bee nprimara perhapd yan s very temporary Norse structure. There would appear to have been a distinct preference on the part of Norse immigrants for settling on old occupation sites; sinc shortageo n ther s ewa goof eo d agricultura lareae lanth n sdi involvede th , preference may reflect an innate Norse habit of land conservation, very necessary in parts of the Norwegian homeland (Munch 1966). t JarlshofA e maith , n line f developmeno s e Norsth n i et farmstead were two-roomed dwelling-houses with separate outbuildings during the 9th to llth centuries, followed at the end llte o th fearlr h o y 12t dwelling-housee h th centurie f o attachmen e d th en y sb e byra on f ,o o ett finallyd 13td e 12tan th an h n i centuries, , dwelling-houses became elongate porched dan d san small outhouses were attache theio dt r side-walls (Hamilton 1956, 93-189). This broad pattern of development is supported by the evidence from other Scottish sites, and it compares well with the typology of house-forms in other Norse colonial areas; the site at L'Anse aux Meadows in 19 | 0PROCEEDING S OF THE SOCIETY, 1976-7 Newfoundland is particularly useful, because a linear sequence of farmstead types is there sup- porte seriea y dradiocarbof b so n dates (Ingstad 1970) Norse Th . e phase t Buckquosa y belono gt firse th tdevelopment e stagth f eo : dwelling-houses with separate outbuildings. termn I f sizeo s , Scottish Norse houses fall intgroupso otw e firsth : t average mn 5 i 1 s interna widthn i m l,6 lengtwit 5- secon e averagn hd th a hd an dan averagee, floom q s r 0 ares7 f ao m-4-4 widthn y 10-i b 5m 5 m , wit averagn ha e . Anthropologicafloom q rs spac5 4 f eo l datr afo societies practising mixed farming suggest sshoul m tha averagq n s allowet a e d0 b 1 f eacr eo dfo h person (Naroll 1962,187-9), which would indicate unit sevef s o foud nfivo an t r e people occupying the two size-groups of Norse houses. Such figures would be consistent with the impression of single-family farmsteads sagase giveth y n.b Phase : 10th-centuryVI Norse grave (fig , 113 s ) Before excavation began, sligha ther s moundee t wa crese humth th f o tn po ; immediately below the turf was an inhumation burial, and it was the minor disturbance of the ruined debris of phas orden i einserV o rt t this burial thacaused tha humpe d th simplgrava e s .Th ewa e shallow attempo n d pitbeed an ,ha t n mad utiliso et disturbee eth d stone mako s t sor y f formaean o t l grave setting or mound. The burial itself had suffered some disturbance from ploughing, and the bones were in poor condition (Appendix 5). The shattered skull lay on top of the pelvis, and the body appeared to have been buried in a crouched position on its right side, facing SE, aligned S-N with the feet to the N. The silver penny (no. 119), knife (no. 121), whetstone (no. 122), bone mount (no. 124 buckld )an e (no. 125) wer l founeal d close togethe skeletone waise th th f t to a r , apparentl righs it situ n yi n t o side , whil ring-pie eth n (no. 120 acrosy )la lowee sth r vertebrad ean spearheae th d (no. 123disturbes )wa d besid uppee eth bodyrrighte s parskeletoit th e f n to .,o Th n has been identified as that of a male, probably more than 40 years old (Appendix 5). Simple unprotected graves are a common form of Viking-age burial, and a crouched position for the body, though not the most common position, is not unique in N Scotland; several skeletons in the Pierowall cemetery on Westray had their knees drawn up (Thorsteinsson 1965, 165), and t Reaya Caithnessn i , bod e describes th , ywa 'possibls da doubled-ua n yi p position' (CurieJ , 1914,295). Insertio buriae th f no l int existinn oa g moun commoa ds i n economy, ofte Orknen ni y into the mounds of ruined settlements (Thorsteinsson 1965, 163; Charleson 1904, 562, 565). As a group, the grave-goods are most closely matched in N Scotland with those associated wit inhumation ha lona n i g cis t Skaillta , Sandwick, Orkney, wher perforateea d whetstonen a , iron knife with the remains of a handle and a large iron spearhead were found together with a comb and comb-case and an iron arrowhead (Watt 1888). The coin (no. 119) was kindly identified by Mr R B K Stevenson as half (deliberately cut) of a silver penny of Eadmund (AD 940-6). Although other Anglo-Saxon coins have been found in Orkney, this is the first occurrence of a coin of Eadmund; the only other coins of Eadmund in Scotlan those dar e from Isla lond yan a (Stevenson 1951, 173) s owneIt . r must have regardet di either as bullion or as a luck token, certainly not as currency. The coin shows little wear and its date, together with the dating in the first half of the 10th century of the ring-pin, ought to imply buriale th data r e fo withi thire nth d quarte 10te th hf o rcentury bronze Th . e ring-pin (no. 120) was kindly examined by Mr Thomas Fanning, whose report forms Appendix 8. The pin is in excellent condition and belongs to a type of Hiberno-Norse ring-pin with polyhedral head and plain ring whic s beehha n dated elsewher firse th e 10tt o th halet hf e silveo fcentur Th . r yAD hoard from Skaill, Orkney, contain versa y similar probabls ring-piwa d nan y burie Adc 0 D95 (Graham-Campbell 1976, 120). The general form of the knife-blade (no. 121) is very common at this period, characterised RITCHIE: PICTISH AND VIKING-AGE FARMSTEADS AT BUCKQUOY, ORKNEY | 191 by marked shoulders, thick curving bac substantiad kan l tang lengte tane curvTh d . th h gan f eo are perhaps less common features, but this may reflect nothing more than the preference of the smith or the owner (Alcock 1963, 113). The bone mount (no. 124) may have strengthened the sheath for the knife, particularly necessary if it covered the long handle as well as the blade. The latter appear muce b o sht wor whettingy nwhetstonb e th d an , e (no. 122 certainls )ha y been well used iroe nTh . buckle (no. 125havy )ema fastene bel e knife th dt th o ratheet r than fastenee dth belt itself, for its diameter at 27 mm is narrow for a belt. The spearhead (no. 123) might perhaps be better describe javelia s da smallens i headt i r rfo , than most Viking-age spearheado to t sbu substantia arrowhead.n a e b o t l spearheaA d closely matchin Buckquoe gth y exampl founs ewa d in a grave at Lyking on mainland Orkney (Grieg 1940, 80).

ECONOMY assessmeny An economie th f o t site ctermn th e i lif f efooo f so d resources mus limitee b t d even more by the conditions of partial recovery than is cultural assessment; complete middens may have been resula los s f a coastat o t l erosion. Wit exceptioe hth phase th Nors f V eI no e midden that accumulated in the ruins of house 3, there were no distinct midden horizons, and food debris as well as artefactual debris appeared to have been scattered at random over the site. Mis NoddlesB identifieo wh , animae dth l bones draws ha , n attentio additionae somo nt th f eo l hazard f survivaso l peculia fooo rt d debrireporr he n si t (Appendi. x1) Based on a sample of about 7,000 bones overall, it would seem that there was little difference animae inth l bones between Pictis Norsd han e periods l animaAl . l bones were recovered dan identified from the area excavated, but it is possible, for example, that the phase II collection is more complete than that froNorsy man e phase, simply becaus excavatee eth d havarey ama e containe morda e complete samplphase th I f settlementeI eo t survivesi s A . proportioe th , n throughout the occupation of the site was about 50 % cattle, 30 % sheep and 20 % pig. About one-third of the animals in all genera died in their first year of life, indicating that the main functio livestoce provisioe th th f no s k wa hide f n o meat d unlikelihoose an vien th .I f wo d thae tth sheep were truly wool-bearing facthe , t that only five spindle-whorls were found fro levelmall s becomes less surprising importance th ; bonf eo e pins amongs othee tth r artefacts does, however, suggest that clothes were mad textilef eo s rather than leather. Examinatio bire identificatioe Bramwelth D d th f r no boneo t M d speciey 4 sb 5 le l f no s (Appendix 2); most were wild, and domestic geese and fowl appear to have been insignificant in Norse time irrelevand san pre-Norse th n ti e economy. Bird meat doe t appeasno havo rt e played an important part in the diet in any phase. Fish appea havo rt e been more importan sourca s ta foof eo Nors n di e times tha Pictisn i h times. Identificatio fise Wheeleth A h f r boneno M y rsb implies that, although line-fishing from boats must have occurre Pictise th n di h period, deep-water line-fishin r morfa s e frequengwa n i t Norse times (Appendi . Evex3) n allowin e possibilitth r gfo y that f sievino , withoue us g e th t techniques, only the bones of the larger fish species were recovered, it does not seem likely that fishes were ever a dominant element in the diet. Marine mollusca, identified by Dr J G Evans (Appendix 4), were represented most pre- dominantly by limpets and winkles, but neither, in terms of meat weight, can have contributed to any significant extent as food sources in Pictish or Norse times. A decrease in the size of limpets from Pictish to Norse times may indicate the effects of localised over-collection. There is no mean f determinino s g whether limpets were collecte s foo da r huma fo d n consumptios a r no fish-bait. 19 | 2PROCEEDING S OF THE SOCIETY, 1976-7

CONCLUSIONS moune Th t Buckquoda y covere remaine dth fivf so e distinct phase domestif so c occupation. The house-types and artefacts from phases I and II indicate a date of occupation in the middle centurie firse th t f millenniuso m ADogam-inscribee th ; d whorl associated with phas I suggesteI sa date in the first half of the 8th century and, since there was no great time-lapse between phases I centurh 7t e data yth , n seemei II d s an likel r phasyfo . BoteI h phases fall inthistoricae oth l Pictish period, and the ogam-inscribed whorl and the painted pebble can be seen as specifically Pictish artefacts. The distinctive plan-forms of the houses, especially house 4, should help in the future identificatio f Pictiso n h settlements. Despit e lac f th ediagnosticallo k y Scandinavian artefacts house-plane th , phasef so indicat V d s IIIan e V ,NorsI e occupation; dating evidencs ei provide 10th-centure th y db y buria ruine phasf Th housf o becomld so . ha eVI e 1 mounea f do earth and rubble by the time that the burial was made in the third quarter of the 10th century, and there was no trace of deliberate destruction. Allowing an interval of about 40 years after the abandonment of house 1 for the mound to form, and estimating backwards from the date of the burial, phase V must have come to an end around AD 900. The initial establishment of the norse farmstead is unlikely to have taken place much before about AD 800, and therefore all three Norse phases should be contained within the 9th century. It has been argued elsewhere (Ritchie 1974) that ther s evidencei r considerablefo e integration between native Pict incomind an s g Norsemen in Orkney during this initial settlement period. This integration is highlighted at Buckquo face th t y tha yb arteface th t t assemblage fro Norse mth e level dominates si nativy db e products cleas i t I r. thaabsence th t f importeo e d artefact commoa s i s n situatio Norsn no e colonial sites (e.g. L'Anse aux Meadows, Newfoundland (Ingstad 1970), where the only foreign objects were an Icelandic stone lamp and a Hiberno-Scottish ring-pin), but the explanation can vary from area to area in conjunction with other factors. Megawhas drawn attention to the lack of 9th-century Norse women' argumenn a s a n 'continuitsr grave tfo IslMa e f distafth e o n th si n yfo criticae sidth t ea l settlement-stage' (1976, Orkney25)n I . , where there were 9th-century female Norse burials, the explanation must be more complex, but there can be no doubt that some form of social integration between PicNorsemad tan n centurexisteh 9t probabl e d least da yth an n ti y int 10te oth h century.

CATALOGUE The finds have been deposited in Tankerness House Museum, Kirkwall, Orkney; museum registra- tion numbers are given in brackets (THM). Bone pins (figs 4, 5) 1 Perforated bone pin, flat section, 85 mm long, phase I (THM 1976.40). 2 Bone skewe longm m r, 4 pinphas9 , eI (TH M 1976.86). 3 Bone skewer pin, 69 mm long, phase I (THM 1976.44). 4 Bone skewe lonm gm r (incomplet1 pin9 , t tip)ea , phas I (THeII M 1976.70). 5 Bone pin, flat triangular head, 50 mm long (incomplete shank), phase III (THM 1976.71). 6 Bonepin, flattriangularhead,hippedshank,29mmlong (incomplet tip),phast ea (THV eI M 1976.78). 7 Bone pin, flat ovoi dlong m headm , 0 phas7 , I (THeI M 1976.37). 8 Bone pin, flat fan-shaped long m headm , 7 phas4 , (THV eI M 1976.85). 9 Bone pin, expanded flat-topped head, ova section i l n witdecorativo htw e ova longlm dotsm ,0 4 , phase IV (THM 1976.80). 10 Ball-headed bone pin, hipped shank, 29 mm long, phase IV (THM 1976.76). 11 Ball-headed bone pin, head decorated overall with dots, hipped longshankm m , 0 phas4 , V eI (THM 1976.81). RITCHIE: PICTISH AND VIKING-AGE FARMSTEADS AT BUCKQUOY, ORKNEY 193

10 11

15

17 14 19 21 23 16 20 FI G4 Buckquoy: bone pins (scal :1e1 ) 19 | 4PROCEEDING S OF THE SOCIETY, 1976-7 12 Ball-headed bon lonem pingm 2 (incomplet3 , e shank), phas (THeV M 1976.96). 13 Bone pin, broken ball-head, hipped shank, 40 mm long, phase III (THM 1976.72). 14 Bone pin, flat-topped ball-head decorated overall with dots, hipped longshankm m , 1 phas4 , V eI (THM 1976.88). 15 Bone pin, flat-topped ball-head decorated overall with dots, collar between head shankdan , 20 mm long (incomplete shank), phase V (THM 1976.89). 16 Thistle-headed bone pin, bands of cross-hatched decoration on head and lower shank, decoration on flat top of head, 57 mm long, phase IV (THM 1976.83). 17 Nail-headed bon longm e m pin , 0 phas4 , I (THeII M 1976.69). 18 Nail-headed bone pin, hipped longshankm m , 3 phas3 , e IV (THM 1976.84). 19 Nail-headed bone pin, hipped shank, 48 mm long, phase IV (THM 1976.77). 20 Bone pin, squat cylindrical head, hipped lonshankm gm 5 (incomplet5 , e shank), phas I (THeI M 1976.39). 21 Bone pin, cylindrical head with collars at top and base, 48 mm long (incomplete shank), phase IV (THM 1976.82). 22 Bone pin, cylindrica lonl m headm g 0 (incomplet3 , e shank), phas (THeV M 1976.98). 23 Bone pin, flat-topped circular head, collar between head and shank, hipped shank, 50 mm long, phas I (THeI M 1976.45). 24 Animal-headed bone pin, carved animal' ssmaln o hea t ldse ball-head , facin t righga t anglo et shank, 37 mm long, phase II (THM 1976.38). 25 Animal-headed bone pin, animal's head set at right angle to shank, four bands of cross-hatched decoration incise longshankn m do m , 7 phas4 , (THeV M 1976.99). 26 Animal-headed bone pin, carved animal's collarn heao t dse , facin t righga t angl shanko em t m 0 2 , long (incomplete shank), phase V (THM 1976.95). 27 Lower part of shank of bone pin, 26 mm long, phase I (THM 1976.36). 28 Unfinished hea upped dan r par shanf to bonf ko e pin, flat ovoi dlong m headm , 1 phas,2 (THI eI M 1976.75). 29 Shank of bone pin, indications of a flat ovoid head, 41 mm long, phase IV (THM 1976.74). 30 Lower par f shano t f bonko longm e m pin ,8 phas1 , (THeV M 1976.92). 31 Central par hippef o t d shan f bonko longm e m pin ,9 phas3 , (THeV M 1976.93). 32 Lower part of hipped shank of bone pin, 38 mm long, phase V (THM 1976.94). 33 Upper par f ?unfinisheo t d bon longm e m pin , 9 phas3 , , (THeV M 1976.90).

