ON EXCAVATIONS CARRIED out in 1937 and 1938. by A. O. CURLE, C.V.O., LL.D., F.S.A.Scot., F.S.A

A VIKING SETTLEMEN FRESWICKT TA , CAITHNES1 7 S

A VIKING SETTLEMEN FRESWICKT TA , CAITHNESS. REPORT ON EXCAVATIONS CARRIED OUT IN 1937 AND 1938. BY A. O. CURLE, C.V.O., LL.D., F.S.A.ScoT., F.S.A.

Anyone interested in the science of place-names "who studies the map of Caithness from that poin f vieo t w will observe thae nameth t e ar s divisible into two groups, on a philological as well as on a geographical basis. If the county is divided into two sections by a line drawn from Cross Kirk Bay, some 6 miles to the west of Thurso on the north, to the tow f Lybste no e south-eas th n o r t coast, thus separatin e hilth gl country fro lowlande mth t wili ,foun e b l examination do n that wherea placee sth - namee formeth e witexceptionn w i sar r h fe s Celti d Gaeliccan , thosn i e lattee th r have their root Scandinavian si n speech. Suc a hdistributio n pointa ver o yt s considerable displacemenf o t the native population during the period of the Norse settlement; for it can hardl supposee yb d thae Celtith t c people willingly relinquishee dth coasts and fertile tracts of land for the barren moors and mountains that for westere mth n portio countye th f no . Notwithstanding this abundant evidence of Norse occupation, no y buildinan trac f o eg referabl o that e t perio s observewa d d whee th n survey of the Antiquities of Caithness was made for the Royal Com- missio Ancienn no t Monument 1911n si . The earliest Viking settlements would probably be situated by the estuaries of the Wick and Thurso Rivers, on the sites of the present towns, where long since all traces of their existence must have been removed, or buried beneat e streetth hd housesan s . Attractive landing-places elsewhere on the coast, which is for the most part rocky and precipitous, are few and far between, but the bay of Freswick, some 6 miles south of Duncansby Head, with its shelving beach, is a marked exception to the rule. From as early as the time of Earl Sigurd, towards the close of the eleventh century e namth , f Freswicko e e for th f Thraswickm o n i , , flits across the pages of the Sagas. In the Saga of Burnt Njalx we read how Sigurd, having learnt that his brother-in-law "Havard in Thraswick" had bee ne Scot slaith y sn b Earls , Hund Melsnatid an i , gathered together a mighty host from all the isles and fought a battle at Duncansness, in whic e earlhth s were defeated. a Thelate t a nr date when Kari Solmundson, who had escaped from Njal's burning hall, struck off the

e Story f Th 1Burnto Njal, translated fro e Icelandicmth Saga f Burnto r GeorgSi Njaly b e Dasent, chap. Ixxxiv. : 72 PROCEEDING E SOCIETYTH F O S , FEBRUAR , 1939Y13 .

hea f Grunnao d r Lambi' t Eara n l so sSigurd' s fellowhi s d boardsan e h , s shi farehi o Freswickp t n i d , wher e s housabod th hi e too h en p ei eku o"wortha f y man, whose nam Skeggi,s ewa witd "an h whom they stayed a ver" y long time."1 Eventually, after Karmadd ha i a pilgrimagee , and obtained absolution, he returned to Freswick to the house of "Master" Skeggi gavo "m awh hi e, shi f burden,po d witan "h eighteen o n me n boar sailee dh d bac Icelando kt . At a still later date, in the Orkneyinga Saga, we find Freswick again coming into notice in connection with Sweyn Asleif's son, whose father Olaf had an estate at "Dungalsbae" (Duiicansby), where the former frequently resided. Sweyn, who was a very notable Viking, looked afte e estatth r f Freswic o e s stepsons d himselhi ha r d kfo a fortres f an , s in the neighbourhood, called Lambaborg, for which identification has been suggested both for Bucholie to the south of the bay and the Broch of Ness to the north. Incidentally, from this Saga we also learn that there were thickets in which men hid themselves at that date (1153) not far fro e "halmth f Freswick, o l greao n thad t an "ta t distance froe mth hall there was a farmstead. In the centre of the wide bay the sandy beach is backed with irregular banks of sand, which rise sharply to a height of 20 or 30 feet before attaining the general land level, and as yearly these banks are eroded and driven farther back by the tempestuous winds which prevail on that coast, their contours now probably differ considerably from those presented at the time of the Viking settlement. Back fro e beachmth , generationwherw fe ea thero meadows sag ewa - land, ther s nowi e , e combineowinth o t g d actio f burrowino n g rabbits and tearing winds, an area several acres in extent, reduced to an arid deser deef o t p hollow sanf occasionaso d dan l resistant ridges. From this area sherd f coarseo s , unglazed potteryt , no whic s wa h analogou y recognisean o t s d ware found elsewher Scotlandn ei d froha ,m tim timo et e bee nMuseume senth findin e o t Th potterf .go similaf yo r character at Jarlshof in Shetland, associated with a Viking settlement, furnishe e means identificationth it d f o se sand th a visi o d yt t an are, a wala f o l emerginp e discoverto th lee o dt th gf yo e fro e sanon m th n di e hollowe osoutherth fth n o s n e characteristilimitTh . c featur f thio e s e explaineb potteryy ma dt i e ,numerou here,th s i , s impressione th n o s surface huske th strawd f s an o ,oatsf o , which have been employex mi o dt with, or temper, the clay, in order to bring it to a proper consistency for potting. Over most of the sandy area there were exposed evidences of occupation in the shape of traces of building, midden refuse, hearths, and burnt broken stones suc wers ha e use Norsn di e cookinproducinr fo d gan g stea bathsr mfo . 1 Ibid., chap. cliv. • A VIKING SETTLEMEN T FRESWICKA T , CAITHNESS3 7 .

The exploratio s commencenwa Junn di e wore th 1937 kd carriean . d a furthe r onfo r perioweekx si f do s durin e summegth e f 1938o r Th . outcropping wall above referred to made a suitable starting-point. It lay approximately east and west along the edge of the hollow, and was formed with large boulders from the beach. Eventually it proved to be the north face of a typical Viking wall, measuring a metre in width with a core of compacted earth in the centre. The remains of buildings which were revealed following this discovery have been grouped under thre e, accordin C heads d , Ban A ,, o thei gt r apparent age, commencing with the latest.

GROU. PA The buildin f whico g e walth h l mentioned above forme e soutth d h fron oblongs t wa measureplansd n (No, o an 2) , 1 . d , figsan d 1 .interiorl y beed ha nfeet 4 t I 1 subjecte .3y 0b somo dt e reconstruction walle th r s ,fo at the east and \vest ends were of different character. Where best preserved, the building had an elevation of from 2 to 3 feet. The area contained within the walls was covered deep in kitchen-midden refuse, and there was no definite floor recognisable over the greater part of it. The doorway fee 3 ,width n i t places norte th wa , n hdi walpoina t a l t fee1 1 t distant fro wese m th ker o t n silr endbo d l an remaine, d betwees nit jambs t abouA . foo1 t inche6 t s inwards fro e doorwaymth largo tw ,e uprigh tedg n protrudind slabo et an se s g abov floor-levee eth heigha o t l t of a few inches extended for a length of 6 feet 6 inches—evidently the bas somf eo e arrangement employe screeo dt fire nth e behind them froma rush of wind. In the centre axis of the house, towards the west end, and covered partiall e projectinth y yb g flagstones e hearthth y la ,, measuring some 11 feet in length by 3 feet 9 inches in breadth. There were no remain surroundina f so gpavinf o ker r bgno beneath. smala ease Ad tth lten rectangular construction with built sides occupied the centre of the wall. It measured 2 feet in height, and its upper surface wa suggeso leveo t s s a l t tha e levellinth t beed gha n purposely effected. As may be seen from the plan (fig. 2), the structure had not been laid on a square foundation, the north side being 3 feet shorter than the south, so that the spaces on either side were not symmetrical. On the floor of the dwelling, between the hearth and the front wall of this interior structure, heavy flagstones had been laid in two distinct rows from an area of scattered paving at the west end (PI. XXXVIII, 1). As wil subsequentle b l y explained northmose th , coverew ro t o ddraina n t bu , purpos discernibls ewa othere th r ,efo which , directioe howeverth n i d nle , of the intake of the vent to be afterwards mentioned. On examination it 74 PROCEEDINGS OF THE SOCIETY, FEBRUARY 13, 1939.

%''?t - pi V1J-/^S T^H= . -.1?^I **2 ^U1 Z) I«8- J|g yr^ > *Q Ili t/i 1/1 uj e> o O_ -I Q '• z z Z LU a a ~ <5 * in O z> -SI;) p 03 CD giS o+ -05 1 A VIKING SETTLEMEN T FRESWICKA T , CAITHNESS5 7 .

sto 76 PROCEEDINGS OF THE SOCIETY, FEBRUARY 13, 1939.

FIRST BUILDING SECTION C-D E222 SECON. DDo E3 THIRD Do. DDE BATH BUILDING

SECTION A-B

, showinI Fig f . o Wes 3 . d gen t Bath Chamber, Drain Ventd an , , with sections, also Walf o l earlier Building beyond. .A VIKING SETTLEMEN T FRESWICKA T , CAITHNESS7 7 .

became apparent realitthan i e constructio s th t ya wa e wesd th en t na chamber of which the walls had been reduced all round to an even level, and the interior carefully filled up with stones and turf, so as to ensure levea l surface. Whe filline nth g materia l beeal nd removedha l chambea , exposes rwa d measuring 4 feet 10 inches along the back and front walls, 4 feet 5 inches along the north, and 5 feet 10 inches on the south (fig. 3 and PL XXXVIII, 2). entrance Th fros lattee emwa space th th rf eo directio lefy t wa betwee y nb n e mai th soute nhouse th walth hd e chamberf ewalo an lth f o l , whics hwa crossed by a line of stones forming a kerb at its commencement. The entrance, approached by a step and over a projecting sill, was placed some- centre wallwese e wha measureth d th th f f o tan ,o e o t d som feee2 inche4 t s in width. Directly opposite, through the north wall, was another opening, 1 foot 8 inches wide, not furnished with a kerb, and in lieu of a step, with two thin slab stonf so e sloping downwards int enclosen oa d space r closeto , , betwee e nortnth h e chambee outeth walth f o e d rmail th walan r f no l building e detaileth , d descriptio whicf no h wil furnishee b l d later. The walls of the chamber on the inside remained to a height of about 14 inches. The floor was carefully paved all over, except for a small area measuring about 12 inches by 6 in the south-west corner, where there had been a fireplace. A heap of levigated clay, amounting to about a barrow-load, was piled up in the north-west corner. Beneath the surface the fireplace was filled for a depth of 16 inches with burnt broken stones, and nons a , thesf eo e fragments corresponded evidens wa t i , t that thed yha been broken before being employed s afterwarda , s explained e fireth . n i , Among the stones a small quantity of peat-ash was observed. When the clay had been removed, a stone, such as could be comfortably held in the hand s noticewa , d projecting fro e pavingmth bein n O .g s liftedwa t i , foun havo dt e been usepluga s da , fillin neatlga y formed hole inche3 , n si diameter surrounded an , largy db e pebbles. Further removinn o , f o e gon the paving stones near the centre of the floor a well-fashioned drain was exposed, which found its exit through an opening 5 inches wide in the west e chamberth wal f o l , thence continued beneat e coversth h , previously mentioned on the floor of the house, to discharge by way of an offshoot into a sump, dug in the floor towards the north wall. On examining the fireplace after removal of the stones with which it had been filled, a vent was discovered passing through the adjacent west innee th walln wale r sinkine O facth .th lf linteeo a blockef d go lha d this openingoutee th n r O face. , however intake th ,stils elwa clear, flanken do either side by stones sunk in the floor, one level with the vent, and the other rising above it, as if to direct the air into it. There was no doubt, from the facts above related, that this was a well- preserved example of a Viking bath, but before dealing further with it, a 78 PROCEEDINGS OF THE SOCIETY, FEBRUARY 13, 1939.

