Proc Soc Antiq Scot, 113 (1983), 554-593

Excavation t Scallowasa y Castle, 197 198d 9an 0 LindsayJ HallW D d *an t with contribution CaldwellH D J Daveyy I Hodgso s b P , W G , n an SmitdC h

SUMMARY This report describes exploratory excavation fivsn i e areas near Scalloway Castl Shetlane th en i d Islands. Although a few finds of pre-medieval date were recovered, the earliest features date to the 17th centurpossible ar d yan y associated with eithe constructioe rth r noccupatioo castlee th f nstratifieo .A d sequence of archaeological deposits from the 17th to 20th centuries was found in one area.

INTRODUCTION The town of Scalloway lies on the W coast of the Shetland Islands, some 11 km from Lerwick. The castle is situated on Blacks Ness on the W bank of the East Voe of Scalloway 3935 (NG40 ) (figU RH s 1,2). The construction of Scalloway Castle was undertaken in 1600 by Earl Patrick Stewart, Earl of Orkney and Shetland. He succeeded his father to the title in 1592 and required a suitable residence for his representative in Shetland as well as an occasional home for himself, although his principal base was the Palace at Birsay in the Orkney Islands. The choice of this site for the castle may well indicate that there was a thriving community of traders and merchants whom Earl Patrick wished to control at first hand. Patrick taxed each parish of Shetland obliging them to send as many men as possible to carry out the building work on the castle. These men were not paiprovido t dd wageha e d foodthein st surprisinglyan ow rNo . , many refuse obeo d t Earl' e yth s dictates and several were tried for the crime. Soon after its completion, the castle was the setting for several meeting Althine th f so t whicga h Patrick would decre law taxesw d esne an . n 160I 9 Earl Patric imprisones kwa Jamey db I followinV s gdeputatioa f Shetlanderno s king'e th o st court describing their ill-treatmen Earl'e th t sa t hands. Bisho thew npLa took over power in the islands and records survive describing his courts in Scalloway Castle, the first of which was held on 21 August 1612. Earl Patrick was beheaded along with his father in 1615 at Edinburgh. The castle continued to be used as a legal centre until 1653 when Cromwellian troops were stationed there. When John Brand visited the castle at the beginning of the 18th century he recorded that it was in a bad state of repair. He wrote, The sclaits have for the most part fallen from the roof, and are daily falling with every storm, so tha timbere tth , muc whic f hvert o ye y hs i goo fresd dbeginninan s h i t (Branro go t d 1883,137-8). * 136 Spynie Street, Bishopmill, Elgin 136 Ercall Road, Brightens, Falkirk HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AND 1980 | 555

km 30

Fi o1 Locatio n map Scallowaf so y

In 1733 Thomas Gifford noted '. . . now nothing remains but the stone walls' (1879,8). By 1803 the Earl of Morton had given Andrew Mitchell permission to remove the gateway and window ornament houss hi r t Sandea s fo documentaro N . y record relate ancillaro dt y buildings survives although it may be assumed that there must have been several-perhaps similar to those at the Earl's Palace at Birsay. The castle was finally placed under the guardianship of the Com- missioner WorkM H f so (now Scottish Development Department 1908n )i . Areas to the N and S of the castle, which were threatened by the construction of a knitwear factory and an access road for the Blacks Ness Harbour Development respectively, were investi- gated maie Th .n objectiv discoveo t s ewa r whether ther beed eha n outbuildings associated with the castle. The excavations were carried out during July and August 1979 and January and Feb- ruar ye Urba th 198 y n0b Archaeology Unit (no e Scottiswth h Urban Archaeological Trust) with a team of four excavators. Most work was carried out by hand except in Area G, where mechanicaa l excavato uses removo rdwa t e overburden. The original site context numbers have been used in this report except that '4' prefixes Area A numbers to avoid confusion with Area B numbers, and all numbers relating to Areas C, D and F and E have been omitted as few features of archaeological interest were encountered. detaileo N d plan r sectiono s f theso s e area includee same ar s th r e dfo reason l pland Al . an s sections are in the archive which will be housed in the National Monuments Record of Scotland, 54 Melville Street, Edinburgh. Area A deposits have been divided into a sequence of three phases, 556 I SOCIETY OF ANTIQUARIES OF SCOTLAND, 1983

EAST VOE OF SCALLOWAY

FI G2 Site location plan firs e baseth t n publishedo d editio Ordnance th f no e Survey map, 1880

the first of which, from the evidence of the finds, relates to phase 1 activities in Areas B and G. Phasin othe e t useth no n dri s g i area s becaus featurew efe s were encountered.

THE EXCAVATIONS, 1979 and 1980

AREA A (figS 2, 3, 4) Although some wor carries k wa t durin dou 197e gth 9 season mos excavates Aref o twa aA d betwee JanuarFebruar1 9 investigateds n2 2 wa d an ym y5 t onl1980y b bu , y a stri0 A .3 pc small area, 6 by 8 m, towards the W end of the site had deposits of any antiquity. The removal of turf and topsoil from the remainder of the site revealed natural Laxfirth Limestone at a depth 0-c . 3f m o Phase (fi1 ) g3 earliese Onth f eo t features locate ditcha s Aren di wa , 4551aA t int, whiccu o s naturahwa l limeston placen i d sean throug hlayea f browo r S n n clayera t I y. subsoim 1 depta c o t f l ho HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AN7 D55 198 | 0

AREA A PHASE 1

METRES

4527

DITCH

FI G3 Are , phasaA e 1, plan sectiond san s 558 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1983 from the main N section for some 6 m, where it turned W and continued for at least 12 m until t disappearei d under Castle Street maximue .Th m surviving widt thif ho s L-shape. dm 3 ditc s hwa The W end was some 0-38 m lower than the N end which merely reflected the slope of natural down to the W. Fills appeared to respect the change in direction of the ditch and there was no indication that the N-S length extended further S or the W-E length further E. The basal fill consisted of a grey brown silt, 0-05 m thick, which was overlain by layers of decayed limestone which had been deposited from both sides although the greatest quantity had come from the E side of the N-S length. At this level the ditch had been recut at least once t face beed cu ha se n th eroded t bu , probabl actioe th f watey nyo b r flowing nex e alonTh t . git major ditch fill, 4543, consisted of angular limestones and some dark brown clay. This appeared havo t e bee ndeliberata e backfilling and, fro evidence m th clay-pipe th f eo e finds, would appear to have occurred in the mid-17th century. Layer 4543 was overlain by a deposit of dark brown clay soil some depthn 0-1i 8m . This turn n i covere,s wa , layey db r 4538 consistin limestonf go e rubble and some dark brown clay. This deposit, the final backfilling of the ditch, was probably necessitate subsidence th y db f loweeo r fillslimestone Th . eindicat y rubblfile ma lth n ei e that t leasa t originae somth f eo l upcas beed ha t n include backfillinge clay-pipth w n di fe e Th e. frag- ments recovered appear to be slightly earlier in date than those found in layer 4543. No features other than a small pit, 4549, which slightly cut the upper fills of ditch 4551, were found within the area partially enclosed within the L-shape. A gully, 4539, with a width which varied between 0-25 and 0-66 m and a maximum depth of c 0-28 m, lay parallel to the ditch, some 0-24 m from the E edge. The feature could only be followed for c 2 m because of an electricity substatio . AlthougS e th o nt hdireco n ther s t ewa relationshi p between ditch 455d 1an gull coulyo 4539tw de welth , l have been contemporary irregularitieo .Tw gulle base th th yf eo n si may have been cause somy db e typ post-in-trencf eo h palisade which, together wit ditce hd th han associated upcast would have forme sizeablda e fortification thin i s arewa t aI . tha greatese th t t quantit decayef yo d limeston falled eha n back int ditche oth , possibly fro upcasn ma t bank. It is likely that the ditch is Cromwellian in date although it is possible that it could have been constructe earls da s 160ya 0 whe castle beinns th ewa g built. Indeed ditce only th , hyma function-te on hav d eha o drain water fro foundatioconstructioe e mth th r fo castle e t th ncu f no . S oditce littlgulld th hf an e o y were recovered tha functione tth botf o s h must remain conjectural although the evidence does possibly indicate the presence of a palisade and bank at the E side of the N-S length. There could have been some type of enclosure in the area of the castle pier as Earl Patric knows ki hav o nt timbed eha irod an rn stole 160n ni 4 (Donaldson 1958, 178). The final major backfilling of the ditch, layer 4538, represents the beginning of a complete re- organizatio mid-17te th Are f nn o i aA h century. Phase (fi2 ) g4 4536, the surface layer of the E-W length of the backfilling of ditch 4551, was composed of tightly packed limestone rubble which must have been used as a yard or path during phase 2. Immediately to the N of this cobbled area lay a mortar-bonded wall, 4529, which was aligned composes 0-7c wa wids E-W8m d wa eroughlf an t do I . y faced limestones which survivea o dt height of two courses for a length of c 2-8 m. Both E and W ends had been destroyed by more recent building activities. A spread of stones, 4532, may well have represented flooring although this did not extend as far N as the main N section because it had been cut away by a more recent feature (fig 4). Flooring to the W, composed of a peat-ash and clay mixture, 4541 and 4547, was overlain by a charcoal spread, 4528, some 0-1 m thick. Although it appeared in the N main section, a similar charcoal spread to the NE of 4528 was not excavated as, within the area investi- HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 9 AN55 D| 1980

4537

'•'-'•I 4528 v;y:ax 4532^,1 •"My«•»•»»••.. ofen J-

AREA A PHASE 2

a*»-^?S!v:--'ft.0-.-'rt

METRES

- - °

AREA A PHASE 3

iiw^^Lyyj

FI G4 Are , plan 3 , phase ad A san s2

gated, all evidence of this layer had been totally removed during the construction of a later wall, 4503. Numerous pieces of smithy waste consisting of iron hammer flakes mixed with small pieces of coal, san clad dyan were recovered fro charcoae mth l layer, 4528 mor.A e detailed repore th n to metal-workin gSpearma M debri R y sb n wil storee archivee b l th n di . Pottery sherds from floor f considerabl o 454d 454e an 71ar e interes s warea t s from Holland, Germany, China, England and possibly Scotland were recovered. These date to the late 17th or 18th centuries. Another | SOCIET 0 ANTIQUARIE56 F YO SCOTLANDF SO , 1983 mortar-bonded wall, 4527, was laid directly upon the charcoal layer 4528 abutting wall 4529. It s onl considerabls wa y0-5c wa wid 2m d an e y distorted wherslumped ha t ei d int t 4549opi . It is probably best regarded as an internal dividing wall. No associated flooring levels were found. A laid gravelled surface, 4537/4540, to the E of the building may have been a contemporary yard. Domesti industriad can l finds would suggest tha buildine tth beed g ha n cottaga r smitheo y an s constructiodit beginnine th t na f phasgo encroachmen n appearea 2 e b o t s t towarde th s latcastle th e n e17ti h century .t surprisin Thino knows s si i t i s g na intentl thaal purposed o t san s the castle had gone out of use by c 1700. Occupation appears to have continued till the last quarter 18te th hf o centur shera s ywhita f do e salt-glazed stoneware, possibl f Scottisyo h manu- facture, was recovered from flooring level 4547. Phase 3 (fig 4) A laye f daro r k brown clay soil, 4522/4531, some 0-1 thick5m , separated phas featuree3 s from mos f thoso t e belongin previoue th o gt s phase. Phas consistee3 f parbuildina do f o t f go which three main walls, 4503, 451 4526d 2an , were discovered. Although muc 452f ho beed 6ha n otherrobbee th t sdou survive heighto dt 0- c wit 8f m so h width betweef so e nTh 0-d . 0-86m an roughly faced limestones were mortar-bonded outwity t locatela no t e i wals S th s hda e lwa Th . excavated area narroA . w entrancwallE e ,th somn ei e 0-7 wide2m , provided access fro mcoba - bled yard, 4525/4535. Within the structure the soil layer 4522 was compacted, possibly because of construction activities, although more likely because of use as an earth floor. A thin wall, 4517, some 0-4 m wide, with traces of associated timbers, divided the building into N and S rooms. Immediately partitioe th surfacp f slightee o to th S e nf eo th e wal dth y wallla lo t , 4529, fromt i f phasi s a e2 beed ha n usesuppora s da staircasa r tfo thin ei s later phase. Thiindicaty sma e tha t leasa t t some of the phase 2 building was standing when the phase 3 house was built. The remainder of the rooS s covere e mfloowa th n i dr with neatly laid squared flagstones, 4514 VictoriaA . n half- penny, dated 1861 recoveres wa , d fro beddine mth g laye slae r 452th b r floo1fo anotherd an r , dated 1862 beddine founs th wa ,n i d g layer 4510 unde more th r e poorly laid slab floor, 4518, roomN e inth . No dating evidence for the construction of the building was recovered from the foundation trenche pottert e manth sbu f yo y finds fro initiae mth l earth floor, 4522 earln a ,f yappeao e b o rt 19th-century date. Obviousl slae yth b floors coul t havdno e been laid e 1860stil th e lonl Th .y domestia f o fint dno c almosn natura s etwa circular stone slab, some diameter0-4n i 5m , incor- porate floon di r 4518. Thihavy sma e bee cooper'na s barrel-forme floorina s a r e beforgus s it e slab. The cottage is known from a print of 1867 and is shown on the Ordnance Survey map of 1891. onle th ys trencArewa aA h where features coul phasee exteny db an o dtt although phases 1A and IB in Areas B and G do appear to be roughly contemporary with phase 1 in Area A. The main value of work in Area A was the late sequence of deposits excavated and finds recovered. These are comparatively rare in towns on the Scottish mainland where Victorian and more recent buildin oftes gha n destroye mucho ds .