Bone implements ) (fig6 , 5 s 34 Bone pin or awl, cleft head, 47 mm long, phase IV (THM 1976.79). 35 Double-ended bone implement r fish-gorgeo l longm aw , m , 8 phas4 , (THeV M 1976.97). 36 Bone point, 105 mm long, phase I (THM 1976.46). 37 Bone point, long11m 3m , phas I (THeI M 1976.43). 38 Bone point, rounde longdm tipm , 3 phas5 , I (THeI M 1976.42). 39 Bone point, 94 mm long, phase II (THM 1976.41). 0 4 Bone point, rounde longm d m tip , 5 phas9 , I (THeII M 1976.73). 41 Bone point, 75 mm long (incomplete), phase V (THM 1976.91). 42 Bone implement, broad rounded tip, cylindrica longl m headm , 6 phas9 , (THV eI M 1976.87). 43 Bone spoon, pointed stem, shallow bowl, 152 mm long, phase I (THM 1976. 103). 44 Bone handle, socket 32 mm deep, decorated with three concentric incised lines round end of handle, longm 92m , phas (THV eI M 1976.48). 45 Bone mount, incomplete remains of two peg-holes, 59 mm long (incomplete), phase I (THM 1976.47). 46 Bone mount, two peg-holes retaining bone pegs and one incomplete peg-hole, 58 mm long (in- complete), phas (THI eI M 1976.51). Bone combs (fi) g7 7 4 Single-edged composite bone comb, four plates, three iron rivets suspensioo tw , n holes, double- ring-and-dot decoration, 70 mm long, phase V (THM 1976.108) (pi 13b). 48 End-plate from single-edged composite bone comb iroe non , rivet, ring-and-dot decorationm m 3 2 , long, phase V (THM 1976.107). RITCHIE: PICTIS VIKING-AGD HAN E FARMSTEAD BUCKQUOYT SA , ORKNE I Y19 5

34 35

42

24

FI G5 Buckquoy : bone artefacts (scal :e1 excep 24-s :12 )t no t6 a

FIG 6 Buckquoy: bone spoon (no. 43) as found with handle crushed on to stone, and bone handle (no. 44) (scale \ : 2) 19 | 6PROCEEDING S OF THE SOCIETY, 1976-7 49 End-plate from single-edged composite bone comb, one iron rivet, 20 mm long, phase V (THM 1976.114). 50 Double-edged composite bone comb, three plates, three iron rivets suspensioo tw , n holes, ribs decorated with incised outlin longm dotsd m ,e an 4 phas 7 , eI (TH M 1976.49). 51 End-plate from double-edged composite bone comb, incomplete rivet-hole longm m ,9 phas1 , I eII (THM 1976.102). 52 Plate from double-edged composite bone comb, iron rivet retaining fragmen longm m ribf ,o 4 t 4 , phase III (THM 1976.100). 53 Double-edged composite bone comb platesx ,si suspensio e ,on n hole, four iron rivets, ribs decorated with incised crosses, end-plates shaped, 116 mm long, phase IV (THM 1976.104). 54 Plate from double-edged composite bone comb e rivet-hole on long, m m ,7 1 phas, (THV eI M 1976.105). 55 Double-edged composite bone comb, one end-plate and one central plate missing, ribs incised with pair diagonaf so lonl m linesm g 5 (incomplete)9 , , phas (THeV M 1976.120). 56 Two end-plates from double-edged composite bone comb, one iron rivet, one plate retaining fragment of rib decorated with two incised diagonal lines, 21 mm and 23 mm long respectively, phase V (THM 1976.116). 57 End-plate from double-edged composite bone comb, one iron rivet, 20 mm long, phase V (THM 1976.111). 58 Plate from double-edged composite bone comb, one rivet-hole, 9 mm long (incomplete), phase V 1976.110)M TH . 59 Plate from double-edged composite bone comb, 21 mm long (incomplete), phase V (THM 1976.110). 60 Rib fragment from composite bone comb, one rivet-hole, ring-and-dot decoration, 44 mm long (incomplete), phas (THeV M 1976.112). 61 Rib fragment from composite bone comb, incomplete rivet-hole, ring-and-dot decoration, 22 mm long (incomplete), phas (THeV M 1976.113).

47 48 49

53

54 56 57

FIG 7 Buckquoy: composite bone combs (scale 1 :2) RITCHIE: PICTISH AND VIKING-AGE FARMSTEADS AT BUCKQUOY, 7 ORKNE19 | Y Cetacean bone 62 Squared bloc f cetaceako n bone with thickconcav m , 31m mm , e5 uppe12 y b r surfacem m 2 14 , phase II (THM 1976.50).

Human bone (fi) g8 63 Spindle-whorl made from human femur-head, 36 mm diameter, unstratified (THM 1976.174). V

64 69 77 82

87

FIG 8 Buckquoy: pottery, spindle whorls and painted pebble (scale 1 : 2)

Pottery (fig 8) All sherds have a fine sandy fabric containing mica and a little small stone grit: outer surfaces are buff to red in colour, inner surfaces black. 64 Four wall-sherds and one flat-based sherd (base diameter c 150 mm), 9-10 mm thick, phase II (THM 1976.60). wall-sherdso Tw 6 5thick m m ,9 , phas I (THeI M 1976.61). 66 Wall sherd, 10-1 thickm 2m , phas I (THeI M 1976.62). 67 Wall sherd, 9 mm thick, phase II (THM 1976.130). 68 Wall sherdthickm m ,6 , phas I (THeI M 1976.130). 69 Wall sher flat-based dan d sherd (base diamete mm)0 7 rc , thick6-m 9m , phas (THI eI M 1976.131). 70 Two wall-sherds, 7-9 mm thick, phase II (THM 1976.132). 71 Two wall-sherds, 7 mm thick, phase II (THM 1976.129). 72 Wall sherd, 6 mm thick, phase II (THM 1976.127). 73 Wall sherdthickm m ,,9 phas I (THeI M 1976.126). 74 Sixteen wall-sherds, 7 mm thick, phase II (THM 1976.125). 75 Wall sherd, 7 mm thick, phase II (THM 1976.124). 76 Wall sherd, 9 mm thick, phase HI (THM 1976.121). 77 Flat-based sherd (base diameter c 50 mm), 7 mm thick, phase IV (THM 1976.59). 78 Wall sherd, 5 mm thick, phase IV (THM 1976.123). 79 Wall sherdthickm m ,6 , phas (THV eI M 1976.128). 80 Wall sherdthickm m ,9 , phas (THeV M 1976.133). 81 Wall sherdthickm m ,7 , unstratified (THM 1976.175). 82 Spindle-whorl, roughly made from pottery sherd, 35 mm diameter, phase IV (THM 1976.122). 198 I PROCEEDINGS OF THE SOCIETY, 1976-7

90

- - • ~. -^ ---"£•' -

FIG 9 Buckquoy: stone gaming boards (scale 1:2) RITCHIE: PICTISH AND VIKING-AGE FARMSTEAD BUCKQUOYT SA , 9 ORKNE19 | Y Stone (figs 8, 9, 10) 83 Limestone spindle-whorl, 34 mm diameter, phase II (THM 1976.57). 84 Limestone spindle-whorl, ogam inscription incised round perforation on one face, 36 mm diameter, phas I (THeI M 1976.56) (pi 13a). 85 Limestone spindle-whorl, 31 mm diameter, phase V (THM 1976.162). 86 Haematite-stained siltstone spindle-whorl, 35 mm diameter, phase V (THM 1976.161). 87 Quartzite pebble, painted overall with smalm m l5 2 circle , brown mm si 5 3 ny pigmentb m m 0 4 , thick, phas I (THeI M 1976.58).

92

FIG 10 Buckquoy: stone gaming board (scale 1 :2)

88 Siltstone thick whetstonem m ,5 1 phas , (TH,e V mm wor0 M3 bot n ny o b 1976.159) h m sidesm 3 ,8 . 89 Siltstone pebble, polished face on one side, 56 mm diameter, 15 mm thick, phase IV (THM 1976.157). 90 Flagstone gaming-board, incised with 36 cells, circle incised round central intersection, 235 mm by 175 mm, phase IV (THM 1976.158) (pi 13c). 91 Sandstone gaming-board, incised wit cells6 h3 , circle incised round central intersectiony b m m 3 ,25 , unstratifie172mm d (THM 1976.177). 92 Flagstone gaming-board, lightly incised with 36 cells and circle at central intersection, overlain by , unstratifieincoherenmm 0 17 dy b t (TH line m circlesd m M san 0 1976.176)19 , . Shell 93 Limpet, two perforations at apex, 55 mm diameter, phase I (THM 1976.55). Glass 94 Fragment of blue glass bead, 7 mm diameter, 4 mm long, phase V (THM 1976.163). 95 Blue glass diameterbeadm m long m 7 , m , 3 ,phas (THeV M 1976.184). Lead(&g) 11 96 Lead weight, base diameter 12 mm, weight 9-1 g, phase V (THM 1976.154). 200 PROCEEDINGS OF THE SOCIETY, 1976-7

Iron (fig 11) 97 Fragmen single-edgef o t d knife-blade lonm gm 0 (incomplete)7 , , phas I (THeI M 1976.67). 98 Knife-blade, 76 mm long, phase II (THM 1976.65). 99 Fragment of knife-blade, 45 mm long (incomplete), phase II (THM 1976.66). 100 Fragmen knife-bladef o t lonm m g 1 (incomplete)3 , , phas I (THeI M 1976.144). 101 Two fragments of single-edged knife-blade with part of tang, together 62 mm long (incomplete), phase III (THM 1976.137). 102 Fragmen knife-bladef o t lonm gm 6 (incomplete)6 , , phas I (THeII M 1976.136). 103 Fragmen smalf o t l single-edged knife-blade lonm gm (incomplete)0 4 , , phas I (THeII M 1976.135). 104 Fragment of knife-blade, 78 mm long (incomplete), phase IV (THM 1976.138). 105 Two fragments of single-edged knife blade, together 55 mm long (incomplete), phase IV (THM 1976. 139). 106 Four small fragment knife-bladef o s , phas (THV eI M 1976.140). 107 Two fragments of single-edged knife-blade, together 70 mm long (incomplete), phase V (THM 1976.151). fragmento 10Tw 8 knife-bladef so , togethe lonm gm 3 (incomplete)r7 , phas (THeV M 1976.146). 109 Fragmen knife-tangf o t , rectangular sectionlonm gm 3 (incomplete)3 , , phas (THeV M 1976.153). 110 Fragmen (incomplete) m f iroo tm n3 1 strip iroe y b on n, m ,rivet phasm 5 3 ,(TH eV M 1976). 111 Iron double-heade dlong m rivetm ,8 phas1 , I (THeI M 1976.68). 112 Nail fragment, 22 mm long, phase III (THM 1976.134). 113 Nail, 40 mm long, phase IV (THM 1976.143). 114 Nail fragment longm m ,4 phas3 , (THV eI M 1976.141). 115 Nail fragment, 17 mm long, phase IV (THM 1976.145).

96

3 12 120 2 %12 ? FI 1 GBuckquoy1 : metalwork (nos 96-7, 110-11 phasd )an gravegoodI eV s (nos 120-4) (scal :2e1 ) RITCfflE: PICTISH AND VIKING-AGE FARMSTEADS AT BUCKQUOY, ORKNEY | 201 nail-fragmentso 11Tw 6 , rectangular section, phas (THeV M 1976.149). 117 Nail fragment longm m , 2 phas4 , (THeV M 1976.150). 118 Nail fragment longm m , 7 phas6 , (THeV M 1976.148).

Grave-goods associated with phase VI burial (fig 11) 9 Hal11 f silveo f r penn Eadmundf yo , deliberatel t (THycu M 1976.166). 120 Bronze ring-pin, Hiberno-Norse type, decorate shankn do lonm , hea gm ring d o fro6 t dan p 16 ,m ti head (THM 1976.165). 121 Single-edged iron knife with tang, fragment bonf so e handle adherin lonm gm tango g t 0 (TH12 , M 1976.167). 122 Black shale whetstone, much worn, perforate lon m ende gm on 5 (TH ,8 t da M 1976.172). 123 Small iron spearhead, leaf-shaped, socketed lonm gm 7 (TH8 , M 1976.169). 124 Fragmen bonf o t e mount iro2 , n lon rivetsm m g 5 (incomplete4 , ) (THM 1976.171). 125 Fragment of iron buckle, max diameter 27 mm (THM 1976.170).

APPENDIX 1 animae Th l bones from Buckquoy, Orkney by Barbara Noddle, Department of Anatomy, University College, Cardiff Ther essentiae ear l limitations inheren assessmeny an n ti excavatef to d faunal remains. This material derives almost entirely from waste middens, and these structures have a natural history of their own depending on their content and their rate of accumulation. Isaac (1971) showed that under African conditions considerably less than 50 % of a pile of bones could be recovered after 6 months, and that some bones are more readily destroyed than others. A midden may be protected from scavengers by the rapid accumulatio f non-ediblno removee b e y materialma d t i entirel r builders o ,it aesthetin y a yb s sa a r co space-occupying nuisance. This may account for periods in the excavation in which bones were fewer than others or entirely lacking, and it may not be a result of changes in agricultural practice or diet. Also the materia renderee middea b n i y l nma d unrecognisabl e mechanicath y eb l force f lateo s r inhabitants carryin t theiitf o g;ou frorp bone liveto m th n eso poin vief o t w this result larga n si e numbe f durablro e loose teeth which survive after the fragmentation of other bones, and this was sometimes the case at Buckquoy; indeed greao s bee,d weae ha tn th r that som thesf eo e teeth were sand-polished. The bones from Buckquoy were in an excellent state of preservation, as frequently occurs in Orkney, and, although fragmentary, about 80% were identifiable. The material from the different genera was analysed in several ways, where the number of fragments justified this (in the case of the main domestic genera, cattle, sheep and pig). In the following tables the material will be considered under these headings: numbe f identifiablo r e fragments, minimum numbe f individualso r , numbe f fragmento r r pe s minimum individual, anatomical analysis of the different bones present where numbers warrant, and age- individuarange th f eo l where possible bonee .Th s were measured whenever suitable; dimension comf so - plete bones will be presented as tables (and frequently occurring measurements of bone widths are availabl archivs ea e material efforn a determino n t )i t e whethe changy r an ther s typf eo e animaf wa eo l between the two cultures represented during the period of the excavation. This method of presentation is adopted from that of Chaplin (1971) who discusses the validity of the statistics employed. The item number of fragment individuar s pe t bee no ns widellha y used previously seemt ,bu s particularly usefu thir fo ls site wher t seemei s tha bonee th t s from whole carcasses (comprising som identifiabl0 e14 e bones excluding ribs) were originally placed on the midden, with certain exceptions deduced from the anatomical analysis; the number of fragments per individual then gives some idea of how much material has been lost, for the reasons discussed above. The ages of the animals are given in groups to allow easier presentation than exact chronological ages 'text-booke th ; ' data usually employed (Silver 1963 derive)s i d from improved animals of the 18th century and later, and is not thought to be applicable to these animals of a considerably earlier date and living in a less favourable environment. alln I , some 7,000 bone-fragments were identified. Thes groupee ear d unde headinge rth cattlef so , sheep, pig and 'others'; this last group comprises goat, horse, deer, seal, dog, cat and rodent, and a few fragments from non-mammals, amphibia crad alse nban ar o included. Tabl summara s i eI y table showing distributioe th differene th f no t gener differene th n ai t periods, both fro poine mth vie f o t numberf wo f so 20 | 2PROCEEDING S OF THE SOCIETY, 1976-7 identified fragment minimud an s m numbe f individualsro . This las derives i t d fro mose mth t frequently occurring fragment, combined wit age-range hmeasurementth e th d ean fragmentse th f so assumes i t I . d that the diiferent layers contained separate groups of individuals. Fro seee b m nn thaTablnumbee ca t th ti e1 f bone-fragmentro s fro diiferene mth t periods varied widely. The numbers were highest in the ploughsoil, Norse V and Pre-Norse Ib, and lowest in Norse III and Pre-Norse Ilii, whic havy hema obscured some small differenceculturae th o t e l changedu s alsd o,an TABLE 1 Proportion bone-fragmentf so individuald san differene th r sfo t periods firse Th .t figurnumbee th s ei identifief ro d bone-fragments, followed by the %. The second is the minimum number of individuals, followed by the %. The numbee thirth s di fragmentf o r individualr pe s . Period Cattle Sheep Pig Other Topsoil (1) 833 (45%) 611(33%) 297(15%) 97(7%) (2) 21 (42%) 17 (34%) 12 (24%) (3) 39 32 20 Norse V (1) 973 (44%) 672(31%) 382 (17%) 168(8%) (2) 28 (47%) 19(31%) 13 (22%) (3) 35 29 27

Norse IV (1) 303 (62%) 112(23%) 47(10%) 26(5%) (2) 16(55%) 8 (27%) 5(9%) (3) 19 13 9

Norse III (1) 120(47%) 84(33%) 37 (12%) 21 (8%) (2) 6(29%) 10 (50%) 5(21%) (3) 20 9 8

Pre-Norse Ilii (1) 150(38%) 135 (34%) 92(23%) 28(5%) (2) 5 (22%) 10 (44%) 8 (32%) (3) 30 12 10

Pre-Norse Ilib (1) 201(53%) 95 (25%) 70(19%) 9(3%) (2) 10 (42%) 9(37%) 5(21%) (3) 20 9 13 Pre-Norse Ilia (1) (56%4 1 ) 11 (44%) Nil Nil (2) 5 (50%) 5(50%) — (3) 3 2 — (50%2 67 )Pre-Nors ) (1 b eI 285(21%) 296 (22%) 20 (7%) (50%6 2 ) ) (2 7(13%) 20 (37%) (3) 26 14 13