description must be furnished of the small closet which opened out of it on the north side. The angle formee junctio th e nort y th wesd b df han no t e wallth f o s chambe beed rha n constructe bevela n frod d o bast e ther i an m , th f n eo era the foundations of a lightly constructed wall, which had the appearance of being secondary, and which crossed the floor obliquely to the north wall of the house. The space, or closet, behind it measured in greatest length and e westh t breadtA end . ,3 lyin7 feehy b tg tilted agains e nortth t h wall, were two large flagstones, with a smaller one between (PI. XXXIX, 1), which, however, were not of sufficient length to have reached the opposite o havt wald e an formel a shelfd s migha , t have been suggestee th y b d angl t whica e h they laye flooTh .r beneath the s coveremwa d wita h deposi f middeo t n refuse simila thao t r t whic s spreahwa d generally over e house. e flooth th f o r Below this ther s foun e remainwa th d f whao s t appeared to be a small, open gutter, formed with thin stones set on edge obliquely on either side, which, passing beneath the westmost flag, led under the bath-chamber. The back of the closet at the east end was paved. From early times, down in fact to the present day, for the practice still exists in remote districts of Scandinavia and Finland, the method of bathing indulged in by the peasantry, was that of the steam, or vapour, bath. So much in common had it with the bathing practices of Slavonic, Turkish, and Persian countries, that it is a reasonable supposition that the baths of these various regions had all a common origin, and that doubtles northere sth n peoples importe fashioe dth n with them when they cam theio et r present territories fro e Neamth r East durin migratioe gth n periothire fourtd th dan n di h e they o th centuried r n r i era, No ou .f so course of centuries, seem to have introduced any radical alterations into systeme th . Various traveller havo swh e indulge sucn di h baths merelr o , y witnessed the process of bathing, have furnished accounts of it, and among the former, Paul Du Chaillu has described the bath-chamber as he found it, and his experienc a bather s a passaga e n i , e whic s worti h h quotin n part.i g 1 e mosth f to characteristie ''On c institution e countrth f o s y (Finlands i ) e Saumath (bath-house)a smal houseg s i lo lt I . , . buil . . t very tight, wit windowso hn , havin gsingla e apertur e smoketh t abovele outo et ; e centr n oven-likth a n s i ei e structure buil f looso t e stones, under which a fire is kept burning till they are very hot: then the fire is extinguished, e •wometh d nan clea e placnth e thoroughl f ashe yo d soot e smokan s th , e hole havin ge meantim th bee n i n e closed largA . e vessel filled with water is placed within; and a number of slender twigs, generally of birch trees, are put into it, to be used as switches." After describing the assemblin e bathersth f o g , mald femalean e , who, wit a hthermomete r - . - • •- '-- i The Land of the Midnight Sun, vol. ii. p. 201. A VIKING SETTLEMEN T FRESWICKA T , CAITHNESS9 7 .

standing below zero, appeare n costumedi s that reminde f Africo m ahi d minus the colour, and his own embarrassment on finding himself among same theth men i condition describo goet e h ,n o s procese eth bathingf so . "I hastily pushe e doo welcomes dth re wa voice opeth d f aly nan o sldb e compan intensth o s I closeheas e s ya e wa Th t t behindi tha . I t dme could hardly breathe I begge d an , d rais o t themory t ean m eno stear mfo a while. ... At first I seated myself on one of the lower benches built around, after a while getting on the other above. More water was poured t stones ho sucd e o an ,nhth a volum f steao e m arose tha I tcoul t dno I jumpeendur o s , eit d down agai reclined n an half-sittina n di g posture in orde breatho t r e more shormosa a freely n n i I ts timprofus. wa eI e perspiration; again and again steam was raised by pouring water on the d steaan stonesmr ai becam t , ho til t lase a l th extremelt y oppressive. Now and then we poured "water on each other . . . then with boughs everyone's back and loins "were switched till they smarted severely. . . . In about half an hour the people began to depart, at first submitting to a final flagellation, after which cold water was poured upon the body: then all went home as naked as they came. ... I rolled myself in the somd snodi es w a others. " The detailFreswice th f so k bath-chamber poin procedura o t t e having been followed there very similar to that related by Du Chaillu. Burnt broken stones woul buile b dintp u t oa hea p mingled withe th pea n i t corner above the vent, and the fire lighted from beneath. When the heat e stone oth fd bee ha sn raisea glo e o bath-chambewt dth heatd an , r prepare s describeda d e batherth ,o r to batherse o spac ,s th r wa e fo , restricte o holt d d many, would take their place a benc n o sh alone gth back walld watean , r woul e e throwstoneb dth o n product so n e th e necessary steam. The stones were of such a size as would have retained their a considerablhea r fo t ee procestimeth o f s throwin,o s g waten o r to them would be repeated for as long as was necessary, or as their heat remained sufficient flagellatioe Th . n with twigs doubt,o n , "would follow, as that, from all accounts, appears to have been a regular part of the procedure d finallan , a ydouch f colo e d water woul e administeredb d . It is possible that this last act of ablution took place in the closet on the north sidee incline th n froni ,f o t d flags, which would direc e wateth t r to the drain, passing out from beneath them. This last suggestion is put forward tentatively, for the drain was not placed in the position most convenient for such a purpose, nor "would the direction of the water into e bath-chambe e basth th f o e r buildin a soun e gb d arrangement whet ni might "with little trouble have bee e maidirecd th nle no t t drain outside. a bath-hous n I e excavate Dy db r Aage Roussell t Sandnesa , Greenn i s - land,1 a small lateral chamber was likewise found, but in that case it 1 Sandness thed an Neighbouring Farms: Researches into Norse Culture Greenland,n i . 76 - . _ p 80 PROCEEDING SE SOCIETYOTH F , FEBRUAR , 1939Y13 .

actually contained the stove. In the Freswick closet there were no indication fira f e o s ever having been laid, s beeandha ns ,a shown, the stea generates mwa d withi bath-chambee nth r itself. It is highly improbable that the heap of clay deposited in the north- west corner of the chamber had any connection with the bathing arrangements. As it had been levigated, it is likely that it was intended to be manufacture useth n di potteryf eo . The secondary characte e batth hf o rbuildin s evidengwa t froe mth filling up of a doorway in the back wall, which had originally given access to the building adjoining on the east. One other featur f intereso e s discoverewa t e maith nn i d building. From the face of the north wall, at a point 5 feet 6 inches to the east of the entrance, there ran a gutter formed with stones set on edge in a shallow trench, and converging at base, which discharged itself into the sump, and had evidently served as a latrine. Ver relicw yfe s were recovered fro e bathe housmth th r .o e Sucs ha there were consisted of two perforated femur-heads, which had been used as whorls; a discoid perforated object of bone which may have been similarly used; an oblong object of bone', polished and rounded at one extremity; a turnbuckle or snib of bone for a door (PL XLIX, 2); several fragment f mediaevao s l pottery, glaze d unglazedan d s wela ,s varioua l s sherd Vikinf so g cooking pots. Lyin alignmenn gi , timmediatelI wit. hNo ease separated th an t o yt d by the mutual wall which forms the back of the bath-chamber, was another building plan)n (Noo I I ., measurin fee0 g lengt3 n feei 2 t 1 n i t o ht b1 y1 breadth walle th ,whic f so boteasn e ho th h tt enda sides d , an ,wer e much dilapidated. The mutua d originallyha e wesld th wal en s t alread a a l, y stated, been pierced by a doorway somewhat to the north of the centre, and at its sout recesa therd s hen ewa s measurin inche0 g2 s across inche9 , heighn si t from the original floor-level, and 19 inches deep, the back of which was formed with a large upright slab set in the west face of the wall. The e immediatwalth n i l e vicinit f thiyo s recess, which appear havo st e been reconstructed buils i , t with thin flat stones, afte nature interion th ra f eo r flooe wall f thiTh o r .s hous s covereewa d dee midden pi n o refusen d an , feature of interest came to light in the course of its clearance. There were indication firef so s havin t heri thereg d n beet eo theran t bu ,n li e were no signs of a definite central hearth, and it is doubtful if the building had been used as a dwelling. fee2 inchet 6 t A s eastwar mutuae th f do l wall beneatd an , floor-levele hth , earlien e wala th f o lf ro ther exposes buildingp eto wa e dth , lying almost parallel with the former, and returning westward at its south end (see plan, f superioo s wa r t I masonre building e walth . fig e th f 3) th o . l o t yn o s A VIKING SETTLEMENT AT FRESWICK, CAITHNESS. 81

surface d stooan , d a erecheigh feer 2 inches6 fo f tt o t . Unfortunately, exploratioe th house th whicf f neo o t hformei parda t would have entailed destructioe th t considere bathno e s thid th f an sn,wa o d justifiable. No. II only yielded a fragment of the upper stone of a rotary quern of garnetiferous schist a larg, e ovoid pebble chippe e centrth f eacn i do e h side, and an anvil stone. Making contac south-wess it tt wita house I . hth teNo corner shows a , n on the plan, was a range of building in a very dilapidated condition (No. Ill), whic alsd hoha suffered hande spoilermuce th th f t t hsa consisteo I . f do two rooms connected by a doorway in the centre of a mutual wall, and evidentld ha y forme smithyda , with probabl yworkshoa XXXIXL p(P . 2) , The eastmost chamber, the walls of which had been very poorly constructed, measured some 12 feet 6 inches in length, by 10 feet at its east end and 11 feet 3 inches at the west. At one time a doorway had been broken through the west wall of No. I, but subsequently closed, and a door, probably later, opened at the west end of the south wall. flooe coveres Th wa r d with five layer f flaso t stones risin convea o gt x profile, wit e higheshth t e centre poinlowese th n Th i t . t layer consisted of flat, heavy beach-stones, lying on sand; while above them were layers of Caithness stone, neatly fitted to one another, with occasional pockets of midden refuse between. The depth of this paving was 9 inches. Below uppee th r layer there were remain hearta f so h evidently secondarye Th . purpose which this flooring was intended to serve was not obvious, but it would have provide dthoroughla y stable foundatio anviln a r bencr o n,fo h set upon it. south-wese th n I t corner, adjacen doorwae th foo1 o inchet 7 d t yan s from both walls, there was a pit 2 feet 2 inches in diameter and 1 foot deep filled with soil, having countersun centre post-hola th t i n f ki eo e surrounded with packing stones foo 1 inche,0 1 t s deep fro surfac e inche9 mth d n ean si diameter. No corresponding post-hole was found on the opposite side of opposite e doorwayth th n i r e no corne,e chamber th f o r , where thers ewa clayf o d . be a The only relics found in this chamber were three perforated femur-heads, whic beed hha n use whorlss d a whicf o l hal , came fro north-ease mth t corner. The partition wall had been constructed with thin flat stones, and still remaine heigha foo1 o d t inchesf 3 to t . Placed agains t soma t i tfeee2 t north of the doorway was a block of stone, some 8 inches square in section, rising a few inches above floor-level, and firmly set in the sand. Stones, either built int wale or placeth o l this da s one usualle ar , y associated with the position of a seat. The section of the wall to the south of the doorway was founded more deeply than that opposite, and as the latter rested on a midde evidentld ha t ni y subside somo e dt seee th b e n ni y extent ma s a , illustratio XXXIX L doorwae n (P feeTh 2 s t. ywide2) ,wa opened an , n do VOL. LXXIII. 6 82 PROCEEDINGS OF THE SOCIETY, FEBRUARY 13, 1939.

secone pavea th o f t o d de eas d chamberareth en tt aa thi f O s . chamber (PI. XL, 1) the whole of the west wall, and almost the entire length of the south wall, had been removed, and accordingly the exact dimensions were not procurable. But where the walling had survived at the east end, it d measureha fee3 1 dwidthn i t , and, judging e heartfro th e siz mf th o e h extene e soutth th f d ho tprobabld an wallha t i , y measured abou fee1 2 t t in length. The south wall still showed four to five courses of stone rising to a height of 1 foot 9 inches, and towa,rds its west end, as shown on plan beed ha n t i buttressed(fig, 2) . f heavo , w wityro bouldersha lying against its outer face. Occupying the interior of the chamber was a large hearth, measuring som fee0 lengthe1 n i tfee 8 breadtn i y t b westers , it t ha n fee 6 end d t an , at the east where there "were intakes on both sides, covered with a bed of compacted peat-ash to a depth of 13 inches. It had been enclosed on three sides wit hkera longf bo , narrow stoneedgen o t numbea ,s se whicf o r h still remaine hearte situ.n di Th beed hha n distan fee4 t t fro e easd mth an t south walls, and only 2 feet 6 inches from that on the north. The paving in e doorwafron th e divisiof th o t n yi n wale l e edgextendeth th f o eo t p du kerheartho n bd intervenean , t thida e ease sth th end tn t o endA .d ,an sout kerd h en sideb e stoneth , fee2 , inche6 t lengthn si places wa , inched9 s inwards fro generae mth l alignment pavea rean i s thif o rd sdwa an , recess in the hearth, which was free from peat-ash, and measured some 2 feet square in extent. Opposite this, firmly set at the base of the south wall, wa flat-toppesa d stone abou foo1 tbreadt n i t projectind han inchew fe gsa above floor-level, which may have been connected with the anvil. On the north side of the hearth, and at the east end, a flue had been formed, rather ove foo1 r widthn i t , witflao thtw stone t onse s edge paralle kerbe th o .t l When discovered, this had been used as a fireplace, and was filled with kitchen-midde ninnes refuseit t beed rA enha n . t di blocke d wit hnumbea r of thin flat stones standin edgen go , which, whe e trunth e nature th f eo constructio realiseds nwa , wer e originath e foune b l o t dcovere flue th f , o s and were replaced (PL XL, 2). The flue terminated in a slope of hard compacted peat-ash. While no food refuse lay upon the hearth, such material covered the area at the east end of the chamber. Irrespectiv e spacth f eo e occupie e firee characteth th ,e y b th d f o r relics found in this part of the building left no doubt that it had been the smithy of the settlement. Heavy lumps of slag, the residue of bog iroe froor nm whic e irod beeth h ha n n inefficiently extracted, were e slath gnumerousf o y traco bloomerya an n t ther f s d eo bu , di ewa r no , lie among the peat-ash. We found six hones, all but one of the type which have been designated "haunched," from the haunch-like expansions at one end; three objects of iron; and a flat ovoid pebble, faceted on both sides as if by polishing. In addition to these were several objects A VIKING SETTLEMEN T FRESWICKA T , CAITHNESS3 8 .

