A Hut-circle Settlement at Kilphedir, by Horace Fairhurs David an t . TaylodB r

GENERAL These investigation t Kilphedisa Strate th f Kildonan o hi r nconcernee ar (fi ) g1 d wite hth sit fivf eo e circular smale houseth d l aresan cultivatef ao d land around, whic firss htwa occupied in the period about the fifth century BC. Some three centuries later, a single house of a rather different type was constructed on the same site and a second phase of cultivation may have occurred. Peat then blankete settlemene aree th df th a o t which subsequently remained undisturbed presendowe th o nt t time excavationse Th . wha n facn o ,i vera s tti y exposed hillside, were under- taken in an adult field-school under the auspices of the Sutherland Education Committee. The circumstance spapere wervolunteer th f o Ou e. unusuad discussee en ar re d helperth an lt da n si schooe th l numbered between twelv worke e thirtyd w ean d dan , during four season fortnigha f so t each, in 1963-5 and again in 1968. The writers present the results jointly on behalf of the field- school and of the Sutherland Education Committee.

HUT-CIRCLES IN NORTHERN BRITAIN expressioe Th n 'hut-circles perhaps i ' s somewhat antiquate usefullt i t dbu y describe classa s of field monument whic probabls hi y more numerou Northern si n Britain thaothery nan . Sur- prisingly little excavation has been attempted in and very little is known with precision concerning the distribution, differences in form, period and the economy of the inhabitants. Hut-circles vary considerably in size from less than 16 ft (5 m) internally to over 40 ft (12 m) but are generally regarded as the stone foundations of round houses. Some consist of a simple ring-like enclosure delimited by a bank of earth and stones or even boulders, usually much overgrown. Many hut-circles and especially those of northern Scotland, however, were normally located on a slope, presumabl better yfo r drainage thesn i d e casesan , leve,a l platfor firss mtwa constructey db quarrying into the uphill side and spreading the material in a crescentic apron downslope. The stone foundations of the round-house which was superimposed are to be traced as a distinct bank around the edge of the built-up portion of the platform; this bank seems often to merge with the quarry face upslope. Of recent year southern si n Scotland, many platform closela f so y related type which appear simply as a levelled space without the ring-like bank of the hut-circle have been recognised. Excavation has shown that these 'platform settlements' consisted of houses with timber walls, lacking stone foundations, as Feachem has been able to demonstrate.1 Similar floors have also been traced within some hillforts, and seem to represent the dwelling sites of the occupants; the North Eildon is a well-known example where very large numbers are involved.2 Insufficient is known abou perioe th t thesf do e structure decido st e whethe culturara l differenc involves ei r do §hS o ta to o1

H ta o en

typV t e circleHu • s ° / KILPHEDIR t circleHu so with DISTRICT expanded entrance 'Galleried' structures

Based on the Inventories o' THE ROYAL COMMISSION NO ANCIENT MONUMENTS CAITHNESS and SUTHERLAND

FlGl FAIRHURS TAYLORD TAN HUT-CIRCLA : E SETTLEMEN KILPHEDIRT TA , SUTHERLAN7 6 | D whether the hut-circle is normal in upland or rocky areas where stone was more readily available than timber. Marked differences in the use of materials of construction of these prehistoric round-houses may, at least in part, explain some very curious anomalies in the distribution of hut-circles, particularl regioe th primarf n yni o y concer thin i s papenorte th Scotlandf hr- o RCAe Th . M Inventory for Sutherland stressed the fact that very large numbers occurred in the eastern part of county,e th whil beed eha nver w more founyfe th en di rugge damped dan r west; this would

sugges3 t tha environmentan ta l facto significants ri othee th n r handO . Inventorye ,th Caithnessr fo , writte same th en i period,4 emphasise scarcite dth than yi t county which lowlans it n i , d districts, is neither rugge r extremeldno t devoii ys i wet f extensivr do no ; e stretche f uncultivateso d land where hut-circles coul expectee db surviveo dt muce .Th h later Inventory Orkner fo Shetlandd yan 5 similarly mentions only a very few hut-circle sites. Such problems obviously need further investi- gation. There can be no doubt that the sites of these round-houses and in particular the more easily recognisable hut-circles, provid ereservoia archaeologicaf ro l information whic bees hha n much neglected in the past. The few excavations which have been attempted have usually seemed unrewarding unless the related wheelhouses of the Northern and Western Isles are taken into account, and in general, investigators have been more concerned with hillforts, brochs and duns, which seem to represent the dwellings of the aristocratic elements. Of late years, however, current opinion has swung away from facile explanations of cultural changes in terms of mass invasion; what seem particularle b o st y neede t thida s tim fress ei h evidence which doe t over-emphasissno e the contribution of small aristocratic movements, but relates to the common people, many of whom must have bee descendante nth populatioe th f Bronzr e seveo o th e nf no Neolithie Ag c times.