Pre-Norse la (1) 112(62%) 26 (14%) 28(16%) 12 (7%) (2) 6(46%) 3(23%) 4(31%) (3) 19 9 7 lo Pre-Norswn i whatevey B . a eI rIlid criterioaan adopteds ni materiae bule th f th , ko l derives from cattle. The proportion is highest in Norse IV (62% fragments, 55% individuals) and lowest in Pre-Norse Ilii (38 % fragments, 22% individuals). Sheep comes second, ranging from Pre-Norse Ilii (34 % fragments, 44% individuals Pre-Norso t ) (21b eI fragmentsn i : individuals)% 3 1 thirg , pi dd ranginan , g from Pre- Nors (22b eI % fragments, 37% individuals fragments% 0 Norso )t (1 individuals % V eI 17 , ) and nonn ei Pre-Norse Ilia. However, ther littls ei e overall variation t seemi d san , likely tha availabilite th t livestocf yo s kwa dictated by the environment rather than by direct choice. There is one difference which might have significance, and that is the change in number of bones per individual amongst the smaller genera, which increases after Norse III, wherea numbee sth fragmentf ro individuar spe l amongst cattle remains much the same throughout. A variety of reasons for this may be suggested: TABLE 3: Sheep bones - Anatomical distribution

u o g -2 u g 03 ^ 1 1 3 5 52 « M B.8 0 •3 e M r* ex E to 13 c 1 1 i 1 0 S 3 eg «- cd I •5 -S & f 1 i ft> O S « 1a 1 15 In tn > on HM tti H (£ fc O < u f <£ 3 1 _ A . A _ A A A Cfl CO a> • o • * Q U ' | i -o 10 M "O i <" * 2 X O 1 X X 2 4 ^ £ < X S ^ oX ^ ^ o\ ^ * * 2 t x 2 x 3 1 Topsoil 611 17 NW 4-4 3 4.1 4 W 12-N 4 2- 3 9 W 3-N 5 6 0 4-1 4 N W 2-6 5-2 12 N 6-2 12 N J 1-7 0 3-81 1 1-N 5 7 6-2-35 6 0 31 2-17 2J 51 21 W J I 21 M W5I 2W 51 41 M 3 2WJ 5M2E E J M 5M 4M 2M E 2 21M V 664 16 N 3-8 4 5-6 7 J 2 12-N 1 3- 6 9 W 2 4-9 1 6-81 3 N 3W 4-2 4-8 9 2W 8 13 N 1-9 8 J 3-3 12 2-N 2 4 N 6-3 0 61 9-7 18- 6 5 3N2J 31 M W 2J2I J3I 2J2I W2J 41 W 51 21 2M 41 3M E 2J M 3M 41 5M 3M 2M PE PE IV 101 6 N 1 3 4 1 I 11 7 2 — 271 42 t I 1 6 12 3 J J 5 2 2 I2 1 2 42 9 1 N 3 5 1 31 1 W J 21 M M WI I 21 M III 57 5 J2I 7 2 1-8 1 J 18 1-8 1 A 1-4 1 M 3-65 21*I 5-3 10-6 21 71 — 1-8 1 J 1-8 1 — 53 3 J 3-61 19-2 2 M 2M 21 M M 21 Ilii 135 10 3N 5-9 3 5-1 3 J 19-2 0-7 1 — 4-4 4 2N 5-9 3 N 3.7 7.4 4 N 3-73 N N 3-5 5-4 79 3-N 3 7 4-1 7 11-4 8-838 1 W2J I 21 I I W J W2I M 21 ME E 2M M M M E 1Kb 84 8 N W 4-7 2 5-9 3 I 10-7 4 9-5 3 31 W W 3 5-9 31 8-3 5-9 3 21 5-9 3 I 2-421 1-2 — 3-6 7-2 3 N 2-4 — 7-2 2 N J3I MP J J M M 2M M WI P MP Ib (25) 55 4 N J 3-6 1 3-6 1 W 5-— 4 3-1 6 — — — 2 1 — — 5-2 4— 7-— 2 I 1-811 3-6 2 1-8 16-4 3 N 30-9 2 9-1 1 — 21 M M Ib (29) 96 4 W — — 5-2 3 J 17-— 7 6- 3 3 J 4-3 2 3-11 I 7-3 2 4-212 6-2 3 I 4-221 2-1 — — 13-5 4 N 5-21 19-8 2 I 21 M I M M I Ib (31) 125 5 W 3-2 — 3-2 2 J 13-J 6 4- 4 8 J 8-4 84-0 3 I 3-2 8 3 21 7-2 4 W 2- 2 4 21 6-5 4 — 13-6 4 J 8-82 6-4 2 I J2I I 21 M M M JI 21 M M M M

TABLE 4 Pig bones - Anatomical distribution

[rt - n> u -7 33 :S i_ a t- "«•— "s rt co 3 .s is f 1 f 2 S1 "-o§ § "> B JD ^- H W 13 o 1 ! 1 3 a *••' O 0) •-* B TO Wl J3 O & 1 £-> £>• C/5 O OH P-l H H O •* OH J —— > — 1TO f i — CO VI VI CO VI CO •0 bo a •o -a •a I -o a ^ o M ^x o s^ M ^ •o S> -a GO " &0 ™ W) I Z S ^ S S-? S < ^ ^S S <>

o E g •5 w -^ "^ 1 % ~ 8 i § i | | § aa 3 Wl 3 1 tu-i I « E B | E •5 .2 5 Oi gi Wl I *- ! i 1 I 1 1 tl .S *•" *8 S X $ 1 | oi I I u < uft2 S G« C/D c/5 21 Hj H w ( — — *—— — ^ t — • —— ^ ——— i f — ^ f —— - ~ -, t -i f "w? (— •*- — ^ r- J2 ' ^/? "i/? V) CO ' ^ ' CO1 a o o "^ *0 *O QD *O •a a •o a -o 6 o-a .o «, •§ 0 Z Z £ < S-? £ x 1 1 * * x ! * !< N? "§ ™ 1 1 =>5 s< 5 < ^ ° 2Z ^. o <^ N 3 Topsoil 8 2-1 4 7 7 79 3 3 N 1-3 12 4734 N 1 4W 2 , 2 5 6 W 3 4 7 W 2 6 2NN 4 3 1 10 9 5 15 11 2-5 1-5 19 13 3W, W 2W2J W W I, 4M I W 2W IM 3IM J 21 4M 21 2J3I P6E P 6E M M M 5M V N 2 2224 2 14 4 9 2N 0-6 15 4 7 3-5 7 N 3 3 M 42 N8 3 4 10 N 3 10 N 4 14 4 2N41 7 I 1 2 9 3 14 3 N 1 1 18 2D 2J J2I 2J2I W 31 J W 2b MP 41 3M M 3E M 21 2M 41 I J, I 3E P6E M 4M 7M 4M IV N 21 9 02-8 2 N 2- 8 3 1 80-3- 2 58 3-2 8 N 2-8 2 I 3-3121 5 4 N 05115 5 W 7 4 7-6 13 31 1 21 6 NIM 2W J I M M W I P2E I M J 2M P2E I Ili i 1 1-2 53 15 8 01 5 — N J 3 2 3 4 N J 2 4 10 0-6 1 J 06 3 1 3 N N 1 1-5 8 6 4 J 22 46 0-67 2 M W J 1 I I M M Ilib 188 8 2N 52 4 3 N 0-5 10-6 6361 — M 0-1 5 1-5 2-1 1 N 1-5 1 J 0-5 4 N 3-1 4 6-2 9-5 4 J 4 19- 1 65- 12-8— N 4 I W J 1 1 2M I 2M M 2M Ib (25 N )2 612 7 73 2 — 1-64- 6 7 1 — 0-— 7 N 2-4 2 N 6 1-6 N 1 11-8 4 2 NN 3 3-2 9 3-3 2 8-7 10-2 4 N 16-5 3 2-3 — 11 21 2W M J M J 2W M .TI M J N Ib5- 0 4 (2921 )6 47 J 3- 3 90 1 4- 1-5 84 2-4 3 J 4-8 6 N 4-2 1-6 3 N 3-6 4 J 2-5 201 6 2-'l 10 8-8 5 N 20-6 14- 6 0-I 8 14-— 4 41 P I I 41 I I M I 3M 3M E 2M M M M 2M Ib (31) 477 9 NW 1-8 1 2-7 3 E 1-4 15-7 5-4 6 3-4 3 J 3-5 3 N 6-5 2-5 5 N 5-0 4 N 2-7 41 2-57 6-1 4-6 4 W 16 3 4-8 4-8 12 7 2J I JI M I 2M I I 3M I MP P 3E M 3M 2M M 3E N la 6 2-6 2 111 M 6- 2 0-2 8 9-8 7-— 3-13 5 N 4-5 2 I 6-2 4-5 3 J 2-7 3 N N 2 2- 72-2 7 6-2 4-5 1 — 2 6-2 8 3-5 6-— IE2 W J E J 2M W I I M J : PICTISH AND VIKING-AGE FARMSTEADS AT BUCKQUOY, ORKNEY | 203 . befora e this perio middee dth beins nwa g use butchering-wastr dfo largee e th only d r cattlan , e wer poine e th slaughterf bonet o ta t dou , wherea t alwaysno s thiss wa practise d wit smallee hth r genera (as happens in the modern butcher's shop); b. the smaller bones were removed extensively for working, though pig-bone is not very often employed for this purpose; c. the smaller bones were more vulnerable to the mechanical forces exerted upon them by overlying occupation, but this is not borne out by the numbers of loose teeth; d. the small bones were removed by scavengers which had become scarcer by the later periods. compris4 d Tablean 3 anatomican e, a s 2 l analysi bone-fragmente th f so cattler sfo g , sheepi d pan respectively, where there were sufficient number f fragmento s layee th warran o t rn i s t this (over 50). These tables also include the minimum number of individuals estimated for each fragment, and an estima- age-range tioth f no eacr efo h fragment where suitabl sucr efo h deductions analysie .Th bees sha n carried skuln mandiblo d t lan ou e fragments, horn cores (no coursf pig)e o t th ,n e i vertebrae ,bone e mosth f f sto o limbse th , includinonlyg pi ,e ulnwhere th gth n ai e this bon separats ei e fro radiuse mth . Carpald san tarsals have been included, as far more of these bones were found in this excavation than any other in the writer's experience. Sessimoid bone sternebrad san includee ear cattler dfo thes a , y were again unusually frequent. Ther alss ei ocoluma loosr nfo e teeth age-rangee Th . s comprise newborn (N), weanling (W), juvenil abouo t p yeae1 t(u moder y rb n standards) (J), immaturepiphyseae th o t p e(u l closur latese th f teo maturing bones, abou yeart4 modery sb n standards matur d (I),an e possibl(M)s i t I . asseso et furtheo stw r categorie evidence th n so dental-wearf eo , post-matur elderld an ) ye(P (E) chronologicao N . e b n ca e ag l suggeste thesr d fo the s ea lasyo depentw t muco ds h upon environmental conditions boneg pi o s N .wer e identifie perioe th n di d Pre-Norse Ilia t sinc ,bu fe o es w bones were recovered from this perio wholea s da , it is probable that pigs were kept. Among all the genera, the highest proportion of any bones found was that of the phalanx, though of course there are a large number of these bones (24) in each animal. The bone present in the highest proportion migh vertebraeexpectee e b tth e b thest o dt bu , e bone largele sar y cancellou mord san e readily destroyed tha compace nth t long bone. Amon cattle gth e bones proportioe th , f pelvino c fragments i s variable proportioe th , f radiano l fragments comparativel Ilibn femorai d an ,w ylo l fragmentn i w lo s Iliiproportioe .Th metapodiaf no earle l th fragment yn i Pre-Norsw lo s si e periods differencee .Th e th n i s long bones mighdifferenf o explainee e tb us e t th lon y db g bone bone-workinr sfo differene th t ga t periods. When the individuals represented by the different bones are considered, it is obvious that many adult individual t representeno e sar lony db g bones explainee b ; face y thith t s ma y thadb t onc ebona e reaches adult size (tha immaturs ti possiblt shafa older)r eo e no s ti ag fragmentt ,o i et . These individuals are represented by the tarsal bones. A few adult animals are only represented by teeth, and this, coupled with the occurrence of long bones only as fragments, does suggest that adult bone was extensively used as an industrial material, or that adult bones were broken up more than younger ones by natural forces, for processin other glur fo gfo r r eo purposes hige Th h. proportion tarsae th f slo bones also suggests thae th t carcas suspendes swa butcherinr dfo hamstrine th y gb g tendon techniqua , e tha stils i t l employed today, and that when butchering was complete the hock bones were discarded intact. The character of the metapodial fragments suggests that many of these bones were in fact worked, and that many of those found were discarded after unsuccessful splitting. Amongst sheep proportioe th , phalangef no carpal/tarsad san l bones foun lowes dwa r than thaf o t cattle. These bones are much smaller in this genus, which might account for their loss. Far more indi- vidual representee sar lone th gy dbonesb particulan ,i latestibiae e th rth n t,i period topsoif so Norsd lan e . AgaiV n change materiaw ra n sbonr i lfo e working migh invokede b t t anothebu , r explanation mighe b t change cookinn si g method ste;a w prepare smala n di l vessel might resul bonen ti s being broken into less recognisable fragments thapreparee non larga n di e vesse joina r o lt preparee roastingb y db y ma t I . noted that in Norse III there is a low proportion of metapodials, in Pre-Norse Ilii few scapulae and tibiae, Pre-Norsn i d an e Ilib many radi fet wibu phalanges proportione boneg .Th pi fairl e e sar th y f sconsisteno t throughout. anatomicae Th l dat bees aha n summarise compareTabld n di an , e5 d wit proportione hth eacf so h bone in the whole skeleton. The purpose of this is to provide a comparison for other sites, although at the momen t possiblno s i t distinguisti o e t h between change midden si n conten 'input'o t e thosd du te ,an edu to subsequent events. The table includes the proportions of the different skeletal elements in the whole animal. It is assumed that the skull normally breaks up into eight identifiable fragments, and, for this particular site, where preservation is so good, all the coccygeal vertebrae are included. The shorter tail of 20 | 4PROCEEDING S OF THE SOCIETY, 1976-7

TABLES Average anatomical distribution of animal bones

intact Cattle Sheep Pig Recovery goat in

animal /O /o % Africa (Isaac 1971) Skull 5-7 3-7 4 2-6 Mandible 1-4 3-5 4-3 4-2 Vertebrae 33 10-7 13-3 11-7 0-7 Scapula 1-4 4-7 5-2 2-4 Humerus 1-4 3-2 4-2 4-5 100 (distal) Radius 1-4 3-3 4-9 3-7 23 Ulna 1-4 — — 3-4 76 Pelvis 1-4 4-1 5-3 3-3 Femus 1-4 2-2 5-9 5-9 22 (proximal) Tibia 1-4 2-9 6-1 6-0 88 (distal) Calcaneum 1-4 2-0 2-7 — Astragalus 1-4 3-5 3-7 — Carpals/tarsals 5-7 7-1 1-6 8-5 Metapodials 2-8 7-7 10-5 10-9 33 (proximal) Phalanges 17 16-3 9-0 12-5 Loose teeth 23 13-7 14-2 7-1

TABLE 6 Percentage of individuals in different age groups found at different periods Age code Age code M and Period Animal NWJ Age code 1 older Topsoil Cattle 42 16 42 Sheep 30 30 40 Pig 33 33 33 Cattle 33 19 48 Sheep 35 30 35 Pig 46 30 24 IV Cattle 50 10 40 Sheep 50 37 13 Pig 20 40 40 III Cattle 57 14 29 Sheep 30 30 40 Pig 44 40 20 Ilii Cattle 60 20 20 Sheep 60 20 20 Pig 62 25 13 Uib Cattle 50 20 30 Sheep 33 44 23 Pig 40 40 20 Ilia Cattle 25 25 50 Sheep 20 40 40 Pig Nil Ib Cattle 37 25 38 Sheep 37 44 19 Pig 42 42 16 la Cattle 50 17 33 Sheep 33 33 33 Pig 50 25 25 RITCHIE: PICTISH AND VIKING-AGE FARMSTEAD BUCKQUOYT SA , ORKNE5 20 | Y the pig is offset by its possession of twice as many metapodials and phalanges as the cattle and sheep. The dat f Isaaao c (1971 alss i ) o include onle th ys da comparativ e material available, althoug t referhi o t s condition Hottenton si t village presene th f so t day. Part of the difference between the % found and the % in the intact animal will of course be due to the fact that most bones are found broken, but it is thought that any bone will break into two or at the most three identifiable fragments. Phalange usualle sar y found intact pattere Th . recoverf no vers yywa different from that Africafoune th n di n experiment. Tables 2, 3 and 4 include data on the age-ranges of the animals found, but this material is more comprehensible when t whicpresentea e animan hag a e Tabl n dkillei s lTh s along i . die wa e6 t di r do s a , gives some indicatio sore environmenf th o t f no t whic inhabiteht i economie th whicd s o t dan e wa ct hus i put. Thus animan a f ,i killefirss e li th t life s n economidyeae i it th ,f ro c return fro onls i hide mt d i yth ean the carcass not yet fully grown. It has consumed less than an older animal but there is also the return on its parents' keep to be taken into consideration. Older animals may also have contributed their wool, labour of offspring accordin othe o alongenusg n o g t s pi r e ha economi th ; c possibilities other tha carcasss nit . It is not necessary for the animals to have reached the 'mature' stage before breeding, although it is TABLE 7 Bones from animals other than cattleg sheepi d pan Period Goat Horse Dog Cat Deer Seal Whale Otter Rabbit Mouse Vole Toad Rat Crab Total Topsoil 7 (2) 36 (5) 3 (2) 7 (2) Red 5 (1) 2 (1) — — 8 (2) 2 (1) 10 (2) — 7 (1) 4 97 V 30(4) 45(4) 4(1) 34(5) Red 8 (2) 10(2) 7 —2(1) 1(1) 15(3) 1 2(1) 7 165 6 2 2 — — — — — — 1 ) (1 4 ) (1 1 e Ro — — ) (2 8 1 — IV III 6 (1) 4 (2) — — — — — — — 1 (1) — — 2 (1) — 12 Ilii 1 (1) 6 (2) 2 (1) 13 (1) Red 1 (1) 1 (1) 1 — 3 (1) — — — 1 (1) — 29 9 — ) (1 1 — — — — — 1 ) (1 1 — — — ) (2 6 — Ilib Hia ___ — ___-______0 Ib — 50(5) — 1(1) Red 11 (5) 8(1) 1 3 (1) 3 (2) — — 1 1(1) 1 80 la — 10(1) — 1(1) — 2(1) — — — — — — — — 12 The figure in brackets is the minimum number of individuals present.