t necessarilno e founb n sucho i dyt , associations, viz .a hammer-heade d ) foun7 , f bronzo L d n L witpi (P ea hsmall-toothed , single-sided comb (PL XLVII a finge ,d 2)ran ; rin f thingo , flat bronze plate, penannular, tapered e extremitiean th , o 10)d t thess L A L . e(P s three last-mentioned relics all came from the extreme west end, two of them may have been from the site of the wall which could not be identified, while the last- mentioned was from the upcast; they may all thus have belonged to an earlier period. There were also foun a bond e pin ,a sandston e pebble, longitudinally grooved and plano-convex in cross-section, which had probably been used as a plummet. Near it was found a quartz object, which seemed to have served the same purpose. From the vent there came a whorl fashioned from a perforated femur-head. To the northward of the last building, and lying directly east and west, at a distance of some 23 feet from its eastern end', are the remains of a dwellin plan)n go (Nowalle V th ,I whicf . so h have been reduced practi- call o foundatioyt n leveXLIL . (P lThoug1) , h this n i buildin t no s i g parallel alignment with those previously described, the relative levels indicate that it probably was in contemporary occupation in its final stage. Two doorways in the north wall were blocked, which suggests that the house at one time had formed part of another construction, or of a range of buildings extending to the north. t I measured interiorl fee 9 ylengtfee2 n 2 i t1 inche 6 t y hb widthn i s , with a slight reduction in the latter dimension towards the western end. e wallsTh , which appeare periode o belont on l o dal t g, measure fee2 d t 6 inche breadthn si , excep e easth t t wall, which measure feed3 inches3 t . The entrance, 3 feet wide, was through the east wall, a little to the north of the centre of the gable, while in the north wall were two filled-up doorways. As was not unusual in houses of the Viking period, there had been an outbuilding in front of the door, now merely represented by foundations —no doubt to afford protection from the violent winds that frequently north-ease th n o bloa se tfrown i coaste mth . Betwee gable nth thied san a passag beed eha n formed, leadin e maith no gt door, -whic enteres hwa d throug ha doorway fee2 , t wide, just outsid e south-easeth t e anglth f o e dwelling aren fronn I a pavinf thif as .o o t swa g 1 2fee 6 extentfee n i ty b t . A poina t foo1 t inche9 t fron si doorwae mth passage yth crosses ewa y db a massive kerb, which projecte a heighinche0 1 o XLI L t df o (P ts. ,2) At a distance of 3 feet 6 inches from the outer doorway the passage expanded from 2 to 3 feet, and continued at that width to the entrance to the house. kitchen-middena Owinerectioe n o s nortgablth e it d th o g t t f eha en no , there had been a considerable subsidence at that point, which had resulted a stee n i p e passageslopth thibeed n n ei O ha sn . laid several layerf o s paving slabs to reduce the gradient. doorwae norte To th th housee f h o th coverin d f yo an , northmose gth t 84 PROCEEDING S E SOCIETYOTH F , FEBRUAR , 1939Y13 .

portion of the gable, was a box-like enclosure (PI. XLII, 1), measuring 5 feet4 d divide feey an ,b t d lengthwise intequao otw l compartmenty b s two large upright slabs incompletew no , t whicbu , h must originally have risea heigh fee2 o nt f to t abov e box e th floo th ef .o r This enclosure had been formed with a surrounding wall through "which there appears to have been a gap, some 18 inches wide, on to one or both divisions, at the gable side. The inner or south compartment alone was paved. Ther s nothinwa e o indicatt g e purposth e e serve y thesb d e twin com- partments, but they may have been used to contain young animals, or entrance th f o innee d eth stores kerb,a t en ther r s A thia e.wa n slab set on edge and rising to a height of some 6 inches above the floor-level. Withi housee nth , centrae lyinth n gi l axissomewhad an , ease th to t t e "lon th e site centref oth th geo f s fire,wa , " measurin feet 6 fee2 g1 y , b t whic d bee hha e dee n th laipn do deposi f fooo t d refuse that formee th d floor. If it had been originally protected by a kerb, all the stones had been removed. From the face of the north wall, and alongside the hearth, several stones projected, which may have formed brackets for the support obencha f . Occupying e south-easth e doorwat th corner lefe f o th t o yt , on entering n oblona s gwa , area, measurin fee5 g inche6 tfeet4 y ,b s slightly raised d outline e outean , th n r do a settinsid y b ef fla go t stones (PL XLII, 2). This was evidently the site of the bed, and was analogous in position to the platforms met with in the chambers of the Viking buildin t Jarlshofga , Shetland.1 Such bed-platforms indicate thae simplth t e fashio f sleepino n g on bedding spread out on a platform in place of upon a constructed wooden bedstead still prevailedNorsd ol e e worTh . d seng originally meant such bedding rathe itself.d r be tha e 2 nth On the opposite side of the doorway a thin wall extended inwards for a length of 4 feet 6 inches, leaving between it and the north wall a space 2 feet 3 inches wide. There appeared to have been an enclosure of som e south-wes th sor n i t t corner, measuring som fee6 e inche6 t y b s 5 feet, but the remains were very indefinite, and there were also indica- tionpartitioa f o s n wall havin n e wala th f g o lf o bee p nto erectee th n do earlier occupation beneath, which had collapsed when the house fell into ruin. Two short lengths of wall, 1 foot 6 inches apart, projected from the south wall towards its west end, suggesting the remains of a cupboard, objecn a d f cetaceao t an n bone, 5|^ inche lengthn si , founimmedia,te th n di e vicinity, had evidently been half of a turn buckle or snib with which to keep a door closed. othee Th r relics recovered from this house were neither numerous nor important. There were three whorls, two of bone made from femur-heads, stonf o e e on wit d han some simple, nondescript ornamentation around the. 1 Proceedings, vol. Ixx. p. 261. * Aage Eoussell, op. cit., p. 70. A VIKING SETTLEMEN FRESWICKT A T , CAITHNESS5 8 .

perforatio facee on a tapere ; n o n d penannular rin f vergo y thin bronze, of which one-third was amissing; part of a haunched hone; a piece of thin sheet bronze witrivet o , probablit h tw n si y parcauldrona f o t oblono tw ; g pebbles both fractured at one end; an iron rivet with a lozenge-shaped head iron a ; n sam e hooa knifeth d f eko metal an , stona ; e wit hnarroa w longitudinal groove on one face, possibly a sharpening stone; and a piece of uppee th r ston rotara f eo y quern, forme garnetiferouf do s schist found an , d wale th ln i when closeopenin o eastmose tw th de t doorwayth g ou f to e th n so north, als opiercea bonef o r thirA . d whorl made fro mfemur-heaa s dwa discovere northmose th n di t divisio enclosure th f no e outsid doore eth A . certain amount of iron slag was also found in this house. On the level of existine th g wall-hea wes e therd founth s en tt dea wa piecd a mediaeva a f eo l cooking pot, referabl secone datn th ei o et thirteentde th hal f o f h century. Based on the north wall of the house at its western end was a small enclosure in the form of a quadrant of a circle, measuring some 10 feet along norte th entrancee easte hth Th fee bas9 n . o y t eb fee 2 , t wide fros mwa , the north interrupting the contact of the two sides, and beyond it the east wall extende e surfaca distanc th r dn fo feet8 f thi O o ef eo .s enclosure, at the level of the remaining wall-head of No. IV, a paving extended over one-hal areae th f ,fo evidentl latef yo r date. Beneath thi blacs soie sth lwa k closeld an y compacted, with foolittlo n dr e o refus e intermingled. Towards e west th e easth , t wals constructewa l d "without facingd stonesha d an , probably been backed with tur than fo tfeaturee sideTh .thi f so s enclosure, and the condition of the floor, suggested that it had been a small fold such uses ahol o swa dt ewee dth s when they wer emilkede b drive o n t I n .ni which case extensioth e e easth e entranctf th o nwal t mads a l wa eo t e facilitate the herding of the sheep. No relics were found within it. sane th df o hollow ease d e th en tt th f A , o facin p to ge directlth o t n yo sandy bank that ros frop beache u mth , wer foundatione eth larga f o se enclosure (plan 2, No. V), which can only have been the " naust" or boat-shed into which the boats were drawn in autumn by means of rollers and ropes over a slipway up from the shore, to be sheltered from the storms during wintere th tarred an reconditioned , dan followine th r dfo g year's voyages. The "naust," as will be seen from the plan, had been an irregular U-shaped construction open toward seae sth , wit norte hth h wall housef so s Nos. I and II forming the south leg of the letter, and the other lying 26 feet to the north. Forming the centre part of the back wall was a straight lengt foundationf ho fee 2 lengt n 1 fee,i 2 t inche6 d t han breadthn si . Else- where the foundations were irregular and massive, formed with large heavy boulders. Whil e seawareth d portio norte owin, th U f h no e o gt lim th f bo more th lesr eo s level characte heave s surfacit th f d yo r ean boulder s with which its outline was formed, suggested a slip on which a boat might have been constructed positiv,o n ther s ewa e evidenc thir efo s forthcomingn I . 86 PROCEEDINGS OF THE SOCIETY, FEBRUARY 13, 1939.

norte th hopee sidth nd f e o hearthspaca be site f a th eo fron n s n ei o wa t of clay, burnt to a brick red, and from this was recovered an iron ship rivet, much corroded, while remains of several others were picked up in the immediate neighbourhood. At the upper edge of the bank near the centre there was a setting of large stones extending across the area for some 4 or 5 feet, with a single stone on end sunk deeply in rear of them. From the position of these stones it seemed probable that they formed the site of a windlass for drawing up the boats. On the slope in front, and towards its upper end, lay three or four large flat stones in a line, which were possibly the remains of a gangway. At the seaward end of the "naust," adjacent to the site of the hearth, r Simother founs M ewa y n b d Bremner t upright se ,e san th dn ,i wits hit upper edge just protruding e pieceth f d o s an e cookin , th , 4) t (PI, gpo LI . e smal th reconstructedw l no cii , p3) , (PILI . , were subsequently als- ore covered from this spot. On this area there was also found a large quartz pebble, 4 inches in length (similar to others found on the site to be discussed later on), flat on_one side, with a partial perforation at one end, and probably intended for use as a plummet. As previously mentioned, the historical evidence in the Sagas shows that a Viking settlement had existed at Freswick from a period at least as early as the eleventh century, and to judge from the standing of those whose names figur connection ei residence th s n wa peopl f witt eo i , himportf it eo - ance. While nothing has emerged in the course of the excavation that would afford a clue to the terminus a quo, we are able from certain relics to assume a fairly sure date for the terminus ad quern, the latest occupation e grouoth f f buildingo p s which have been discusse e foregointh n o d g pages. Associated with these buildings have been found sundry sherds of mediaeval pottery attributewhice b n secone hca th do t dthirteent e halth f fo h century. Among sher e the on particular n mdi s i , wall-heae founth n do f do No. IV, which presents a close parallel to a sherd found at Bayleigh Castle, Essex e occupatioth , f whico n h ceased some time befor e yeath e r 1277. The other piece of evidence is a silver penny of the reign of Henry III. of England, Moneyer "Willem on Lund," Mint, London, dating from about 1258—1272, 1e surfac \vhice blown-outh th s picke n f o ho ewa p du t hollow. This is slender evidence for dating, but it is strengthened by the negative evidence, supplie face th t y thad b relic o n t , which coul assignee db a o dt later date, was found. occupatione th f o d en ,e somth r Ifefo , date between 125 127d 0y an 0ma be accepted, thepossibls i t ni suggeso et abandonmene causa tth r efo f o t