THE PROBLEM IN SUTHERLAND The county of Sutherland appears to be of special significance in the study of hut-circles. The RCAM Inventory6 was one of the first to be published in Scotland, dating back to 1911, and is far from a complete record but even so it lists about 250 localities where hut-circles occur either as single sites or in groups. The recently published sheets of the new 6 in OS map suggest that this number coul doublede db therd wely somethine an , b ema l ordee th f 2,00f o rgo 0 hut-circles surviving. Unfortunatel ploughe yForestrth e th f so y Commissio vero d yn extensivnca e damage Ordnance wore th th f ko d an e Survey canno completee b t soono sitee dto .Th s occur only rarely alon strathfloore e gth th infrequenseef e d so b s an m o t t above about 1,00 t 0(30f ) OD0m . Much intervenine th f o g ground, which includes extensive stretche plateauf so moorlans ,i d wit hcovea r of peat and it is here that the great concentration occurs. Appearing in close association with most hut-circles, groupings of low overgrown cairns are found which the writers of the Inventory suggested7 were heaps of stones cleared from the adjacent groun provido dt e patche f arableso . Ofte heape nth s were indicateS O d ol 'tumulis da e th n o ' 6 in sheets; they are so numerous that they have perforce been omitted from the new 6 in map. Great number f thesso e cairns with accompanying hut-circle occury sma together r instancfo , e sidE f middle o th o et Strath Naver betweeClearancd ol e nth e settlement f Truderscaio s g (NC 703342), and Rosal (NC 690415). An excellent example is to be seen in an easily accessible localit Bettyhilt ya sanda n lo ymout e Nave e terrac th 700612) th C f f o h o r(N eW . Her groune eth d is free from peat and an old land surface with cairns and hut-circles seems to have been exposed recently from beneath a cover of blown sand. 68 | PROCEEDINGS OF THE SOCIETY, 1970-71 Inventorye Th notes8 that addition i , hut-circlee th o nt cairnsd an s , another much rarer combination occurs in the form of hut-circles with trailing banks of stones and boulders marking out what appear to be minute fields which are devoid of the clearance cairns. These alignments foune b associatioo n t di e ar n wit hrathea r different typ f hut-circleo e whic mors hi e strongly constructed and tends to have a curious thickening of the wall at the entrance; a short souterrain may be attached. Excellent examples are to be seen to the W of the Kildonan Burn (NC 915225). This combination of hut-circles and clearance cairns or boulder alignments occurs in many localities outside Sutherlan Grahad dan contributes mha dvaluabla e paper which discussee sth cairnfield general.n si 9 Abercromby describe dsita e with both cairn alignmentd san Aberdeenn si - shire ; Thorneycroft discussed the association in Perthshire ; Fenton notes the same combination in souther10 n Scotland.12 Nowhere, however, have the hut-circles11 , cairnfields and alignments been reporte sucn do hcentrascala n i s easterd ea an l n Sutherland. Unti r investigationou l s started, only two sites within the county had been excavated and the reports were so vague as to be almost useless.13