considered unlikely that cattle would breed much earlie t thia r s period. primitive Th e Soay sheep will breed young, however Kildalthougt d S ,an e a lamb th presene e islandn hi th ew t y sa tda s (Boy alt de 1964) achieve little success, 2-year-old animals succeed. Table 6 shows that a very high proportion of the animals died young generalisationa s a ; , abou l generthira tal f do a wer etime death ey dtheb y wer eyeaa r old, about one-third were killed between the ages of 1 and 4 years, and about one-third lived longer. There is thus little evidence of any economic function for the animals other than the provision of mea hidesd tan , althoug proportioe hth n reaching mature years amongs sheee cattltth d pan risin s e i th gn i later periods. Nevertheless, over half the animals were being kept for at least one winter, and though there was no doubt considerable killing in the autumn, Childe's concept (Childe 1931) of only a nucleus of breeding stock being retained is again refuted, and some winter fodder must have been available for the majority of animals. This concept has also been discussed by Higgs and Waite (1963). There is, however, little evidenc cattlr efo e being kept extensivel workins ya g animals. numbee Th bonef ro s derived from 'other species', tha wils i t d species, carnivore horsesd san s ha , been summarise Tabln di . Furthee4 r detail thesf so e animal presentee sar smale Tabln di Th . le 7 numbe r of confirmed goat bone alss si o included here problee Th . separatinmf o g sheep from goat notoriousls si y difficult, particularly when the sheep is of the primitive type found in this excavation, and the number of goat presented her doubtless ei underestimaten sa t likelno s yi t tha i goae t th t tbu , mad morp eu e thana small proportio smale th f lno ruminants . appreciabln Thera s ewa e proportio horsf no e bones through- out. Huntin mammalf go s seem havo st e played littl eeconomye parth n ti , althoug deee ro rd h botan d hre t wer seaca w eed l fe scantyfound bonesan a g therd d an ,an ,Do . e wer rodentsw efe . Crab claws were found, particularly from the later periods, and two specimens of toad. Rabbit bones are included in this table, but the modern rabbit is of course capable of inserting itself into any part of the deposit. The whale bones probably came from washed-up specimens; the flesh of these animals might have been used more widely than appears here seemt i s ,a s likely tha carcase tth sucf so hlarga e animal woul fleshee db d where it was found. This completes the economic part of the report. The rest of the material covers the size and type of the animals involved. Thi confines si sheee cattleth d po an dt ther s a , maturw e fe werbone g o epi e to o st afford sufficient data. However, the pigs were very small compared with modern specimens and they had the slender bone'wile th df s o boar ' typ animaf eo onlle whicth ys domestihwa occurring cpi g outside eth 206 PROCEEDING SOCIETYE TH F SO , 1976-7 Orient until recent times. There are two lines of evidence available for the study of the animal type. The most important is the measurement of bone-sizes, and, since whole bones were so few in the excavation, the bulk of the data comes from measurements of bone width. (The few whole bones are listed in separate tables, Tables 8 and 9). Some evidence is available for the estimation of body weight from some of these measurement case cattlth f eo n si e (Noddle 1973). othee Th r sourc informatiof eo n regarding typ animaf eo preciso l (breeto r usefa tere a b s di md o t for this early livestock fragmente th s )i skulf horn-coreso d lan t musI . t alway borne sb minn ei d thae tth size of an animal is not only a function of its genetic potential, but also depends upon whether it received sufficient nutrition durin perios git growtf do achievo ht e that potential; thus variatione th e b y sizn si ema resul periodif o t c weather-change developmene th t stock-typesno w d ne f san o t . measuremente Th whole th f eso cattle bone givee sar Tabln i . e8

TABLE 8 Dimension wholf so e cattle bones (measurement) mm n si Proximal Distal Period Bone Length width width Midshaft width Topsoil 5 4 Metatarsa 2 4 l 1 20 22 V Metacarpal 170 47 25 Metatarsal 197 41 48-5 223 Ib Metacarpal 172 52 26 Metacarpal 172 53 27-5 Metacarpal 178 53 26 Metacarpal 177 52 26 Metacarpal 172 49 48 28 Metatarsal 197 40 47 22 Metatarsal 205 39-5 48 22 Metatarsal 200 43 46-5 23 Metatarsal 207 40 44 23 Radius 247 68 61 42 (level of nutrient foramen) Scapula 258 125 62 42 (minimum neck width)

Histograms were made (and deposited with the archive material) of the bone measurements obtained most frequently. They comprise minimal width scapula neck, humerus width distal condyles, radius maximum proximal width, tibia maximum distal width, metacarpal maximum proximal and distal width (take epiphyseat na l line), metatarsa metacarpas la maximud lan m length first phalanx changee .Th s in size determinee whicb n t largeh ca no t ther e bu , dar e doe increasn a s e seeb sizmn o ei t f scapul eo a nec widtd kan metacarpalf ho lengtd ,an firsf ho t phalanx. Ther discerniblo n s ei e chang distan ei l tibir ao metatarsal, and both distal humerus and proximal radius seem bigger in the Pre-Norse period. Therefore no conclusion can be reached about the change in type of animal for this evidence. It was noted that the anterior posterior dimension of the metatarsal midshaft was always greater than the width; this factor is variable in modern animals, but the significance is not yet known. Approximate body weights have been obtained in terms of modern animals from measurements astragaluse th n o . These range fro mdefinit o ther d n abou e an e, ar e abouto kg distinctiont 13 0 g 3k 23 t s visible between the different periods. The few heavy specimens occurring in all those periods where there were large concentration f bono s e were probably bulls; sexual dimorphis r greatefa ms i primitiv n i r e breed cattlf so e tha modern i n (Walker 1964) f thes.I indeee ear d entire males followt ,i s tha majorite tth y malf o e animals were either castrate thao ds t their remains become difficul distinguiso t t h fro female mth e or the majority of males were killed young. Ver largw yfe e skull fragment completr o s e horn cores were recovered. Polled t likel skullno ye sar to have survived, judging by the condition of the other bone fragments. There is a strong tradition that the Norse kept polled cattle (Wilso polle ne bases i th 1910) t i n dt o Gallowa,bu y bree Scotlanf do whicn di h pollee th d conditio deliberatels nwa y encourage facilitato dt e rail transport south durin 19te gth h century; however, this mutation occurs consistently about once in every odd 50,000 births, and, once having RITCHIE: PICTIS VIKING-AGD HAN E FARMSTEAD BUCKQUOYT SA , ORKNEY 207 occurred tendt dominant,i e b o st specimene .Th s found compris elargea , presumably male specimen from ploughsoil level and a much smaller presumably female specimen from Norse V; both are curved, the mal forware eon downwardd dan directioe th t bu , n canno determinee b t femalee th n di . Fro earle mth y Pre-Norse period derive shorta s , young (base texturen do ) specimen, attache larga o dt e portiof no the frontal bone. Ther rouna s i e dparietae holth n ei l bonfron i e hor emth m abounm core0 5 t ; similar lesions have been observe medievan di l animals (Highworth, Wilts; Noddle unpublishedd an ) possibly result from the stunning procedure adopted in slaughtering. The second Pre-Norse specimen is smooth texture curvedd dan believes i t I . derivo dt e fro mfemalmucs a i d hean large r tha Norse nth e TABLE 9 Dimensions of complete sheep bones (in mm) Proximal Distal Midshaft Period Bone Length width width width Topsoil Radius 132 26 26 16 Metatarsal 124 19 21 — Metatarsal 133 19 22 12-5 Norse V Humerus 130 33 26 15 Humerus 134 22 26 15-5 Humerus 117 28 23 12-5 Metacarpal 124 19-8 23 12 Metatarsal 127 — 21 10-5 Metatarsal 134 18 22 10-5 III Radius 136 27 26 15 Ilii Radius 130 28 27 15-5 Metatarsal 127 18 22 11 Ilib Metacarpal 117 20-5 22 12 Ib Radius 147 29 27 15-5 Metacarpal 125 21 23 12-5 Metacarpal 113 — 22 12 Metatarsal 137 19 21-5 11 Metatarsal 116 20 22 12 Metatarsal 124 17 21 — la Radius 142 29 27 17 Metatarsal 127 18 20 11-5 Humerus 143 36 25 1-5 Radius 150 28 26 15 Mouflon Metacarpal 125 21 23 11-5 (Modern) Metatarsal 137 19 22 10-5 Soay Humerus 129 35-5 24-8 _ (Modern) Radius 137 27 25 14 Metacarpal 119 19-7 22 12 Metatarsal 131 17-5 20-4 9-8

specimen amoune th t f informatiobu ,o t n about 'breeds ' derivee whicb n hca d fro hore mth n shaps ei arguable. The young Pre-Norse specimen is somewhat similar to that illustrated by Platt in her report on the animal bones from Midhowe Broch (Gran Callanded an t r 1934 variabilite , th 515) d an , paralleles yi d by her finding on the Norse Jarlshof cattle (Hamilton 1956, 212-15). sheepe casth e same f th eo th , n eI dat f bonao e measurement hord an sn core availabls i s r efo analysis, but because of the survival of both wild and primitive breeds of domestic sheep to the present day much better comparative materia availables i l . Collection bonef so s from mouflo available nar e from zoological gardens and museums, Soay sheep derived from the feral island population of St Kilda and its descendants from recently imported specimen mainlande th n so Orkned an , y sheep fro islande mth f so RonaldsaN y have been use provido dt t onleno y straight measurement t alssbu o proportions. Table9 list l wholsal e bones founexcavatione th n di s with representative samples fro comparative mth e animals. The bone measurements indicate a sheep small by modern standards, slightly larger than the Soay and thus weighing rather mor Soag k ey0 tha 2 averagee nth . Ther sligha s ei t decrease wit passage hth timef eo , 20 | 8PROCEEDING S OF THE SOCIETY, 1976-7 which could be attributed to decrease in nutrition coupled with increasing exploitation of the environment rather than any genetic change. There is no distinct difference between Norse and Pre-Norse animals. In the case of the first phalanx, there is an abrupt reduction of length in the mixed material of topsoil, although long specimens still persist. When the proportions of this bone are examined, there is a gradual transition from something simila wilthe d mouflorto n animaforan mto l more lik present-daethe y Soay. proportione Th scapule th f so a neck sho same wth e tendency picture th t less ebu ,i s obvious because eth chang shapn ei fros ei m mouflon proportion extremele th o st y long narrow Soay for thed man n back shortee agaith o nt r Orkney (modern long-tailed sheep hav evera y short thick scapula neck). However, the changes are not very great, and the data for all the animals overlaps in part of its range. Horn core evidence is not very helpful. Firstly only one good specimen of sheep horn core was obtained from the excavations; secondly comparative th l ,al e breed sheef so p have similar hornmale verd th en an ysi variable hornfemalee th n si , except tha Soae th t y cross section tendmore b o est oval tha otherse ovanw th fe l .A horn tops were found. The evidence can be summarised as follows. The sheep stock at the earliest stages of the excavations was of a very primitive type not far removed from the wild. Nevertheless, it must have been an imported form originally wils a , dt believe sheeno e phavar o dt e occurre Britise th n di h Isles animale Th . s were subjected to pressures which selected a type more similar to the Soay. Eventually, by topsoil, appears a form indistinguishable fro Soae mth y fro limitee mth d material available t possibl whetheno y s i sa t o e.I t r this was newly introduced or whether there is a long time-lag in this mixed layer during which the sheep continued to change elsewhere but were not deposited at this site. As the Soay sheep is generally considered Norsa e b eo t type appearance th , f thieo s anima t Buckquoa l surprisingls yi y late. Platt recorde sth presenc breedo tw sheef f seo o p durin Norse gth e perio Jarlshoft da , distinguishe theiy db r horn cored san some very long bones of a possible third type which she described as deerlike, but which would also be agreeable wit hmoufloa n type (Hamilton 1956, 214). Ryde s inferreha r presence th d primitiva f eo e pre-Soay typ sheef eo p from fleece evidence (Ryde Stephensod ran n 1968) thoughs i t I . t that this animal would not possess wool, but a hair coat. The age of the animals at death would suggest that they were not kep r woofo t l production. specimee On considerablf o s ni e interest t comprisei ; frontase parth f hore o tl on bonn d corean e oyouna f g polycerate specimen broken-ofe th d an , f horn t leascorothea e f eon to r horn, probablo ytw horns. The polycerate habit crops up sporadically in domestic sheep from time to time, and has been recorded in very early specimens from both Catal Huyuk and early Egypt (Epstein 1971). Elwes (1912) suggests that the majority of medieval sheep in N Scotland had four horns and both the sheep skulls found durin excavatione gth Lunn i s d were polycerate (Bergquis Lepsikaad an t r 1957) e recentlTh . y extinct Hebridean breed was frequently polycerate (Low 1846) and the habit survives in the present day Manx Loughton and the St Kilda Black; this last breed is the only one which seems to have more than four horns witfrequencyy believes hi an t i d originato dt an , e from Skye (Ryder t 1968) habino e s Th . ha t been recorded in present day Soay sheep, but up until the first world war a number of Shetland sheep in westere th n island e partth f so s were polycerate. Two pathological specimens were found mandibla , e exhibiting paradontal diseasabcesn a d sean cavity of such proportions that it had broken out of bone in two places (topsoil), and a humerus with a cauliflower-like exostosi laterae th n so l distal portion, probabl resule yth continuef o t d trauma sucs i s ha encountere modere th y db n Orkney sheep feedinrocke th n ygo shore RonaldsayN f so .

SUMMARY Some 7,000 mammal bones were recovered fro sitee bule m th f theTh .ko m originated froe mth topsoil and the periods Norse V and Pre-Norse Ib. There was little variation in the proportions of the genera throughout, in the anatomical distribution of the remains, or in the size and age range of the animals. proportioe Th n throughou abous twa t 5cattle0% , about 3sheep0% , aboud abou othean % g 5 tt r pi 20% genera. There were no pigs in Pre-Norse Ilia. About one-third of the animals died in their first year of life amongs l generaal t , indicating tha maie th t n functio livestoce provisioe th th f ns o k wa hidef no d san meat. Amongst the other genera the most frequent was horse, but there were also a few goat and deer t (larg ca rare s d ebones wa wilan , g d Do .cat ) more frequent craw bfe clawA . s were also identifiede Th . cattle were small, about 180-20 livg 0k e weight, about halsize moderf th feo n beast t rathesbu r larger tha medievae nth l English animal sheee Th . p wershort-tailea f eo d very primitive race mann i , y respects more lik moufloe eth n tha Soaye unlikelns th i t .I y that they were wool-bearing. Durin perioe gth d covered RITCfflE: PICTISH AND VIKING-AGE FARMSTEADS AT BUCKQUOY, ORKNEY | 209 excavatioe byth n there gradua a seem e b o st l change toward Soauppermose e sth th yn i type d ,an t layers possibls ii t e tha Soae th t y appears possiblA . e polycerate specime founds nwa . greao Thern e t ear dif - ferences from the bones described from the Norse period at Jarlshof by Platt (Hamilton 1956, 214-5) and no marked differences between Norse and Pre-Norse livestock.

APPENDIX 2 Bird and vole bones from Buckquoy, Orkney bBramwell,yD Fulwood, Baslow Road, Bakewell, Derbyshire Identification of birds Fro table mth e greates see e overleae b th nr n thafa ca t y t fconcentratioi tb birf no d bone fros si e mth topsoil; although this material is evidently an unstratified layer, the variety of species is very similar to the other Norse layers and so may be regarded as being the refuse left by those people. A striking fact about this assemblag smale th es i l part playe econome th n di poultryy yb , with vergees w domestid yfe ean c fowl and no sign of domestic ducks at all, while the odd fowl bone from the Pictish levels can probably be dismissed as intrusive, due to rabbit burrowing. It suggests very strongly that the Pictish people did not hav domestiy ean greatese cth bird r fa t alltsa y amounB . birf o t d mea provides wile wa t th dy dbirdb s which these people, Nors Pictsd ean , foun thein di r environment, which provide dricvariea d han d fare though many of the species would not be relished today, e.g. gulls, cormorant and shag. Present habits of the birds show four main environments where hunting must have taken place: (a) moorland, cliffs and scrubland; (b) meadow, marsh and freshwater; (c) sea shore and mud flats; (d) the open sea. These divisions apply to the places where the birds may be found ouside the breeding season. The sea birds, of course, must come ashore to breed and so would at that season utilise cliffs, turfy places and fresh-water lochs. It is likely that most of the pursuit of these species would be carried out at this season. Gannet, guillemot, cormorant and razorbill, for example, are very approachable when nesting on the cliff ledges. The largest seabird in this list is the now extinct great auk, but its sparing remains seem to indicate t particularlno s thawa t i t y common eve t thina s early date, thoug t continuehi breeo dt Orknen do y till 1812 (Lea and Bourne 1975,100-2). Another large bird in the list which no longer breeds anywhere in the British Isles is the common crane, an inhabitant of wild and secluded marshy places. In northern Europe today this bird is a summer visitor but other species in this list are autumn and winter visitors, e.g. knot, jacksnipe and turnstone, showing that wild fowling was spread largely over the whole year. Osprey is represente larga y db e femur presence .Th blacf eo k grous somf o s ei e interes ornithologyo tt bones ,a s also Skaile th turne t a l excavationp du NorsDeernessf n si o e eag , Orkney. Thi sedentara s si y species though nestin seet g ye doemhav o t t sno e been recorde Orkneyn do . The few bones of domestic goose belong to the normal early domesticated form of grey-lag, that is a bird almost identical in size with the wild grey-lag. A few measurements of the Buckquoy birds are given: Left tarsometatarsus, 78-7 mm long, 17-8 mm prox width, 6-7 mm least width of shaft, 18-0 mm distal width. An incomplete tarsometatarsus, 18-5 distal width. An incomplete humerus, 25-0 distal width. The thicker shaft to the complete tarsometatarsus quoted above is the reason for regarding it as a domestic bird, following studie medievan so l geese bones from King's Lynn (Bramwell 1977). domestie Th c fowl bones from Buckquo froe usuae yar m th l small variety whic widels hi y found from Roman to quite recent times, with little sign of selective breeding. There is however an incomplete humeru thin si s collection which betray sturdiesa r variety, though still onl ysmala l birdspecimend Od . s of thi t snecessaril naturno y ema y denote selective breedin probable ar t gbu yfeatura e which appeary sb chance from time to time. The following fowl bones were complete: humerus, 63 mm and 69 mm; femur, 69 mm and 79 mm; tarsus, 61-5 mm and 71-5 mm; coracoid, a pair at 48 mm. The Orkney vole, Microtus arvalis orcadensis Of considerable zoologica skullw l interesfe f thi so a se volear t , fro topsoie mth Vikind an l . gV Normall remaine yth burrowinf so g animal suspece sar archaeologican ti thiln digsi st cas ,bu stat e eth f eo preservation agreed with the other Norse material. The point of interest is that zoologists are still not agreed as to how the Scottish islands acquired their faunas following the sterility of the Ice Age. Some R 21 | 0PROCEEDING S OF THE SOCIETY, 1976-7 TABLE 10 No. of bones Norse Pictish