Freswic Norsemene th yea e y th kb r n I 1264 .yeae th , r subsequene th o t t indebtem a Roberr I M 1 o dt t e Roya Kerth f o rl Scottish Museu e identificatioth r mfo f thino s coin. A VIKING SETTLEMEN T FRESWICKA T , CAITHNESS7 8 .

battl Largsf eo , Alexander III. sen armn a t y into Caithnes exaco st fina t e fro e peoplmth e because thesubmitted yha Kino dt g Haako f Norwano y in the previous year. The army stayed in Caithness until the autumn and

departe seay db , taking wit1 h them treasure th e thecollectedd yha , much of which, however loss twa , durin e coursgth theif eo r voyage Dugaldo t , , King in the South Isles. It seems a fair assumption that at the hands of Alexander's expeditio e Vikingnth s were driven from Freswicks a r fo , seafarers they were probably in a better position than most of the inhabitants of the county to render assistance to the Norwegian king. As stated above, the buildings explored have been divided into three groups according to their periods, estimated by their relative depths from the surface. Group A, the constituents of which have been described in detail above, consisted of (1) two oblong buildings which may at one time have been dwellings, but which in their latest stage had in the one case been use contaio dt nbath-housea othee th n ri possibld an , y alloweo dt become derelict r beeo , n e smithy intern usea a storeth ) s d da -(2 an ;, communicated chamber n isolatea ) (3 ; d dwelling n parallei t no , l alignment, and possibly of slightly earlier construction; and (4) the "naust" or boat-shed. The group appears to have been part of a settlement on a courtyar e dGreenlan e casth complexa plan f th t eo n no ,i s da , farms, and such as also appears to have been the case in Shetland. We have here the bath and the smiddy, situated on the south wing and somewhat isolated, as being both potential sources of conflagration. The boat-shed has occupied the east side, and the dwelling has been erected on the north-west. fare mTh buildings, consistin e stablesth f go , byre, dairy othed an , r out- t housesrepresentedno e ar , , except, perhaps annee e case th th th f o exo t n i , dwelling, and it may be presumed that they occupied the north side, and so as far as any part remains they probably lie beneath a bank of sand, which rise heigha morfeeo d s6 t f an tto e abov hollowe flooe leveth e th ef th o f r lo . The dwelling-house in its details does not conform exactly to the typical Vikine housth f eo g period s revealea , t Jarlshofda , Shetland 1934,n i , r o

thao t t expose Groun di shortle pb Bo t , y described. indicao Thern s -ewa 2 tion along either side of the platform, or dais, on which the benches were placed, and where the inmates sat and took their meals, nor was there found burnte th f o ovey brokey an n an floor e sunr th n no n k,stonei s that accom- pany such cooking arrangements, both features housee founth n di s above referred to. Nor was there any remains of a high seat, either at the side or earlie e hearthe th th n f i r o s indicatiohousy ,a d an en Grou r n ee i no th , t npB a of the hearth having been originally contained within a kerb. The amount of kitchen refuse covering the floor pointed to rather filthy conditions of occupancy during this last phase, and suggested that slum conditions may . Ferguson1J , Alexander III., . 120p . 2 Proceedings, vol. Ixix . 265p , . 88 PROCEEDINGS OF THE SOCIETY, FEBRUARY 13, 1939.

have arise t Freswicna k befor e settlementh e t finally close s existenceit d . paucite Th relicf yo s produced fro carefue mth l examinatio mucf no h kitchen- midden refuse pointee furtheth o t dr conclusion thae residentth t t a s "Thraswick" had not formed a wealthy community.

GROUP B. Group B, situated at the west end of the hollow, was covered by a much greater depth of sand than were the buildings of Group A, and part of it actually lay beneath the foundations of No. IV of the latter group.

FIRST BUILDING X • FIRE-HOLE

LATER Y= KILN

POST-HOLE

Fig. 4. Plan of Buildings in Group B (VI) and C (VII).

It consisted, as shown on the plan (fig. 4), of a dwelling-house (No. VI), a small chamber constructed agains s easit t t wallrepresented an , y b d fragmentary remain f somso e piecegreatea flueo a wallinf T f so o . r d gan extent even than in the case of the previous group there was evidence of reconstruction, and of earlier buildings having existed in the immediate vicinity. The three members of the group lay approximately east and west, with the dwelling-house at the end, a little out of alignment with the others. house Th e was, however, itself secondara y constructio e siteth ,n forno s a , wil observee b l referency db beed plane ha th n t i o ,erecteet d againstd an , with its south wall partially embedded in, the wall of an earlier structure, much more massiv charactern ei -whicd an belonged , hha buildina o dt f go greater length. This wall, which starts at the western jamb of a doorway easte onth exposes i , distanca r feetd6 fo 5 f , eo wit hwidta somf ho feete5 , A VIKING SETTLEMEN T FRESWICKA T , CAITHNESS9 8 . obviouss lattee wa th t i t rs a , a point d an ,, thaVI t . i t No f o wes e d th en to t had been broken through and faced up, exploration was carried on into a high ban e sam sanf th ko en di alignmen "westwardo t t whad an , t appeared e samth e b wallocateds o t wa l sinkingy b , t intervalsa s a farthe r fo , r distanc feet5 2 f , o emakin ga tota l lengtfeet1 8 f . ho Fro e detailmth s remainin walle th doorwae f dooa , o th ease f d rgo th e checen tt th ya n ko west side and a projecting kerb opposite the rebate, it is evident that the building, of which this wall formed a part, must have been situated to the northward seconA . d doorway, also primary, som feee3 inche6 t s wide, had been constructed some 13 feet 6 inches from the west end as exposed, but had been subsequently blocked, as after stated. On no part of the area cleare vicinite th n di y wer furthey ean r remains foun thif do s early building, nor were there any traces revealed by trenching. Any suggestion that it was part of an enclosing wall to the settlement may, I think, be ruled out excavatioe t witth face me n hi th s t y thaGrouf wa b no traco t i n t f peC o at the eastern end of the hollow, as well as by the domestic character of the doorways themselves. It was very apparent throughout the course of the excavation that whe a nfres h buildin ge erected b "wa o t s , stones froy abandoneman d structure were fully utilised. As the material was invariably boulders from the adjacent shore, laid in dry-stone fashion, it suffered no deteriora- tion by use. The dwelling-house (PI. XLIII, 1), except as above stated on the south, was surrounded by walls of an average thickness of 3 feet, and a similar height on the south where highest, but greatly reduced elsewhere. It measured interiorly some 36 feet 6 inches in length, and 13 feet in ease th feet2 t 1 whera tt pre-existine bu eth "wese widt d th en tt ha g wall d beeha n supplemente n interioa y b d r facinge entrancTh d bee. ha en a doorwa y b y immediatel e blocke e th "wes th f o to dyt doorway, fee3 t "wide, crosse projectina y db g sil inche7 l s high backed outside an ,th o dt e larga y b e rounded boulder (PIinside th . n XLIIIt gavestei a O y p e. b 2) , to the floor-level. Before the doorway on the outside was an area of paving e interioe houseth Th f . o r, whe e blowth n n sand beeha dn removed founcovereds e b wa ,o dt , especiall ease th tt yenda , wit hdeea p deposit of kitchen-midden refuse, in which limpet-shells predominated—a barre ns relica deposir sfa wers a t e concerned t i yielde r fo , d practically nothing smalexcepw fe l a tfragment f cookino s g pots. Immediately above floor-level was a stratum, 5 inches deep, of brown humus with a thin layer of sand in the middle of it, representing probably a period when vegetation had spread over the floor of the abandoned dwelling. e middle floorth th fee0 n f 1 I e,o t distan d 1t5 an fro feee easd mth ten t from the west, lay the hearth, 12 feet in length by some 8 feet 6 inches breadthn i , with e stonesomupstandinn th a f f eo o s g kerb still remaining 90 PROCEEDING S E SOCIETYOTH F , FEBRUAR , 1939Y13 .

makine soute o on eastn th d d ghreturan northa e southe an , th th t o nt a ,- west corner. Peat-as depta footupoy 1 o t hf la h o t ,ni representino gtw period f occupationo s , distinguishabl n intercalatea y b e d laye f diso r - coloured sand. Along the north wall there was still traceable, for a breadth of some 3 feet, the site of the pallr, or dais. This was readily recognisabl e e puritsanth th df y yo b ewhic y almose surfacehla th t a t , having been protected durin e periogth f occupatioo d a flooring y nb r o , coverin f somgo e carpeting material, possibl f rusheyo r heathero s . This conditio marken i e sans th dwa f d no contrase e flooth f th thao t o tr n o t dwelling, which was uniformly discoloured to a depth of several inches. Toward lono tw gwes e d flasth en t stonesedgen o t , se wer, e probable yth remains of a kerb that had extended all along the edge similar to the remains found at Jarlshof.1 On the south side, and to the east of the centre, there extended for a distance of 6 feet a line of similar stones, which though only som inche0 2 e s distant froe existinmth g wall-face were the same distance out from the face of the original wall as were those fro e nortmth h walld accordinglan ,e regarde b y s havinma ya d g formed edge-stones of a dais which had existed along that side in the original house. * At the east end, between the kerb of the hearth and the end wall, was platforw lo a m som fee5 e inche9 t s lond slightlan g y les n breadthi s , with a row of boulders along the front, and a narrower platform rising 2 fee 9 inche , te it bac th f e foremos o skt th a broadn O f theso t. e platforms, for the back one seems too narrow, stood the high seat reserved for the master of the house. On the south side of the high seat platform, covere uppee th y rdb ston rotara f eo y quern (PI a post . s XLVI-wa , 1) , hole, some 18 inches in depth and 9 inches in diameter. There were wedge stones at the mouth of the hole, and on one side a packing of clay, while a fragmen f carboniseo t d willo s recoverewwa d fro e bottommth e Th . old Norse name for the high seat was ondvegi, derived from the name of the two thick pillars between which it was placed. These pillars, in Viking pagan times, adorned with carved representations of the gods, were regarde s holy a dd whe e an Norwegia, th n n colonists emigrateo t d Iceland they took their high-seat pillars with them. The position of

the above-mentioned post-hole suggests that it may hav2 e contained such a pillar, while another post-hol e nortth n ho e side inche7 1 , s deed pan 7 inches in diameter, may have held its corresponding fellow. On the other hand t i shoul, e noteb d e nortd th tha n ho t side four post-holes were discovered mor r les o t eequaa s l distances froe anothermon n i , which series occurred that adjacent to the site of the high seat. On e soutth h side, althoug e post-hole hth t similarl no e ar sn alignment i y ,

1 Proceedings, vol. Ixix . 273p . . 2 Falk and Gordon, Scandinavian Arcliaiology, p. 324. A VIKING SETTLEMEN FRESWICKT A T , CAITHNESS1 9 .