THE KILPHEDIR LOCALITY Kilphedir lies some 5 miles (8 km) up the from on the E coast of Sutherlan strate dTh (fi. hers hi . g1) e abouOD ) m t 0 £ (3 mil t f s som i e) wid 0 (•d £km 10 e ean At this point, the main Helmsdale River is joined from the N by the Kilphedir Burn which has cut deea p gorge through fluvio-glacial deposits eithen O . gorgre sidth f edistinco a t shoulder occurs at about 300 ft (91 m) above which the ground becomes more gently sloping, before it rises in great sweeps towards the mountains on the Caithness border. On this plateau which is developed over granite mantla , f peas eformeo heathera ha t d an d y moorland is characteristic (pi 12). Downhill this often gives place to scrubby oak and birchwood, with pasture on the flatter stretches near the river. There is little cultivation; sometimes a shooting lodge or shepherd's cottage occurs along the Strath, as at Kilphedir itself, but apart from these and an occasional farmhouse, the desolation recalls the complete depopulation of the Clearances of 1813-20. The area within a radius of about half a mile (800 m) from the mouth of the Kilphedir Burn seem havo st e forme attractivn da e pocke vera lanf o r t ydfo lon terracgw periodlo n a e o n O . eithe bur re joint i sid th mai s e nf a es th o n Strath smalo remain,e tw therlth e Clearance th ear f so e settlements of Kilphedir and Chorlich. Just above the edges of the gorge on either side are groups of hut-circles with both the clearance cairns and boulder alignments. Three other groups of hut- circle locatee sar dlittla e further upstrea otherd man wherE s denudee o occuth etw o rt d burial cairns are also traceable. The whole district is dominated by a broch, set within a formidable ditch (Inventory no. 307), at a height of 475 ft (145 m). In all, a score of hut-circles can be recog- nised, one of which undoubtedly has a short souterrain attached (Inventory no. 328) and another is probably similar (no. 327) (fig 15). Why this pocke f lano t d proved attractiv settlemeno et t ove lona r g periot f timdno o s ei readily apparent othebut , r highly localised area long-continuesof d settlement occur elsewherein eastern Sutherland. One is to be found at the neighbouring Caen burn to the E (ND 013183), while others appea Edrablrat e immediately acros Helmsdalsthe e Rive 983180 Kilournanat r (NC )and , 3| miles (6 km) upstream (NC 930188).

SITE ETH Detailed investigations were primarily concerned with a well-defined area lying between the steep slopeKilphedie sidth e f th seo , Claivallee E rth e d f glesth y o an Rat o n, t hNE Fhine th o nt FAIRHURST AND TAYLOR: A HUT-CIRCLE SETTLEMENT AT KILPHEDIR, SUTHERLAND 69

IROSETTLEMENE NAG T KILPHEDIR

+2O. +15' , +10, \ 5 '.+0 '-

•,^,,x: Clearance Cairn

\Z'!?\ Excavated Area

FIG 2 ope e (fi p, th W gi2 n S 12a) mooe th . o Tracert whaf so t appea anciene b o rt t cultivation plote sar visible between rough banks and alignments of boulders (Sutherland Inventory no. 325). We are indebted for the general plan (fig 2) to Mr Gordon Petrie of the Topographic Science Sub- Departmen f Geographo t Universite th t ya f Glasgowyo mucs i t i ; h generalise time th en i s da available he was unable to show all the vague spreads and irregularities. Rather more detail has been arable addeplae th f th n eo o d grount d discussed later (fig 14) addition I . alignmente th o nt s of the boulders, Mr Petrie has also shown a number of clearance cairns, mainly in the NE and SW wher alignmente eth lease sar t definite; agai plae nth n omits some doubtful examples. Between the banks and amongst the cairns, the ground is relatively free from boulders. The area cleared of 70 PROCEEDINGS OF THE SOCIETY, 1970-71 stones in this way measures approximately 300 yards (260 m) N-S and 180 yards (160 m) E-W, though the limit is often very indefinite and irregular. When operations commenced, four hut-circles were visible and a fifth was located during the survey; they have been numbered I-V for reference (fig 2). Hut-Circles I and II were 35-40 ft (10-7-12- internan i ) 2m l diameter and, though rather ill-defined, they were clearly simila formn ri . A circular floor had been made by quarrying into the rising ground at the back and spreading the spoil on the downside. The remains of the wall were overgrown but appeared to consist of earth and stones and were not always continuous, especially behind the quarried face at the back. Hut- smallers Circlwa I eII , stoo isolation di n neaKilphedie edge th rth f eo markedls r wa glen d an ,y centree holloweth t a .t Hut-Circldou vers ywa smallV eI , bein diametegn i t (6- onlf ) 0 1m y 2 r ill-defineo internallys s wa d dbeed an ,tha ha n t ti overlooke d previously. Hut-Circl clearls wa eyV distinc verts fro ywa othere mt th strongli s a slona s gy ha entranc build an t e passages wa t I . illustrated in the Sutherland Inventory (no. 325, fig 40) where a Y-shaped wall of boulders is shown running into the centre from the back.