Topsoil V IV III Ilii IIib Ilia I Great northern diver 3 3 1 1 51—3 Fulmar 2 — 5 3 — — — 1 Manx shearwater 8 6 4 1 Gannet 23 39 6 1 8 — 1 3 8 Cormorant 4 1 — 1 1 — — — Shag 6 7 — — Mute swan 1 — — — Whooper swan 1 3 — — Goose, domestic 8 12 1 — Shelduck — 2 2 — Mallard 1 — — — Teal 3 — — — Wigeon — 1 — — Shoveler 2 — — — Eider 2 — — — Common scoter 1 — — — Goldee ney 1 — — — Small duck species 1 2 — — Osprey — — — — Merlin 1 — — — Kestrel 1 — — — grousd Re e 2 — — — Black grouse 2 — — — Fowl, domestic 32 9 1 1 Crane 1 — — — Water rail 1 — — — Corncrake 1 — — — Oyster catcher 1 — — — Golden plover 1 — — — Turnstone — — — — Dunlin 1 1 — — Knot — 1 — — Greenshank — 1 — — Curlew 8 — 4 — —— — 4 Whimbrel 2 — — — Jacksnipe 2 — — — Phalarope, ?grey — 1 — — Skua, ?pomarine 1 1 — — Black-headed gull 1 — — — Lesser black-backed or herring gull — 3 — — Glaucus gull or greater black-backed gull 12 2 — 1 — — — 1 Gulls, small species 1 — — 1 — — — 1 Little auk 3 3 — — — — 1 — Razor bill 10 2 — — Great auk 2 — — — 1 __ __ _ Common guillemot 9 5 1 — — — — 7 Black guillemot — — — — — 1 — — Puffin 5 1 1 1 — — 2 3 Rock dove 1 2 1 — Ring ouse blackbirr lo d 4 — — — Redwin sonr go g thrush 1 1 — — Starling 7 2 — — Roo cror ko w 6 7 1 — Raven 2 — — 1 indeterminate 18 7 3 — RITCHIE: PICTIS VIKING-AGD HAN E FARMSTEAD BUCKQUOYT SA , ORKNEY 211

TABLE 11 speciee Th s arranged under habitats: (a) (b) (c) (d) Merlin Swans Oyster catcher G.N. diver Kestrel Mallard Turnstone Fulmar Red grouse Teal Dunlin Manx shearwater Rock dove Wigeon Knot Gannet Osprey Shoveler Curlew Cormorant Ring ouser o l Goldeneye Whimbrel Shag Blackbird Small ducks Gulls Shelduck Redwin r thrusgo h Crane Eider Starling Water rail C. scoter Roo cror ko w Corncrake Skua Raven Golden plover Little auk Greenshank Razorbill Jacksnipe Greak au t Phalarope Guillemot Puffin

suppor 'relicta t ' theor postulatd yan e 'land bridges' from various directions, thus enablin islande gth o st be repopulated by certain species, but more recently Corbet (1961) has dismissed the land bridge theory i nchannee depte vieth th f f who o l surrounding Orkney regarde H . f accidentavolo e th s ea l human introduction, possibl cattlyn i e fodde turfr ro som y b , e immigrant people timy an e t fro, a mesolithie mth c to Viking age, and since the arvalis species is otherwise a continental mammal, it might be thought that a Viking origin coul withie db nvariete reasonth w y orcadensi.No considerabls si y larger tha continentas nit l counterpart, in other words it is a specialised island form which must have required many generations to develop, but the Buckquoy skulls are already large, so the origin of Orkney voles must be sought further back in prehistory. A bone of the common toad (phase V) merits similar consideration.

Acknowledgment The writer is grateful to Mr G S Cowles of the Dept of Ornithology, British Museum (Natural History), for his help in identifying the osprey.

APPENDIX 3 The fish bones from Buckquoy, Orkney by Alwyne Wheeler, British Museum (Natural History)

INTRODUCTION The remains of fishes recovered from the excavation of the Pictish and Norse homesteads at Buckquoy, Birsay, amounte considerabla o dt e quantit bonesf yo . These wer mosn ei t case broken si n condition, some even fragmentary, and many were not identifiable even to family. No scales, otoliths, or teeth were contained in the samples examined. This was probably because recovery was by hand-picking of excavated material and no sieving was attempted; comparisons of recovered fish remains from sieving and hand-picked medieval site t Greasa t Yarmouth (Wheele Joned ran s 1976) elsewhered ,an shows ,ha n that the former method of recovery produces a very different fauna list, enhanced in particular by small species. The interpretation of the Buckquoy fish bones is therefore dependent on the supposition that they represent only the larger species of fishes caught and consumed at the site. samplee Th s were sorted firs thao s t t vertebral centra (which dominate l sampledal numben si d ran quantity) were segregated from bones fro jaw e craniumd mth san alsd an o, fro fragmente mth s which would prove to be unidentifiable. The jaw bones, principally dentary and premaxillary bones are mostly very characteristic and form a firm basis for identification. From the preliminary species list formed in this wa vertebrae yth l centr othed aan r complet characteristir eo c bones were identified. Only bones positively 212 | PROCEEDINGS OF THE SOCIETY, 1976-7 identifie genuo dt specier so includee sar thin di s report. Where, becaus theif eo r fragmentary naturr eo lac f distinguishinko g features, identificatio doubtfus nwa l these bones wer t furtheeno r consideredt i , being felt tha speciee th t s list constructivelwoule b t dno y increase theiy db r inclusio interpretae th d nan - resulte tioth f no s migh materialle b t y affected.

RESULTS summarA resulte th f ygives o si tabulan ni r form (Table furthe12)d an , r note e givee th sar n no bones identified estimatee th , d fise sizehth f concerneso d wher ecomparisoa possibles ni somd an , e notes on the biology of the species with references to potential fishing methods. Biological details are from Wheeler (1969).

TABL2 E1 Species or groups of fishes represented in the Buckquoy material. Minimum numbers of each species are given; presenc indicates i groupe ) eth (P t r identifiedfo sno generio dt c level Pictish Norse V V I I Il aI II Phaslibe Species or group P P — — — Elasmobranc — h Conger eel 1 1 — — 2 1 1 — — — — — Trout 2 6 — — — — d Co 3 0 1 — — — 2 Saithe? Haddock — — — — — 2 Ling — — 1 — 4 5 Cod-family indet. — P P — P P — — — — — 1 Hake Red seabream — — — 112 Ballan wrasse 1 — — — 21 — 1 1— — — Mackerel Grey gurnard — — — — 11

Elasmobranch - shark or ray. Because these fishes lack bones, their skeleton being composed of calcified cartilage, skeletal remains are few. They are represented by vertebral centra at two periods (IV Norse th f eo ) levelsV d .an Vertebral centr t possibl featurelese no as ar i t identifo i et d san y these remains further. Their size, however, suggests tha fishee th t s were moderately small sharksf i ; , less tha mn 1 i length, and if rays less than 0-75 m in length. Sharks of this size in the area of the Orkneys are the dogfishes (Scyliorhinus spurdo e spp.th d )an g (Squalus acanthias), l commoal n species; severa skatr o l kindy era f so {Raja spp.) coul involvede db sample Th . e from perio containedV d nine centra; those fro othee mth r periods were eac hsingla e centrum. Conger conger - conger eel. Identified from two Pictish periods (I - a right dentary from a fish c 1 -7 m long; II a - anterior part of premaxillary from a fish c 1 -4 m long), and three Norse periods (IV - five vertebral centra, one right dentary from a fish c 1-5 m long, one premaxillary from a fish c 1-3 m long; vertebrax si V - l centra fro mfisa t leas ha t 1- long) 4 largm a conge e s i eTh .l fisee r h which occurn si inshore waters fro intertidae mth deptm lrock0 n zonho 15 yo et grounds habis .It livinf to g amongst rocks and in wrecked ships results in it mainly being captured on baited lines, and by chance in crab or lobster pots; it is very rarely caught in trawls. Salmo trutta - trout. Identified from one Norse period (V - one left dentary from a fish of c 0-25 m length) troue Th . t occurs naturall severa n ylargei e th f ro l Mainlane lochth f so d (Orkney n wels i ) a s a l streams and in the sea, as sea trout. The relatively small size of this specimen suggests that it may have been caugh freshwatern ti troue Th wel.s ti l know anglersn a s na ' althougfisd han alsn caughe oca b ht i n ti net capturs sit baitea n eo d hoo witr ko hlura somf eo e kin mors di e likely. Gadus morhua - cod . Aparently presen Norste onlth n yei levels whenc identifies wa t ei d from three period - thre lefV e se(I t on righpremaxillar d tan y bones from fish estimate 0-4e 0-7, b 0-70, m o d 5t , m 8m RITCfflE: PICTIS VIKING-AGD HAN E FARMSTEAD BUCKQUOYT SA , ORKNE3 21 | Y lengtn 0-8i 6m h (2-5- gutteg 6k d weight) vomerso tw , lefe ton , premaxillary fro mfisa h estimatet da 0-66 m and 2-5 kg, two right dentaries one from a fish estimated to be 0-70 m and 3 kg, two left maxillaries in fragmentary condition righo tw t - premaxillarie ;V s from fish estimate 0-8e e b (4- 0m o don t 5 d kg)an , right maxillary). The cod is a well-known food fish which is common off Orkney. Small specimens can be captured close inshore, but larger fish (above c 6 kg) are usually captured in deeper water. It will take a baited hook readil captures i d yan d toda largn yi e number trawln si seind san e nets t live.I schooln si t no st closbu , eto on, the sea bed. Pollachius sp. - saithe or pollack. The remains identifiable to these species are not sufficiently com- plet alloo et w precise determination. Pollachius representes i onlPictisn de i yon paraho periotw - - I d( sphenoid f smalso l fishespresens i t bu severa)n i t l Norse period represented san numerouy db s bones. Norse Th e- fou periodV r(I parasphenoide sar smal o tw larg lo d fishestw ean f so , three vomers, three right maxillaries, two left premaxillaries of large fish, and eight right and three left dentaries representing three large, three medium and four small fishes; V - one parasphenoid, three basioccipitals, two left, two right maxillaries, two right and two left premaxillaries all from large fishes, one right and two left dentaries all moderately large fishes). Due to lack of comparative material of known lengths it is not possible to estimate the lengths or weight fishee th f sso involved; small fisprobable har y less than 0-4 , larg5m e one 0-9sc . 0m Saith pollacd ean similae kar externan ri l appearanc closele ar d yean related. Both species occun ri the seas around the Orkneys but the saithe is the more common and is abundant inshore when young. It is thus likely that these fish were saithe. Young fish tend to be found in schools close to rocks, although the larger fish inhabit more open sea beds; at all stages they are easily captured on baited hooks or lures, althoug modere hth n fisher r saithyfo prosecutes ei d mainl trawlingy yb . Melanogrammus aeglefinus - haddock . Identified periodfro Norso e mth tw e f eso on levels - V (I , cleithru cleithrume mon fro- parasphenoid o 0-4V mc fis a ; f tw , 5ho m fishef so 0-7d . 0-5sc 0an m) 0m These lengths represent weight f 0-90so , 1-01 2-7d respectivelyg an ,5 k haddoce Th . bottom-livina s ki g species which occurs in schools over open ground; it is today usually caught in trawls but can be captured hookn o s provide t usuall no beda s di se yt .thee I fishefoune th y ar n do d close enough inshore b o et captured from the shore. Molva molva - ling. Present as numerous remains from both Pictish and Norse levels although more abundan lattee th n ri t period. Ling bones have been identifie Pictise th n di h period lib (one left pre- maxillary fro mmoderatela y large fishlefe ton , dentary, three large anterior abdominal vertebral centra; l thesal e bones could have originated fro msingla Norse fish)th cleithrume n eI on . perio - o V dtw ,(I caudal vertebral centra vomerso ,tw , three parasphenoids righo tw , t dentarieso frotw -2, m1 0 m c fis f ho left premaxillarie -41, 1 3 0m d vertebras an fro m m1 c fisl vomers f centrao ho tw - ;,V three parasphenoids, basioccipitalso tw righe on t d lefpremaxillare an ton , y both fragmentary, three right dentaries (all frag- mentary offiso )leftw o 1-5h c 0-5t a lengthtw dentarie 0f fishm 0 d o m e an , ,on s both from fish elm, nine anterior vertebral centra). linlarge a gs Th i familed membeco e y th whic f ro h weigha attain d excesan n ti lengts a m f 2 so f ho 35 kg. It is relatively common in the vicinity of Orkney in depths of 50 m and more; only small fish occur in shallower water. Large specimen smostle sucar s ha y represente Norse th n dei material ten frequeno dt t rocky groundmostle ar d ys an capture baiten do d hooks. Merluccius merluccius - hake. Identified only in Pictish period I (one right dentary from a specimen of c 1 m, five vertebral centra). Possibly all these bones came from a single fish. hake relativela Th s ei y deep-water fish; today specimen thif so s sizrarele ear y foun thed dan n only depth wate m n t leasi a 5 f 16 .rto However, hake migrate into shallowee rb watey summen ma ri t i d an r before the stocks were over-fished (as they are today in the NE Atlantic) that large hake could be caught inshore. Although commercial fisherie trawlingy hakb r se fo ear , thi caughe b s fisn hooksn ho tca . Pagellus bogaraveo seabreamd re - . Presen tNorse onlth n yei level represented san very w db yfe bones. Found in periods (III - one abdominal centrum of a fish c 0-45 m length; IV - three caudal centra offish c 0-40 m length; V - one left dentary of a fish c 0-37 m, three abdominal centra of at least two fish one c 0-45 m, the other smaller, possibly 0-37 m). The red seabream is an uncommon fish north of the Irish Sea today although it has been reported off Orkney. Its presence in these Norse remains might indicate that it was more common there at that period than today suggesting possibly a rather warmer climate. It grows to a maximum length of 0-51 m todas i t i y; caughkg weigha 4 d c trawln i tf an o t s occasionall alsd hooksn yoan o . 214 | PROCEEDINGS OF THE SOCIETY, 1976-7 Labrus bergylta - balla n wrasse. Identifie Pictise on n hdi leve thred lan e Norse levels. Pictish period (I- firs t anal pterygiophore articulated san d spines). Norse perio - firs V t d(I ana l pterygiophored san articulated spines, three caudal vertebral centra from fis lef e 0-3hc tlong on 0dentary m - V ; lowee on , r pharyngeal bone, pai uppef ro r pharyngeal bones). The ballan wrasse is a rock-haunting species which can normally only be captured on hooks or in baited pots. It is uncommon today in the vicinity of the Orkneys but is reported from time to time. The remains identified suggest that it may have been more common in the 7th to 9th centuries than it is today, spars o more to b bue o et ar tclosel y analysed. Scomber scotnbrus - mackerel . Represented only Nors e levelfroo th e mtw n ei s on perio - V d(I vertebral centrum; V - three centra, one abdominal and two caudal). mackeree Th well-knowa s li n food fish whic largels hi y pelagi life-styls it n ci e livin schooln gi s near the surface and coming close inshore in summer. It is easily captured on hook and line, although today take varieta y nb othef yo r methods. Fro materiae mth importann a l t recovereno s t wa foo t di d fish. Eutrigla gurnardus - grey gurnard. Identified from Norse levels in period IV (dorsal fin spine) and V (one right opercular spine, one dorsal fin spine). This material is not sufficient to permit assessment of the length of the specimens involved. The grey gurnard is a common sea fish around the British Isles, most common in 20-40 m although occasionally captured in shallower water. It is entirely a bottom-living species found on sandy or muddy bottoms and vulnerable to capture by line and in trawls.