) (fig5 G , AREASD 2 AN SB werG Aread e excavatean sB d betwee Septembe4 1 Juln2 d yan , r m 1979 5 y .b Are9 , aB fiela n di t castlknowe Castle th ou wa th lint f n ei o s snse a e E witGardenN he shalloa th o st w depression visible within the guardianship area to the S. This trench was completely excavated by hand, but subsequently a mechanical excavator was used to remove overburden to the N and W HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AND 1980 | 561

AREAG & B S

3033

30O7

Guardianship wall

FIG 5 Areas B and G plan

depta 0-. c Boto t 4f m h o h these extensions were labelle . NaturaldG , Laxfirth Limestones wa , reached at depths varying between 0-6 and 1 m in areas undisturbed by archaeological features. Several natural V-shaped drainage channel e surface bedrockth th t f o escu . Thes larged efe r channels some 0-1 wid5m 0- y maie deepeb 1 m th nn I .thes e were aligned NW-SE running down towards the East Voe. Patches of iron panning up to 20 mm thick were present on top of natural in some places emphasizin impervioue gth s nature bedrockth f eo . Natura overlais wa la y nb layer of fine light clayey soil, 523/3018, which was necked with charcoal and contained pieces of butchered bon quartd ean z fragments somn I . e undisturbed areas thi n i 0-so t 2deposim p u s twa depth. A few pieces of clay-pipe stem fragments from 523 are of a 17th-century date but are too provido t w fe morea o smalto ed precisan l e dating. Phase1A Apart from layer 523/3018 through which 17th-centur lated yan r features were cute th , 56 | 2SOCIET Y OF ANTIQUARIES OF SCOTLAND, 1983 earliest evidenc f man'eo s activitie featurs swa e 527, largpara f o te L-shaped depressione Th . S, E and W limits of this extensive cut hollow were not located as they lay outwith the area excavated but it appeared that the portion within Areas B and G was the N limit of a levelling operation of some considerable size. The feature cut 523/3018 and the natural limestone to a depth of 0-4 m at the E end but only 0-1 m in the NW corner of the trench. As bedrock sloped NW-SE t coul flae cu th te d basth hav f eo e provide drelativela y soli leved dan l surfac constructior efo n purposes, although it was over-cut for this purpose by some 0-15 m at the SE corner. Above a primary fill of eroded material from the exposed face of 523/3018 were various layers of builder's waste, 528, 521 and 526. These consisted of limestone rubble and chippings together with loosely packed mortar. Hollow 527 is most likely to have been the N part of a large area cut down to and into bedroc thao ks buildina t buildingr go s coul constructee db souna n do d6 level52 d . an Fill1 s52 were probably spread f wasto s e derived fro constructioe mth , f wallinS no e th go t som y ewa muca h larger area having bee t thafinalls ncu nwa y built upon sucy hAn . building woule dli outsid aree eth a excavated, withi aree nth a under guardianship findso N . , othe rpiece w thafe nsa of animal bone, were recovered from layer 528, but a few sherds of 17th-century Northern European earthenware were found in layers 521 and 526. An L-shaped linear feature, 3034, towards the W end of the excavated area cut the W end of the limestone rubble, 526, and had a width of c 1 m and a maximum depth of 0-2 m. The fill feature depte f soistoneth d i th varie f an s beelhd d o a e sha an n d t aloncu lengte e gth th f ho used indiscriminatel backfillings ya stones w feature fe ill-defines Th enda . ,W t ewa e ,bu th t da 3007, may have represented the line of the feature at the W baulk where the fill could not be otherwise separated fro surroundine mth g soil. There were levelo n s associated wit L-shape hth e which could be considered as being flooring or associated make-up. Neither was there any evi- dence tha t removo t 303 beed g 4ha n du e walling. Therefore t wouli , d appear that, althougha foundation trench had been cut, walling had not been constructed within it and that the feature had been quickly backfilled as there was no evidence of erosion of its sides. No finds were re- covered from the feature fill. A few possible pestholes and a shallow slot, c 1-25 m in length, 3034f o t thes N bu , e coulth la o yt d have been accidentally created durin cuttine gth hollof go w 527 .stratificatio o Thern s ewa thin i trencse areth f enablaho o t e further interpretatio madee b no t . seriea y threcornef la sE o N G e Areae f d parallero th an n sI B l trenches, 531/3012, 532/3013 524/3033d an endS whicf e so th , h appeare roughlo dt y respec presencedge N tth largf e eo th ef eo L-shape t 527dcu , althoug hcleao thern s r stratigraphicaewa l relationship between the 527d man . The trenches were aligned roughly N-S and the only one completely excavated, 531/3012, was widt soilength2n m i 4- cu y l e b 3l m laye Al . r 523/301 somn i d e 8surfacp an place to e f seo th natural. Their bases were comparatively flat. Althoug survivine hth g depths were 0-2ec th 2m features had obviously been truncated by later cultivation activities because a spread of limestone fragment firss wa ts identified some 0- highe2m r althoug featuree hth sisolate e coulb t dt no da that level. The surviving fills consisted of decayed and powdered limestone. The few pottery finds appea indicato rt e tha trenchee th t s were contemporary wit othee hth r phase 1A activities. Although no organic material was preserved because of soil conditions, it is most likely tha featuree th t s wer esuccessioa f latrinno e trenches. These could have been use workery db s during construction of the castle and associated outbuildings in the early 1600s or, perhaps more likely, by Cromwellian troops in the 1650s. When backfilled, they would have been ideal agricultura t bedsho l . Similar layout parallef so l flat-bottomed trenches have been found during excavations in Elgin (Lindsay, forthcoming), Lanark (McGavin & Wordsworth, forthcoming) and Forfar (Spearman 1982). HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AND 1980 | 563 Phase IB Evidenc phasf eo e IB activity became visibl t mucea same hth e level thas s a f phas o t A e1 and consisted of a few features in the SW corner of Areas B and G. These either cut or overlay phase 1A features. In nearly every case so little of the features lay within the excavation that interpretatio difficults ni . Parstona f to e wall, 3002cornee withiW y th S la , f e angle ro nth th f eo site probable . Onlth f faco 0- d yN 1-a lengt 8facm 2e an E m th e f hwero e fount seemei t dbu d from the single surviving course that the roughly faced limestones were mortar-bonded and pinned with smaller angular pieces of the same kind of stone and that the core was made up of unbonded limestone rubble. A narrow strip of the associated foundation cut, 3010, existed along the N side of the wall but no finds were recovered from the earthy fill or from the make-up of the wall itself. Although only 0-8 m of the presumed E side was visible it did appear to have bee ntrua e built fac mortar-bondes ea d trimmed pinningd an s wer S fule t withiN- ese lTh . nit widthW E- s wer t determineeno t thesdbu e could have bee mucs na f simila i 1-s ha 6 m r widths of mortar-bonded facing stones, as those on the external faces, existed on the internal sides of the walls. From the small length visible it appears that the method of construction is unlike that used in the castle where the walls are mortar-bonded throughout. The technique is more reminiscen thaf o t t use Munesn di s Castle, Unst, wher walle eth s have core claf rubbled so yan . pestholeA , 3035, immediatel walf o lN 3002e t foundation-cuth cu ,o yt 0-2c s 4m t wa 301 d 0an wide by c 0-2 m deep. It could have been created during scaffolding activities related to the buildinfindo n wal e t sth werlbu f go e recovered fro sande mth y fill. Wha shallosida N tf appeareee o th w e b trench o dt , 3009d t trencen cu ,E e h th 303 d 4an of the foundation trench, 3010, associated with wall 3002. Trench 3009 continued the NW-SE alignment of wall 3002 and its fill consisted of a dark grey clayey loam which contained a con- siderable numbe f angulao r r schist limestonesd san fule lTh . flat-bottomee widtth f ho d feature wat determinesno excavatesids N onls a e wa yd th dan saide b littl dn otheeca r than- thaap t ti peared to be a robber trench cut to remove a wall which had once abutted against the NE corner of wall 3002 t coul.I d possibly have bee basae nth l remnant foundatioa f so n firstrenche th t t bu , explanatio probabls ni more yth e feature likel th greate0-a s dee s 2ym a d e wa pan r frequency of larger stones would have been expected in the fill of a foundation trench. absence Th mortaf eo r fragment greay an t n i squantity coul takee db indicato nt e that this lengt t beef wallinho no nd constructegha same th en di manner t leasa , t thia t s depth wals a , l 3002. The presence of a single clay-pipe fragment of the early 19th century is the only indication of whe postulatee nth d robbing occurred. Other features Features belonging to phases 1A and IB were overlain by topsoil to a depth of 0-6 m in some areas. Loam even forme duppee parth f o rt filhollof o l w 527, whernumberes wa t ei d 519. Although this was not a sealed horizon it was possible to differentiate between it and higher similar deposits becaus containet ei d fragment marinf so e shel charcoald lan welnumbes a .A s la r of 17th-century pottery sherds and clay-pipe fragments, a sherd of residual coarse hand-made pottery and the ends of two stone ard points were recovered from fill 519. The presence of the pointd indicatar y sma e that this typ f implementeo relativela , y commo Shetlane n th fin n di d Islands havy ema , continue til e t leas a l17t e us th tn dhi century. However, they coul residuae db l as was the pottery sherd already mentioned. Outwith the E half of hollow 527 features of a 19th- century date were isolate t mucda same hth e dept thoss ha IBd d phasef eo an an . Pit7 A 1 ss51 3006 are good examples of these later features. Pit 517 not only contained ceramic finds of the 19th century but also material earlier in date, including a Dutch clay-pipe bowl of the mid-17th century. | SOCIET 4 56 ANTIQUARIEF YO SCOTLANDF SO , 1983 Although little dating evidenc s recovereewa d fro e excavatiomth f phasno e d I1BAan features in Areas B and G, it appears that both phases date to the 17th century as do the phase 1 Cromwelliafeaturef o likele e activitiee b ar Areo yn Th st i G . Areaad d A Are n sni an an saB A date although it could be argued that they were associated with the construction and early occu- patio f thno e castle. Probably ther littls ewa e time betwee initiae nth l activitie f phaso s e d 1Aan wore th k undertaken during phas . FroeG shor e IAreamd Bn i th an t sB lengt f wallinho - gex cavated it seems that at least one outbuilding was aligned differently from the castle itself and that a different method of construction was employed. The functions of outbuildings compared with that of a strong retreat may well explain these dissimilarities. ARE A(fic ) g2 Are , excavateaC d durin 197e gth 9middl e Castl e summeth th n f i e o et ou r t seasonse s wa , Gardens between two concrete floors for World War II Nissen huts. The trench was aligned roughly N-S and was 11-4 m long by 2 m wide. All work was carried out by hand. Natural lime- e overlyinth ston l s reacheal ewa d depta g t an layerda f 0- ho 9m well-cultivatea , d soils wa , removed in spits. The only feature located within this soil layer was an E-W linear spread of angular poolimestonew fe r a qualit d brickan sd yre depta t sa f 0-4 ho belo6m existine wth g ground level. The feature was c 0-64 m wide and lay across the trench c 0-9 m from the S end. lengtfeaturOnle m th 2 seen s f ya ho e t wa appeare I . footinge b wala o d t r lsfo probabl y associated with gardening. Find 18th-centurn a f so y date were recovered from spit soif so l below this feature so the wall could be of any date after this. Clay-pipe finds from the various spits of soil indicate that there was stratification according to depth although this was not so obvious with regard to the pottery finds.

(fiF ) gAREA2 D AN SD werF Aread e an excavate sD d durin summee gth 1979f ro trenche e Th . s e werth eo t lait dou E of Castle Gardens on a flat low lying area where there was a break in slope towards the East Voe of Scalloway, some 7 m to the E of the garden wall. The position of the trenches was also determine larga y db e sub-rectangular cropmar aerian see, a m n knlo 6 photoanomalyc y b 0 -2 c , graph. Area D measured c 12 by 5 m and Area F, 1 m to the S, measured c 11 by 5 m. Both areas were excavate hany db d dow naturao nt l betwee levee e th th lnn n are o o 0- d 42m m aan brea slopef ko . earliese Th t features wer eseriea f naturao s l drainage channels simila thoso t r e observed cuttin gt thes naturabu eG wer Aread n i l ean morsB e substantial wit largese hth t being 1-7m long by 1-4 m wide. It was up to 0-2 m deep and had a fill of brown silty clay and large pieces of fractured limestone up to 0-36 m in length. The bottoms of these channels were uneven, but in general they slope towardE d e dow th Ease o sth nt t Voe findo .N s were recovered from these natural features layeA . browf ro n soil between 0- d 0-overla4m 2an y natural. Within this general deposit was a large spread of water-rounded limestones which stretched from the W side of the trench some 8-9 m and extended beyond the N and S limits of the trenches. Some of the stones were up to 0-25 m in length. A number of possible pestholes and slots were excavated and found to be negative. It appeared that stones, together with turves and earth, beed ha n timdepositee on e wite th ht d a area s occupie softee th y rd b material takinsurface th n go e appearanc archaeologicaf eo l features. This dumpin resulted gha levea n di l surface being created for some 6 m from the garden wall before it fell away towards the East Voe. No finds were re- covered from the soil below the layer of stones; finds from gaps between the stones include som 19th-centura f eo y date indicating tha levelline th t thif o ss g i date earlierr o knows i t I . n that HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AN5 D 19856 | 0 a bowling green was built in the 1850s and the spread of stones is probably the foundations for thif o s sitd e stile featureth een ar l f E knowo e .N th Fielde n th locall Bule o st r Booth o ls ya l Green Park. The only other feature discovered was an earthen bank and spread of stone chippings running N-S at the W end of the site. This must have been connected with the construction or repair of the E garden wall. ARE (fiAE ) g2 Are , excavateaE d durin summee gth f 1979 t acros o rpositione s cu wa ,d sen thaE o ds s it t acrosridgy S thla eN- d e moderslina a whic en f e o almos n W hra ns whol e it pat tth d hean length of the Castle Gardens. This trench measured 7 by 3 m and was hand dug. Turf and brown top- soil, 0- deep4m , covere onle featureo th y tw aree d beneatdy th aan la s ha thi s s wa layer e On . modern pipe trench which cut diagonally across the W end of the trench and the other was a stone revetment which formed the E edge of the path.

CONCLUSIONS t Scallowat possiblA no s dato wa et t yi e 17th-century features wit degrey han precisiof eo n so it is impossible to say whether features located related to the construction of the castle in the early 1600s Bishoo t , occupatioe buildinge th pth f o Law't o r e o ,Cromwelliay sus nb n troops in the 1650s, although the general impression given by the finds is that many of the features are likel mid-17te dato yth t o et h century. More conclusive results coul obtainee db d fro excavae mth - tio f areano s nearecastlee th o t r. Withi guardianshie nth p area evidenc f outbuildingeo s could castle th f efoune o b betweeE dN buildin e thesf nth o t leas e A e . on tg G itsel Aread d an an f sB building differena n o s si t alignmen maie th no t structure origine Th .functiod ditce san th hf no found in Area A could be resolved more satisfactorily if more of its length to the N and its re- lationshi castle th eo pt building were investigated. The stratified sequence of pottery and clay-pipe finds recovered from Area A is important becaus t coverei s periomuce th f ho d from 160particulaf o s 1900o i t d 0an r interes t relatei s a t s lato t e deposit ofteo ss n missin urbagn i nmainlane siteth n so Scotlandf do .