SOIE TH L PROFIL STRATIFICATIOD EAN N A brief comment is necessary on soil conditions from the point of view of natural stratifica- tion, as few records exist for Scotland of excavations in strongly podzolised soils with peat, such as characteris site t Kilphedireth ea . Broadly speaking surface th , e everywhere withi site nth e consiste f 3-1do f (8-3o 2n i ) 0cm peat, normally overlying a layer of dark grey earth with bleached grits. Outside the hut-circles this layer gave place to a dark, rather mottled chocolate-brown coloured earth at the base of which, peate th f , o cam p fro to ) eweakl a e mcm th 0 perhapy(3 developen i 2 s1 d iron-pan lattee a .Th s ri characteristic phenomeno thit na s altitud Kilphedie th n ei r distric should tan e referree dth b s a o dt regional iron-pan. Below came a sandy clay with much gravel, which had become indurated by natural agencies to such an extent that considerable force was required to break through it. In fact, the inhabitants of the hut-circles had rarely penetrated it except very locally when making post- around holesan hut-circlese n I dth . , disturbed eart occupatiod han n material alik beed eha n subjected to renewed podzolisation after the settlement had been deserted and a 'local' iron-pan was often traceable, above the horizon of the regional pan and frequently at the level of the old floor. Usuall e 'localyth ' iron-pan appeare thina s da , dark brown deposit which roots could penetrate stronglHut-Circlo n s i t s bu ,wa y t i develope eI resemblo t s da metaea l plate; here, the root of a bush or tree which had been mineralised in situ had formed a cast resembling the socket of a tool. This is the secondary iron-pan as indicated on some of the sections (fig 4). Under- neath it, the lower lying deposits of occupation material, which might be still flecked with carbon, could appear as a bright yellow or orange layer above the regional iron-pan. In Hut-Circle V, wher loosers soie eth l wa iron-pa e ,th barels nwa y traceabl occupatioe th d ean n materia dars lkwa and more normal by southern standards. It is to be emphasised that these purely natural horizons - the grey-gritty layer, the iron-pans, the bright yellow and orange colours and the indurated layer - completely dominated the soil r profilsectionou n i e s excep Hut-Circln i t . TheeV y could traverse obliquely both occupation materia floor-levelsd lan onld lowese an y,th t indurated laye mucd rha h significanc archaeon a s ea - logical horizon, correspondin 'bed-rocko gt ' under other conditions. Nowhere, howevere w d di , find evidenc blankea f eo peaf o t existencn ti occupatione time th th f et o e a pea e obviousld ;th tha y formed over the hut-circles and the cultivated plots subsequent to the occupation. RomanC C J r s M fro Macaulae mth y Institut Soir efo l Science, believes tha iron-pae th t n FAIRHURST AND TAYLOR: A HUT-CIRCLE SETTLEMENT AT KILPHEDIR, SUTHERLAND | 71 forme relativeln di y open moorlan tha d absence ) dan iron-path e tm th 2 f (6 eo t f n0 belo20 e wth contour, probably indicate presence sth woodlanf eo d alon Strate time gth th en t ha whe pa e nth was forming on the moor above. This point is of significance in considering the location of the settlement on its very exposed site beyond the upper limit of cultivation as it has existed in more recent centuries.