DISCUSSION Some interest centrecomparisoe th n so made b o net betwee speciee nth s recognise Pictise th n di h and those in the Norse levels which can be seen in Table 12. The relative abundance offish bones is much greater in the Norse levels than in the Pictish suggesting a greater reliance on fish in the diet and improved fishing techniques, althoug possibls i ht i latteee tha th greate e f rto th occupatio e rag n could have accounted for such a difference, as also could different-sized populations on the site. The list of species in Pictish remains contains three type f fisso h which coul caughe db shallon i t w inshore waters. Ballan wrasse, conger eel, and saithe or pollack are all rock-haunting fish relatively easy to catch on hook and line when fishing from the shore as for example the craig fishing as described by Fenton (1973). The use of a poke net (a large circular net, baited and drawn swiftly to the surface) would also be a possible means of capturin gsaithe thesth s ea e caughfisd han woul thiy n i t swa d mainl smalle yb wers a ,Pictis e eth h fish, thiviabla s si e alternative otheo five Tw .r th e specie f determineo t sou d (lin haked gan ) are, however, exceptions, both are relatively deep-water in their life-style and are more likely to have been caught on lines fished from (noboata se tt snecessarila offshore)r yfa . The fishes identified from Norse remains are more numerous. Again, conger eel, saithe and ballan wrasse coul capturee db craiy db g fishing couls a , d mackere occasionn o l summern si . Cod large ,th e saithe or pollack, haddock, ling, and gurnard are more likely to have been caught on long lines set in deep water. Cod, saithe, and ling are mid-water or near-bottom fish such as might be captured on the middle hook linea f so , while haddoc gurnard kan bottom-livinge dar fish which would tak ebaia t fro msunkea n line. Many elasmobranchs (e.g. the rays and dogfishes) are also bottom-living. Both red seabream and mackerel coul caughe db mid-waten i t sucn ro hlinea . These notes suggest that durin Norse gth e period fishin prosecutes gwa deepen di r water using many-hooked lines, requirement whicf so h would have beena larger, more sea-worthy fishing boat than close inshore fishing needed. speciee Onth f eo s identifie alss di interest f seabreaod o re e Th .mrar s i e todaregioe e th th yn f i n o Orkneys yet it is represented in three periods of the Norse occupation which suggests that it may have been more common then than it is today. This may be due to a less favourable climate today (it is a warm-water fish founMediterranean)e th n di limitee vien th i f t wo d bu , material availabl sparsitd ean y of supporting evidence amongs fise th ht species (the ballan wrasse remain lese sar s positively capablf eo same th e explanation), this interesting suggestio furthet no s ni r pursued. Additionally interestins i t ,i g that severafishese th f o l, notabl hakee yth , linhaddocd gan r whickfo h length estimate possiblee sar e ar , considerably larger than would be captured today in inshore waters. This suggests that larger fish may have been foun centurieh shallo9t n d i o t h sw 7t tha watee nth todayn ri possibl.A e explanatio thir nfo s s i that fishing pressures on these species may have reduced the stock of large fish close to the shore. In other words the present-day distribution of these (and other species) of small specimens inshore and large ones in deepe resula re exploitatiof b wate o ty rma n rather tha selectiof no habita f nfishe o th .y b t RITCHIE: PICTISH AND VIKING-AGE FARMSTEADS AT BUCKQUOY, ORKNEY | 215 APPENDIX 4 mollusce Th environmentd aan , Buckquoy, Orkney b yEvans G J Penny d an Spencer, Departmen Archaeologyf o t , University College, Cardiff Marine mollusca All large shells of marine molluscs were collected during the excavation. They came from eight phases of Norse and Pre-Norse occupation; in addition, shells from the topsoil were collected thus making a total of nine samples. The shells were identified and counted (Table 13), and the results presented grapa s a h showing change relativf so e abundance throug deposie hth t (fig 12) l specie;al consideree sar d to be human food debris. TABLE 13 Buckquoy, Orkney. Invertebrate fauna, largely human food debris fro Pre-Norse mth Norsd ean e middens. 4= - t countedpresenno t probable= bu tf c ; possible= ? ; . Phase Pre-Norse Norse Top- la Ib Ilia IIib Ilii III IV V soil Patella vulgata Linne 1130 2728 35 284 516 431 6834 22543 5480 1 — — — — — — Patella — aspera — Lamarck Patina pellucida (Linne) — — — — — — — — 3 Gibbula cineraria (Linne) 24—2 7 6 515 3 — Calliostoma zizyphinum (Linne) 4 14 — 3 — — — 28 1 Littorina littoralis (Linne) 27 128 12 17 26 10 45 406 559 Littorina littorea (Linne) 542 2364 36 785 1400 651 5401 26854 17748 — — — — — — — — Trivia f arcticac — (Montagu 4 ) Nucella lapillus (Linne) 70 410 19 56 99 226 109 1414 456 2 — — — — — — Buccinum — undatum — Linn6 — — — — 1 — — — Cepaea 2 hortensis — (Muller) — — — + — 1 — 7 Mytilus 2 edulis Linne + — — — — — f Modiolus c — modiolus— (Linne— ) Ostrea edulis Linne — — — — — 1 1 — — — — + + — — — Pecten + maximus — (Linne) ? Lima hians (Gmelin) — — — + — — — — — — + — — — — — c f Cardium— edule— Linne Cancer pagurus — — — — — — + — — Spirorbis borealis + + + + + + + + + + — — — — — — — ? Flustrella— hispida Limpets (Patella vulgata) and winkles (Littorina littorea) predominate with dog whelks (Nucella lapillus) flad tan winkles (Littorina littoralis) constants a t subordinatebu , , elements shellp To . s (Gibbula cineraria and Calliostoma zizyphinum) and mussels (Mytilus edulis) are sporadic, while all other food species are represented by one or two occurrences only. The species are all characteristic of rocky shores being common in the intertidal zone, and there is no reason to suppose that they were collected from other tha immediate nth e vicinit sitee th .f Theryo virtualls ei suggestioo yn sandy-shorf no mud-flar eo t species- e.g. Cardium and other burrowing bivalves (cf the list in Coles 1971, 354). The histogram (fig 12) shows that during the Pre-Norse phases limpets gradually gave way to winkles as the most important species eaten. During the Norse period this trend is reversed, while in the topsoil, winkles once more assume dominance. To estimat weighe eth meaf o t t represented modera , n limpet sampl measures limpetewa 0 10 d- s from Lavernock, Glamorgan, were collected and weighed as follows: Weight (kg) Shells 0-66 Flesh, raw 0-31 Flesh, boiled in water for 5 min 0-18 Flesh, boiled in water for 20 min 0-15 216 | PROCEEDINGS OF THE SOCIETY, 1976-7 Using the figure of 0-15 kg/100 limpets, the meat weights of the Buckquoy samples were estimated, and are listed in fig 12. The total is 58-3 kg, which, applying the data of Shawcross (1967) of 650 calories/kg, an days0 d1 r 270 .dayr Althougfo sufficiens 0pe i n suppor ,o caloriet n ma n ma e r ow ht ton s meape it n to clearly a rough calculation, side-stepping many sources of error (such as the differences between the modern and ancient limpet populations), it nevertheless serves to demonstrate the wholly subsidiary natur f shellfise o diete th .n hi In order to trace changes of size in the limpet population through time, a sample of 100 shells from each phase (excluding C) was measured, the measurements taken being the external length, breadth and height. Plot lengtf so h against breadth yielded straight-line graphs wit hslopa 45°f eo . Eithe thesf ro o etw measurements can thus be used as an index of size (height not controlling the meat content to any great extent, mosventrae whicth f o tn i animal)e ls h i parth f o t . Distribution curve r lengtsfo h were plotted, showing that the limpets in phases E, D and B are larger than those in phases A, III, II and I. The shells

24322 TOPSOIL 51301 NORSE

PRE-NORSE

0 10 30 50 0 10 30 70 0 10 O 100°/o FIG 12 Buckquoy: changes of relative abundance of shellfish (percentages of total fauna)

topsoie inth intermediate ar l e betweegroupso tw e n.th Thi sintensivo t diminutio e du e eb y sizn i ema collection by man. A similar phenomenon, possibly attributable to a similar cause, has been recognised in Danise th h kitchen-midden sites (Brothwel Brothweld an l l 1969, 64). Plots of height against length showed that the majority of the limpets were of normal profile, i.e. they were neither excessively flattened nor excessively conical (cf the modern population from Lavernock which is composed largely of markedly conical individuals). The significance of this is that the Buckquoy shells were collected throughou intertidae th t t frol no zonem d eithean , r clos high-water-maro et r ko low-water-mark exclusively, for limpets in the upper part of the intertidal zone are generally conical (Lavernock sample), those from near low-water-mark, flattened. Notes certainon species Patina pellucida (Blue-rayed limpet). This species lives on the sea-weed, Laminaria, and the few examples from the topsoil may derive from sea-weed spread on to the land as manure. Littorina littorea (Edible winkle) smalw fe closelle .A example th e b y y relatesma d Littorina saxatilis, but their degree of preservation made certain identification impossible. Trivia arctica. foue Th r shell thif so s species were identifie basie th thei f s n o absenc e do rth sizd ean e of dark spots. However fossie los th e therefor, s i b lthesn ti t state y i d ema an ,e possible tha shelle th te sar Trivia monacha (da Costa). Nucella lapillus (Dog whelk) shelle Th .thif so s species varie giando considerablytw tr o e on d an , specimens were recorded. All, however, were clearly referable to the one species. Cepaea hortensis. lanA d snail, whic havy hma e been use r food dt whicfo equallbu y , hma y have bee nland-snaie parth f o t l fauna sitee livinth . n go RITCHIE: PICTISH AND VIKING-AGE FARMSTEADS AT BUCKQUOY, ORKNEY | 217 Cancer pagurus (Edible crab) smale .On l claw fragment (dactylopodite), possibl thif yo s speciess ,wa recovered. Spirorbis borealis. This species is a small tube-worm (Annelida, Serpulidae) which secretes a calcareous tube rolled in a spiral (diameter of spiral, 2-4 mm), and which often occurs attached to stones, seaweed, etc, in the intertidal zone. A few specimens were found encrusting limpets and, on at least one shell inside,e werth n eo , indicatin havo t t gi e been collected dead occurrence Th . thif eo s speciee th n si midden is, of course, incidental. Flustrella hispida. A small compound animal (Polyzoa) which secretes a calcareous skeleton and lives on sea-weed, rocks, etc. One colony, possibly of this species, was found on one of the shells of Patina pellucida.

THE ENVIRONMENT archaeologicae th f o N clife A th f o sectionlt sitesamples m 0 ,wa lan,20 r dfo d snailhope th en si that these might give some clue to the recent environment of the area, and in particular, its reaction to human occupation. The section showed the following stratigraphy (fig 13): Depth below surface (cm) 0-20 Topsoil. Dark, sandy loam. Modern vegetation consists of unkempt grass and weeds. 20-45 Midden. Compact, sandy loam with numerous stones, bone fragment marind an s e shellse th , latter mainly limpet edibld san e winkles. 45-62 Pale, orange-brown sand; very compact. Calcareous. burief o 62-6p dTo 7soil . Dark-brown clay loam with intermixed shell sand. 67-75 Bas burief eo d soil. Non-calcareous, dark-brown clay-loam fre shelf o e l sand. 7 5+ Solifluxion debris. Pale yellow-brown stony loam with large angular stones pitche l direcal n di - tions; stoneuppee th n sri part (75-9 relativel) 0cm y horizontal. Five samples through this section were analyse r landfo d snails usin methoe gth d describey db Evans (1972) resulte Th . s (Tabl hav) e14 e been presente histograa s da relativf mo e abundance (fig 13). The snail species have been grouped into three broad ecological categorie s- woodland , intermediatd ean grassland. Lauria cylindracea is often put with the woodland species, particularly in the S of Britain, but in the W and N it is more sensibly included with the intermediate or even open-country species. In the Orkneys today it is often extraordinarily prolific in dry, open sand-dune and fixed-dune pasture habitats. A consideration of the stratigraphy of the section described, together with the results of snail analysis, enables the following tentative sequence to be put forward. There was first a phase of solifluxion, indicating a cold climate and open, treeless landscape, which probably relateclosine th o st g stageLase th tf sGlaciationo . A period of soil formation then ensued under the milder, temperate conditions of the Post-glacial. parene Th t materia non-calcareous i l soie thus th i l d s san devoi lanf do dlowes ratsnailsit y n ean i r t a , uppee levels centimetreth w t rfe .Bu buriee th f so d soil contain fragment shelf so l sand derived ultimately winy b da transporse fro e mth t (Evans 1972) incorporatioe Th . thif no s material probably took placy eb earthworm mixing when the cover of blown sand was minimal. A similar phenomenon occurs at Skara Bra numbe a Orkne n et i a d placef o ryan s coasalonN Cornwalf e o gtth l (Perranporth, Newquad yan Dayme lattee r th Bay) rd sitan , ehavI e discussed wit Keitr hM h Valentine (Dep Soif to l Science, Reading University agreeo )wh s that depositio thia f no laye wind-blowf ro n sand followe earthwory db m mixing mose th te likelar y mechanisms involved. The land snail fauna from the buried soil (62-67 cm) is unequivocably a woodland one and closely similaabundance th n (i r f Discus,eo Carychium Zonitidaee th d an woodlano )t d faunas from southern Britain. In the shell sand above (45-62 cm), the grassland species, Vallonia excentrica, appears, suggesting a degre deforestationf eo . Whether thinaturaa s si l process brought overwhelmine abouth y b t g actiof no the wind-blown sand or whether it is in part or wholly of man-made origin is not clear, but it should be pointed out that marine shells, presumably human food debris, were present at the base of this horizon indicatin) earlcm n a 2 ( ct 6 y presenca e clearancstage gn th th ma n e i f eo e process significance .Th e th f eo marked increase of Clausilia bidentata in the blown sand is uncertain. A sample from the midden (20-45 cm) was too poor in land snails for a count to be made, but the WOODLAND SPECIES GRASSLAND SPECIES 5 Q a g 5 w 1 o b "2 I I < y 5 o o a Q en Q; si i >:(A?S CO ^,^s i tKUJ =J ~j eg Ui 1 LU(jg 1 EID^ LU ^t ff J :m §&l3 5Q.^3> Lu !±:i ^v1!••^ 0-r 1382 I I UNKEMPT i GRASS 10- 2) 20- 1 ol 4

^ d» 30- A rf MIDDEN /, r% 40- o)> • 70 50- CLEARANCE '. 199 PHASE 60- '' 329 "WOODLAND" ______70- 0 10 0 10 0 10 0 10 30 0 10 30 50 70 0 10 0 0 O% ,'j OPEN 80- ^'i GROUND FIG 13 Buckquoy: lanla d snail diagram from cliff section 200 m N of archaeological site RITCHIE: PICTISH AND VIKING-AGE FARMSTEADS AT BUCKQUOY, ORKNE9 21 | Y few shells which were extracted (Clausilia, 5; Cochlicopa lubrica, 1; Vallonia, suggesrelate) e 1 b o t o dt t i t layere th s below rathe soilp to .re thath o nt The topsoil is characterised by the predominance of Lauria cylindracea, a reflection, perhaps, of the somewhat more specialised environmen presene - particularlth f y to tda absence yth woodf eo y vegetation - by contrast with that which obtained in the past. The age of the midden in this section is uncertain, and it is not possible to tie in this sequence with that from the Pre-Norse and Norse archaeological site, though the close proximity of the two suggests dealine same ar th tha e ed w tgsite an wite . Moreoverhon smala , l collectio lanf no d snails from Norse phas I (takeeII n fro earte mth h withi limpete nth winklesd san ) suggested affinities wit faune e hth th n ai blown sand (45-6 rathe) 2cm r than wit topsoie hth l (Table 14). TABL4 E1 Buckquoy. Land Mollusca. Norse III is a small sample of earth from marine shells in the archaeological site. All samples, except Norse III, weighe dryr dai 1-.+ =g 0k non-apical fragment uncertai= . cf ; n identification.

i———————————*———————————> Norse 62-67-5 53-5-62 45-53-5 I II 0-10 — — 7 Carychium 5 2 tridentatum 8 5 (Risso) Cochlicopa lubrica (Muller) 445 — — — — 1 Cochlicopa 5 lubricella — (Porro) — 1 2 9 4 1 Cochlicopa spp. — — Columella— edentula— (Draparnaud 2 . cf ) Vertigo pusilla Muller 2 1 — — — Vertigo substriata (Jeffreys) 8 1 — — — Vertigo pygmaea (Draparnaud) 1 — — — — Papilla muscorum (Linne) 1 — — 77 — 5 116 6 2 31 Lauria 3 cylindracea2 Costaa (d ) Lauria anglica (Wood) cf. 2 — — — — Acanthinula aculeata (Muller) 4 + — — — Vallonia excentrica Sterki — 45 — — Clausilia bidentata (Strom) 46 51 28 4 — — — — — + Cepaea 1 hortensis (Muller) Helicella caperata (Montagu) — — — 1 — — 2 2 Punctum — pygmaeum 6 (Draparnaud 5 ) 2 1 2 1 9 3 Discus rotundatus 0 11 (Muller) Euconulus fuhus (Muller) 1 — — — — Vitrea contracta (Westerlund) 3 4 — — 5 5 4 6 1 5 Oxychilus 1 alliarius1 (Miller) Retinella pura (Alder) 21 4 3 — — Retinella nitidula (Draparnaud) 11 5 — — — Vitrina pellucida (Muller) 1 9 1 49 — — — — — 1 — Limacidae SUMMARY shell-fise Th h debris indicates total dependenc rockn eo y shore species limpete Th . s were collected throughou intertidae th t l zone. Ther decreasa limpet e s esizi th e f th e o sn ei fro Pre-Norse mth Norso et e phase occupationf so , possibl over-collectingo t e ydu quantite Th . shell-fisf yo insignificans hi termn i t f so bulk in the diet of the occupiers of the site. Land snails indicate a formerly woodland landscape, with a subsequent phase of deforestation possibl associatee b o yt d with early human settlemen areae th .n i t

APPENDIX5 The human bones from Buckquoy, Orkney by J Hillsdon Smith, Department of Forensic Medicine, University of Edinburgh LONG CIST BURIAL This skeleton was poorly preserved and showed disintegration and disappearance of some of the bones comprising the right wrist and hand, right foot and left hand. 22 | 0PROCEEDING S OF THE SOCIETY, 1976-7 Age Alavailable lth e long bone epiphyses were fully united thire .Th d molar teeth were presenn o t both sides of the upper jaw and on the right side of the lower jaw. The sternum was not fused to the manubrium .lippin o glenoie Thern th s f geo wa d cavity symphysie Th . s pubic characteristics, together wit above hth e findings, suggeste range ag 25-30 f eo n d a years. Sex On the basis of pelvic characteristics and the shape and formation of the frontal bones, mastoid processes and orbits of the skull, the skeleton was assessed as being male. Build Thiassesses swa beins da g 'average'. Stature onle yTh intact bone lefpresene overals tth it s tibi d twa a lan length applie Trotteo dt d ran Gleser's equations gav elivina g heigh 67-68f o t inches. Teeth These were all healthy and showed only slight wear on their occlusal surfaces. Other characteristics Lumbar vertebrae 5 Therd collapss an ewa 4 splittind bodiee an th f f thesgo so e vertebrae with consequent 'splaying spine th t thif ea o ' s level righe .th Lumbabodte d rotatesids th an yf e4 o wa n r5 d o of 5 was eroded. Right femuranterioe Th r surfacthi f uppee o sth d bonf ren eo e bore three, what appeare, be o dt sinuses. Thes levea e n wero l ) wit e(i greatee hth r trochanter belom ; m (ii wlowee 0 )2 (iiith d (i) n an ;)ri part of the neck of the bone. The edges of these 'sinuses' were smooth and rounded, and the lower part of this portio bone th ef no (whic s separathwa e fro shafte alss mth owa ) rounde smoothd dan e Th . mid-shaf missings wa t . Tarsal bones There was marked asymmetry between the tarsal bones of the right and left feet, the former being smaller, distorted and, when re-assembled, giving the appearance of a hyperextension (equinus) deformity. These characteristics, together, indicate a septic condition of the spine, probably tuberculous in nature, giving rise to a psoas ('cold') abscess and subsequent osteomyelitis of the femur with foot deformity. The right hand, right foot and left hand There was no feature present to indicate that any of these joint beed sha n amputated featuree Th . s were thos soif eo l erosio bonf no naturad ean l decay. Cause of deathlikele Th y caus deatf eo t thiha s age, assumin interpretatioe gth tuberculoua f no s disease is correct, is generalised or miliary tuberculosis with wasting and toxic anaemia.