APPENDI POTTERXE 1:TH Y (figs 6-8) W J LINDSAY Apart fro pre-medievao mtw l sherd locaf so l hand-made coarse pottery, ric steatiten hi smala d lan , rimsherd, possibly from a medieval greyware globular pitcher from the Low Countries, the Scalloway Castle ceramic materia madpost-medievaf s i lo p eu l wares. These includ Countriesw eLo , Germad nan British wares together with Chinese porcelain and a single sherd of Persian origin. It is noticeable that few Scottish or English products datable to before the 18th century are present. There are only a few sherds from vessels comparable with those discussed in George Haggarty's report on Scottish post-medieval pottery excavated at Stirling Castle (1980). Material belonging to recognized Scottish types consists of seven small sherds probably from small open bowls possessing either one or two small handles in a sandy micaceous fabric with green glaze internall externallyd yan , together witsherdo htw s froe mth shoulder jugf hara so n si d grey reduced smooth fabric with incised wavy decoratio greed nan n glazn eo external surfaces. The small bowl sherds are reminiscent of a 15th- and 16th-century vessel type recovered authoe byth r during excavation Elginn si 17the earlTh d . -an y 18th-century sherds appea confiro t r m available th e documentation showing that Scalloway Shetland an , wholea s stronda d ha , g trade links with Northern Europe at that time (Gifford 1879,25; Smith 1978). As the assemblage contains the largest number of Northern European earthenware imports so far recovered in Scotland, these are described in some detail. A total of 975 sherds are of Northern European manufacture with many, both earthen- stonewaresw fe warea d san , probably product Weser/Werre th f so a valleys unfortunats i t .I e that, apart 37 I SOCIET 6 56 ANTIQUARIEF YO SCOTLANDF SO , 1983 FTP) r V 14 / * 8 4 r

) 10 \ 1 y 7 12

61 r 13 1

^ 15 " S/16 •'C'l^s -(>^I^9

\ -27-

28 FIG 6 Northern European earthenware pipkins (scale 1:4) HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AND 1980 | 567 fro materiae mth l recovered from Areremaindee , mosth aA f o tobtaines wa r d from unsatisfactory archaeological horizons.

NORTHERN EUROPEAN EARTHENWARES Tripod pipkins (fig 6, nos 1-28) mose Th t common typ potterf eo y vessel presen assemblage th n i t fora s eearthenwarf i m o e tripod pipkin, wit hglobulaa r body roundea , d base, thre tubulaea legd san r handle. Although three fragments from pinched rod handles of a type often associated with Dutch vessels were recovered, 45 fragments are parts of tubular handles, a form as yet unknown in Holland but common in some areas of Germany Stephan G Verhaegh F ( H r D , perd ean s comm). However, this general typ vessef emanufactureo s wa l d throughout Northern Europe and in England from the 16th century onwards. It is not a known Scottish type. Examples comparable with the Scalloway material have been found during excavations at Kirkwall (MacAskill 1982), Pittenwee MartinmC ( , pers comm), Skail GellingP ( l , pers comm), fro wrece mth k Kennemerland t Skerriea MartinP s( , pers commt Kelsa d TabrahamC o( )an , pers comm). Stray finds have also been recovered at Muness Castle (D Arthur, pers comm), Fair Isle (C Martin, pers comm) Cairsto Unsd Talbotnan E ( t , pers comm Aberdoud )an r Castl Leitd eHaggartyan G h( , pers comm). It is somewhat surprising that so little of this distinctive type of ware has been found in Scotland but it significancf mao e yb e that mos bees ha tn recovered from coastal areas noteworths i t I . y that Scalloway, Kirkwal Pittenweed an l m hav t beeeno n subjecte massivo dt e Victoria r morno e recent development. It may only be in towns relatively undisturbed since the early 17th century that archaeological deposits containing this ware surviv extenty an o et . Therefore falsa , e impressio distributioe th f no thif no s import may exis Scotlanr tfo t presentda likels i t .I y that tripod pipkin thif so s type were exported widely through- out Northern Europe as material similar to that recovered at Scalloway has also been found during excavations in Trondheim and Bergen (I Reid, pers comm). Considerable time and effort was spent in attempting to divide the 868 pipkin sherds into fabric types subsequentld , initiallan e ey y y b usin y yb gmicroscopa e with magnification Thi. 30 sx prove o t p su d to be a fruitless exercise. Even fragments from the two distinctive handle types could not be divided on a fabric basis. All that can be said is that under oxidizing conditions the fabric is orange red in colour, tha containt i t largsa e proportio sanf ntharico s d i subanguladn t han i ti r whit translucend ean t quartz grits which can be up to 0-75 mm long, although they average c 0-5 mm. A small amount of mica and occasional rounded ferruginous inclusions are also present. Under reducing conditions the fabric is grey. Normally pipkin completele sar y glazed internally with some external coverag rim n necksed o san . The colour of glaze varies considerably from orange and ochre through various shades of muddy green almosbrowo o t t d nan t black. Dark brown specklin noticeabla s gi e characteristi somn co e ochrd ean brown examples. A few sherds are white slipped internally and covered with a thin, but fine, rich mottled green glaze. welThiy lsma have bee attempn na produco t emora e pleasing gree usiny nb g better colour- mora n o e t sympathetian c ground singlA . e sherd from Pittenwee larga d eman sectio pipkia f no n recovered from the Dutch ship Kennemerland, which sank off the Skerries in 1664, also possess this internal decorative treatment secone .Th d exampl houses ei Lerwicdn i k Museum. e ScallowaTh y pipkin e thiar sn walled, well mad neatld ean y finished. Different form f rimso s , legs and handles are present but, unfortunately, these parts seldom joined with body sherds to provide large profiles. The variety of glazing and of rim forms together with the several variations of handles and the two distinct leg types may well indicate that more than one production centre is involved although, alreads a y stated, pottermane th f yo y finds wer t weleno l stratifie lengtha d dan y time span could equally explain many of these differences. Artistic licence or boredom could account for minor differences of relief decoration, especially on handles. The rims The 150 rimsherds represent at least 111 vessels. In cases where joins occurred or where it was clear that sherds were fro msingla e vessel they were counte vessele diameteron m s dRi .a s vary considerably, , evenmm wher0 16 betweeesingld a foran m 0 eri minvolveds i n10 . Five mai typem nri s were identified; uprighn a t for mr threo wito e htw groove externae th n so l surface; three everted varieties whicf o e hon , has lid seating; and lastly an inverted form. Rim form 1 (fig 6, nos 1, 2). This upright form with two or three grooves on the external face of the rim is the most common type present in the Scalloway assemblage with 63 sherds representing at 568 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1983 leas vessels0 5 t . Man thesf yo e sherds were recovered from unstratified deposits significans i t i t bu , t that none was recovered from the various ditch fills in Area A indicating that at Scalloway the type may not have been in common use until the 18th century. Stratigraphically, the earliest sherds were obtained from the Area A, phase 2 gravel yard surface, 4537/4540. Although rim diameters vary between 100 and 160 mm greatea r proportion nea e tenmaximub e th ro d t m size. Shoulder f pipkinso s possessing thiform sri m have finely executed external rilling which extends down vessels beyond handle junctions. Twenty-eight of the sherds have brown or almost black glaze internally while the remainder possess glaze of an ochre colour. This rim form with brown glaze is present in the material from Pittenweem. One atypical small sher decorates di d internally wit hseriea f paralleo s l white slipped lines whil othero etw e treate ar s d in the same manner externally. Decorated areas on these three examples are covered with clear glaze. Rim form 2 (fig 6, nos 3-6). The second grouping consists of 44 sherds which represent at least 30 forsimplm a vessels ri ms i e e Th .everte d type with internal 'lid seating'. Twenty-eight sherds havn ea additional external groove which may only be a potter's whim. Internal surfaces are covered with light brown glaze speckled with dark brown. As there is no evidence of wear on the glazed 'lid seatings' and as no sherds from lids were recovered, it would appear that the 'seating' facility is a vestigial and non- functional feature. It appears that at Scalloway this form may be somewhat earlier in date than form 1 as a sherd was recovered from the Area A, phase 1 ditch fill, 4538, which was probably deposited in the mid-17th century although the layer also contained some redeposited material from earlier in the century. souno Thern s de wa archaeologica l evidenc wheo t fore s ena th m cease usede b o d.t Rim form 3 (fig 6, nos 7-10). The third form is slightly everted and the 30 sherds represent at least 19 vessels. Sixteen examples have a single external groove while two others have double grooves. Eleven examples have single cordons at the bases of short necks and two have double cordons. Although other vessels could have possessed the latter feature, the sherds were often too small to include it. Glaze colour varied considerably on internal surfaces being ochre, muddy green, brown and black. Form 3 appears same th eighhavs o t a t e a , time5 te beeformsherds d us ea an n n i 2 ss fro msingla were) vesse10 eo (n l recovered from ditch fill 4538. Although this layer seem havo st e been deposite mid-17te th n di h century it did contain a few clay-pipe finds datable to 1620-40. Rim form 4 (fig 6, no 11). The four sherds in this small group, representing three vessels, have slightlya everte form dmri wit hpronouncea d upper external cordoridga neck/bode d th e an t na y junction. Glaz eithes ei r brow muddr no y green. sherdNone th founs f eo s wa seale n di d archaeological deposits. , 13)five 12 form eTh .m s sherd(fiRi 5 no , g 6 thi n si s grouping posses inverten sa formm l dri Al . have an upper external groove and a lower single or double cordon. Internal glazing is either ochre, gree brownr no fore mpresenTh .s i t among material recovered from ditch fills 453 454d 8an wels 3a s a l from higher levels indicating that it was in use at Scalloway from the early or mid-17th century until at leas beginnine tth 18te th hf g centuryo . This for alsms i o presen Pittenweee th n i t m assemblage. Body sherds Body sherds were counte studied dan d along with basa maine l th sherd n ,i base, sas roundee sar d and merely thicker continuations of the sides. However, six sherds from flat-bottomed vessels in the same orange red fabric were also recovered so some rounded body sherds from these vessels may be errone- ously include totale th n d.i Upper body sherds were easily separate thes da commonle yar y finely rilled externall considerable ar d yan y thinner than those fro lowee mth r portion f vesselso s 5 tota60 A . f o l body sherds was studied and 150 of these show some evidence of reduction, generally on the thicker lower areas of vessels. Reduction is seldom complete and may have been caused, at least in part, when vessels were used over fires. Basal areas are often smoke-blackened externally. Internal surfaces are completely glazed wit same hth e variation colouf so r presen mentiones a t d with regarrimsherdse th o dt . However, many have glaz dirta f eo y green colour possibly produce beiny db g continually subjecte heao dt t from cooking fires. The handles (fig 6, nos 14-19) Forty-eight fragment f pipkio s n handlef throw f theso o e presene 5 ear 4 snar d tubulaan t r con- struction t leas A .slightl8 2 t y conical tubular handle representee sar assemblagee th n di mose Th . t com- mon form has a decorative terminal composed of various grooves and ridges which are similar in character externae th o t l decorative treatmen form mri join e n 1o t.Th s between handle bodied san s appea havo rt e been relatively weak as many breakages occurred at interfaces. It is apparent that handles had been origin- ally well-thumbed on to rilled body surfaces. The longest handle appears to have been c 110 mm, but HALLINDSAYD LAN : EXCAVATION SCALLOWAT SA Y CASTLE9 , 56 197198D | 90AN

many others seem to have been shorter which would not have made for comfortable or easy handling. Nine fragments show trace browf so r blacsinglthesf a no o s e keha e on glazspo d whitf o etan e slips i t I . unlikely that either featur s intentionali e . This for f tubulamo r handle wit e decoratehth d terminas i l present in material excavated at Trondheim (I Reid, pers comm). Two other variations of hollow handle desig presen e Scallowae nar th n i t y assemblag singls ea e examples. Bot e morhar e cylindrica forn i l m flangea s ha de terminalon d an simplee .Th r for presenms i Pittenweee th n i t m material. Three handle fragments of a form often associated with Dutch pipkin material are also present. All possess the familiar pinching towards the tops of handles but, as has been stated above, it was im- possibl separato et e these fro tubulae mth r version fabria n so c basis fore alsmTh s .i o know Germann i y (Dr H G Stephan, pers comm). No 14 is unusual with regard to size and is unlikely to have been a cooking vessel althoug smoke h ar part t i ef s o blackened mors i t .I e likel havo yt e bee nmuga .

The legs (fig 6, nos 20-24) Fifty-six fragments of pipkin legs were recovered and it was possible to estimate that, assuming each vessel possessed three legs, these represent at least 27 vessels when additional factors of size, form and colour of glaze on the interiors of pipkins were also considered. Legs are well fixed to vessels, with few clean breaks at body/leg junctions unlike handles which show more evidence of breakage at points of application. Obviously the latter were subjected to more daily stress then the former. Two distinct leg types were identified but it proved impossible to differentiate between them on a fabric basis. It is be- writee lieveth y rdb that vessel f threo s e specific volume representee fragmentg sar le e th s sda readily divided into three groupings accordin unfortunats sizeo gi t t I . e that more complete profile f vesselo s s reconstructee coulb t dno tesvolumo de t tth e theory. Leg type 1. This is a simple pulled conical form, the end of which has been trimmed and angled e horizontalth o t ° 45 . t c surfacThirty-seve that o cu s a t e standi tg eth le n wit so e nth h fragments representing at least 18 vessels are present and it appears that there are three large, seven medium and eight small vessels. The number of small vessels could be considerably larger as the 20 fragments are very small. Nearly all the sherds showed some evidence of having been smoke blackened. Leg type 1 is present in Area A ditch fills, 4538 and 4543, both as abraded and unabraded sherds indicating that although the form was in use in the mid-17th century it must have had earlier origins. It is likely that leg type 1 fragments belong to vessels possessing rims of form 2, 3 or 5 as examples of these were also recovered from the ditch fills. Leg type . Thi2 alss si o basically conica s bee shapn ha i l n t benebu t inward restd en thao ss s it t completely on a horizontal surface. As a result, the weight of a vessel and its contents is directly borne throug lengtlegse e hth th .f ho Ninetee n fragments, representin t leasga t nine vessel presente sar . Only two examples are relatively complete but the writer is certain that the other pieces are indeed from legs as oppose solio dt d handles becaus alignmente th f eo f throwino s g impression insidee th n f vesselso so . As with typ possibls legse1 wa t i , divido et fragmente eth s accordin sizego t thin ,i s case intgroupingo otw s which appear to be equivalents of the large and medium type 1 examples. At least two large and seven medium-sized vessel e involvedar s vere yOn . abraded fragmen s trace whita ha t f o s e slipped coating. materia2 g le obtaines None lwa th f eo d from well-sealed archaeological levels.