THE DETAILED SURVEY Hut-Circle I (fig s 3-5; pis 12b-13b) Hut-Circle I with a diameter of as much as 40 ft (12-2 m) from bank to bank could easily be mistaken for a stock pen; indeed, the new 6 in map of the Highlands, Sheet NC 91 NE, indicates 'enclosure'n a . Excavation showed, fro presence mth hearta f edomestio d han c refuse including numerous potsherds, tha substantiaa t l round-hous existed eha d here t stooI gentl.a n dE o eS slop highese somn eth o f eo t grounsitee th . n Naturado l drainage befor peae eth t formed must have been good, but there appears to have been no shelter from the wind. Outside the circle to the W, there was a shallow superficial depression in the peat suggesting an underlying drainage channel. A section, however, showed that the regional iron-pan which predates the house, followed the profilunderlyine th f o e g hollow which must therefor naturale eb . Inside peae 6-1s th , wa n 0ti (15-2 thic) occasionalld 5cm kan cm)0 ver(3 y .A n ovei y 2 wel1 r l developed iron-pan occurredn i the E quadrant, and both here and above the peripheral wall in the W, incrustations of bush or tree roots with the iron-pan showed that at least shrubby vegetation and possibly trees had colonise remaine housdth e th f esbeeo d afteha n t i rabandone t befordbu e ther beed eha n time peae fo formth ro t . The floor measured 37 by 34 ft (11-2 by 10-4 m) internally, with the long axis on the line of slopine Th . gentrance SE beesitd th e eha n th levellen ei quarryingy db bace th kt a e dow th o nt indurated layer which at its maximum depth below the surface was about 18 in (46 cm); the soil had been spread over the opposite part of the interior and also had been laid down as a shallow apron outside the doorway. This crescent-shaped area in front of the entrance had then been outward) cobblem 3 ( distanc a t f r provid o t ,d0 fo 1 f eo ehara d surface where traffic would have been concentrated impossibls wa t I . determino et e precisely where quarrying gave plac fillino et g as the original surface had obviously been irregular. At least in the final stage of the occupation, the cobbled floor of the hut had lain approximately where the gritty, dark grey layer had sub- sequently formed just below peat, but flecks of carbon and potsherds were found for 6-12 in (15-3 belo) 0cm w that horizon rathea n i , r sandy earth which coul brighe db t orang yellowr eo - brown. Wher floo e bee d induratedowee g th rha th n du no t d layer cobblino ,n beed gha n necessary. The entrance, which was about 3£ ft (1-07 m) wide, was roughly paved and walled on either side by lines of large slabs somewhat casually placed in position. The enclosing bank of the hut, whic sectionee hw thren di e places, represente remaine dth freestandina f so g wal f bouldero l s and earth. It was some 4-7ft (1-22-2-13 m) wide at the base and was traceable all round except behind the deeper part of the quarried sector opposite to the entrance. Here, only a single row of boulders rose above ground level but the quarried face seemed formerly to have had a stone revetment to replace the free-standing wall elsewhere. Our section across the bank in the SW, where the stonework was apparently intact, gave the clearest indications of the wall structure. maximue Th m heighplacee th f o dt moro stonen cm)s bas0 e e (9 thaswa Th .e t f consisten3 a f do laye f boulderro s extending acros fule sth l width, withou foundatioa t n trench lyind an , g either mounw housee lo th friabla f do f n o o , groun d E e ol directle earthe th dth n i surfacn .ys o a , eor Rising above this base and forming both the outer and inner face to the wall, there was a double 72 | PROCEEDINGS OF THE SOCIETY, 1970-71

KILPHEDIR HUT CIRCLE I

i<"*' .,^

£ M (MIIMH"1'1^

5 4 3 I I O

FIG 3

ro f slabwo r bouldersso , standin eacn gi h caseslabo tw , s high. Between these double rowf so facing stones and above the boulder foundation, there was a mass of brown friable loam forming walle core th th .eo t Thi s earthen core appearewale th ll sectional dn i Hut-Circlen si Td Van I I , sI alike coult .I d perhaps represen remnanwale d somth n so ti l y risina ef wa to g abov stoneworke eth , but it seemed to lie just where the rafters of a conical roof would have rested and the earth may have originate this da n sods used instea r togetheo , dof r wit hheathea r thatch slighA . t earthy FAIRHURST AND TAYLOR: A HUT-CIRCLE SETTLEMENT AT KILPHEDIR, SUTHERLAND | 73

KILPHEDI HUR- T CIRCLI E

SECTIONS

SPaf^p^^7^^

PH. P.M.

onqt-Crtf Sub-Soil (INDURATED)

FIG 4 accumulation within the house, immediately below the peat and above the cobbles in the dark grey-gritty layer, may also require some such explanation. Between the inner face of the wall and the centre of the house near the hearth, numerous post-hole r stone-socketso s were found . (fiEac3) g h consiste settina f do f stonego s arounda patch of relatively soft earth penetrating either into the indurated layer or into the stone packing of the floor. The post-holes varied in depth from 5-14 in (12-36 cm) and seemed designed for a pol muct eno h more tha (10-1diametern n i i ) abou5 r 2cm o t4 . They wer t firseast ea no to y t recognis occurrence th s ea e seemed haphazard soot i t n,bu became clear tha majorite tth held ydha in place the vertical poles forming a ring of supports half-way between the outer wall and the centrhousee th raftere f holconicaeth o a o t , f p dso u alsls roofowa obviout I . s tha rine d tth g ha KILPHEDIR O O ra HUT CIRCLI E fa O