VIKING-AGE BURIAL, PHASI EV This was a poorly preserved partial skeleton. Age All epiphyses of the long bones were united. There was insufficient skull, dental or pelvic material present to enable an age estimate to be made. However, the presence of osteoarthritic 'lipping' thoracie oth f c vertebrae frane bodieth kd osteophytesan lumbae th f o s r vertebran ei suggese ag n a t excess of 40 years. Sex On the basis of the shape of the obturator foramen of the pelvis, the digastric groove of the skull and the general bony configuration, the indications are that this is a male skeleton. Stature Measurements of the left ulna, left radius and right humerus, using Trotter and Gleser's formulae give a living height of 5 ft 9 in to 5 ft 11 in. Cause of deathfragmentatioe Th boneskule decae d th th nan f f lys o o mak t difficulei asseso t t s possibilite th f heayo d injury. However, examinatio f thesno e fragments showe cleao dn r evidence that ther bee d sucy eha n an h injury findinge Th . s were more consistent with collaps pressur o skult e e th f ldu e o e and decay after burial.

INFANT BURIAL IN HOUSE 1, PHASE 5 The appearances and measurements of these bones are consistent with them being those of a newborn or neonate infant of undetermined sex. The cause of death was not apparent.

MISCELLANEOUS INFANT BONES followine Th g bones, initially collected together with animal bones fro middee mth n deposits, have been identified by Dr C Oyler (29 Church Road, Longlevens, Gloucester). RITCfflE: PICTIS VIKING-AGD HAN E FARMSTEAD BUCKQUOYT SA , ORKNE1 22 | Y Phase V Fragmen infant'n a f o t s cranium, newbor r earlierno ; long bone, probably tibia, from newborn or earlier infant; fragments of humerus and probable femur from infant, newborn or earlier. Phase V InfanI t clavicle. Phase Ilib Part of infant's humerus, newborn or earlier.

APPENDIX 6 The ogam inscription on the spindle whorl from Buckquoy, Orkney by K H Jackson, Department of Celtic, University of Edinburgh Instead of the normal straight base-line, this feature is engraved as a crude circle, roughly concen- tric with the central hole. The only other circular ogam inscription known to the writer is that on the symbol grounde stonth n ei f Logiso e House, Aberdeenshire whicn ,i h case thersuco n hs ei rational explanation circlee foth r . Apar whic, t froM e clearls hmi th y diagona base-linee th o lt othee ,th r stroke evidentle sar y intendegroupl right-anglet A al a r therefore o d b , s an B do (aicmi) t , , it H belon o lettersef o t se t o th go t . facn Mosi te morar t r les eo t right-angle sa poine th o tst where they reac line e t somhth more bu , ear e sloping, as at Logie, due partly to the fact that their direction is related to a circle, not to a straight line; caso n thers e i n doub i y t ean bu t which aicme they belon . Almosgto t every strok quits ei e clearly madt eou whe whore nth hels strona i l o dt g angln ligh a examined t ea tan d wit hmagnifyinga glass. Since the base-line is a circle, it is impossible to say where the inscription is supposed to begin, or reao t whethe t frode i outsideinside e ar me th th r w eo . However marx e kth , describede belob y wma intende 'feather-symbol'e th s da , particularl thers y a sligh a s ei t base-linbreae th d n ken i e e on here d an , ot curvefi s toward . Beginninx e sth g therefor t thiea s point assumind an , reao t de g ar frotha e e mtw th inside followine th , transliteratione th s gi : smalA abovx l base-linee 1 levee th eth f o l , whose left-han beford den slighe eth t break curvep su almost to touch the x. Either the feather-symbol, or the forfid for E. 2 Three strokes of the H aicme, sloping very diagonally down to the base-line but not crossing it; the gap between the first and second strokes being slightly larger than that between the second and third. Probably T rather than HD. longA 3 , clear stroke passing diagonally down from righ lefto t t , clear acros e base-lineth s . Unquestionably M. 4 Five stroke t righsa t anglebase-line th o sclearlt d ean y crossin . Certainlgit . yI 5 Five strokes abov base-line t crossineth no verl t al ey, bu g it diagonal firse ,th t three close together, base-line convergint th bottomfurthee laso e n bu th o th tw tt p e V a firs e to r a ;e th apart n gi th t a t strokfive eth f doee o t reac sbase-linno e hth endt eagainsbu p su uppee th t lasre parth t f stroko t e precedine oth f . gProbablI ylettea fivf ro e strokes, therefor , ratheeQ r tha (whico n thretw d h ean woul TD)e db . 6 A single stroke right across the base-line, sloping down diagonally from left to right; the direction slope ofth et mus i show d t an therefor t cannos, i . M A e b te eb 7 Three strokes below the line but not crossing it, therefore belonging to the B aicme; the first sloping extremely upwards diagonally toward leftothere e sth th , s doin decreasinn same i gt th ebu g order as to slope. This must be V. 8 Four clear strokes in the same position and of the same type and slope, but all parallel and sloping abou thire mucs a th t ds ha strokprecedine th f eo g letter thire Th .d stroke slightly crossee th s base-line. An S. 9 A very clear upright stroke crossing the base-line at right angles, therefore A. It has a sort of long uppes seriit t a fr end; compar inscriptione th similae n eth i A r t Aboynen 'serisa a n o f , Golspie, and Latheron. 10 Two strokes below the line but not crossing it, sloping upwards diagonally towards the left; L. 11 Two more of the same, rather closer together and sloping less, converging slightly towards the bottom but not joining. There is a line running diagonally down from the bottom of the first stroke which/seem continued an o tw joi o sse t welnth l beyond thit apparentls bu ,si yfortuitoua s scratch. Anothe. rL j PROCEEDING 2 22 SOCIETYE TH F SO , 1976-7 12 Four, or perhaps five, strokes above the line but not crossing it, sloping downwards towards the left; the doubtful fifth at a slightly greater angle. Either C or Q. After this there is a gap with room for one letter of five strokes, before the x is reached again. This gap further supports the view that the inscription begins with the x. The whole inscription is therefore either TMIQAVSALLC (or TMIQAVSALLQ) or ETMIQAVSALLC (or ETMIQAVSALLQ). Alternatively, if we read the inscription the other way on, looking at it from outside the circle and not from inside, we shall have SDDACTANIMV (or NDDACTANIMV) or ESDDACTANIMV (or ENDDACTANIMV). datine Th f Pictisgo h ogams unaccompanie decorativy db e motif always i s s difficult t thie bu son , peculiaritie e nons th ha f eo s suc 'bind-ogamss ha fancifur o ' l letter-forms which characterise late inscrip- tions in Scotland, and an 8th-century date, perhaps early rather than late in that century, seems probable. The A with serif would be in keeping with this. All the readings are wholly unintelligible and cannot be Celtic, but this is of course perfectly familiar in the ogam inscriptions of Pictland, and ours must be simply one more of the fairly numerous ogams in the unidentifiable language used by the Picts for this purpose (Jackson 1955, 138-42, 152-5). In case anyone is tempted to suggest that the possible ETMIQ consists of Latin et 'and' and the genitive of the Gaelic word maq 'son', the following considerations should be held in mind. First, such an inscription woul t begidno n wit presence h th 'And' s Latii a r f eo nno ; word probabl inscription a n ei spindla n no e whorl readiny .B g something like Allc miqt e Avsa assumind an g doethaq tmi s mean son'e 'o mighfth e ,w t get something like '(of) Allc and of the son of Avsa' (i.e. belonging to these two people with wholly unknown names) thit pretts si .Bu y nonsensical spindla s i r eno , whorl likel announco yt e tha belongt i t s mana o t , eve 'Allcf ni ' cannocaswergenitivy q e wifes truth an es mi i e hi e n t eI b t.I maq.f ethao t this Gaelic word does see mo occurfore t th m n i ,maqq meqq,r o fivn i e other Pictish inscriptions (Jackson 1955 expectec e , th 140)r t thithougdd a fo s,q an date h th vers ei y archaic thiprobabln shelca e b dt yno to be an objection, since it could well be due to deliberate archaisation on the part of the engravers. The objectio differena f o s ni insuperables i t t i kind d an , Irishd genitive 'sonc males Ol th : ma ,wa n ' i f eo sometime earln i c sy meic meicd Middlme r mec,r an o o e Irish stile lTh . later form mic, whic alss hi e oth modern one in Irish and Scottish Gaelic, did not appear until the late llth century at the earliest; and moreover it was apparently unknown in Scotland till even later, since the Gaelic texts in the Book of Deer, hailing from Bucha middle 12te th th n hi f eo century , have exclusively meic mec,r o never mic. muse W t be content to write off this inscription as unintelligible, like all the other 'Pictish' inscriptions.

APPENDIX7 Chalk spindle-whorls from Buckquoy, Orkney Collins,H byG Institut Geologicaf eo l Sciences, Edinburgh Three spindle-whorls were identified by macroscopic examination: No. 83 A fine-grained cream coloured limestone, weathering white and powdery. It contains broken fossil shell fragments, unidentifiable t almosbu , t certainly from lamellibranchs. No. 84 A fine-grained cream coloured sandy limestone, with quartz grains up to 0-5 mm in diameter. No. 85 A fine-grained cream coloured limestone with a few rounded quartz grains, about 0-25 mm in diameter. The three spindle-whorls are made from a closely similar rock type, varying only in detail. This rock-type is not to be found in situ in Orkney, but it can be reasonably well matched with chalk erratics to be found in the boulder clay of both Orkney and the mainland of Scotland. We have a specimen from South Ronaldsa 3285)C y(M , reputedl f Cretaceouyo s chalk, which matches your specimen5 8 d an 4 s8 quite well alse oW . hav especimea n from Moreseat Farm, southwes Peterheaf o t 3286C d(M ) whica s hi shelly cream coloured limestone comparing favourabl. y 83 wit . hno lons ha g t I been believed that Mesozoi Morafoune e b cth Nort e rocko n dt i th y e n hFirti s ar d han Sea south of Orkney. Recently our Continental Shelf Unit obtained samples of chalk from the Moray Firth. Thes creae ear m coloured fine-grained limestones, with occasional quartz grains specimee On . s ni quite sandy, having quartz grains up to 2 mm diameter. In addition, I have compared the spindle-whorls RITCHIE: PICTISH AND VIKING-AGE FARMSTEADS AT BUCKQUOY, ORKNEY j 223 with specimens from the Daneian chalk in the Royal Scottish Museum. The Daneian chalk does not occur in the British Isles, and is the uppermost part of the chalk. Your specimens compare favourably with the matrix of two fossil specimens (RSM cat nos 1911.5.12894, and 1911.5.12919) from Faxe, Denmark. However, from my inspection of the chalk specimens mentioned above, I believe that the spindle-whorls could well have been made from chalk pebbles, obtained from local glacial deposits.

APPENDIX 8 bronze Th e ringe fron dpi m Buckquoy, Orkney by Thomas Fanning, National Park Monumentd san s Branch, Offic Publif eo c Works, Dublin The pin is a fine well-preserved specimen, the upper shank circular in cross-section, the lower shank assumin grectangulaa r cross-sectio taperind nan poina go t t whic slightls hi y curved. This portion is decorate broas it n ddo faces wit incisen ha d step pattern delimite verticao minoe tw th y db n l r o line d san faces wit hdoubla e vertical linepolyhedrae Th . l hea outlines dha d lozenge-shaped panels ornamented with a simple interlaced knot on one side and a saltire motif on the other. It is pierced to take the narrow grooved terminal circular-sectionee th f so d ring which also bears light groovin placeo tw t sga alons git circumference e pin-hea fixeth w n e rinno i d s Th gi d. through corrosio t originallnbu t woulyi d have swivelled freely. Lengt mm1 h17 ; thicknes f shano s mmk5 ; widt f heaho mmd9 ; thicknes f heaso d 6-4 mm; external diamete rinf ro g ; thicknes2-05mm . rinf so mm g5 The Buckquoy pin belongs to a type of dress-fastener usually referred to as a ringed pin or ring-pin, dividewhice b n hca d int onumbea classef ro subdivisiond san rin e pin-heabasid e th g th an f so n so d form (Fanning 1969; 1975). This specimen clearly belong plain-ringede th o st , polyhedral-headed variety which have a relatively long pin with a small close-fitting ring held in a facetted head. Ornamentation, when present, is typified by the example under discussion - the interlaced knot and saltire motif on the pin-head step-pattere th d lowee an th n rn o shank . Also characteristi groovine ringth e s th ci ,n gparticularlo y wher t enterei pin-heae sth d givin sligha t gi t zoomorphic appearance. Although this f frequentypo ringef s eo i n dtpi occurrenc areae th sn ei settleVikinge th y db f so Norse origi majorite nIristh e th hf yo example numbea Scottise d th s an f ro h pins were recovered from native Celtic sites. Close parallel botr fors e fo ornamend hth m an citeBuckquo e e b th dn f o tfroca n mypi Lochlee crannog, Ayrshire (Munro 1882, 131, fig 144) and Lagore crannog, Co Meath (Hencken 1950, , 1430)14 72g .fi , Both these pins have small close-fitting rings with outer surfaces slightly groovede Th . Lochlee specime decorates ni d wit three hth e motifs Buckquo e founangulae th th n d- o n r yinterlacepi d knot, outlined saltire and step-pattern - while the Lagore pin has the interlaced knot design. Unfor- tunately, neither pin can provide any help as to dating, lacking as they do associations and stratification. furtheo Tw r parallels Limericn ,i k City Museum, although unprovenanced alse ,ar o worth mentiof yo s na they share the same decorative designs as the Buckquoy pin (Fanning 1969, 10, fig 1, 3, 4). However, a small polyhedral-headed pin from a ringfort at Lissue, Co. Antrim, was found in a level dated by the excavato 10te th h o rcenturyt , largelevidence th n slatya o f eo e trial-piece which contained, amongst other designs numbea , attemptf ro interlacen a t sa 11 g fi ,. , pd1 1) i . knono , t (Bers10 g fi u, 194750 , Comparative examples from Viking site more sar e helpful with regar datingo dt Skaile Th . l hoard, Orkney, deposited c AD 950, contains a silver pin of this type decorated with a lightly incised step-pattern. bees Thiha nn incorrectlspi y liste fish-hooa s da k (Grieg 591940,129g fi ; , Proc49 . Soc,no Antiq Scot,3 (1857-60), 249, no. 49). Another, undecorated, bronze pin from Jarlshof, Shetland, was found in a midden dated by the excavator to 10th-century Viking levels (Hamilton 1956,152, no. 35, pi XXIX, 4). A number of male Viking graves provid wits eu h similar dating particularn i , , three burialn s fro Ma Isle mf th eo each of which contained a ringed pin in fragmentary condition but undoubtedly of this form. These grave datee sar withio dt , pi centure s(Bers87 n0 th VIII, Wilso95 d 79 , yuo , an ,e t A62 0 nD, 85 196643 , XII, d, XTV, c). A floruit, therefore, in the 10th century is acceptable for this type and is strengthened by associatioe th n wit Buckquoe h th silve a f o rn pennypi Eadmunf yo Mane d(940-6)I th f x o leas t a o n t.tw I graves there was clear evidence to indicate that the pin had been used to fasten the woollen cloak in whic man'e hth s bod beed yha n reasonabls i dresse t i d dan supposo et e that suc case ht th ea alss owa Buckquoy (Bersu and Wilson 1966, 50, 68). Vikine Th g settlement Faeroese th n si , Icelan indeed dan d North America have yielded pin thif so s 22 | 4PROCEEDING S OF THE SOCIETY, 1976-7 type. The majority of the Icelandic pins are polyhedral-headed (Eldjarn 1956, fig 140) and the pin dis- covered in the excavations at L'Anse aux Meadows, Newfoundland, is an undecorated example of this form wit hsmala l worn pin-head (Ingstad 1970,134,135, fig 29) .bes e Probablth tf parallelo e e yth on r sfo Buckquo comen ypi s fro mfemala e grav Vikina n ei g burial moun t Tj0rnuyida Faroese th n ki . Thin spi is decorated wit interlacen ha d knot moti stild flan retain portiosa core th df nuseo securdo t fastenine eth g (Dah Rasmussed an l n 1956, figs 6-8) occurrencs It . examplee Faroee herth th d n ei san s from Iceland may possibly represen lina t k with settler Hiberno-Norsf so e origin , moror , e likely, Viking trade along Atlantie th c sea-route from centres suc Dublis ha Riordai6 n( n 1971, 76) absence .Th thif eo s form from Scandinaviae th n homelands highlight Hiberno-Scottise sth h association thid an s assumes additional importance when one considers the evidence of the 10th-century Hiberno-Norse coin hoards which indicates tha Vikinge th t Irelandn si thay b , t century, looke furtheo dn r than Scotlan Westere th d dan n Isle supporr sfo thein i t r campaigns (Dolley 1966, 18). Finally origiplain-ringee e ,th th f no d polyhedral-headed for bounms i witp ddevelopmene u hth f o t the ringed pin series as a whole. In Ireland, a number have been found on sites dated to pre-Viking times woult i d dan seem likely, therefore, tha typadoptee s Vikingstth e e wa th y db , probably durin secone gth d centuryh 9t e fore halth m.f Th fo became increasingly popula marketw ne s employe ra s th opened an - p du ment of designs such as the interlaced knot and step-pattern probably sprang from this exchange and contact between Irish, Vikin perhapd gan s late Saxon workshops (Bersu 1947, 50).