Decorated pipkins (fig 6, nos 25-28) addition I e materiath o nt l already discussed ther smala s ei l grou f ninpo e slip-decorated rim- sherds representing five vessels whic alsy pipkinse ohb ma l havAl . e slightly n evertei m m d 0 rim18 sc diameter which appear to be from globular vessels. The fabric is similar to the general pipkin material. Three rimsherds are decorated with bands of white slipped horizontal lines, on one example internally and on the others externally. One has additional slip-painted geometric designs on the shoulder of the vessel. Clear glaze renders slipped areas cream and body surfaces light brown. Two examples are glazed botn o h interna externad an l l surface wite s whilon h e onleth y internal decoratio glazes ni than do t face alone generae Th . l decorative treatmen identicas i t thao t l t employe three th en do smal l sherdm ri f so thao t td forusean sommn 1 do e earthenware dishes. Anothe rime th thin i sf so r categor whits yha e external slip covered with clear glaze finae .Th l vessel, represente singla y db e sherd similas i , forn ri d man coveres fabrii t cbu d wit externan ha l milky wash. Glaze render washee sth d area yellow-greee th d nan internal surface brown. Trace a brow f o s n overglaze design survivoutee th rn eo surface e abovTh . e examples three th , e decorated sherdform ri m f abradeso n 1 a , d slip coated a fragmen d typg an le e2 f o t 570 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1983

29

30 44'

IP 31 45 If 32

\ 331 46 34

47 35 36

37 J

38

39

40

50

51 7

52l

FIG 7 Northern European earthenware vessels (scale 1: 4) HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AND 1980 | 571 tiny flake from a slipped tubular handle may be parts of chamber pots. None of the illustrated rims was found in well-sealed archaeological deposits. The colander otherd an similar vessels (fi , nog7 s 29-32) Four rimsherds from bowl colandersr so fabria n i , c identica pipkie thath o t lf no t material, were recovered colandepara s i f e o t On . . Botr withmm hdiametea 0 interna 34 externac d f an lo r l surfaces are covered with transparent creamy orange glaze which has occasional brown speckles, probably drawn from iron in the vessel fabric. The external rim surface of the colander shows evidence of having been smoothed, possibly the result of some type of horizontal handle having been applied further round the vessel. There are parts of two holes which taper from 8 to 6 mm on the internal surface. These were made applicatioe prioth o rt f glazen o thre e Th . e other sherd thin si s grouping have more upright rimse On . sher bees d ha least na t partially slip-coated internall appeard yan s cream throug show 0 glazee 3 h th o s.N evidence of a horizontal handle having broken way from the rimsherd. The colander and the decorated rimsherds were foun ditcn di h fill 4538 indicatin gmid-17ta hrecoveres centurwa 0 3 yo d n datfro t embu layer 4522 which contained late-18th century material. Slip-decorated small vessels 33-36s (fino , g7 ) Sixty-one tiny sherds from slip-decorated small vessel presene pottere sar th n i t y assemblage. Most of these sherds are so minute that positive identification of vessel types proved to be impossible although small open bowls and globular vessels have been identified. The orange red fabric is very similar to that of the pipkin material although the globular vessel sherds appear to be in a finer fabric with smaller quartz grits. There are 13 sherds probably from small open bowls. The seven rimsherds are all of a simple upright form withavino htw gcordoa 33)o externae (n nTh . surfacem belom ri ri l e wdecoratee th sar d with series of vertical, or almost vertical, white slipped lines about 10 mm in length. Both internal and external surface coveree ar s d with glaze which renders body areas orang r ligheo t brow slipped nan d areas cream smale Th . l handle fragment simple sar e looped types, probably applied horizontallo tw s ya examples have traces of similar decoration to that used on rims along what were presumably upper edges. This typ f smaleo l bow knows i l havo nt e been manufacture Nortn di h Hollan Germand dan y in the late 16th to 18th centuries (Hurst et al 1975). The second grouping of 48 sherds is made up of 38 rim and body sherds together with 10 handle fragments from little globular vessels which coul smale db l pipkins, jar cupr so s (nos 34-36) vessele .Th s are delicately made with thin walling of between 2 and 3 mm. Body sherds are often finely rilled externally in the same manner as some of the pipkin material. External surfaces are much adorned with various types of white slipped decoration. Both internal and external surfaces are clear glazed producing orange- brown body areas and cream decorated areas. Two collared rimsherds possess the same type of decoration thas a t presensmale additionath n a n l s o tope ha ne bowllon seried scloself an so y spaced white slipped dots on the body of the vessel (no 34). This example, together with eight others, has some dots highlighted in green. There is a tiny sherd in the Pittenweem material with the same type of decorative treatment. Only one sherd is completely slipped on the external surface and it is overlaid with brown slipped dots. Some sherds, possibly from more open vessels, have trace largef so r white m slippem 3 d15 spoto t p su in diameter while others have part f zigzaso othed gan r type f lineaso r motifs largese Th . t body sherd (no 35) possesses a stub of a horizontally applied small loop handle and has a white chevron design aug- mented with small green glazed dots on the outer body face. The handle fragments are all of rod and strap construction. From the slight evidence it appears that most were applied verticall vesselo yt 36)o (n s , althoug exampled ro h e somth f seo were applied horizontally lattee th f ro typ o coveree ear Tw . d with badly decayed white slip sherd1 6 . e Nonsth f eo in this general categor recoveres ywa d fro msealea d horizon groupine .Th g contains sherds from several types of pottery vessels, each of which could well have had a specific function but the Scalloway material is too fragmentary to merit more detailed categorization. Little can be added other than that this 'family' of materia probabls lwa y mad Hollann ei Germanr do y durin 17te gth h and/or 18th centuries.

Earthenware dishes 37-52s (fino , g7 ) Asherd4 tota19 f so l from earthenware dishe presene pottere sar th n i t y assemblage. There would | SOCIET 2 57 Y OF ANTIQUARIES OF SCOTLAND, 1983 appear to be 23 sherds from plain vessels and 171 sherds from slip decorated dishes. It was seldom possible differentiato t e between fabric mosd an s t sherd vere ar sy similapipkie th o t nr material maine th n ,I . dishe groupee sar d accordin presence th o gt absencr eo decorativf eo e treatment. Group 37-39)s (fiA no , g7 . Grou consistpA undecorate3 2 f so d sherds whicf o rims2 e 1 h,ar t A . least nine dishe representee sar d with either simple beade hammer-header do d rim diameterf so s between 220 and 260 mm. Glazing is similar to that on pipkins with only the internal surface covered. Two sherds have creamy orange glaze with a few brown speckles which is similar to that on the colander. No 39 was recovered from ditch fill 4538 which is of a probable mid-17th century date. Another small sherd was found in ditch fill 4543. Group B (fig 7, nos 40-43). This grouping has been subdivided into two categories, 1 and 2. Both have a white internal slip and B2 has additional superimposed decoration. Group Bl consists of 22 sherds including 11 rims, probably all from different vessels. The rim forms are similar to those of Group A and have diameters between 140 and 280 mm. Some, presumably accidentally applied, white sli exteriore presens pi th n o f trims o s . Transparent glaze renders slipped areas yellow possibls i t I . e that materiae somth f eo l include Groun di l belongp B becaus 2 e B th o st f eo small sizes of the sherds involved. Group 32 representes i sherds0 6 y db , includin rimx gsi s from separat eforms a dishesm ri e e sar .Th described for the Bl material. Most of the sherds in the B2 grouping have spots and dashes of orange to light brown slip applie whitee th o dt slipped ground foud an sr examples show evidenc f featheringeo . Internal surface coveree sar d with transparent glaze, making white slipped areas yellow likGroue eth l pB material. Eleven sherds possess additional superimposed blobs or smears of green glaze. Two rimsherds have had holes bored through them subsequent to the application and firing of glaze. This was presum- ably don tha disheo e s th t s coul decorativr hune fo db p gu e purposes holee Th . s taper toward centre sth e of the vessel wall from 5 mm at the surfaces to 3 mm. Obviously, they were bored from both directions to avoid chipping the glaze and the vessel wall. No 43, from layer 4547 in Area A, indicates that this type of dish could have been in use during the late 18th century. Group s 44-48) (fino C , g7 . Grou consistpC sherd4 5 f o ss from dishes which have decoration consistin eigho t p tu concentrif go c white slipped lines fabrie glazd Th . can e isolatee coulb t dno d from that of Groups A and B and the rim forms and diameters are also similar to those described above. Eleven rimsherds l froal , m separate vessels, possess decoration reminiscen three th ef o tdecorate d pipkin sherds of rim form 1. One Group C dish sherd has white slipped vertical strokes on the outside surface of the rim similar to those present on some small bowls and globular vessels (no 44). The internal decora- tion on this example consists of a wavy line bounded by two concentric circles. Another sherd has white slipped dots and streaks of green glaze like that on some Group B2 material. One rimsherd was recovered from ditch fill 4538 showing probabls thai t i t y contemporary with e 17th-centurmucth f ho y pipkin material. No 48 is in an atypical fabric which contains lenses of lighter clay in the main orange-red matrix. Group 49-51)s (fino D fourt , e g7 .Th h groupin madeighgs f i o p etu slip-decoratef do x sherdsi d san these are rims. The orange red fabric is harder than that of much of the pipkin material but has the same inclusions. Rim forms are as already discussed but decoration appears to be slightly more sophisticated. internae Th l surface agaie ar s n white slipped with concentri r wavco y line thesd san e appear off-white, crea r palmo e orange under transparent glaze. Unslipped areas appear brown with some darker brown speckling occurring basae On . l angle sherd from ditch fill 4538 is covered with zoomorphia par f o t c bowa pardesige b f 49)sher e o o ty l (n nTh rathe. dma r tha includes i ndisa t hbu d here becauss it f eo apparent affinity groupinge wit th rese f hth o t . This typ decoratiof eo commonls nwa y use Germann di y and North Holland in the late 16th and in the 17th centuries (Hurst et al 1975). As the Scalloway example is abraded it could be residual in the mid-17th-century deposit. The only other sherd from a stratified deposit is a small fragment from a rim (no 51) which was recovered from layer 4541, the smithy floor. This deposit also produced material of a late 18th century date. Group (noE t illustrated). Grou consistpE f eigho s t sherds from disheorange-pinn a n i s k fabric whic hardes hi fined an rr tha npipkie thath f o nt materia somd an l e other dishes. Concentric slipped lines in an intense white decorate the insides of vessels and one sherd is rouletted. Clear glaze makes body areas orange but it has been rejected from most of the white slip. Where glaze is present, slipped areas appear cream. Again some type of green staining gives additional interest to vessels in the form of con- centric circles and dots which have bled into adjacent areas. None of the group was recovered from mane th f yo waree stratifieon se b produce dy 16te deposit17td ma th t han i hn d i t centuriesbu e th n si Weser/Werra region of Germany. HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AND 1980 | 573 Weser-type ware (fig 7, no 52) Nine tiny sherds including one hammer-headed rim fragment (no 52) in what is generally termed Weser slipware were found fabrie Th . har s ci pald dan e pin coloun ki r wit whithd smalan ed inclusionslre . Upper surface completele sar y covered wit hthia n white slip. Finely executed slip trailed designs consisting of orange and brown concentric circles lie parallel to rims while there are zigzag and spiral patterns toward centree sth vesselsf so . Almost clear glaze makes undecorated areas yello r palwo e green, while green staining provides additional decoratio fore th dot f m zigzagsd o n n i s an shere recoveres .On dwa d from layer 4538, the backfilling of the ditch. One small sherd in a similar pale and hard fabric has orange feathered slip decoration. Weser-type war generalls ei y latdatee 17tth ed hdo 16tt an hcenturies . Although s beeiha t n recovered from various excavation Englandn si writer'e th o t , s knowledg onle eth y other comparable material from Scotlan s recoverewa d d during Urban Archaeology Unit excavationn i s Kirkwall (MacAskill 1982). Discussion In spite of considerable research, none of the Northern European earthenwares found during the excavation Scallowat sa assignee b n y ca specifi do t c production centres wit degrey han certaintyf eo . Even apparentle th y distinctive Weser-type ware coulDutcf o e db h origin earls a , y 17th-century kiln material of this kinrecoveres dwa t Enkhuizeda ZuideVerhaeghee F ( th e n no rZe , pers comm). Even greater doubt exists about the remainder of the earthenwares although the fabric does appear to be too coarse consideree b o t dGermaf welDutcho y e lb ma t n.I manufactur coul d produca eproductan e e db th r o t s of small town villaged Wesee san th Werrd n si an r a valleys. Althoug unlikels i t hi y thamateriae th t l was mad exporr efo coult ti d hav Scallowao t ey foun wa by-produc a s s dit ya tradf to morn ei e important significane b goodsy ma t I .t that during this perio greatee dth r part f Shetland'so s trad withiy ela e nth Hanseatic sphere, with Hambur Bremed gan n know Germas na n ports (Gifford 1879, 25). Bremee th t na mouth of the Weser would have been in an ideal geographical situation to gather this type of material for export if it was indeed manufactured up river. It may not be fortuitous that a few stoneware sherds believe Duingee b o dt n Ware produca , Wesee th f ro t valley, were also recovered durin excavatione gth . The presence of a considerable quantity of 17th-century Dutch clay-pipe fragments may indicate that at leas t earthenwaree somth f eo s were mad Hollann clay-pipe i th t dbu e tradhavy ema e been totally divorced fro importatioe mth potterf no y vessels. GERMAN STONEWARES (not illustrated) Only eight sherds from German stoneware vessels were recovered from stratified deposits. These came from an Area A, phase 2 feature and the soil horizon between phases 2 and 3. Another 21 small sherds of probable German origin were recovered from unsealed levels. earliese Th t stratified material, from layer 4536 Aree th , phas, aA yare2 d surface, consista f so small sherd fro mBellarmina boda d yean sher grea n dyi fabric with dark grey, almost black inclusions which are up to 2 mm in length. The inner surface of the sherd is orange while the external surface is metall- ic brown. Dr G Stephan believes this sherd, together with eight others, to be Duingen ware. At Scalloway it appears to be an 18th-century import. The Bellarmine sherd is part of a stamped neck and includes facpara f eo t mask wit eyen ha , eyebrow, frown furrow lowea d ran s eyelid fabrie Th . gres ci y witn ha orange internal surface while the exterior is covered with grey mottled brown salt glaze. The sherd probably dates to the first half of the 17th century. None of the stonewares merits illustration. It is surprising that so few sherds of German stonewares, compared with sherds of Northern European earthenwares, were recovered. However, thei indication a r e absencb y n ema tha t pottery imports arrive t Scallowada by-produca s ya tradf o t morn ei e important good thaa d t san t no ther s ewa sufficient marke merio t importatioe tth bettef no r quality ware quantityn si .