there is a series also equidistant, and accordingly the pillars they containe n botdi h seriey havma se been merely use n suppori de th f o t roof. alln I , alon e nortgth h side four post-holes were recovere fee7 o tt d6 apart, and along the south side five. The three in the centre were more or less equidistant feet7 o t , whil6 t e post-hola ,eth e already referreo t d besid hige th site hf eth eo somseas wa tfeee 8 t distant fro neighbours mit , and quite out of alignment. The fifth was only 4 feet west from the fourth. The post-holes on the south side, with the exceptions of the firsd thirdan t , were placed neare e walth r e lopposit th tha n no e side, while the fourth and fifth were situated one on each side of the blocked doorwaysomd ha f the ei d connectios ya ha , n wit. hit In a number of instances the holes had been purposely preserved, casee s on alreada , n i y mentioned coveriny b , e moutgth h wit ha quer n stone, and in others by the insertion of pointed boulders, thus indicating that when this hous abandones ewa d ther minde ideth n e a n ai s stil wa l s owneroit f s tha t mighi t t agai roofee nb occupiedd an dexistence Th . e of such an intention may have saved it from the demolition that overtook other building s neighbourhoodit n si . On the north side, some 12 feet distant from the west wall, and just in front of where the edge of the dais had been, was a cooking oven. It was much dilapidatede stoneth f o s thaw onlformed d fe an ha t,ya e dth sides remained, while the sole at the bottom had evidently been removed. However o dispet , douby an l t there might hav e purposth e o beet es na of the hole, several burnt broken stones remained within it. Its dimension d beeha sn apparentl foo 1 s dept inchesfeey2 5 it y to t b t d h7 an , e simila8th inchesn i r s oveA . n foun t a Jarlshof,d e slath 1 b forming the back was sloped backwards. The method of cooking was that still practise y somb d e primitive tribese brokeTh .n stones, heatee th n i d fire nearby, were deposite e botto e holeth th n di f m. o They were then covered wit ha laye f grasso r r otheo , r vegetable matter, suc s leavesha , f whic o s place p e cooked hwa b e footo dth o e t do nth . Further layers of grass and heated stones were superimposed, and on the top, to keep in the heat, was placed a covering of turf. On the south side of the house, at 5 feet distant from the east wall, and little more than 1 foot out from the face of the south wall, was another stone-lined pit whicn ,o e slabhth s remaine thren do e sides t measureI . d 2 feet square, and had a depth of 1 foot 4 inches. It was not floored, nor did it contain peat-ash, nor any burnt broken stones to indicate that t alsi beed oha n usecookina s da g numbeovena t bu ,piece f o r brokef so n pot 'were found within it. In its vicinity, along the margin of the central hearth, wher kere a bee d bha n removed, ther substitutes ewa elongaten da d 1 Proceedings, vol. Ixix. p. 277. 92 PROCEEDINGS OF THE SOCIETY, FEBRUARY 13, 1939.

hea f burnpo t broken stones, whic d probablhha y come fro ma cookin g hole nearby (PI. XLIV. 1) , In the south-west corner there was a suggestion, from the placing of the few flat stones remaining, that there may have been a bed-platform some 5 feet square on this spot. For a distance of 7 feet 7 inches from e westh te nort wallth n ho , side only , hara ther s d ewa compacte d floor of peat-as d clay e characteristihan th , c floorin e Vikinth f o gg dwellings, and the absence of this flooring from the south-west angle gives strength to suppositioe th n that that arebeed aha n occupie bed-platforma y db . In the centre of the "west wall there was an opening, flanked on either side wituprighe hon t boulder, formin e mout gth venta f ho , pavedd an , extendin e westh thearte o g t edg th f eo h (PI side . XLIVeTh stone . 2) , s e venoth f t remaine r almosdfo s fulit t l distanc e southth t excepn ebu o , t stoneo e openinth tw t r sa o ge fothe on beed r yha n entirely removed from the north side. Judging from the complete portion, the breadth must have been about 1 foot and the depth some 5 to 6 inches. Some arrangement must have been provide r closindfo e intakee openinggth th r fo ,s a , remaining at present, is quite unnecessarily large. At some 20 feet distant from the west end the foundations of a secondary wall, 2 feet 6 inches wide, not bonded into either side wall, crosse a distanc t e interio a fee2 th d d XLVf L to ean (P fror , s m,it 1) junction with the north wall a narrow doorway had been made through it to the interior. It ran from a point on the south just where the reduction in the height of the wall occurred. In the interior in connection with this later wall lengte e heartth ,th beed f ho hha n somewhat reduced squara y b pavinf eo g laid upo. nit In the inner face of the south wall, towards the interior, two distinct style f masonro s e apparenyar XLIVL e west(P , tth tha 1) ,o , t t formed with roughly shaped blocks , e originabein th e fac e earlief th o ge th d l an r wall; that to the east, consisting mostly of boulders, being the construction of the later house and built against the former. In the original wall fee9 1 t t froa e wesmth tfoo 1 end inche5 d t an , s above floor-level, there wa neatla s y fashioned recess, measuring acros e fronth s , inche7 t5 y b s and extending inwards with gradually diminishing breadth to a depth oinches6 f . e examinatioTh e mas th f kitchen-midde f o sn o n refuse that fillee th d eastern hal thif o f s house eved an ,n extended northe ove e walth th ro t l, yielded almos o relicsn t , except some fragment f cookino s ge th pot f o s usual characte e relic a belt-chapTh sd . founan r5) , f bronzdL o e L (P e e floor-levelonth immediateld an , y abov , consisteeit f para com do f o t b (PL XLVII, 3); three broken pieces of hones of the haunched variety; bono tw e whorls made fro e headmth f femurso s shoro tw ;t metacarpal bone f pigo s , perforaten a d; nea ) e centr 14 th rXLVIIIL d (P e an 2 1 , A VIKING SETTLEMEN T FRESWICKA T , CAITHNESS3 9 .

awl of bone; an iron knife-blade and two or three unimportant objects of iron; three pieces of thin bronze, two of them with rivet holes in them, parts of a cauldron; a small cubical block of quartz, grooved across each face for a cord, probably to be used as a plummet, or weight; three stone pot-lids of different sizes; and the upper stone of a rotary quern of unusual form, being approximately pear-shaped, with the perforation for the handle in the narrow end, found as stated, covering a post-hole. At the e earlth bas e f soutyo e th wal n ho l ther s founewa a piecda larg f o ee steatite urn, and the polished tine of a red deer antler. The further remains of buildings in this group are fragmentary, and difficul determine.o t t - Immediatel e dwelling-hous ease th f th o t o yt s ewa aren a a enclose thren do e sides only, measuring som feee8 t from norto ht south expandind an , g fro mwidta fee5 inchef bace 9 tho th nortr kt o s a h end to 8 feet at the front. The west wall, which had been reconstructed in part, to the extent of the reconstructed portion, formed its west flank. The north wall was an earlier building than the last, and the east side was formee westh e tthiry b dwalth f do l constructio e groupth n ni . There was no indication of any wall, or doorway, having closed the south end, nor was there any access to the area from either of the adjacent buildings. It may possibly have been intended for an open shed. On the floor, however, toward e soutth s h end, there were indication n interioa f o s r constructio f somno e remain e kindlineo Th largtw f s.o f so e flat boulders, som 3 feee te see b n plaapart o ny n ma ,(fig ) projectin.4 g froe mth east and west walls, enclosing a space some 7 feet 6 inches by 3 feet, which beed ha n paved. f roundeo Beneat d be e pavin da hth water-wor s gwa n stone r drainagee siltefo sth dd an materia, l between them showed that e arrangementh beed ha tn effective. Possibly this bee a e sit sha f nth eo store closet. Still more perplexing tha e lasth nt wer e remainth e s adjoininn o g the east. Here again there had been walls on three sides of an enclosed arearepresentew no , fairla y db y good wal foo1 inchel0 1 t se th hig n ho west, a much reduced wall on the south, and a mere foundation on the east. It "will be observed on reference to the plan (fig. 4) that the south-east corne beed rha n rounded, \vhic unusuals hi . This are crosses ai foo1 inchet 6 tda s back fro fronte mth soutr o , h wall, wala heavief y o lb r construction fee3 , t wide, which join wese sth t wall with a butt-joint air-ventn A . foo1 , t widd traceablan e a distanc r fo e f o e 1northwarde 4 th fee o t t carries i , d through this wal alignmenn i l t witha corresponding openin soute th n hg i spac e wallTh e. betweewallo tw se nth is paved t indicatiothero n bu , s ei n tha a constructet d duc acrosd le t , it s noeasweste e opes rth th t tha t t . wa endna a t certainl i t t i ts ,a no s ywa The vent is formed with flat boulders on either side, and is still carrying its covers for a few feet at the north end. Adjacent to its termination 94 PROCEEDING E SOCIETYTH F O S , FEBRUAR , 1939Y13 .

there was a bed of peat-ash, but not of sufficient extent to induce the belief that it had been associated with the flue, but the similarity between this duc thad an tt proceedin adjacene gth f fro o wes e md th t en tdwelling , which is believe o havt d e bee a nflue , indicate that thed serveyha a dsimila r purpose. There has been so much reconstruction, however, of the buildings on this site that it seems impossible to determine the original purpose of any remaining fragment. Onl relifounde s yon thacd wa an ,t among midden nortrefuse th f o ho et the westmost section of cross-wall. This was the portion of a, penannular brooc bronzf ho e wit hzoomorphia c termina XLVIIL l 1(P , 6). Unconnected with the constructions forming Group B were other fragmentary remains in the immediate neighbourhood. From the front of the doorway at the eastern extremity of the early wall against which hous . VI eNo "ha d been erected, ther uncoveres ewa dshora t lengt f wallinho g runnin a south-easterl n gi y direction t absolutelbu , n yi isolation owinremovale ascertainables a th r o gt fa s a ,l otheal f ro , portions of the structure to which it had belonged. To the west of this lay a large kitchen-midden which, on examination, produced the greater part of a bone case for a long comb (PI. XLVII, 1).

GROU. C P This group, lying at a lower depth than Group B and so probably of earlier date situates n extreme i wa , th hollowy e t la dth a f d eo an eas,d en t a nort souto ht h direction t consisteI . e southerth f do n portio "whaf no t appeared to have been a range of buildings (fig. 3), much dilapidated, and passing away to the northward beneath the high bank of sand which forms boundare th holloe th f thaywo n i t direction. It will be observed from the plan that the walls had been thicker than in the later buildings, measuring 4 feet across, the same width as the early wall utilise parn di foro t te sout mth h . sid Thef housVI d o e. yha No e been constructed in characteristic Viking fashion with a core of compacted earth between two facings of stone. At a point in the west wall the skull

of a small whale had been inserted; Underneath the building lay the remains of an earlier structure which had been formed with walls of wattle and daub erected on a bed of clay 5 inches thick. The carbonised wattles and baked clay, which lay to a depth of 9 inches between the later wall and the primary clay foundation, showed that the house had been destroyed by fire. An examination of the carbonised wattles revealed that they had been of willow, a fact which supplie e nature thicke th a clusth o f t eeo t mentioned above s havina , g existed in the neighbourhood in Saga times. southmose Th t buildin range th f ego consiste a chambe f do r measuring 1 See p. 100. A VIKING SETTLEMENT AT FRESWICK, CAITHNESS. 95

18 feet by 14, which had probably been a barn. A doorway, which appeared havo t e been som feee3 t wide gived ha , n accesinterioe th o st r fro wese mth t by a paved passage through the wall. The north side was formed by a broad partition wall, through which there had been a narrow doorway leading to an adjacent chamber, and subsequently closed. The floor consisted of clay and peat-ash beaten to a hard consistency, and on it there -were no remains of a central hearth, nor of any of the characteristic features of a dwelling. In the south-east corner a kiln had been constructed, obviously for drying grain or fodder, and the rounding of the corner in which it had been placed, both externall internallyd yan , indicatee para th df s o ttha wa t i t original building .adjacene th Ove f remaine o th kile rd t th n wallsan f so , as well as over the greater part of the interior, lay a mass of clay, which seemed to have been used as a core to the walls, or for some other indefinite purpose in the structure. The kiln measured som feee9 t fro bace intake vene th mth th k o t f t eo of the chamber, and had expanded from 14 inches or thereby at the former position to 3 feet 3 inches at the widest point at the inner end, where the width appeare havo dt e been reduce linina y side e df b stoneg o s on a , n so shown in the illustration (PL XLV, 2), possibly for the purpose of carrying s coveree flu a th e floofloorwa e a deptf Th o o rdt f .severa ho l inches with peat-ash. In the south-west corner of the "barn," directly opposite the mouth of the flue, a small rectangular basin had been formed by excavation, with large stones place floore paves d th wallswa o againsan n , do t tw I .e th t e sideth s were luted with cla thao s y t i coult d contain e waterth n I . opposite north-wesr o , t corner, smalthera s elwa basin-shaped hollow, sunk som inchee6 s below floor-level, and, wit hdiametea f abouo r feet2 t , lined l rounal d with clafloored yan d with flat stoneinterioe XLVIL Th s(P . r2) , was filled with peat-as d kitchen-middehan n refuse sigo t n ther ns bu , ewa of the action of heat on the clay. Beyon e partitioth d e nortn th e buildin walth hn o s mucl wa g h dilapidated wese Th t . wall was" traceable furthea r fo , r distanc fee6 3 tf eo fro e cross-wallmth a lin f tumbley eo b , d boulders e easth t d walr an , fo l a shorter distance. From the latter, at a distance of some 2 feet 6 inches fro built-ue mth p doorway vera , y narrow wall projecte shora r dfo t distance terminatios int it interiore t oth a d an , n appeare havo dt e abuttee th n do circumference of a round hearth, or the remains of a second kiln, constructed against the north face of the partition wall. All that remained of this construction was an outline of flat stones set on the circumference and not contiguous. In the course of the exploration of this last-mentioned building, e earlth y wall whic- un h bate s passeth ease hwa f th o t o dt I acrosI . sNo covered and was found to be part of an earlier building than No. VII. 96 PROCEEDING E SOCIETYTH F O S , FEBRUAR , 1939Y13 .