O

s o

FIG 5 FAIRHURST AND TAYLOR: A HUT-CIRCLE SETTLEMENT AT KILPHEDIR, SUTHERLAND 75

Stona walling

Original surface

DIAGRAMMATIC RECONSTRUCTIO EARLF NO Y HOUSE (HUT-CIRCLE KILPHEDIT IA ) R

FIG 6

been replaced more than once thert ,bu roos ei differencr mfo opiniof eo recognisinn i gpatterna . Three ovalpossibilitya e t gapsar bu s, 5 showt seems occu i a , g d fi n rsan i necessar visualiso yt e sockete somf o th e f eo us s e morth e tha nthrer pattere o once th o et fi outliert all nTw .t a ,no o sd and yet they are unusually clear specimens. One to the left of the entrance just inside the hut-circle correspondina d ha g outlie Hut-Circln ri could an dI e I indicat woodeea n doorpost reachino gt head height and set back into the sloping roof, inwards from the passage through a very low encircling wall (see reconstruction, fig 6). A central post does not seem to have been employed to support the roof but the hearth was nevertheless off-centre. It was roughly rectangular, measuring 4 ft 8 in by 3 ft 6 in (1-42 by 1-07 m); the cobbled surface was bounded by a low kerb. Immediately around the hearth, the earth was reddened and slightly carbonised, while flecks of charcoal occurred commonly over the floor of the house. Nowhere, however, was there any appreciable ash accumulation and only a very minute quantit charcoaf yo noteds wa l . renewae innee Th th f rlo rin postf go s more than once would indicat occupation ea n appreci- ably longer than just one or two seasons. Moreover, the number of small finds was substantially more than might have been expected fro mvera y temporary occupation e verth y n I aci. d soil conditions prevailing, unburnt bon irod ean n objects would disappear peate th ; , which might have acte preservativa s da lesr efo s durable objects, doe t sees no m havo t e formed until afterwards. The distribution of the find spots within the house (fig 5) may show no more than a concentration near the hearth, but could indicate a high frequency within the innermost ring of posts as though this wer threee latese th eth f .o t Pottery Abou sherd0 14 t s were recovere alln di , though many were small fragments whicd hha been trampled int floore oth ; minor difference fabrie th n ci s suggested that eigh r eve o tpotn nte s could have been involved. vessele Th s see havmo t e been fairly straight-sided, expanding upward simpla o st e rim. Excepr fo t one which is rounded, the rim sherds have been markedly flattened by pinching between the thumb and first two fingers, creating a kind of flange, usually on the inside but sometimes on both sides (fig 7 and pi 13b). Decoratio confines ni shere on do dt wit h diagonal scratchin. g2) acros. (fi no fla e m g7 sri th t The most distinctive group consists of about a dozen sherds with a slightly mottled grey fabric, hard and smooth, and soapy to the feel by reason of a plentiful addition of steatite grits. A nearly complete flat base, 3-3 in (84 mm) in diameter, with parts of the lower wall was recovered (pi 13b), together with two rim sherds, one of which was decorated. The flattened rims and 76 | PROCEEDINGS OF THE SOCIETY, 1970-71