TIN-GLAZED EARTHENWARE 53-55S S (finO , g)8 Thirty-six sherds fro t leasma tin-glaze 2 2 t d earthenwar r Delftware-typeo e vessels were recovered. Unfortunately, it has not been possible to ascertain whether the material is Dutch, English or Scottish in origin but it does appear to date, on stylistic grounds, to the late 17th and/or the first half of the 18th century. The cream open textured fabric is soft and powdery, containing white subangular quartz grits wit0-m o t 5 m h p occasionau l red-brown inclusion similaa f so r size. Interior exteriord san f vesselso s 574 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1983

56 ~7

FIG 8 Tin-glazed earthenwares (nos 53-5); Chinese porcelain (nos 56, 57); Persian ware (no 58) (scale 1:4)

coveree ar d with whit eundersidee glazeth n O . f plateo s thickle sth y applie slightld dan y blue-tinged glaze is pitted with pin-head sized holes. Glaze flaking occurs on nearly all the examples. Three sherds, probabl chambea f o m y ri rpart pota f so , sho evidenco wn f decoratioeo n other than overall glazing whil remaindee eth e eithear r r decorated with blu r witeo h various combination f bluo s e along with yellow purple,d greenan d .re , Twenty-five sherds, representing nine plates, a small bowl and five large bowls are decorated in dark blue alone or with the addition of light blue (nos 53-55). On plates the simplest form of decoration consists of a blue band or bands parallel to the rims but more complex geometric and scroll and foliate design presene sar othersn o t smale .Th l bow 55)o (n l, probabl bowa te cupr y a ldecoratioo s ,ha n similar in style to the scroll and foliate designs on some plates. Eight sherds, from three plates and three large bowls, possibly punch bowls, possess decoration using additionasome th f eo l colours listed above. Muctin-glazee th f ho d earthenware materia obtaines lwa d from Are , phasaA deposites 2 wa d san found in association with white salt-glazed stonewares of a possible late 18th-century date. However, the smithy floor, incorporating the phase 2 deposits, could have had a long life span so at least some of the tin-glazed earthenwares could be considerably earlier, perhaps even late 17th century in date. Some of the more basi lesd csan decorative Northern European earthenware dishes were probably hun walln go s cherishes a d possessions brightee th o mord s , ran e colourful Delftware types could well have been heldn i even higher esteem.

ORIENTAL WARE Ss 56-58 (fino , g8 ) Chinese porcelain) 57 , 56 s (fino , g8 Twenty-one sherds, probably representing 21 Chinese porcelain vessels, were recovered most of which appear to date to the late 17th and early 18th centuries. All are either decorated internally or extern- ally with underglaze blue, sometimes with red or gold overglazing. Small bowls and saucers are the most common types of vessels present. The two illustrated examples were recovered from Area A, phase 2, smithy flooring levels 454 4547d materiae 1an Th . l probably arrive Scallowan di eithea yvi r Hollanr do Germany as indirect trade.

Persian ware (fig 8, no 58) singlA e rimsherd fro msmala l bow f Persiao l n origi recoveres nwa d durin excavationse gth e Th . fabric is hard and a pale brown colour. Internal and external surfaces are decorated in cobalt blue and black unde transparena r t alkaline glaze exterioe bowe Th .th f l o rdepict animal-possibln sa yhorsea . shere Th d appear late dato th est o 17tet t recovereh no century s wa dt I . fro msealea d archaeological horizon. HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AND 1980 | 575 BRITISH WARES (not illustrated) Staffordshire-type slipwares Six small sherds of Staffordshire-type slipwares are present in the Scalloway pottery assemblage. The fabric is a light buff colour and often contains lenses of red-brown or brown clay together with hard red-brown inclusions up to 0-75 mm. All the sherds possess brown or red-brown slip trailed decoration and one example, part of a press-moulded dish, has been feathered. Pale transparent amber glaze covers surfaces that have been decorated. Five of the sherds are from press-moulded dishes and one of these, fragmentm ri a bees ha ,n impresse producdo t edelicata cruse epi t finish. Dish sherd glazee sar internan do l surfaces only. One small body sherd from an upright vessel, possibly a mug, is in a fabric with less than normae th l inclusions shere slis Th . dpha trailing externall glazes i d botn dyo an h surfaces. This sherd, from Are , layeaA r 4541 smithe th , y flooring, appear 18th-centurn para e b f o to st y vessee l th type t bu , other unstratified dish sherds coul somewhae db t earlie rperhap- s datin late eth 17tgo t h century. Whether the late yar e 17t r 18tho h centur datn yi earliese e th the e yar t English products discovere t Scallowayda . sherdNone th f esufficientlo s i y larg merio et t illustration.

White salt-glazed stonewares Seventeen sherd f 18th-centuro s y white salt-glazed stoneware e presentar s s impossibli t I . o et state whethe Englisrf o the e y ar Scottisr ho h manufactur thef i t y ewerbu e mad Scotlanen i d the likele yar y to date to late in the century. The ware is known to have been produced in large quantities at Prestonpans and there are records of William and Jo Cadell sending this ware to John Urquhart in Kirkwall, Orkney in 1784 (Cruickshank 1976 . Therefor,3) t woulei d seem possibl alses thaowa t beini t g sen Shetlando t t , perhaps via Hay & Company whose premises were at the old harbour in Scalloway. Sherds from cups or small bowls, a possible chamber pot and press-moulded decorated plates were recovered decoratioe .Th lattee th rn n o consist rathef so r austere embossed scrolls, basketword kan other geometric patterns around rims. In addition to the British wares described above a quantity of English and Scottish wares of the 20td 19tan hh centuries were also recovered. These hav t beeeno n discusse thin di s repor moss a t t were recovered from unsatisfactory archaeological horizons.

CONCLUSIONS It must be stressed that although the Scalloway Castle pottery assemblage is relatively small it is , mose th t comprehensive collectio thif no s typ materiaf eo recoverer fa o s l excavatioy db Scotlandn i t I . woul dangeroue db drao st w firm conclusion onls sArea e yth group aA s were recovered from stratified deposits. As little of the 17th-century pottery can be assigned to specific production centres and, as the only other dating evidence consists of clay-pipe fragments, this report is of necessity a descriptive record varioue ofth s wares wit hope hth e that future wor Scotlankn abli e b provideo y t dma e more accurate dating. Within Area A, Northern European earthenware pipkins with rim forms 2, 3 and 5, leg type 1 tubulad an r handles wer common ei durine nus middl e 17te gth th hf ecenturo thesd yan e probabld yha earlier origins, as several abraded sherds were recovered from the earliest feature, ditch 4551. A pinched handl well-knowe th f eo n Dutch for alss mowa foun thin di s feature. Colanders, plai slip-decorated nan d , togethedisheD d Groupf san o r C wit , sA h Weser-type ware were also presen mid-17th-centure th n i t y ditche unfortunats th filli f t s.o I e that apart from obviously abrade t possibld no sherds estimatwa o e t t si e how much materia residual s probablt e leasa th wa lf to d slip-decoratee some an ,th e us f eo d dishes sa decorative objects may have ensured that they had an artificially long life span. It is noteworthy that only singla e sher Scottisf do h potter recoveres ywa d from this early feature, fill 4543, indicating tha t thia t s period Scalloway pottery needs were not being met to any great extent from the mainland. It is only with phase 2, late 17th or early 18th century till the late 18th century, that pipkins with rim form 1 appear - apparently superseding rim forms 2, 3 and 5. Dutch-style pinched handles continue in use as do dishes of Groups A, C and D and colander type vessels. Group B dishes first appear at this shera timd dean fro mBellarmina sherdd ean s probabl f Duingeyo earliesne th war e etar stratified stoneware sherds recovered from Area A, although the Bellarmine sherd is probably residual. Delft-type ware e alssar o present along with sherd f Chineso s e porcelainn a presence e b lattee Th . y th f ma reo indicatio strone th f ngo trading links between Hollan Chind dan a durin late gth e 17t 18td han h centuries alsy oma perhapd an s indirectly indicate thaDelft-type th t e wares were manufacture Hollandn di e Th . 576 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1983 earliest English material, Staffordshire-type slipware, appears during this period. White salt-glazed stonewar Englisf eo Scottisr ho h origi alss ni o possibilit e presenfirse th th r t r timehavint fo i tfo t f yo .Bu g been made in Scotland in the late 18th century the writer would have been inclined to date the end of phase 2 to the middle of the century. However, its presence may be the only indication that the cottage or smithy had a long life span. Although body sherds from pipkins were recovered from phase 3, 19th-century features, these residuae coulb l dal thiy b l ssherd w timefe f Delft-typso A . e waralsy oema hav earlied eha r origins. Victorian coin related san d pottery provid morea e accurat phase datth r occupatione3 e fo . pottere Th y sampl smalo to meris o eli t t discussio vagariee th f no coastaf so l trade.

CATALOGUE OF ILLUSTRATED MATERIAL (figs 6-8) Area code, context number and phase, where applicable, are given after a brief description of each piece. 1 Pipkin rim form 1, brown speckled internal glaze, A, 4540,2. 2 Pipki form nri m 1, dark brown mottled internal glaze , 3001G , . 3 Pipkin rim form 2, brown internal glaze, G, 3001. 4 Pipki form n, browri m 2 n speckled internal glaze , 3001G , . 5 Pipkin rim form 2, brown speckled internal glaze, A, 4538,1. 6 Pipki form n, darri m 2 k brown speckled internal glaze , 1007C , . 7 Pipki form n, darri m 3 k brown internal glaze , 3001G , . 8 Pipkin rim form 3, ochre green internal glaze, C, 1003. 9 Pipki form n, ochrri m 3 e internal glaze , 503B , . 10 Pipkin rim form 3, ochre internal glaze, A, 4538,1. 11 Pipki form nri , orangm 4 e brown internal glaze, C, 1003/1007. 12 Pipki form nri , ochrm 5 e internal glaze , 4538A , . 1 , 13 Pipkin rim form 5, ochre orange internal glaze, B, 503. 14 Small globular vessel, orange-brown speckled internal glaze , 4541A , . ,2 15-19 Tubular handle , 503sB ; B, 503; C, 1003 , 3019;G , 3001;G . 20-22 Pipkin leg type 1, A, 4538, 1; A, 4538,1; G, 3006. 4 Pipki2 2 3, & 3006typ g G n le , , 3001 e2 G ; . 25 Decorated rim , 3001,G . 26 Decorated rim , 1006/1007,C . 27 Decorated rim , 503,B . 28 Decorated rim , 3001,G . 29 Colander, interna externad an l l creamy orange glaze , 4538A , , 1. 30 Shallow bowl or colander, internal and external creamy orange glaze, A, 4522, 3. 31 Bow colanderr o l , interna externad an l l creamy orange glaze , 4538A , . 1 , 32 Bow colanderr o l , interna externad an l l creamy orange glaze , 1006C , . 33 Slip-decorated small open vessel , 3001G , . 34 Slip-decorated small globular vessel, B, 503. 35 Slip-decorated small globular vessel , 519B , . 36 Small strap handle wit spoto htw whitf so e slip , 3001G , . 37 Dish rim, orange internal glaze, G, 3001. 38 Dish rim, brown speckled internal glaze, G, 3008. 39 Dish rim, orange internal glaze, A, 4538,1. 40 Dish rim, overall white internal slip with clear glaze , 4536A , . 2 , 41 Dish rim, overall white internal slip with clear glaze, A, 4522, 3. 42 Dish rim, overall white internal slip with clear glaze , 3001G , . 43 Dish rim, overall white internal slip with orange slipped spots underlying clear glaze with green streaks, part of a pierced hole, A, 4547, 2. 44 Dish rim, white slipped decoration under clear glaze, C, 1006. 45 Dish rim, white slipped decoration under clear glaze wit hgreea n streak , 4536A , . 2 , 46 Dish rim, white slip trailed decoration under clear glaze , 3001G , . 47 Dish rim, white slipped decoration under clear glaze, A, 4538, 1. 48 Dish rim, white slipped decoration under clear glaze with green streaks , 503B , . HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AND7 19857 | 0 49 Dish sherd, white slipped zoomorphic design, possibly part of a cockerel under clear glaze, A, 4538, 1. 50 Dish rimsherd, white slipped decoration under clear glaze, B, 503. 51 Dish rimsherd, white slipped decoration under clear glaze, A, 4541, 2. 52 Weser-type ware dish rimsherd, white overall slip and orange-brown slipped decoration under clear glaze with green areas, G, 3001. 53 Delftware-type plate, dar lighd kan t blue decoratio blue-tinged nan d white glaze , 4547A , . 2 , 54 Delftware-type plate rim, dark and light blue decoration and blue-tinged white glaze, A, 4521, 3. 55 Delftware-type bowl, dar lighd kan t blue decoratio blue-tinged nan d white glaze , 3001G , . 56 Chinese porcelain dish, blue underglaz overglazd re d ean e decoration , 4547A , . 2 , 57 Chinese porcelain bowl, blue underglaze decoration, A, 4541, 2. 58 Persian ware bowl, blue and black decoration, C, 1003.

APPENDIX 2: BRICK AND TILE (not illustrated) W J LINDSAY Two piece f brickso , three fragment floorinf o s piece0 3 g f tild pantilso ean e were recovered from stratified deposits durin excavationse gth . pieceNone th f larges so i e enoug merio ht t illustration.

BRICKE TH S pieceo tw bricf se o Th k were recovered fro finae mth l backfilling, 4538 ditcf o , h 4551, phasd an e1 fro mfloorina g level, 4541, associated wit phase hth smithye2 . Both largee feature Aren Th i . re aA s ar piece from ditch 4551 is 102 mm wide by 44 mm thick but only 95 mm of its former length remains. The cream-yellow soft fabric contains occasional rounded sandstone inclusions up to 3 mm in length with irregular rounded void smald san l claylensed brice re Th f .lighs k i so surfaces weighn it i t d an ts wore ar n apart fro brokee mth n end. Fro excavatioe mth n evidence fragmentw fe a , s from clay-pipes, it seems that it was finally deposited in the mid-17th century but because of the worn faces it could have been made earlier in the century. The only other stratified fragment of brick is well-worn and no estimate of its original dimensions could be made. The orange-red soft fabric is slightly micaceous, contains iron inclusion roundes ha d sdan voids surfacee .Th s associated wit voide hsometimeth e sar s charcoal-coated. brice Th k fragmen founs wa tlaye n di r 4541 which contained other material latdatee th e o d18tt h century and earlier but again because of its abraded condition it could have been made considerably earlier. There is no knowledge of brick being used in the construction of Scalloway Castle although Douglas Simpson guidebooe drafa th n f i , o t Munesko t s Castle, bricreferd re ko st bein g incorporate coree th sdn i of walls. He was of the opinion that the material could have arrived at Muness as ballast from Holland (information from Dr R Fawcett).