Very few relics came from this group. Such as there were consisted o numbefa sherd f usuae ro th f slo hand-made pottery stria , thif po n bronze, obviously a patch for a cauldron, and which had been affixed with paper- fastener rivets, thre whicf eo h still remained attachedf o lona n d gpi an ; bone piecesn i , onld an ,y partially fashioned.

THE RELICS FOUND THROUGHOUT THE EXCAVATIONS. The relics found wer t e wholnumerousth eno n eo werd an ,e characteristic of the later Viking period in Scobland.

Bone. The objects fashioned from bone form the largest category. Combs (Small).—There three ar e combs (PI. XLVII, 2—4) l incompleteal , . Two of them (Nos. 2 and 4) have the teeth on one side only, but the third a shor w tbo subsidiare th f o p yto combe th n , o furnishe d ha s dha (No ) 3 . with finer teeth, and which appears to have extended for only a short distance along the centre. The beak-like ends of this last-mentioned comb, both above and below, are ornamented with a single double concentric dotcircla d .ean Of the other two, one (No. 2) is ornamented with two incised parallel lines extending the length of the bow, enclosing between them the row of rivets that hold both sides together teethe th placd n i , an , e between them. The other t comdecorateno ways y bi an .n di wale th f f o l Noo camp 2 . to e founs froth e smithywa n mdo th 4 . No ; house No. II; and No. 3 came from the hearth in No. VI. type completf Th eo e double comb whict representeno s hi d her probs ei - ably referabl thirteente th o et h century founs however, a wa n As do 3 . . No , earlien a sit f eo r period doed t conforan ,s no e double-toothe th mo t d comb type, it is probably referable to an earlier date. Its bowed or beaked ends, with doubl decoratiot e circldo d ean n engraved upon them forn i , m resemble those feature larga n o se single comb foun Bergen,n i d d datinan 1 g from e commencementh e thirteentth f o t h century. Both singl completd ean e double combs fro e Vikinmth d post-Vikinan g g periodf frequeno e ar s t occurrence owing to the fashion then in vogue for the men to wear their hair coma long t thif bo bu , s form appearOsle th unusuale ob n i o s t A . comb, the larger teeth are rounded, not rectangular in section as is usually the case in the later combs. Comb-case.—This object (PI. XLVII , whic pieces1) n ,i s hincomd i an , - plete, has consisted of four tapered strips of bone each 5^ inches in length, of which two broader pieces, ^ of an inch in breadth at the centre, formed uppee th othee stripsth o rcasee pard rincth tw n a f an ,breadt , n to f h i -fgo h 1 Sigurd Grieg, Middelalderske By fund fra Oslo og Bergen, p. 232.

A VIKING SETTLEMEN FRESWICKT A T , CAITHNESS7 9 .

at centre and tapering more than the larger pieces to either end, formed the lower parts. The strips forming each pair were placed opposite to one another with the larger strips uppermost, and fastened at one end by single iron rivets to a plate of bone about \ inch square, inserted between the ends so as to leave space for the comb, and projecting to the extent of one-half its length beyond platee On ,. imperfect, remainss measureha d an , d about 1 inch by \ inch. The other part of this plate has been pierced by two oval f whico holes e s stilhi on , l complete d showan , s considerable evidencf o e s outeweait t a r edge. Through these hole passecora s dwa enablo dt e ownee th carro rt coms ycasa hi n bei suspended fro necks mhi . All four strip plano-convee sar section xi decorated nan d wit hseriea s of lightly incised parallel lines, between the outer pairs of which, in each d circlan et caseornamentdo f a o poin, t occurw A tro . som a s e f •o § incn e uppea th h f baco r strid k en pfro e whermth e remainineth g plate occurs, both strips have been perforated with circular holes, the one

measurin incn g a diametern -i hf 3%othee o th incho n d ra T ^f an ,. o prevent the comb falling out of the case when it was worn, some temporary fastening was necessary, and this was probably supplied by a tapered pin of bone or wood, which was passed through these holes. The size of the rivet hole still remaining at the opposite end of one strip makes it quite clear that the purpose of these larger perforations was not to hold rivets, an otheo thadn e us rn that suggested appear meeo st case e th t . complete Th e lengte comth f hbo withi e cas s beenth eha n 4-g- inches. Among the combs found at Jarlshof 1 there is one with straight ends measuring 4| inches in length and pierced at one end obviously for such a pin, while perforations at one end of other combs of the period are not uncommon. a smalcas A r 2 fo el comb, foun n Osloi d , shows also tw o perforations for a cord in the bone-plate at one end.3 In the long comb found on the Links of Skaill, Orkney, and preserved in the Museum, there e perforationar s throug e plateth h t eithea s perforatioe r on end d ,an n only at one end of the comb-case. Such an arrangement would obviously save troubl makinn ei t gimmateriai - re e com s th whico t y bwa s a lh wa placed in its case. As mentioned above e remainth , f thio s s comb-case were founa n i d kitchen-midde f earlo n y date, whic respecn hi t tha t appearei t e b o t d situated within the limits of an earlier house was probably not referable to the earliest occupation of the site. It was, however, obviously earlier datn i e tha founds whicnn i wa dwellin, come 3 . . hth VI . bNo gNo Dress Fasteners.—Four object XLVIIIL (P s , 11—14), thus generally designated, made from metacarpals of pig, and perforated in the centre of their length, were found. Three of the four came from the earlier

1 Proceedings, vol. Ixix. p. 298, figs. 31-34. 2 Sigurd Grieg cit.,. op ,. 185 p , figs. 182, etc. 3 Ibid., . 226p , fig. 187. VOL. LXXIII. 7 98 PROCEEDINGS OF THE SOCIETY, FEBRUARY, 13, 1939.

buildings—twI VI . froe oNo e floo on mfroth f o d rm an dwellinI V . gNo —whil e fourta surfac th es wa h e find. Though such object usualle ar s y termed "dress fasteners," there is a difficulty in accepting this explanation of their use thin ,i ratey s casan ,t efroa absence msigny th an weaf so f eo r around the edges of the perforations such as might be expected from the friction of a cord. Similar objects made either from tubular fragments of bone, or from complete bones as in this case, have been found on sites ranging over a long period of time from the Grlastonbury Lake Village 1 to Viking times. s recovereOnwa e d froe Vikinmth g hous t Jarlshofa e , Shetland,d an 2 another, found in a woman's grave of Viking period at Carn nan Bharrick, Isl f Oransayo e s preservei , e Nationath n i d l Museu f Antiquitiesmo . This last was found associated with a pair of iron shears, and, typical details of a woman's attire, a pair of oval brooches of bronze and a pin wit hmovabla e ring head. If they had really been dress fasteners, it is singular that they do not occu Vikinn o r g sites more frequently onle ar y d representean , isolatey db d examples. Pins, Piercers Bodkins.r o Pins.—There e amonar e relicgth s three small pinXLVIIIL (P s , 7-9), with rounded heads, varyin lengtn gi h from IYS inch to lf-^ inch. Two of them taper directly to the point, while e thirth d show a definits e swelling toward e pointth s , indicatins it g probabl n somi e e us e textile fabric shorA . t baluster-shaped object (PL XLVIII, 10)n inca length,n f hi -fo - s apparentlha , y bee e heanth d of another pin. All these objects were surface finds. Piercers Awls.r o —Of these ther f whiceighte o e ar illustrate e x hsi ar , d in PL XLVIII. Of the others, one is represented merely by the upper othee partialla th y endd b r an , y shaped objec procesn i t manufacturef so . The largest example (No. 4), made from the cannon bone of an ox, or deer, measures 5-g-f inches in length. It has been much smoothed by use at the pointe founs d wa end dd e closewhil,an th f eo d openine doorwayon t gou s on the north wall of No. IV. Another, with a certain amount of artistry in its production (No. 3), has been likewise fashioned from a cannon bone founs higa wa t r deerh da o t I , leve. ox . VIIclearinn i l n oNo a .f t gou No. 5, measuring 2-g- inches in length, made from the cannon bone of a , measurin6 . No sheep . ge floor-leve founs VI th 4^ . -wa n , incheo dNo f o l s lengthn i founs kitchen-middea wa , n di . entrance IV fronn nth i . f o tNo o et Included wit e bonth h f polisheeo piercere on d s i sdeer-horn t no , illustrated, made fro broe mdeer'th d wre tincoarsea f s o f antlereo s i t rI . form, wit hlargea rfouns poine otherswa th d d f to clearinan tha , y nan g off sand outside on the north-east of No. VI. Eyed Piercers.—There are two eyed bodkins, or needles (PL XLVIII, 1 Bulleid and Gray, Glaslonbury Lake Village, vol. ii. p. 406. 2 Proceedings, vol. Ixix, p. 317. A VIKING SETTLEMEN FRESWICKT A T , CAITHNESS9 9 .

1 and 2). The larger, No 1, is imperfect, and measures 4f inches in length. No. 2 is of a form frequently found on Viking sites, and measures 3^ inches in length. Both were surface finds. Turn Buckles or Door-snecks.—Three of these objects, made of cetacean bone o complet tw ,representee on ed hala (PIan y ) fb d.2 XLIXd an 1 , •only, were found. The complete examples measure respectively in length 4f inches and 4^ inches. No. 1 is pierced with a circular perforation, for the pin on which it revolved, near its centre, while No. 2 has a similar perforation at l^g- inch from one end. Both perforations are about ^ of n inca h diameter e imperfecTh . t exampl s beeha e n broken acrose th s holen pi , whic probabls hha y bee. 2 n . towarde casNo th f endeo e n i son s ,a

Fig. 5. Object of Cetacean Bone. (]. These objects resembl e woodeeth n snecks formerl ye door th use f n o sdo cupboards thed an ,y possibly serve dsimilaa r purpose. Two similar objects, one complete and one represented by one-half only, were recovered from a kitchen-midden in the immediate vicinity of the Viking dwelling of Jarlshof, Shetland.1 Object of Indeterminate Use.—An object made from cetacean bone {fig. 5), which in general appearance resembles a turn buckle, on closer examination appears to have served some other purpose. It measures 5^ inches in length, 1^ inch in breadth at one end, and 1-j at the other. At 1^ inch fro e narrowe me blocth th s bee d kha nen r pierce hola r y efo db a bolt, ^ of an inch square in section, and the upper surface has been .slightly lowere t thida s poin receivo t t ecirculaa r iron washer beneate hth bolt head, the rust of the iron being visible on the surface. It is thus •obviou t intendeno s s tharevolvewa o dt t undersidi te Th e objec. th f eo t s bee t baclevea ha ncu n ko l fro mpoina t jus advancn i t e holth o et f eo the end, so that if the block was bolted to a flat surface there "would be a space of ^ an inch between it and the latter at the extremity. It is clear from the size of the perforation and the evidence of the washer that it was intended to be firmly fixed. Possibly it was a cleat. 1 Proceedings, vol. Ixix 293. p , . 100 PROCEEDING E SOCIETYTH F O S , FEBRUAR , 1939Y13 .