KILPHEDIR - THE: POTTERY

HUT CIRCLE I BACKING

1-2 Steatite

8 Coars3- e sand

HUT CIRCLE H

HUT CIRCLE

J Inches

FIG? FAIRHURST AND TAYLOR: A HUT-CIRCLE SETTLEMENT AT KILPHEDIR, SUTHERLAND | 77 steatite admixture might recal sherde th l s fro pre-broce mth h round-houses with souterraint sa Jarlshof lattefabrie e th th t t rca 1 sitbu 4 thics ei coars d kan comparison ei scarceln ca d e ynb an cite parallela s da t Kilphedi.A surface sherde rth th f eo s with steatite backingi finelo ss y smoothed tha t almosi t t suggest dulsa technique l th glaz d ean e could have been use purelr dfo y aesthetic reasons. Steatite is a fine grained talc which is rather rare in occurrence but is found in small quantities in a number of different localities in Sutherland. It occurs in upper Strath Naver and also some 30 miles (50 km) across the hills from Kilphedir on Loch Shin. The bulk of the pottery from Hut-Circle I is brownish buff in colour 1an5 d contains a plentiful mixture of what appears to be coarse sand derived from the local granite. Most fragments are as roug abrasiven a s ha , though some examples have been wet-smoothed. Several cooking pots with representede ar m ri thic) e wit(1wall£d n th o i t 2kt . mm an hdiamete a p a s) u cm 0 nearlf ro (3 t f y1 Another grou abouf po sherd0 t5 thinnee sar smoothed ran r ofteoutside ar nd blackeneean d with soot. Finally ther anothes ei r grou f abou po hearte foun0 th 5 tf o h d almosE togetheS e th t o rt naturae th f o l p soilto n , o whic vers hi y thin walled, 0-175-0- (4-4-2n i 5 mm), rather smootd han with finer sand. Three very small rim fragments (fig 7 nos 6, 7, 8) were poorly formed and the diamete restimated e coulb t dno . Bone minut5 e fragment burnf o s t bone. Flint 9 small shapeless flakes, all about % in (12 mm) across. Used Pebbles A rounded, coarse grained sandstone pebble 3£ in (9 cm) in diameter and 1-J in (4 cm) thick, was well smoothed on one flat face. Another elongated pebble was cracked across and may have been used as a smoother. Pebbles cracked by heat, which often occur in number on Iron Age sites, were noticeably absent. Hut-Circle //(figs 8, 9; pi 13c) The second house lies 70 yards (64 m) to the SSE of Hut-Circle I on a well drained but exposed sitedimensione .Th slightle sar y greater tha Hut-Circln i , ebeinI g approximatel- y39 (12-12-5t f 1 4 internall) 0m y (fi. Superficiallg8) y stronthera s ewa g resemblance betweeo tw e nth mucd saman e hth e natural horizons were traceabl soie th l profiln ei e beneat hlayea f peato r . A freestanding wall appeared behind the quarried face opposite the entrance; on the W, the encircling bank was barely perceptible. Our excavations were only partial; two 14 ft (4-27 m) squares were opened in the first year and a section was cut through the back wall (fig 9), but small finds were absen attentiod an t thes nwa n concentrate Hut-Circln do hearte subses Th . ehI wa - quently located by a special excavation to make sure that the circle was more than a stock pen. The section through the freestanding wall is difficult to interpret as the inner face had fallen forwar somd dan e slip appear havo st e occurre outsidee th n do . Eartcore walthith f ew o n h i slo l was again characteristic. Clearly, however, the house floor had been quarried to a depth of between cm)0 hearte 9 verd .s Th an hy wa 0 fragmentar(6 t f 3 smala d t an lefybu doubh lo 2 aren tas f ao t identificatios ofit almosn ni t precisel same yth e relative positio Hut-Circln i s n a entrance Th . eI e slabbes wa d rather than 5-Jcobbles t (1-6f - wa wide) 8d m dan , thoug poorls wa t hyi outlined e atth sides. The position of an outlying post-hole on the left on entering again emphasises the likeness Hut-Circlo t littl e . ThereI b en doubeca t that occupatio mors nr leswa e o s contemporaneous. Scarcely any domestic rubbish was observed, and bearing in mind the fragmentary hearth, this might indicate a secondary usage as a stock pen. Pottery One rim sherd was found against the inner face at the back. It was from a cooking pot, diamete cm)r0 abou (3 thic) , t (1witf £d n i 31 tkmm an hflattened a vessee Th . l, 9) flange . (fino m g7 dri had been ill-fired and the sherd is crumbling, but there is no doubt that it is identical with some of the sherds from Hut-Circl. eI Bone Part of a burned tooth, probably of a large mammal. | PROCEEDING 8 7 S OF THE SOCIETY, 1970-71

KILPHEDIR HUT CIRCLE E

5 4 3 2 I O 5 IO 15 3O L_I...I • '______1______1____ ' _ _ i______1

FIG 8 FAIRHURS TAYLORD TAN HUT-CIRCLA I E SETTLEMEN KILPHEDIRT TA , SUTHERLAND 79 KILPHEDIR-HUT CIRCLE U

SECTION

Brown Sandy Loam CreY Sub-Soil i.----i Bleached Gritty [\\] Dark Turfy Layer Iron Pan Horizon f/,/ >,\ Dark Brownish Loam Black Loam Scale I inchr I Foot