THE FLOOR TILES Six fragment floof o s r tile, three from stratified deposit Aren si , weraA e recovered l belonAl . o gt phas featurese2 , layers 453 453d 6an 8 t likphasbrice bu eth , ke1 fragments the well-wore yar e ar d nan almost certainly earlie daten ri . The wely ylma have come fro castlee mth . e samfragmentx th si f el o fabri Al e ar s c whic s hardhi , orange, sand slightld yan y micaceous. Fine rounded quart partialld an z 0- m o gritt y5 m p burnesu t cinderdou y ferruginous inclusiono t p su 6 mm are also present. In one example rounded inclusions composed of iron, mica and quartz grits are thae b tnoticeably thes ma ferruginoue t ei ar d ean s inclusions which hav t beeeno n subjecte same th eo dt degree of heat. Irregular rounded voids are also present, often in association with the cindery formations. A small quantity of shell or fossil shell is also present. The tiles are incomplete but are c 24 mm thick floor-tile wide15s th m i l 5m e Al . on e fragmentd an covere e face sar eon witn do hwhita e slip; some have slip adhering to their (original) sides. The quartz-rich slip is covered with green glaze on five examples and clear glaze on the other. No other form of decoration is present but one glazed surface has a stacking scar.

THE PANTILES Of the 77 fragments, 30 were recovered from Area A, phase 2 features. Several came from the smithy 578 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1983 debris and the associated yard and are therefore of a probable 18th-century date. The sandy fabric of the stratified fragments is relatively consistent being hard and orange with rounded transparent quartz grits up to 0-5 mm in length. Angular orange fragments, occasional lenses of white clay and brown, cindery inclusions are also present. Only occasional flecks of mica are present.

APPENDIX 3: THE CLAY PIPES P J DAVEY Five hundred and twenty-three fragments of pipe-clay artefacts were recovered from the excava- tions, from 31 contexts in the five areas examined. Whilst many of the larger groups are from mixed 'garden' deposits (eg 503 with 122 fragments and 3001 with 181) and numbers from archaeologically critical deposits are generally too small for firm conclusions to be drawn, the pipes do provide dating indications fro mmora e secure base than othe rperioe find th take d f do san n together give some idea of the nature of pipe use and loss in Shetland. The pipes have been recorded and studied according to 'Guidelines . . .' (Davey 1981). Complete site lists and record sketches of all bowls and decorated frag- ments are held in the excavation archives. This report is divided into three parts. First, a summary of the information the pipes provide toward understandine sth eacf go h area excavated; second descriptioa , illustratiod majoe nan th l ral f no finds, and third, a discussion of pipe use in Shetland as represented by this one excavation.

EVIDENCE TH E PROVIDE PIPEE TH SY DB Area A This was the only area of the excavation where the contexts could be phased with any degree of confidenc groue pip8 th 2 ef o p es o fragment s fro contextsx msi , though small, does require some special consideration. The evidence can be summarised as follows: Bore (in 64ths in) Bowl/decorated stem Context 87654 Stems Bowls Total date range Comments 4000 111 4521/2 2 124 1 5 1800-50 Mixed 4525 1 1 1 2 4536 2 17th-centur1y 1stem s 1 1 1630-54530 81 1 Consistent45; fig 29 no5 4543 62 718 1640-60 Consistent; fi3 4 g o 12n , Total 4 14 6 2 2 26 2 28 Two of the upper ditch fills, which were the lowest layers in the sequence to contain pipes, included 19 fragments, eight from 454 1d 13 an fro m 4538 uppetwo,e e decoratee whicth th f s .Th ro hwa d Dutch moulded stem (fig 9, no 5) is typical of the period 1620-40, while the bowl in 4543 (fig 12, no 43) should fall within the range 1640-60. There is no sign of a late 17th century element. If the stem bore dating metho measuro t applies di w witd fe t shouledo i an h s d clearl usee yb d with great cautio f Oswaln(c d 1975, 92-5), 4538 gives a date of 1657 and 4543 a date of 1673. Even with a large enough group, ideally more than 1000 stems, these date sstatisticall e coulb t dno y separated evidence Th . e ther suggests ei s a slightly later date for 4543 than for 4538. The five pieces from 4521/2/4531 provide an interesting contrast. stemse th f Whilo , includino etw gburnishea d example ,evidence areth n , o f boreo e diameter, 17th century in date, the other three certainly belong to the 19th. The single stem from the yard surface, 4525, is likely to post-date 1700. Areas B and G The major group in B is from a cultivated soil, 503, with 122 fragments. Although around 100 of the pipes are probably of early 19th century date, the rest appear to date from the middle to end of the 17th. The single bowl in pit 517 (fig 11, no 29) is a Dutch example of the period 1610-40. The four stems from 519/523 indicat ea 17th-centur y dat r thesefo e fills. Unfortunately therbowo n r decorates ei o l d stem present which might allow these small pre-19th century context chronologicalle b o st y distinguished. HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AN9 D57 198 | 0 topsoie aren th I aG l (3001) provide fragment1 d18 s and wits ,a hvase areth t, majoritaB 19thf o e -yar century date thin .I s case, however, ther significana s ei t element belongin 18te th , ho g 10 t centur g fi g y(e nos 17, 19, 20, 22) as well as to the 17th. The absolute chronology of the pipe-containing layers is made more difficult because, whilst soil layers 300 d 3004an 8 include earl o mid-17th-centurt y y material, 300 3009d 3an , whic stratigraphicalle har y intercalated, produced earl mid-19th-centuryo t y find wells sa . Agai numbere nth s involve smalo to firmr e fo l dar conclusion drawne b t thio st bu ,s evidence doet sa least suggest that the 3009 robbing took place in the early 19th century. F and Areas E D, contextx si e Fivth sf efroo m this area included materia botf o l h 17t19td han h century manufac- ture. The sixth, soil layer 2507, has one 17th-century fragment. AreaC The pipe finds from spits of soil excavated in this area suggest that, although there were no sealed deposits steada , y build-u materiaf po l occurred evidence .Th e fro foue mth r largest pipe deposite b n sca summarized as follows: Context Mean bore (64ths~in) Bore 'date' Bowl range Latest date 1002 6-07 1699-61 1620-1800 1800 1003 6-23 1693-49 1640-1740 1740 1006 6-25 1692-73 1640-1710 1710 1007 7-10 1660-20 1620-80 1680 While the bore dates (cf Oswald 1975, 92-5) represent a mean of all the pipes in the context, the bowl range shows how the latest material becomes steadily earlier as the spits are deeper. This apparent stratification may have implications for the interpretation of other artefacts from this area. General Apart from the specific dating indications suggested by the pipes, a number of more general points madeaboue b nature site n th t.eth ca Firstf e o fac e th ,t that there wer cross-contexo en t joind san that only small proportions of any one pipe were present suggests both that rubbish was well scattered ove thasitd e th rean t only minor proportions were archaeologically investigated. This impressios ni heightened by the lack of bowls to go with very distinctive stems, eg the moulded stems (fig 9, nos 1-5) and the Jonah stems (fig 9, nos 6-8) are not matched by bowls of these types. Secondly, taking all the pipe sdate e whicb n d hca together , som intensite ideth f ao f occupatioyo t differenna t e b time n sca suggested. Pipe depositio dividee b site n eth dca n intno o four main periods: Period One 1635-60. Thirty datable fragments, out of a total for the site of 63, relate to this period. Thus 48% of such finds from the site date to the short span of 25 years, an average of 1-2 datable pipe fragments recovere eacr dfo h yea occupationf ro likels i t .I y that this relatively intense activity represents Cromwelliae th n occupation .j?ipe o Thern e s ear whic h clearly date fro beginninge mth site n th ei f so the first years of the 17th century. Period 1660-1700.o Tw Thirteen databl totale eth fragmentf )o com% 1 e(2 sfro m this periodn a , averag f 0-32eo r eac5fo h year covered. Mos withie ar trang e nth e 1670-9 presumabld 0an y represent a declining intensity of occupation during the second half of the century - a period when pipe smoking was probabl heights it t ya . Period Three 1700-70. Only 11 fragments (17%) were recovered for this 70-year period; an average of 0-15 yearr 7pe . This suggests occasiona visitsr o e ,lus rather than continued occupation. Period Four 1770-1920. nine Th e pieces from this 150-year period (14%) mostly date froe mth 19th century. With an average of 0-06 fragments per year, this suggests occasional use or tourist visits at a much lower level of intensity than during the 18th century. In general pipe smoking recovered during the 19th century from a decline during the 18th. Thirdly, some estimat qualit e pipee th th f ef o sy o durin g these period madee b n s.ca During Period One a range of qualities is in evidence, with the highest finish and best moulds, in general, on the earliest pipes (eg fig 9, no 13, and fig 12, no 39). In Period Two the pipes are almost uniformly of poor quality typified by the Dutch moulded rose examples (eg fig 11, nos 24-28). This may be the contrast between a socially stratified domestic occupation and a military one. In Period Three, the pipes are of very good 580 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1983 31 3>

I k^ 8

6 10

12 14 11

Fia 9 Dutch clay pipes: moulded stems (nos 1-8); fleur-de-lis and related stamps (nos 9-14) (scale 1:1 with detail at 2:1) HALL AND LINDSAYI EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AND 1980 | 581 quality but Period Four pipes never rise above the average mass-produced example made in Glasgow and Edinburgh in such profusion (fig 12, nos 48-54).

CATALOGU ILLUSTRATIOND EAN S (figS 9-13) possibls a Ar s fa followin e eth entrien i e gsar order: description; possible source where known; stem bore in 64ths in; comparisons where appropriate; suggested date range; area code; context number and phase where applicable.

Dutch moulded stems (fig 9, nos 1-8) and fleur-de-lis and related stamps (fig 9, nos 9-14) Floral and leaf decoration (fig 9, nos 1-5) [cf Duco 1981,251-3,379, nos 132-59] 1 Probably Amsterdam; 1630-40; ^in 503, ;B . 2 Dutch; T; 1670-80«sin , 3004;G . 3 Probably Amsterdam; &; 1630-40in , 1500;D . 4 Probably Amsterdam; ^ in; 1630-50; G, 3001. 5 Probably Amsterdam ;1630-50 ; ^in 4538, ;A . Jonah pipes (fig 9, nos 6-8) [cf Duco 1981,254,380-2, nos 160-75 (1630-50)] 6 Vestigal scales survive only; probably Gouda or Leiden; ^ in; B, 503. 7 Very worn mould; probably Gouda or Leiden; -fa in; C, 1002. 8 Unusual form of scales; probably Gouda or Leiden; /¥ in; G, 3001. Fleur-de-li related san d stamp 9-14s s no (fiFriedericf , )g[c 9 h 1975, 69-71; Duco 1981, 248-9, nos 107-22] 9 Single stamp applie t leasda t four times; possibly Gouda; -/-% in Ducf ;c o 1981 110o n , ; 1625-40; B, 503. 10 Identical or very similar stamp applied at least twice; /T in; 1625-40; B, 503. 11 Stamp with four-par appliee di t d once, very faint, probabl fleur-de-lisa t yno ; ^ in Friedericf ;c h 1975, 69 bottom; 1630-50; G, 3001. 8 12 Single stamp applie t leasd a times x tsi , ¥1625-40; ; differenTin 10 d froe , an 1003di mt; C 9 . 7 13 Very high quality four-part stamp applie t leasda t three times; ^ ¥ in Friedericf ;c h 1975,69 bottom; 1630-50 503, ;B . 14 Very faintly applied different four-part stamp, probably not a fleur-de-lis; /T in; 1630-50; B, 503. Dutch roller-stamped stems twistsd s 15-23an no (fi, )g10

15 Parroller-stampea f to d stem wit 'rinha Pearlsf go ' border Ducf c ; ; T\oin 1981,246 1670; 46 o -,n 1700;G, 3001. 16 Part of a stamp as 15;/T in; 1670-1700; G, 3001. 17 Parrollea f to r stamp with zigzag border Friedericf ;c , •/?in h 1975,66, bottom half page; 1700-30; G.3001. 18 Part of astamp as 17;^\ in; 1700-30; C, 1006. 19 Complete stamp as 17;¥ \ in; 1700-30; G, 3001. 20 Complex twist; probably Gouda; ^ in; cf Friederich 1975, 63, third from bottom; 18th century; G.3001. 21 'Rin f Pearlsgo ' roller stam twistd pan ; probably Gouda Friedericf c ;; -fain h 1975, left column, fro6 m bottom; 1720-50 , 1006;C . 22 Rouletted stem roller-stampe GOUDdn i A twice; ^ in; 18th century 3001, ;G . 23 stampeSte2 2 ms a dVERZYLF probabl; -fain y Frans Verzij workins Goudf lo wa o gawh 1753-74; cf Duco 1978,48; B, 503.

Dutch bowls bowld an fragments 24-38s 11g no fi , ( ) Moulded ros 24-8s eno (fi, ) g[c11 f Duco 1981,244,377 31-8s ,no ] roset do ,x onlSi 2y4 left par bowf to l survives, bettere milledt dno ; 1640-60 , 3001;G . 25 Seve roset ndo , just visiblrighe th tn esideo ; •&1670-90; in 2001, ;E . 26 Six dot rose with five sepals, bettered and milled on the front at least; ¥\ in; 1640-60; G, 3004. 27 Six dot rose with five sepals; & in; 1640-60; C, 1003. 38 582 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1983

18 16

\L/ 17 19

22 23

FI 0 GDutc1 h clay pipes: roller-stamped stem twistd san s (scale 1:1 with detai t 2:1a l ) t rosdo ex witSi 2h8 five sepals, mould very wor damagedd nan , poorly bettered, apologetic milling on front only; •& in; 1680-1700; B, 503. Heel stamps relien i l f,al (fig 29-33s 11,no ) 29 Rose consisting of seven closely spaced dots, bowl partially milled but damaged; -/± in; cf Fried- erich 1975,97 15201b o n , ,Ab ; 1615-40; B, 517. 30 Bowl fragment with indistinct stamp, possibly the axe of Gouda; -ff in; cf Duco 1978, 145, no 373; 18th century , 3003;G . 31 Bowl fragment with good quality, very high milling, bottered, indistinct heel stamp, possibly a milkmaid Ducf c ; ; /¥oin 1978,105,141 294o ,n ; this example probably 1670-90, thoug stame hth p is known from 166 18980o t , 1500;D . 32 Damaged bowl with milling on rim, rose and crown heel mark, there is a letter F to the right and a 7 missing letter to the left; ? ¥ in; Friederich (1975) has a number of similar stamps marked IF (Abb 114,122,124,135)s 15,no , datin 1610-40go t alsf ;c o Duco 1981, 53-58s 247no ,503, ;B . 33 Bowl fragment with finely executed crowned L heel stamp; -£% in; This Gouda stamp is first recorded belonging to Cornelis Luijnenburg in 1726 and was in use until the demise of P van der Want Gzn in c 1930. It was held by the Verzijl family from 1753 to 1821 (NB no 23 above); cf Duco 1978, 48, n ; 1750-70o9 , 1506;D . HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE^ 1979 AND. 1980 | 583

o 38 FIG 11 Dutch clay pipe bowls: moulded roses (nos 24-28); heel stamps (nos 29-33); spur moulded stam 34)o p(n : plain (nos 35-8) (detail 1: 1) 584 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1983

D. UNCAN «wo oo 1 I I i I 51 LEIT2 5 H 3 5 IBURG4 5 H

FIG 12 British-type clay pipes: bowls and stems (scale 1:1) HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AND 1980 | 585 Spur moulded stamp (fi) g34 11o ,n 34 Bowl wit hstaa windmilr ro botn o l h spur e side t ofth ff ,cu s o rim , crudely trimmed casting seams; gooo n ; d•fain comparison Friedericf c t sbu h 1975,129 ; 18t44 hb century,Ab , 1002;C . Plain bowls (fig 35-8s 11,no ) 35 Damaged bowl, bottere burnishedt dno ; &1620-50; in 300, ;G 1 36 Well trimmed, lightly polished bowl, bottered, millefrone th tn do only ;; 1670-90 -fain 503, ;B . 37 Poorly trimmed bowl, bottered, milled on front; -fa in; 1660-80; B, 503. 38 Elegant bowl, very smoothly finished, bottered, & in; 1670-1700; C, 1006.