Whorls.—Hemispherical whorls made from the heads of ox femurs (PI. XLIX, 3 and 4) were found throughout the area excavated, and totalle numbern i a rul5 1 s de A the. y have been rudely fashioned with little effort to give them any elegance of form. Spindle whorls from femur- heads are of frequent occurrence in prehistoric and later excavations, and are by no means confined to Viking sites. They were found by Pitt Rivers in Romano-British excavations at Woodcuts Common, also by Sir Henry Dry den at Hunsbury or Danes Camp, near Northampton. They were among the relics from the Keiss broch in Caithness, and the broch of Burria Nortn ni h Ronaldshay. Thoug probablw hno ligho yto effeco t t t their original purpose, they would be sufficiently heavy before the osseous matter, which they contained, perished. ' Bronze. Belt-chape.—A belt-chape (PI. L, 5), 2i\ inches in length, with a rivet round which the end of the strap has been passed across the open end, was found abou foo1 t t abov floor-levee eth . norte th VI . n ho l sidNo f eo Brooch.—About ^ of a penannular brooch of Irish-Celtic design (PI. XLVII, 6), and which had originally been coated with silver, was found among kitchen-midden refuse covering the floor of what appeared to have been a small chamber to the north of the inner cross-wall, and to the "west of the vent which traverses the indeterminate foundations lying onle Th y indicatio . e eastwardVI th . o t No f dat no f so e which this find- spot confer s derivei s d fro e facmth t tha e westh t t wal f dwellino l . gNo IV, presumably of thirteenth-century date, actually lay on the top of it; moreover perioe buildine th ,th f do whicn gi objece hfounth s wa t d appears to have been coeval with that of No. VI adjacent, which in its turn was secondary to the massive "wall which in part contains it on the south. The dat deposif eo brooce th f o t h migh twelfte t lats th thu a s ea e hsb century. The brooch has measured when complete some 2 inches across in either direction s wil e seeA b l. n fro e illustrationmth e terminath , s mori l e realistic in conception of a dragon than that of any other example of the style so far recorded. The head is outlined by a rounded moulding, whic wels hi swelle l junctiocollae a defineey th o e bowse t t ra th Th df .no apointea s d oval, and enclosed wit hnarroa w mouldin represeno gt e th t eyelid e surfacTh . s ornamentei e d wito row f tw hherring-boneo s r o , feather ornament, extending respectively to the ends of the jaws. The outer end botf so h jaws terminat uppee th f voluteso rn e i p to frod e ,an m th jaw spur-lika e projection, also hatched with herring-bone ornament, suggests a horn, or an ear. In the centre of the bow there is a sunk oblong panel withi a nraise d moulding n inc a n length,hi f -o | , containine th n i g centre a rounded boss, while the surface of the panel is enriched with threads of interlacing ornament indifferently conceived and executed. A VIKING SETTLEMENT AT FRESWICK, CAITHNESS. 101

The bow is plano-convex in section, measuring -j of an inch in breadth by ^ of an inch in depth. Though a possible twelfth- century date has been suggested for the deposi thif o t s brooch, that date has coursef ,o , little bearin date th f eo n go the object itself, other than to supply a limit beyond which it cannot have been made. The possibility is that the brooch belongs to a considerably earlier e dateevolutio Th e Iris. th hf n o brooc s beeha h n treatey b f o d various authors, and it is agreed that the source from which it sprung was a penannular fibula with zoomorphic terminal a stylise f o s d form, which mad s appearancit e n Romano-Britisi e h times d examplean , f whico s h have been found at numerous sites in Britain. It is not intended here to follo e variouwth se evolutiostepth n i s f thio n s brooch through phases which sho wfarthea farthed an r r departure fro zoomorphie mth c character, till it is almost lost in the splendid achievements of such brooches as those from Tara, Ardagh, Hunterston, etc. The history of the development over the five centuries from A.r>. 500 to 1000 has been demonstrated by Mr Reginald A. Smith, F.S.A.,1 and other authorities have likewise treated of the subject. At a certain point in this evolution, somewhere about the ninth century, there appear havo t s e been some more definite returo nt zoomorphism in the treatment of the terminals, and we are fortunate in having an approximately dated example, the Croy brooch (PL XLYII, 5), found at Croy in Inverness-shire, as illustrating this process, and supplying us wit htypa e from whic Freswice hth k brooc havy hema descendede Th . Croy brooch, which was found associated with a coin of Coanwulf, King of Mercia, who ruled from 796 to 822, has been dated by Mr Smith to about the year 820. It will be noted that the terminals consist of two motives, a triangular figure which occurs at the end of the bow, with a spherical motif in front of it. If we eliminate the latter we are left with a figure which might well have suggested r itselo , f have been suggestee th , dby gaping head of a monster, and which has several features in common with e Freswicth k head. Here have w , e indicated e profilth , e head with open jaws, the ridge or collar that cuts off the bow, and the moulding carried along the edge of the jaws, and turned up at each respective end so as to form a slight volute. There is also an eye, a triangular figure which, though quite suggestive, lack e developeth s d realis e otherth f mo , •while e surfacth hatcheds i e , thoug t exactl e samhno th ee n yi manneth n o s a r Freswick reliccomparo T . e broocheth e s farthere Croth yn i brooch: , ther w panea s bo ei e centre l th th containin f n i eo ga centra l boss, surrounded by interlacing threads, executed in competent style, and in that respect differing fro e degeneratmth e renderin e Freswicth n go k example. This tendency to evolve a head with gaping jaws has been suggested by Mr Reginald Smith, and in the Freswick brooch we may well have 1 "Irish Brooches of Five Centuries," Archceologica, vol. Ixv. p. 223. 102 PROCEEDING E SOCIETYTH F O S , FEBRUAR , 1939Y13 .

an example of the complete evolution occurring at a late date in the series. Finger-rings.—Two penannular finger-rings with tapered ends were found (PL L, 9 and 10). They are both made from thin strips of bronze. One is imperfect, but the other measures f of an inch in diameter. The imperfect example was found to the right of the doorway outside No. IV, and the other came from the upcast at the west end of No. III. Bracelet.—A bracelet, evidentl a child' r yfo s "wris , formet 6) (PI , L .d fro mpieca f bronzeo e wire incn a ,thicknessn f hi ^o s founwa , d among the upcast soil when the deep wall crossing No. II at the back of the bath, was being exposed. When recovered it was slightly out of shape, but beew nno s restore ha s originait o dt l form, showin interion ga r diameter of lj^ inch. It is slightly penannular, and has remaining at one terminal a collar of bronze, the corresponding collar being amissing. One surface is decorated with a series of short transverse notches. Plates Patches.d an e Vikin—Asth n i g hous t Jarlshoa e f there were found a number of fragments of plates of bronze, most of which had been used as patches on large bronze vessels, of "which the largest still retained in them the rivets of paper-fastener type by which they had been attached smalo Tw l piece. 8) eace (PI , ar sL h. perforated witcirculao htw r holes of small diameter, evidently for pin rivets. The distribution was general. One piece with paper-fastener rivets was recovered from beside the bed- anothed an , platforIV r . fro middee No mth m n i n . VII fillin.No n gi Pin.—A hammer-heade f well-knowo n pi d n Viking typ) e7 (PI, L . was found at the west end of the smithy No. Ill, from what was probably the site of the west wall of the building -which had been torn out. The completes i n pi , excepactuae th r lfo t point whic wantings hi measured an , s 4^ inche lengthn si . Glass. A small barrel-shaped bea f greedo n glass, measuring YQ inc longesn hi t diamete weathered an r e surface th e surfac e s founn th d o th wa n ,f o do e hollow. Coin. A single specimen of coinage, a silver penny of Henry III. of England, inscribed "Wille Lund,n mo " fro e Minmth t of • Londo d datinnan g from 1258—1272 s pickewa e ,p fro th u e dsurface seawar f mth th o t d a e en d area. Iron Objects. Knives.—Four knives were found. Two were of the pronounced hog- narrow-pointe" a s i ) bac 2 e (PI , kL .On for d. m1 . showNo y PI, nb L . blade, triangular in section, with a comparatively broad back. One blade was too much decayed for determination of form. Both the former are A VIKING SETTLEMENT AT FRESWICK, CAITHNESS. 103

type f bladeo s s characteristi e earlth f yo c mediaeval period d similaan , r blades appear amon e relicgth s from Dunadd,1 Argyllshire, preserven di the National Museum. Identical examples were found recently in the stone For Cahercommaif o t Countn i . O'NeilH y r ClarD ly Hencken.eb

same th f eo 4 remain e . , appearth 3 . No e keya b f d No so o PIs , t an L . 2 plat remaine eth butt a socketa f r so o , . Ther alss oewa foun remaine dth s of a pair of shears consisting of the spring and upper portion of one leg.

Pottery. The sherds of pot recovered were numerous and of two distinct classes: first limitea , d quantit f wheel-madyo e mediaeval ware, unglazed, or onl treateo ys smala o dt l extent secondd an ; considerabla , e quantity of potsherds of the quality which, from finds in Shetland and elsewhere, regardee b w no s characteristid a y ma Vikine th f co g period. MedicBval Ware.—The mediaeval ware was, with the exception of one small piece which "was found abov l referablal e, floor-leveVI o . t e No n i l latee th r occupation t consisteI . sherdf do somf so e thre r foueo r vessels, unglazed r onlo , e yshoulde th partiall n d necko an , r so y. Wite on h exception it is of a reddish-brown colour, and the two bases that occur amon e remainth g e sagging e ar sothe th e markedl d ron , an , yso slightly. Therindicatioo n s i e f thumbino n g aroun e basese th dTh . remains of a large bag-shaped pitcher with an interior diameter at base of 1\ inches were found for the most part on the floor of No. I, and in the passage leadine bath-chambeth o gt norte th thaf n ho o r t structured an , wha s evidentlwa t ya smal l e samsherth f edo vessel recoveres wa , d from e bath e flooth f th ,o r beneat e heae north-wesh th f clath po n yi t corner. A portion of what is evidently the handle was found on the floor of No. I. singla s ha e t groovI e dow centre nth e lengthwise deed an ,p finger impression eithen o s e poinr th sidjuncturf t o tea e wit necke hth . This vessel s evidentlha y bee holdinr nfo wely g waterlma havd an ,e beene useth n di bath-chamber. The base of a cooking pot, sagging, and much encrusted with soot, "was found in a kitchen-midden in front of No. IV. The interior diameter of the base is 4-j inches. A portion of a handle, circular in section, but with a ridge along the top, "was found while clearin. floor-leveo gt IV . No n i l e onlTh y piec f mediaevao e l ware wity distinctivhan e charactes wa r e sherth , previousld2) , (PILI . y referre . 86)(p , o comprisint d g parf o t rime th lighf o ,wared re t , soot-encruste e outsidevidentlo s th n d do e an y part of a cooking pot, with a diameter at mouth of some 9^ inches. The rim is everted with an approximately rectangular section, and is

1 Proceedings, vol. Ixiv. p. 118, fig. 5. 2 Royal Society of Antiquaries of Ireland. Extra volum r 1938efo . 104 PROCEEDING E SOCIETYTH F O S , FEBRUAR , 1939Y13 .

ornamented on the upper surface with three wavy, narrow, parallel, continuous ridges. The outer surface, as far as revealed by the fragment, is decorated with zone f paralleo s l line n reliefi s . This sher s foundwa d "whewall-heae V I levee wale o beins th . n th th nf wa o llNo g f removedo d to explore the structure beneath. It must therefore obviously refer to e latesth t occupation verA .y close founde paralleb o s previouslt a , s i l y stated shera n i , d from Rayleigh Castle occupatioe th , whicf no h ceasen di e latteth e thirteentr th hal f o f h century whicd an , s preservehi e th n di Prittlewell Priory Museum, Southend-on-Sea.1 A late thirteenth-century datfairly eclaimee ma yb thesr dfo e few mediaeval sherds. Viking Pottery.—Plate LI, 4, illustrates a cooking pot, complete except for some slight damag rime th , t ea whic Simor founs M hwa y n db Bremner, when the excavations were not in progress, set upright in the sand, in e Museumth o forwarded t t I e an "naust,m .fron, th hi V f y . o tb d "No is of typical Viking hand-made ware as found in Scotland, with numerous impressions of vegetable matter in the body, and it is heavily encrusted with carbooutsidee th n forglobularn s no i I .t m i , wit hflaa t basd ean an everted rim. In height it measures 6 inches, and in breadth at the bulge 6^ inches. From an adjacent spot there were recovered sufficient fragment a muc f o s h enablo t smalle ) 3 s reconstructio t eit , (PI po rLI . n s beeeffectede b ha n o t t I fashione. d wit ha ver y uneven surfaced an , from the encrustation of soot on the surface, appears also to have been used as a cooking pot. It has measured 2>\ inches in height, by 3J inches in diamete bulgee th t ra . Bot above hth e vessel flat-bottomede sar . The general character of the ware, all of which is hand-made, is, besides the use of grass in the paste, a coarseness of technique which is displayed in the very uneven surface of the exterior. The body is hard and well- fired, varyin n colougi r from buf o grey t n fd blacki e potan e , Th ar s. general encrusted with soot only on the upper portion of the exterior surface, indicating that the beed yshoulderha e n ember e th sun th o t n ki p su . There is invariabl encrustation ya carbof no ninteriore alsth on i . Whil generae eth l character of the ware is uniform, there is a considerable variety in the treat- e rimsmenth s e showsectionf a ,o th t n i n s illustrated (figs . Finge6) . r impressio s beenha o n tw applier o a decoratio e s a de on rim th f o sn o n vessels (fig. 6, No. 4), while a single rim, part of a pot of large size, appar- ently of 9 to 10 inches in diameter at the mouth, has been ornamented wit ha serie f smallo s , irregularly placed notche t acroscu s t i s(fig . . 6No , 18). Some oblique cuts placed one inch or so below the rim of another sherd may have been part of a decorated scheme of the complete vessel . 20)No .(fig , 6 Anothe. r sherd t illustratedno , s decoratei , d wit hseriea s of deep vertical grooves, probably finger impressions, extended aroune dth

1 For this information I am indebted to Mr Gerald C. Dunning of the Royal Commission of Ancient Monuments, Wales, etc. A VIKING SETTLEMEN FRESWICKT A T , CAITHNESS5 10 . rim. A sherd of a pot with an indicated diameter of some 6 inches (PL LI, 1) has been decorated with a series of broad streaks, in some dark

Pig. 6. Sections and decorated surfaces of Rims of Viking Pottery found at Freswick. brown colouring matter. Several pieces show perforations evidently made to effect repair. The bases appear to have been flat-bottomed. The cookin s measurabla g r potsfa s a , e fro e sherdsmth , show indicated diameter moutinchese 1 1 th t o frosf t a h o . 0 m1 Whil e morth e e sophisticated form d decoratean s d rims fount a d 106 PROCEEDING E SOCIETYTH F O S , FEBRUAR , 1939Y13 .