ACKNOWLEDGMENTS The excavation was undertaken as a rescue project on behalf of the Department of the Environment, and every assistance was afforded by Mr P R Ritchie, Dr David Breeze and mem- Inspectorate berth f so f Ancieneo t MonumentOrknee th y b y d stafsan f undeJ Drever r M r . r IslD a Mclnnes acte assistans da t director during both seasons verd an ,y special thanke ar s owed to her for her unfailing support. The excavation photographs were taken by Mr IBM Ralston, Mr Arthur MacGregor and Mr IA G Shepherd, and the writer is greatly indebted to l thos tooo al o ekwh t excavatione the pard th mn an i t , especially Mis TuckwelsA l (nos wMr Shepherd)G A I YorsteArir J r M e M , d Schermernan . Grateful thank alse sar o offerel al o t d thos helpeo post-excavatioe th ewh n di n work, particularl RobertsoN W r yr conservinM nfo g StevensoK artefactse B th R r identifyin r nfo M ; coine r gBrycth M ; e Wilso f Tankernesno s House Museum, Kirkwall; Mr Graham Sandison; Mrs Heather Thompson of the Royal Botanic Garden, Edinburgh RomanJ r DurnMacaulae E M S th d f r so ; an D y Institut Soir efo l Research, Aberdeen Claud;M e Sterck Universite th f xo Haute ed e Bretagne, Rennes Anthonr ;D y Jackson of the University of Edinburgh. The writer is deeply indebted to the specialists whose reports for appendicee mth thio st s paper, particularl r AlwynyM e Wheele generouslo rwh y dealt with the fishbones at short notice. Mrs Cecil Curie, Mr lain Crawford and Mr J Graham Campbell have been most helpful with informatio advancn ni publicatiof eo n abou Brouge th t Birsaf ho y Udae Th l d respectivelyan artefacte .Th s fro excavatioe mth n have been deposite Tankernesn di s House Museum, Kirkwall. BorthwickT r wore figuree M th f Th ke o ar s , Departmen Environmene th f o t t (figs 4-11), ScotG I r t (figM d s 1-3)an boto ,t whof ho m special thank offerede sar .

REFERENCES Addyman P ,197 6 'Archaeolog Anglo-Saxod yan n society' Sievekingn i , Longworth G e d G , , Id Han Wilson (eds)E K , , Problems economicn i sociald an archaeology, London, 309-22. Anderson, J 1873 'Notice of the excavation of the brochs of Yarhouse - in Caithness', Archaeol Scot, VI (1873), 131-98. Anderson J , 1883 Scotland Pagann i Times: Iron Age. Edinburgh. RITCHIE: PICTISH AND VIKING-AGE FARMSTEADS AT BUCKQUOY, ORKNE5 22 | Y Anderson J , 1901 'Notice ninf so e brochs alon Caithnese gth s coast from Skirzo t Keis y asBa Head excavated by Sir Francis Tress Barry', Proc Soc Antiq Scot, 35 (1900-1), 112-48. Anderson, A O and Anderson, M O (eds) 1961 Adomnan's Life of Columba. Alcock L ,196 3 Dinas Powys. Cardiff. Almqvist, B O and Greene, David (eds) Proceedings of the Seventh Viking Congress, Dublin, 15-21 August 1973. Dublin. Bergquist Lepiksaard an H , J , 1957 Medieval animal bones found Lund.n i Lund. Bersu, G 1947 'Excavations at Lissue', Ulster J Archaeol, 10 (1947), 30-58. BersuWilsond an M 196G , D ,6 Three Viking Graves Islee th n oi f Man. London. Blomqvist Martenssond an R , 196W A ,3 Thulegravningen 1961'. Archaeologica Lundensia, I (1963)I . Lund. Boyd, J M, Doney, J M, Gunn, R G and Jewell, P A 1964 'The Soay sheep of the island of Hirta, St Kilda stud A .feraa f yo l population', Proc Zool Land,c (1964)2 So 14 , 129-63. Bramwell D ,197 7 'Bird bone' Clarken Carterd i , an , ExcavationsH ,A , King'sn i Lynn, 1963-70, London, 399-402. Brothwell Brothwelld an D , P , 1969 Food Antiquity.n i London. Calder, C S T 1939 'Excavations of iron age dwellings on the Calf of Eday, Orkney', Proc Soc Antiq Scot, 73 (1938-9), 167-85. Chaplin E study197 e R , Th 1 of animal bones from archaeological sites. London. Charleson, M M 1904 'Notice of some ancient burials in Orkney', Proc Soc Antiq Scot, 38 (1903-4), 559-66. Childe G 193 V , 1 Skara Brae. London. Coles 197M J , 1 'The early settlemen Scotlandf o t : excavation t Mortonsa , Fife', Proc Prehist Soc,7 3 (1971), 284-366. Corbet, O 1961 'Origin of the British insular races of small mammals and of the 'Lusitanian' fauna', Nature, (1961)1 19 , 1037-40. Crawford, I A 1973 Excavations at Coileagean an Udail (The Udal), N Uist, 10th interim report. Christ's College, Cambridge. Crawford, I A 1974 'Scot (?), Norseman and Gael', Scot Archaeol Forum, 6 (1974), 1-16. Cruden S ,196 5 'Excavation t Birsaysa , Orkney' Smalln i , (ed)A , , Fourth Viking Congress, 22-31. Curie, A O 1939 'A Viking settlement at Freswick, Caithness', Proc Soc Antiq Scot, 73 (1938-9), 71-110. Dolley196M Hibernoe H Th 6R , Norse BritishCoinse th n i Museum. London. Dunbar G 196 historicJ ,e Th 6 architecture of Scotland. London. Eldjarn K 194, 8 'Ey3ibygg 5Hrunamannaa a fretti', Arbok hins islenzkafornleifafelags, (1943-8), 1-43. Eldjarn K ,195 6 Kuml Haugfeg o heidenumr u Islandi.a o si Reykjavik. Elwes J ,191 2 'Note primitivn so e breed sheef so Scotland'n pi , Scottish Naturalist, (1912). 1 , Epstein, H 1971 Origins of the domestic animals of Africa. London. Evans, J G 1972 Land snails in Archaeology. London. Ewart C 191 J , 3 'Domestic shee theid pan r wild ancestors', Trans Highland Scot Agric Soc, (1913)5 2 , 160. Fanning, Thomas 1969 'The bronze ringed pins in the Limerick City Museum', North Munster Antiq J, 12(1969), 6-11. Fanning, Thomas 1975 'Some bronze ringed pins from the Irish Midlands', / Old Athlone Soc, 1 (4) (1974-5), 211-17. Fenton, A 1973 'Craig-fishing in the northern isles of Scotland', Scot Stud, 17 (1973), 71-80. Forde D ,195 4 African Worlds. Oxford. Graham-Campbell, J A 1976 'The Viking-age silver and gold hoards of Scandinavian character from Scotland', Proc Soc Antiq Scot, 107 (1975-6), 114-35. Grant, W G and Callander, J G 1934 'The broch of Midhowe, Proc Soc Antiq Scot, 68 (1933^), 444. Grieg S ,194 0 'Viking antiquitie Scotland'n si Shetelign i , (ed)H , , Viking Antiquities Greatn i Britain and Ireland, Oslo, vo. II l Goodier, R (ed) 1975 The Natural environment of Orkney. Edinburgh. Hamilton, J R C 1956 Excavations at Jarlshof, Shetland. Edinburgh. Hamilton C 196 R J ,8 Excavations t Clickhimin,a Shetland. Edinburgh. Hencken, H O'N 1938 Cahercommann: A stone fort in County Clare. Dublin. S 226 | PROCEEDINGS OF THE SOCIETY, 1976-7 Hencken O'NeilH , l 1950 'Lagore Crannog Irisn :a h roya 10to t l h residench centurie7t e th f eo s A.D.', Proc Roy Irish Acad, 53c (1950), 1-247. Waited Higgs an P 196 S J , E , 3 'Autumn killing', Antiquity, (1963)7 3 , 274. Ingstad S 197 A , 0 'The Norse settlemen t L'Ansa t Meadowsx eau , Newfoundland preliminarA . y report from the excavations 1961-8', Acta Archaeol, 41 (1970), 109-54. Isaac, G 1971 'Diet of early man: aspects of archaeological evidence from Lower and Middle Pleistocene sites in Africa', World Archaeol, 2 (1971), 278. Jackson, K H 1955 'The Pictish language', in Wainwright, F T (ed), The Problem of the Picts, Edinburgh, 129-60. Jones 197M A , 5 'The marine environmen Orkney'f o t Goodien i , r 1975, 85-94. Kaland, S H H 1973 'Westnessutgravningene pa Ronsay, Orkn0yene', Viking, 37 (1973), 77-102. Kendrick D 193 T , 3 'A gaming Vikinboare th f do g period foun Ireland'n di , Acta Archaeol, (1933)4 , 85. Lea, Davi Bourne d P d197 an R 5,W 'The bird Orkney'f so Goodien i , r 1975, 98-121. Lethbridge, T C 1952 'Excavations at Kilphedir, South Uist, and the problem of brochs and wheel- houses', ProcPrehist Soc, (1952)2 8 1 , 176-93. Long, C 1975 'Excavations in the medieval city of Trondheim, Norway', Medieval Archaeol, 19 (1975), 1-32. domestice M 184 th D Lown 6O , animals Britishe oth f Isles. London. MacGregor, A 1974 'The Broch of Burrian, North Ronaldsay, Orkney', Proc Soc Antiq Scot, 105 (1972-4), 63-118. MacLaren, A 1974 'A Norse house on Drimore Machair, South Uist', Glasgow ArchaeolJ, 3 (1974), 9-18. Marwick H ,197 place-namese 0Th of Birsay. Aberdeen. Megaw, Basil 1976 'Norsema nativd Kingdoe nth an Islese n ei th reassessmenf mA .o Mane th f xo t evidence', Scot Stud, (1976)0 2 , 1-44. Munch, G S 1966 'Gardshanger i Nord-Norge', Viking, 30 (1966), 25-59. Munro R ,188 2 Ancient Scottish Lake Dwellings Crannogs.r o Edinburgh. Naroll R ,196 2 'Floor aresettlemend aan t population', American Antiquity, (1962)4 , 27 , 587-9. Nicolaisen H 196 F 9,W 'Norse settlemen Northere th n i tWester d nan n Isles', Scot Hist (1969)8 Rev,4 , 6-17. Noddle, B A 1973 'Determination of the body weights of cattle from bone measurements', in Matolcsi, J (ed), Domestikationsforschung und Geschichte der Haustiere, Budapest, 377-89. <3 Riordain, Breandan 1971 'Excavation t Higsa h Stree Winetaverd an t n Street, Dublin', Medieval Archaeol, (1971)5 1 , 73-85. Padel, O J 1972 Inscriptions ofPictland. Unpubl thesis, University of Edinburgh. Radford, CAR 1962 'Ararchitectured tan : Celti Norse'd can Wainwrightn i , (ed)T F , , The Northern Isles, Edinburgh, 163-87. Ritchie A ,197 2 'Painted pebble earln si y Scotland', Proc Antiqc So Scot, (1971-2)4 10 , 297-301. Ritchie, A 1974 'Pict and Norseman in Northern Scotland', Scot Archaeol Forum, 6 (1974), 23-36. Roussell A ,193 6 'Sandne neighbourine th d san g farms', Meddelelser Gr0nland,m o (1936)2 , 88 . RCAMS 1946 Royal Commission on the Ancient and Historical Monuments of Scotland, Inventory of the Ancient Monuments of Orkney and Shetland. Edinburgh. Ryan 197M , 3 'Native potter earln yi y historic Ireland', Proc Irishy Ro Acad,, 11c , , (1973)73 , 619-45. Ryder L 196 M , 8 'Evolutio Scottisf no h breed sheep'f so , Scot Stud, (1968)2 1 , Ryder, M L and Stephenson, S K 1968 Wool Growth. London. Shawcross 196,W 7 'An investigatio prehistorif no c dieeconomd tan coastaa n yo l sit t Galateea a Bay, New Zealand', Proc Prehist Soc, (1967)3 3 , 107-31. Silver A 196 I , 3 'The agin f domestigo c animals' Brothwelln i , Higgsd an (eds)E D ,, , Sciencen i Archaeology, London, 250-68. Skaare, K 1976 Coins and Coinage in Viking-age Norway. Oslo. Small, A 1966 'Underhoull, Unst, Shetland', Proc Soc Antiq Scot, 98 (1964-6), 225-48. Spencer, P J 1975 'Habitat change in sand-dune areas', in Evans, J G, Limbrey, Susan and Cleere, Henr ylandscape:e (eds)th effecte n o Th ,Highlande n th o ma f zone, London (CBA Research Report no. 11), 96-103. Stenberger 194M , 3 Forntida Gardar Island.i Copenhagen. RITCHIE: PICTISH AND VIKING-AGE FARMSTEADS AT BUCKQUOY, ORKNEY | 227 Stenberger, M 1961 'Das graberfeld bei Ihre im Kirchspiel Hellvi auf Gotland', Acta Archaeol, 32 (1961), 1-134. Sterckx C ,197 0 'Les jeu damiere xd s celtiques', Annales de Bretagne, (1970)7 7 , 597-609. Sterckx C ,1973 a 'Les trois damier Buckquoe sd y (Orcades)', Annales Bretagne,e d (1973)0 8 , 675-89. Sterckx, C 1973b 'Les jeux de damier celtiques', Etudes Celtiques, 13 (1973), 733-49. Stevenson K 195 B 1R , 'A hoar Anglo-Saxof do n coins foun t londa a Abbey', Proc Antiqc So Scot,5 8 (1950-1), 170-5. Stevenson, R B K 1955 'Pins and the chronology of the brochs', Proc Prehist Soc, 21 (1955), 282-7. Thomas, C 1963 'The interpretation of the Pictish symbols', ArchaeolJ, 120 (1963), 31-97. Wainwright, F T 1962 'Picts and Scots', in Wainwright, F T (ed), The Northern Isles, Edinburgh, 91-116. Walker, C 1964 'Growth and development of meat qualities in Angoni cattle', / Agric Sci Camb, 63 (1964), 155. Waterman, D M 1959 'Late Saxon, Viking and early medieval finds from York', Archaeologia, 97 (1959), 59-105. Watt, W G T 1882 'Notice of the broch known as Burwick or Berwick - Orkney', Proc Soc Antiq Scot, 16 (1881-2), 442-50. Watt, W G T 1888 'Notice of the discovery of a stone cist - at Skaill Bay', Proc Soc Antiq Scot, 22 (1887-8), 283-5. Wheeler A ,196 fishese 9Th British e oth f Islesnorth-westd an Europe. London. Wheeler, A and Jones, A 1976 'Fish remains', in Rogerson, A 'Excavations on Fuller's Hill, Great Yarmouth', East Anglian Archaeol, , reporNorfolk2 . no t , 208-24. Wilson, D M 1971 'The Norsemen', in Menzies, G (ed), Who are the Scots! London, 103-13. Wilson Edwards, V G , , Knox Stephensd ,W an , Jones,J B V 193C Geology,J e R , 5Th Orkneys.e ofth Mem Geol Survey Great Britain, Edinburgh. Wilson, J 1910 Evolution of British Cattle and the formation of breeds. London. Young, A and Richardson, K M 1960 'A Cheardach Mhor, Drimore, South Uist', Proc Soc Antiq Scot, 93 (1959-60), 135-73. The Society indebtedis Scottishthe to Development Department granta for towards cost the this of paper