British-type bowls and stems (fig 12, nos 39-54) 39 Well-made elegant bowl with very fine light milling, pink body, damaged and in three pieces; bore not measurable; incuse IB stamp on heel; source not known, but not unlike early Plymouth forms and stamps; cf Oswald 1969,129, fig 53, no 4; 162(MO; C, 1003. 40-4 rang4A mid-17th-centurf eo y bowls, probably Scottis origihn i n 40 Damaged bowl, yellowish fabric with many inclusions, milled; -fa in; 1640-60; G, 3001. 1 4 Upper par bowlf to , very poorly made, pink/orange outer surface; 1640-60 , 3001;G . 42 Upper par f bowlo t , very micaceous pink fabric, uneven wall thickness, heavily bottered rim, mille roundl dal ; 1640-60 503, ;B . 43 Complete bowl, bottered milledt ,no ;; 1640-60 -fain 4543,1, ;A . 44 Damaged bowl, bottered, milled on ?front; 1640-60; C, 1002. 45 Spurred bowl wit remaine hth mouldea f so d letter ?I abov e spur vera , y unusual form; -fa in; 71640-60; C, 1006. 46 Large bowl with pronounced 'chin', pink/orange outer surface, bottered and milled; -fa in; the form is similar to Scottish pipes from Linlithgow, especially those bearing the star heel stamp or the initials of James Colquhoun of Glasgow; cf Laing 1968, 127, fig 7, nos. 1-4 and Martin 1977, 221; 1680-1700 1006, ;C . 47 Elegant bowl with cut off, 'horizontal' rim; -fa in; cf Oswald 1975, fig 4, G: 21; 1730-50; C, 1002. 48 Plain spurred bowl, damaged, casting seams poorly trimmed ; 1800-50 -fain , 3001;G . 49 RAOB moulded bowl fragment; common throughou e Britisth t h Isles; -fa in; 1850 onwards; , 1501D . 50 Fragment of moulded, fluted bowl, very pure white clay; 19th century; B, 511. 51 Moulded rim fragment, including sailing ship; 19th century; D, 1501. 52 Stem stampe . DUNCAN/LEITHdJ ; probably John Duncas workinwa f Leito o nn gi wh h Edinburgh in 1837; fa in; cf Oswald 1975,205; G, 3001. 53 Stem stampe WOOD/GLA. . d. ; possibl. . S. y Willia WoomC d whworkins owa Glasgon gi w 1857-75; & in; cf Oswald 1975,206; F, 2500. 54 Mould-decorated stem marked THOs. WHIT.../... NBURGH; probabl Whyty recordeT o C e& d in Edinburgh directories 1832-64; -fa in; cf Oswald 1975,206; B, 503.

Fragment of a pipe-clay figurine (fig 13, no 55) Lower part of a pipe-clay figurine probably representing a Madonna and Child. The bottom of the otherwise solid statuett bees eha n hollowe trimmed an t dou d wit knifha thao e s t wouli t e d stanTh . dup effec f clothino t gaines gi verticay db l burnishin mannee th n gi f cla o r y pipes mose Th . t likely source for this object is Utrecht in Holland, where a major industry was active from the 15th to the early 19th centuries. The Scalloway find is very close in size and decoration to a complete Madonna and Child excavated in Amsterdam (Baart et al 1977, 472, no 902). There is no reason to suppose that this object coul havt dno e been made durin occupatioe gth Castlee th f no , sometim 17te th h n ecenturyi , 3001.G .

DISCUSSION Scallowae Th y pipe importane sar t because they provid firse eth t stratified sequence fro mlanda - based excavatio Shetlann ni allod dan w a preliminar y pictur f pipeo e source theid san r chronology sketchede majoo b tw o t re sourceTh . Hollane sar mainland dan d Britain, almost certainly Scotland. e problemTh f identifyino s g Dutch pipes have been discussed elsewhere (Atkinso Oswaln& d 1972; 58 | 6SOCIET Y OF ANTIQUARIES OF SCOTLAND, 1983

FIG 13 Pipe-clay figurine (scale 1:1)

Dave burnishinf o y e 1982a)us e fraughs gTh .i t wholwite pipee th th h f n sei difficulto collectiow fe s ya n are burnished: 69 out of 523 (13 %). The realization that some 17th-century Scottish pipes were burnished (cf Perth; Davey, forthcoming) means that great care must be taken to distinguish the very narrow and parallel marks made by agate tools in Holland from the broader and more unevenly executed Scottish examples. In the case of the Scalloway material all the burnished pipes appear to be Dutch. Of the 53 identified. Dutch bowls and stems (nos 1-38, plus 15 not illustrated) only 10 (19%) are burnished. On this basi whole sth roughlf eo collectio% 0 y8 Dutchns i preliminar.A y fabrice studth f yo s unde hanra d lens suggests that almost 90% of all 17th- and 18th-century pipes are of Dutch type. A further caveat, given the geographical position of Scalloway, is that some of the pipes identified here as Dutch may have been produced further north in Europe. In Denmark, for example, at least one early maker was Dutch (Ahlefeldt-Laurvig 1980 Swede,n i 249 d firse )an nth t makers were also Dutch (Bonds 1980, e 276)th t .A moment we are ignorant as to what these first pipes looked like. While some of the more clearly identified Dutch pipes ,earle sucth ys h a Amsterda Jonae , th 3-5 1 d hs ) an m pipeno , product9 s g (fifi , g9 g s(e 6-8s certainle no )ar y Dutch products remaint i , seee b no s t whethe productioe rth poore e somf no th f eo r quality pipes, such as the moulded roses (fig 11, nos 24-8), was confined to Holland alone. The chronology of the use of the two sources is of interest and can be tabulated using the pipe deposition periods described above:

No of Pipes No Dutch Britisho N Perio 1635-6e dOn 0 30 (48%) 23 (77%) 7 (23%) Period Two 1660-1700 13 (21%) 12 (92%) 1 (8%) Period Three 1700-70 11 (17%) 10 (91%) 1 (9%) Period Four 17704- 9 (14%) 0 (0%) 9 (100%) Totals 63 (100%) 45 (71%) 18 (29%)

These Scalloway results complement wha alreads i t y known from site Scotlandn si . Those from Aberdeen, wher t leasea t l pipe80al %f so recovered fro perioe mth d 1620-1720 wer Dutcf eo h origin, provid closese eth t comparison (Davey 1982a), although EnglisScallowaE N o n h s pipesyha contrastn .I , t Stirlina g Castle (Davey 1980 Pertd )an h (Davey, forthcoming) onl 17th-centure ythira th f do y pipes were Dutch and in Edinburgh itself no Dutch products were recognized from the excavations in St Mary's Street (Lawson 1980) arrivae .Th British-typf lo e pipe Scallowan si y soon after represen165y 0ma t an interruption in Dutch trade patterns caused by the wars with the English. If this was the case, they recovered quickly and dominated Scalloway pipe imports until the end of the 18th century. Further light HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AND 1980 | 587 is thrown by a comparison with the material from 17th- and early 18th-century Oslo (Skrel 1980). Ratios of Dutc Britiso ht h pipes: Scalloway Oslo 1620-50 19:1 50:1 1650-80 6:6 6:7 1680+ 11:1 6: 1 In both cases the decline in Dutch pipes after the middle of the century may be due to wartime con- ditions (cf Davey 1982b).

APPENDI SMALXE 4:TH L FINDS

BONE AND METAL OBJECTS (fig 14) CALDWELH D L Area code, context numbe phased an r , where applicable givee ar , n afte a descriptior f eacno h object. slightlA y1 curve polished dan d bone comb, with broad central reservatio find nean teeth, scratches on both faces below the spaces between the teeth indicate that the teeth have been cut by a knife. Length 43 mm; width 44 mm; length of teeth 13 mm, 18th or 19th century, A, 4522, 3. 2 Openwork bronze corner-mounting and nail - possibly from a book cover, 17th or 18th century, D, 1501.

FIG 14 Bone and metal objects (scale 1:2) 588 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1983 3 Sheet bronze embossed gauntlet, probably part of a doll or letter holder, 19th or early 20th century, F, 2501. 4 Bronze finger-ring with hexagonal collet lacking its stone, 17th or 18th century, F, 2501. 5 Trigger guar pistola r d fo 18t, , earllengtr hmm o 1 y h13 19t h century , 2501F , . 6 Decorated bronze buckle, late 17th or 18th century, A, 4531, 3. 7 Part of a bronze buckle, late 17th or 18th century, A, 4545,2. 8 Bronze buckle, late 17t 18tr ho h century 4531,3, ,A . 9 Bronze buckle, late 17t 18tr ho h century 3001, G , . 10 Bronze buckle, late 17th or 18th century, G, 3001. 11 Parhingea f to d loo bronzf po e fro mshoa e buckle 4528,2, ,A . 12 Parhingea f to d loo bronzf po e fro mshoa e buckle, A, 4528. 2 , THE COINS (not illustrated) DCALDWELH L Thirteen coins were recovered durin excavatione gth t Scallowasa yunidentifiable Castlear o Tw . e onle th y d onean s recovered from stratified horizons were three Victorian halfpennies (nos 7-9). 1 Mary, Lio hardhear no d (Scottish l£d), billon, 1558, countermarke obversn do e with heart con- taining a star (in 1572 it was ordered to be used to determine true coins from false) (Burns 1887, 503, B , . 7) 317 o n , 2 Charles I Turner (Scottish 2d), copper, group III, worn, B, 511. 3 Charles I Turner (Scottish 2d), copper, 3rd issue, C, 1006. 4 Charles I 'Stirling' Turner (Scottish 2d), copper, group IV, G, 3008. 5 Russian coin, Elizabeth, Denga, copper, 1749, G, 3001. 6 Dutch coin, Dui f Weso t t Frisia, copper, 18th century, worn, pierced through centre, possibly a resulsa beinf to g reuse buttona s da , 1506,D . 7 British coin, Victoria, halfpenny, copper, 1861, A, 4521, 3. 8 British coin, Victoria, halfpenny, copper, 1862, A, 4510, 3. 9 British coin, Victoria, halfpenny, copper, 1862 , 4522,A . 3 , 10 British coin, Victoria, halfpenny, copper, 1886, E, 2000. 11 British coin, Victoria, halfpenny, copper, C, 1000. t identifiedNo 12 , copperdiameterm m 9 1 , 3001 G , . 13 Not identified, copper, 20 mm diameter, G, 3001. THE GLASS (not illustrated) LINDSAJ W Y Fifteen pieces of glass were recovered from the Area A, phase 2, smithy floor and the soil horizon separating phase 2 activity from that of phase 3. Thirteen small fragments of green blown-bottles are typical product 18te th hf so century piecee Th . s come from vessels having thick, domed bases, almost vertical sides, sharp shoulders and long necks. A small piece of green window-glass from the smithy originao tw floo s lha r toola edge y fragmenb A . fin a sf i f cleanles o a clea t m t ri yfrocu r e winmth e glass was found in the soil horizon 4522. It is part of an upright bowl with a slightly concave profile which terminate slightla n si y thickened rim average .Th . e thicknesmm glas1 e c th s sf i so

STONE OBJECTS (fig 15) LINDSAJ W Y Three points from ston sharesd ear , probabl ySandstond madRe d e froOl ee m rockfounth e b o st d in the E and S of the Shetland Islands were recovered during the excavation of Areas B and G. Two were found in layer 519 along with two tiny sherds of coarse pre-medieval pottery. It is therefore probable pointd ar thae sth tare , likpotterye eth , residual althoug possibilite hth y remains that cultivation could have continued using ards at least into the 17th century. They are a common find on the Islands. The Scalloway examples are slighter than many, possibly indicating that the motive power was human rather than equin 1-3)s eno (fi, .g15 Twelve dark brow r greno y flake flintf so stona , e which doe t occusno r naturall Shetlane th n yi d Islands, were found during the excavations. Many of these were recovered from the upper levels of Areas HALL AND LINDSAY: EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AND 1980 | 589

i-

FIG 15 Stone objects (scale 1: 2 except no 5,1:1)

rectangulaA . G d Ban r cher fine-grainer to d quartzite, probably fro mflintloca k musket showed evidence of having been worked. There is some evidence of it having been reused after it had broken (fig 15, no 4). largA e quantit whitf yo e quartz pebble fragment smald an s l chippings were presen uppee th n ri t werw fe esoi a recovere l d horizonan F d d Areaf froso an lowee mth sD r levels, includin naturae gth l wate unlikels i t ri t channelsybu thaG Areatd n i ,the an ysB shoul consideree db evidencs da wastf eo e products from a quartz tool-making industry. The only worked tool is a small scraper from a soil horizon fee wth f pre-medievao e on is ni Aret I . la D find s fro excavatioe mth n (fi. g 5) 15o ,n Collin InstitutH e th G gratefuf m r s o a Geologicaf D eo I o t l l Sciences (NERC identifieo )wh d stone specimens nos 1-4. Catalogue Area cod contexd ean t numbe givee rar n followin descriptioga eacf no h object. 1 A wedge-shaped ard point, 155 mm long, in a fine-grained micaceous sandstone of a brown colour. 590 ) SOCIETY OF ANTIQUARIES OF SCOTLAND, 1983 The upper unworn surface is covered with pecked impressions but the underside, parts of the sides uppee th f ro surfacp ti e smoote th ear d considerablt an hbu y scored generan .I l these striae radiate doubo n d t an hav p ti efro centre e beemth th f neo causeimplement'e th y db s passage through stony ground, B, 519. 2 An ard point, 155 mm long, fractured along its length, in a bluish fine-grained siltstone. There is no surviving evidence that the stone had originally been pecked and only the rounded end is smooth wit hfea w scores from 519use, B , . 3 An ard point, only 50 mm of its length surviving, in a bluish fine-grained siltstone. It shows faint evidence of having been pecked originally but only on the upper surface near the cross break. Uppe lowed ran r surface smoote sar scored han d from striae use deepee Th .e ar r and more numerous lowee onth r side , 3001G , . 4 Part of a chert or fine-grained quartzite for a flintlock musket. It is 32 by 20 mm and is off-white in colour, C, 1003. 5 Part of a white quartz scraper showing some evidence of retouching, D, 1501.