Freswick were entirely absent from the pottery found in the Viking house excavated at Jarlshof in 1934, yet in both respects they recall ceramic styles of an earlier period found in a prehistoric site there. Pottery with char-

acteristics very similar was recovered by Professor 1 V. Gordon Childe in the excavation of a promontory fort at Larriban, or Larry Ban Head, on the coast of Antrim, in 1935,2 in regard to which he remarks on its bein a g"ver y characteristic North Irish fabric. e typicaTh " l cooking t founpo t Jarlshofa d barrel-shapea , d vessel representes i , d most closely f figo . 6 .0 1 . No y b t founno s d possiblwa t I restrico et fory periode eithean o mtt th f o r s represente three th ey d b group buildingsf so wholee th sophisticatee n O th , . d rims appeare belonlato e dt th e ratheo gt rearliee thath o nt r periodsA . number were e materiafounth n i d l with beed whicha n I h VI hous . No e fille. dup Stone. Hones.—Twelve sharpening stones or hones, of quartzose schist from the Moine schists, were found on various sites, five of them coming, as migh expectede b t , fro smithye mth . The alle yar , wit exceptione hon f o , the haunched type, the exception being a straight-sided hone of black phyllit clayer eo y schist differen,a t material from othery thaan f .to Except e lasth t mentioned e mucar l h al , wor d incompletean n typicaA . l example is shown in PL XLIX, 8. A sharpening stone, probably of Caithness flagstone, to be used for some narrow, round-pointed metal instrument a dee s pha , groove th e length of each of the two opposite faces, and is considerably abraded at one end. Plummets of Quartz, etc.—A number of small blocks of quartz, varying in weight from \ Ib oz.1 2 .Ib o t ,. wer "nauste eth sit e f founeo th n "o d d elsewherean . Each block present e flaon ts surface d towardan , e on s edge there has been an attempted perforation for a cord, in one case only completed. A specimen is illustrated (PL XLIX, 5). The process of perforating a block of quartz must have been difficult, for several of these blocks have been brokee attemptth n ni . Quartz woul e selecteb d r dfo use as plummets owing to its high specific gravity. In one instance an ovoid sandstone pebble, similarly with one flat side, show n incompleta s e perforatio e edgeon ,n o nowin a fractur o t gL (P e XLIX, 7); and another, with a fractured groove on the flat side, has a groove continued across the opposite face. A small block of quartz, also flat on one face, has been grooved round the centre of the other sides, evidently to hold a cord. 1 Proceedings, vol. Ixviii . 287p . , fig. .54 2 Professo Gordo. V r n Childe PromontorA " , y For thn to e Antrim Coast, Antiquariese "Th Journal, vol. xvi . 188.p . A VIKING SETTLEMEN FRESWICKT A T , CAITHNESS. 107

Polishers.—There are two polishers, an ovoid pebble of porphyry, flattened by use on each side, which was found in the smithy, and a larger water-worn ovoid pebble of quartzite. The latter at one end has been reduced to an angular section by rubbing, and at the opposite end has been slightly abraded. Pot-lids.—Three pot-lids of stone, from 2-| to 3f inches in diameter, were found, all in No. VI, two of which are illustrated (PI. XLIX, 9 and 10). Querns.—Several e piecegrindinth f o s g stone f rotaro s y querns were found fashioned from garnetiferous mica schist. This material, whics hwa also used for querns by the Vikings in Shetland, where it is plentiful, likewise occurs in the west and south parts of Caithness. One of the pieces, which amount somewhao st t less than one-hal uppen a f o frd stoneha s ha , countersuna k circular area abou inche2 t s wide aroun centrae dth l opening, d thougan h much weathered still showe socke e th bas th f so hol o et t d the vertical wooden handle with which to turn the stone. The upper stone of a rotary quern, complete but for a chip off one end, was found abovs a , e related, coverin. VI . No gpost-hol a f o ease d th en tt ea s ovoii t outlinen I di , wit handlende a h hola measure d r on e fo an , t ea s 16^ by 13 inches. Steatite.—A piece of the rim, slightly curved, and 14 inches long, of a very large steatitanothed an , eVI r vesse. e nortfouns No th f wa l o ho dt fragment on the floor-level of the house. Weights.—A number of larger ovoid pebbles, with grooves cut in opposite faces acrosd an ,e end th s holo ,t dcora r ropedo , were also found. Such stone stile sar weightinl presenr e usefo th t y da da t g lobster creels. Sinker.—A pointed oval pebble of steatite, with a perforation at either Ill. corda en founs r ,analogNo e d fo andwa , n th leade i n th ,do f yo objec t presene useth t da t day probabls ,wa y"lina e sinker" (PI. XLIX. 6) , addition I foregoine th o nt g there were foun numbeda flat-sidef ro d oval pebbles, some 6 to 8 inches in length, much chipped and indented at the purpose th e sidese edger whicr th th Fo f efo n so .o hend d thessan e were used, the ingenious theory was advanced by Mr Simon Bremner, a corresponding membe foremaSocietye e excavatione th th th f d o t r a nan , , that they had been employed in boat-building to '' hold on'' against the point of rivee th tbeins whewa g t n i drive . nin

Vikine Th g house discovere t Jarlshoda Shetlann i f 193n di probs 4wa - ably a long house, as being of earlier construction than the well-preserved example, No. VI, found at Freswick, but in respect that the former no doubt continued in occupation until a date contemporary with the Caith- ness example f intereso s i t contraso i t t, e culturath t l evidence founn o d both sites. The Freswick settlement discloses by its relics a class of occupant pooren i s r circumstances e findth s a ,wer e fewe proportion i r n 108 PROCEEDING E SOCIETYTH F O S , FEBRUAR , 1939Y13 .

to the area uncovered, and as a rule, also, they were ruder in character. a completTher s wa e e absenc e scribeth f o ed slates which were suca h remarkable feature of the Jarlshof finds, nor were there found at Freswick any bone pins to compare with the finer examples found in Shetland. While hones were proportionately numerous amon e relicgth s from Fres- wick, the haunchee yth f wero l eal d form therd entirels an , ewa y absent from among the examply m an smale th f elo black hones, quadrangula section ri n and perforated at one end for suspension, which occurred in considerable number t Jarlshof—a s a type t whicinfrequentlno s i h y found associated with Viking relics t JarlshoA . f loom weights were foun grean di t numbers, at Freswick they were conspicuously absent. Whereas at Freswick the most common relics were whorls made from the femur-heads of ox-bones, not a single example of such a whorl was recovered at Jarlshof. In both places remains of querns made of garnetiferous schist were found, and while at Jarlshof remains of vessels of steatite were common, at Freswick they were very rare. The most remarkable contrast in the finds from the two sites is to be e pot-rimsfounth n e wari d Th e. from both place s identicali s d an , distinct from other r discoverewarefa o s Scotlandn di e evidencth y b , e t i bear e numerou th fullo s f o y s impression e husk th d stra f an o s f wo rime e form th e bodyth s f th t oat o sdiffe n bu ,so r materially Shete Th -. land rims of the Viking period are very simple, being as a rule either slightly everte r curvindo g inward e manneth barrel,a n i sf o r 1 while eth Caithness examples supply, as already mentioned, a variety of forms. Whether such early forms were extant among the native population of e Vikin e counte th datth th f o egt y a invasion onln learnee ca syb y b d excavations on native sites. There is a probability, however, that they "beed ha n imported from Ireland. In both settlements numerous fragments and patches of sheet bronze were found, showing that large bronze cauldrons were stil usen i e l Th . objects plummeta r whice fo sus ha s suggeste si d were peculia o Frest r - wick, as also the pebbles supposed to have been used by boat riveters. The characteristic combs of the period are common to both. The interior arrangement principae th f so l house t Freswica (No ) .VI k evern i characteristiy e ywa ar Vikinf co g e culturegenerath d an l, impression derived from the excavation is that the Norwegian settlers in Caithness brought with the mdistinctiva e cultur f theieo r own, which througe hth period of two or three hundred years in which they flourished in the county, remained entirely unaffecte e indigenouth y b d s culture th f o e region e conclusioTh .adducee b o nt d from these circumstance s thai s t immigrante th s either live completn di e isolation amon native gth e inhabitants, or, as seems more likely, that they ousted them from the plains and 1 Proceedings, vol. Ixix. p. 306. A VIKING SETTLEMENT AT FRESWICK, CAITHNESS. 109

fertile regions to seek out a scanty subsistence among the moors and mountains that for e westermth d southeran n n confine e countyth f o s , and where the place-names and the people are still for the most part Celtic.

REPORE ANIMATH N TO L REMAINS Misy B .s MARGER . YI PLATT , Naturae oth f l History Departmen Royae th f o tl Scottish Museum. The majority of the bones are of a small but mature Ox of shorthorn variety. The remainder of the species present are given in order of their numerical importance: Pony, very small variety likShetlane eth d race. Dogs, remain threef so vere on ,y smal fox-liked an l . Sheep, extremely heavy-horned variety. Red Deer, Cervus elaphus scoticus Lonnberg. Grey Seal, Halichoerus grypus (Fabricius). Pig. Gannet, Sula bassana (L). Cod, Gadus callaria. sL

In conclusion it gives me pleasure to acknowledge my indebtedness to Admiral Sir Edwyn Alexander Sinclair, G.C.B., etc., the proprietor of Freswick r havinfo , readilo s g y accederequesy m r permissioo t dfo t o nt excavate, and also to the Messrs Gullick, the farmers, for according their permission, and helping wherever possible. Also to Mr Edwards, the Director of the National Museum of Anti- quities r havinfo , g directe y attentiom d o Freswickt nr muc fo hd an , assistanc productioe th n ei thif no s paper. My thanks are due to Miss Margery I. Platt, M.Sc., of the Royal Scottish Museum, for her report on the animal remains; as well as to Mr David Balsillie, e B.Sc.samth f eo , Museum r determininfo , e chargth - acter of mineral specimens, and Mr M. Y. Orr, of the Royal Botanic Garden, for having identified the vegetable remains. Dr Sigurd Grieg, of the University Museum of Oslo, helped me to date the combs, while Mr Gerald C. Dunning, of the Royal Commission on Ancient Monuments in Wales, etc., gave me valuable assistance in dating sherde th medievaf so l pottery. I am grateful to Mr R. B. K. Stevenson, Keeper in the National Museu f Antiquitiesmo e valuablth r fo , e assistanc e gav h e ove m ea r period of two weeks on the spot. My excellent team of local workmen, unde r SimoM r n Bremner, tackled their job with such zead interesan l t beyond the mere terms of their employment as to deserve special recognition. 0 PROCEEDING11 E SOCIETYTH F O S , FEBRUAR , 1939Y13 .

I owe thanks to Mr F. W. Robertson, Ph.D., of the County Library at Wick, for having lent me a tent and furniture, and other acts of kindness. While, lastly, to the Council of the Society for having allotted to me e generouth s contributions froe excavatioth m n funds that made th e excavation possible, I am under a deep debt of gratitude. The relics have been kindly presented to the National Museum of Antiquitie Scotlanf so proprietore th y db .