APPENDIX 5: REPORT ON THE ANIMAL REMAINS CATHERIN I HODGSO W E SMITG D NHAN

METHODOLOG MEASUREMEND YAN T The bone material was identified by direct comparison with modern material. Ribs and vertebrae, other than the first two neck vertebrae, were not identified or recorded. Bird remains were identified only as to bone and not to species. Fish bones were not identified as to species. Measurements were taken follow- inscheme gth e propose Driescy db h (1976, Sheldric19-100)C Britise M th r f D .k o h Museu cons mha - firmed thabono tw te fragments, whic togethert hfi froe ar , m larga e sea-mammal. They compare favour- ably wit hwhala e vertebra reported fro mRomaa n sit Corbridgt ea e (Hodgson 1980,15).

THE SAMPLES Bones reporteB/G e frod numbe e areao ar an ) m Th tw A . s ( identifie f dron o d bone smalls si : Are, aA 90; Area B/G, 81. The samples may indicate how man interacted with domestic and wild animals, with regard to meat supply and animal husbandry. There is some evidence that the samples represent both domesti commerciad can l refuse, therefore they canno useguide a tb s d a dietar o et y preferences.

RELATIVE FREQUENCIE SPECIEF SO S PRESENT relative Th e frequencie f specieo s s were estimate basee th n sd o of) tota (a : l numbe f identifieo r d bone eacf so h species presentminimu) (b d an , m number animalf so s present. Tabl givee1 numbere sth s and percentages of bones identified from each area.

TABLE 1 Numbers and percentages of identified bones recovered from each area, classified as to species (excluding fish and sea mammal) Sheep/ Area Date Cattle g Pi goaHorst e Bird Total A, phase 1 mid 17th century No 42 11 4 - - 57 % 73-7 19-3 7-0 5 2 1 1 - 7 A, phas e 2 6 1 17th/ o l 8tN h century % 64-0 28-0 - 4-0 4-0 5 - - 1 - 4 A, phas o e3 N 19th century % - 80-0 20-0 B and G 17th century soil No 1 2-1 1 5 % 20'0 40-0 - 20-0 20-0 8 4 2 - 2 4 3 0 1 G o d BN an 17th century % 20-8 70-8 4-2 - 4-2 BandG 19th century No 1 3 1 - - 5 % 20-0 60-0 20-0 HALL AND LINDSAYI EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AND 1980 | 591

The incidence of fish and sea-mammal bones from each area is given in Table 2. The highest inci- dence of fish bones is from the 17th century, ie the time at which the castle was being built or of its early occupation. TABLE! Incidence of fish and sea-mammal bones f sea-mammafisf o o h o o boneN N s l bonesArea A, phase 1 - - A, phase 2 - 1 - 2 A, phas e3 , 17tG 1 hd centurBan y soil B and O, 17th century 19 3 , 19tG hd Bcenturan y The minimum numbers of animals present, estimated on the basis of the most frequent bone of a specie givea n si n area listee ,ar Tabln di e 3. TABLE 3 Minimum number animalf so s present Area Cattle Sheep/goag Pi Horst e Bird A, phas e1 422-- A, phase 2 13-11 A, phase 3 1 - 1 - - , 17tG h centurd Ban y soil 11-11 B and G, 17th century 141-1 , 19thG centurd Ban y 111-- date Th a presented reflec economn a t y concerned with cattle, sheep/goat, pigs, small horser o s ponies, fish, bird largd mammala san ese s (whale . Surprisinglys?) , perhaps sample th , e lacks remainf so cats, dogs and smaller sea mammals (such as seals, walruses and otters). As at the nearby Bronze Age sit Jarlshoff eo , Sumburgh, dee absene rar t (Platt 1933).

AGE OF ANIMALS ON DEATH Cattle cattle On e half mandibl assesses ewa r tootdfo h eruption thin I . s case state e th ,f wea th eo f o r third thire pillath df o rpermanen t molar indicated thae animath toves wa lr five t deatyeara d hsol (Grigson evidence 1974)th n .O fusiof eo distaf no l articulatory surface cattle th n esi long bones examined, onldistae yon l metacarpa totaa f fivf o o l t e lackeou l epiphysess dit , implying tha tanimale 80th %f o s from which these bones came survive 2-2f o e didistao yearsag unti tw e distaf le th l femurO . on ld an s radius examined, none lacked epiphyses, thereby suggestin gthesf o tha l eal t animalf s o reache e ag n da 3^-4 years. It would, therefore, seem that there was little preference for veal or calf skin at this site, and that cattle were successfully overwintered. Sheep/goat Four sheep/goat half mandibles were complete enough to be assessed for tooth eruption and wear pattern, accordin Payne'go t s scheme (1973) numbee .Th mandiblef ro s examine smalo to alloo s dt li wa killing curvdrawe b t evidenco et n bu provides ei f singldo e followine sheeth f po groupe ag g s being killed or dying: 2-6 months, 2-12 months, 2-3 years, 4-6 years. Pigs Because osteologica pigf o t deatsa e ag lunreliable s hi evidence th o t dentae s ea th , l evidencs ewa examined. One half mandible bore a worn third molar and is, therefore, assumed to come from an animal aged over three years (Silver 1963). Horses A single horse radius from fuseArea , d phasdaA ha dista, e2 l epiphysi assumes i d san como dt e from a small animal older than three and a half years. 592 | SOCIETY OF ANTIQUARIES OF SCOTLAND, 1983

CARCASS ANALYSIS A comparison was made of the numbers of cattle and sheep/goat foot bones (low meat yield) with leg bones (high meat yield) because this may be a guide as to whether the bones represent industrial or commercial refuse higA . h proportio foof no t bone indicaty sma e that meat-rich bones were being sent fro sitee mthar th o , t animal-based industrial/commercial activit takins ywa g place . Witneait o hrt only one exception (Area A, phase 3), cattle foot bones outnumbered leg bones, and in half of the contexts studied sheep/goat foot bones outnumbered leg bones. The preponderance of foot bones in the samples of 17th-century materia carcaso t merelproductioe e y th du s lma o e dressint y b e neat's-foof du no r go t oil. thesf I e remain e boneth e smealf ar s o s site eateeth t thena y represent low-prestige cut meaf shino d an tt standarw lo a livingt f da o . Thireflecy diee sma th t t builder e eateth y n b f Scallowa so y Castl wero eewh werd pressean e n forcedme woro dt k withou paymenty tan .

PATHOLOGY AND BUTCHERY Wit single hth e exceptio arthritin a f no c cattle metatarsal ,sign o thern diseasf e so ear e amone gth bones examined. Most of the bones show butchery marks but none appears to have been sawn.

SIZTYPD ANIMAEF AN E O L size eTh range bonef so s measured indicate thacase w onln i tfe sy a wer animale eth s larger than those reporte fron do medievae mth l level f Perto s h High Street (Hodgson, forthcoming). Onle yon cattle metacarpa sheee on pd metacarpaan l largee ar l r than those foun t Perth dhorsa a t ebu , radius si significantly smaller and must represent an animal as small as a pony. In a 19th-century context, the Rev William Stevenson, minister of Northmavine parish, Shetland (NSA 1845,15, 78), reported on the types of cattl sheed ean p presen Shetlandse th n ti : 'The sheep and cattle reared or bred are for the most part of the pure Zetland breed; but little or no attention whateve pais ri theio dt r improvement. ScotcSome th f eo h kinds have occasionally been introduced, but were found not to answer well, as they are too tender or soft for the climate.' There is little reason to suppose that standards of animal breeding and stock improvement were better 17te inth h century.

ACKNOWLEDGEMENTS writere Th s wis thano ht followine kth g peopl havo eewh helpeproductioe th n di thif no s report f Hayfield.o Capy r LaurencHa t M , eSmitJ Shetland han d Islands Council kindly gave permission to excavate the sites. Funding was provided by the Scottish Development Department (Ancient Monuments). Miss J M Scarce identified the Oriental wares and Dr G Stephan the Duingen stonewares CollinH G r s D identifie. stone th d e CaldwellspecimenD r D J d P , an s Davey WG , I Hodgso SmitC d hnan have provided reports. Figwer5 s1- e execute CannelJ y db l and G Harden. Various staff members of the Scottish Urban Archaeological Trust have made valuable comments on drafts of the text.

REFERENCES Ahlefeldt-Laurvig, J 1980 'Clay pipes from Denmark', in Davey, P (ed), The Archaeology of the Clay TobaccoPipe, 4(1980), 219-51. (=Brit ArchaeolRepInt Ser, 92.) Atkinson Oswald& R D ,A ,197 2 'A brief identificatioe guidth o et f Dutcno h clay tobacco pipes found in England', Post-Medieval Archaeol, 6 (1972), 175-82. Baart alt e ,J 1977 Opgravingen Amsterdam:n i twintigjaar stadskernonderzoek. Amsterdam. Bonds G ,198 0 'Clay pipes from Ryssviken' Daveyn i , Archaeology e (ed)P , Th , Claye th f o Tobacco Pipe, (1980)4 , 65-88. (=Brit Archaeol Ser,t In p 92.Re ) Brand J , 1883 Description f Orkney,o Zetland, Pightland - Forth Caithness.d an Edinburgh. (First edition, 1701.) Burns, E 1887 The coinage of Scotland, ii. Edinburgh. Cruickshank R 197D G ,6 Scottish Pottery Society, Archivist's Newsletter, (1976)1 . Edinburgh. HALL AND LINDSAYI EXCAVATIONS AT SCALLOWAY CASTLE, 1979 AN3 D 19859 | 0 Davey, P J 1980 'The clay pipes', in Ewart, G, 'Excavations at Stirling Castle, 1977-1978', Post- Medieval Archaeol, 14 (1980), 46-50. Davey, PJ 1981 'Guidelines for the processing and publication of clay pipes from excavations' Medieval Laterd an Pottery Wales,n i (1981)4 , 65-88. Davey J 1982P , a 'The clay pipes' Murrayn i , (ed)C J , , Excavations medievale th n i burgh of Aberdeen 1973-1981, Antiqc So Scot Monogr Ser, (1982)2 , 216-23. Davey, P J 1982b 'Clay pipe fragments', in Hunter, J R, 'Medieval Berwick-Upon-Tweed', Archaeol Aeliana, 5 ser, 10 (1982), 91-9. Davey, P J forthcoming 'Clay pipes from recent excavations in Perth'. Donaldson, G 1958 Shetland life under Earl Patrick. Edinburgh. Driesch, A von den 1976 A guide to the measurement of animal bones from archaeological sites. Har- vard University. (^—Peabody Museum Bull, 1.) Duco, D H 1978 Goudse Pijpen: de geschiedenis van de pijpmakerij te Gouda vanaf haar ontstaan tot merkene bijlagen d heden, n ee het n t va va Goudse me eigenaren.n glidehu n e Amsterdam. Duco, D H 1981 'De Kleipijp in de Zeventiende Eeuwse Nederlanden', in Davey, P (ed), The Archaeo- logy Claye ofth Tobacco Pipe, (1981)5 , 111-468. (=Brit Archaeol Intp Re Ser, 106.) Friederich, F H W 1975 Pijpelogie-vorm, versiering en datering van de Hollandse kleipijp (A WN monographic no 2). Voorburgh. Gifford T ,187 historicalA 9 description Zetlande th f o Islandsyeare th n i 1733. Edinburgh. (First edition 1786, London.) Grigson, C 1974 'The craniology and relationships of four species of Bos , J Archaeol Sci, 1 (1974), 353-79. Haggarty, G 1980 "The pottery', in Ewart, G, 'Excavations at Stirling Castle 1977-1978', Post-Medi- eval Archaeol, 14 (1980), 36-40. Hodgson, G WI 1980 A comparative analysis of metrical data derived from mammalian remains from selected historic sites in Britain, PhD thesis, University of Dundee. Hodgson WG , I forthcoming 'A general animarepore th n o tl remains excavated during 1975-1977 from the medieval levels at the High Street site, Perth'. Hurst, J, Neal, D S & Beuningen, H J E von 1975 'North Holland slipware', Rotterdam Papers, 2 (1975), 47-65. Laing R 196L , 8 'Excavation t Linlithgosa w Palace, West Lothian, 1966-1967', Proc Antiqc So Scot, 99(1966-7), 111-47. Lawson G ,198 0 "The clay tobacco pipes' Holmesn i , McQM N , , 'Excavation t Mary'S n si s Street, Edinburgh, 1974', Post-Medieval Archaeol, 14(1980), 171-4. Lindsay, W J forthcoming 'Excavations in Elgin'. MacAskill N , 1982 "The pottery' McGavinn i , , 'ExcavationA N , Kirkwalln si , 1978', Proc Antiqc So Scot, 112 (1982), 405-13. McGavin, N A & Wordsworth, J forthcoming 'Archaeological work in Lanark 1979'. C Martin197 e d 7F P "Th, e Dartmouth Britisa , h frigate wrecke Mullf dof ,cla e 1690yTh . pipes'.4 , IntJ Naut Archaeol Underwater Explor, (1977)) (3 , 6 , 219-23. A Statisticalw 184NS Ne 5 Account of Scotland. Oswald ,A 196 9 'Marked clay pipes from Plymouth, Devon', Post-Medieval Archaeol, (1969)3 , 126-42. Oswald, A 1975 Clay pipes for the archaeologist. Oxford. (=Brit Archaeol Rep 14.) Payne, S 1973 'Kill-off patterns in sheep and goats', / Anatolian Stud, 23 (1973), 281-303. Platt I 193M , 3 "The animal bones' Curie , n 'Furthei , O A , r excavation t Jarlshofsa , Shetland', Proc Soc Antiq Scot, 67 (1932-3), 127-36. Silver, IA 1963 "The ageing of domestic animals', in Brothwell, D & Higgs, E S (eds), Science in archaeology, London, 250-68. Skrel D ,198 0 'Clay pipes fro excavatioe mth Revierstreden ni t 5-7, Oslo' Daveyn i , (ed)P , e ,Th Archaeology Claye ofth Tobacco Pipe, (1980)4 , 299-318. (=Brit Archaeol t Ser,In p 92.Re ) Smith D 197 H , 8 "The Scandinavian influenc makine th n emoderi f go n Shetland' Baldwinn i , R J , (ed), Scandinavian Shetland, Edinburgh, 23-33. Spearman, RM 1982 'Excavatio t 17-2na 1 Castle Street, Forfar, 1979', ProcAntiqc So Scot,2 11 (1982), 457-64. Tudor, J R 1883 The Orkneys and Shetland. London. This paper is published with the aid of a grant from the Scottish Development Department