I 2 3

Feet FIG 9 Hut-Circle ///(fig 10, pi 14a) This hut-circl exceptionas ei l becaus smals it f eo l sizmarkedld ean y ovat f 7 l 2 shape y b 2 3 , (9-75 by 8-23 m), and also it is located almost on the lip of a precipitous slope down to the Kil- phedir Burn. Moreover, it was constructed within a natural saucer-like hollow between morainic hillocks without the usual artificial platform for the hut walls. Circles of this appearance seem rare in Sutherland and there was some doubt as to whether it was contemporaneous with the others. Little attention was paid to it until 1968; one of the reasons for excavating it was, if possible, to obtain charcoal for dating as only very small and possibly contaminated samples were available from Hut-Circle. II d an sI Perhaps because of the better drainage conditions on the morainic material and on the edge of the glen, the soil profile proved to be somewhat different from that of Hut-Circles I and II. usuae Therth s l thicknesewa 6-1f so peat(15-2f 0n i o ) , 5whiccm deepls hwa y covered witd hol heather beneatd an , h cam stainee eth d grey earth with bleached grits. Iron-pan formation proved to be very patchy indeed and scarcely any 'local' iron-pan had been deposited after the close of occupation. The older 'regional' iron-pan occurred in restricted but quite thick accumulations at uppee th rinduratee limith f o t d coars e layeth n ri e morainic material formin bottoe e gth th f mo hollow. This upper limi exceptionalls twa y irreguladiversifies wa d rseveraan y db l large boulders. photogrape Th plad han n show tha r excavationfo t , Hut-Circl divides wa I deII into four quadrants. The NE and SW were excavated down to the indurated layer, but two 2 ft (61 cm) baulk ssector werSE e,the lef whil in t N untouchedWethe was entrancThe . excavateewas dby extendin quadrantW S e gth . The grey gritty layer beneath the peat had been somewhat problematical in the two pre- vious circles; partl charactes yit r depend face th t n thas leachedi roundeso e t th i t t bu ,d grits rather differentiate t froi d m normal occupation material possiblA . e explanatio d seemenha d to be that it was the remnants of roofing material, perhaps of turves placed on the rafters. In 80 I PROCEEDINGS OF THE SOCIETY, 1970-71

KILPHEDIR

HUT CIRCLE HE t N

FIG 10 Hut-Circle III, however, this gritty layenoticeabls wa r y thick agains peripherae th t l walling and consequently may have been the remains of an earth luting of the dry-stone masonry. The floor of the hut measured 21 ft EW by 25 ft NS (6-40 by 7-60 m) and had been constructed by smoothing out the irregular surface of the natural hollow. There was little sign of quarrying and the source of the filling material is rather problematical. The floor had eventually consisted o cobblfa eartd ean h surface markedly dishe centree muc s th fillin depte a e t ds a th Th .h f ghwa o as 1 ft (30 cm) in one or two pockets. Potsherds occurred under the filling, presumably because they had been thrown away during the actual levelling operation; other sherds of the same type came from the floor itself. The irregularity of the top of the indurated layer and also of the filling material made the recognition of post-holes unusually difficult. Only three or four rather doubtful ones were plotted sectorE N e in.th Clearly, however centrao , n ther s elwa post-hole granito tw r efo , slabs placed sid sidcrackey d eb ean d with heat represente dcruda e heartvere th y n hi middl floore th f ;eo remain peaf occurreso h kerbalonsideo d tas N thert n flaan e s ,bu g.A n th tde o wa slab, presum- ably forming a working surface, had been placed in position half-way between the central hearth . E e anwale th dth o t l FAIRHURST AND TAYLOR: A HUT-CIRCLE SETTLEMENT AT KILPHEDIR, SUTHERLAND | 81 Charcoal, mainly fragments from small branchet hu twigsr e sfloo o e foun th s th f o r,wa n do and especially aroun hearth e interpenetrates dth wa t .I d with rootlet thest sbu e must represene tth vegetation which colonised the floor at the end of the occupation and cannot differ seriously in age from the charcoal. This point is important because we are dependent on carbon from Hut- Circle III for dating purposes (see below, page 89). entrance Th simpls everwa d ean y roughly constructed with large angular boulders most of which seeme havo dt e slippe positionf o t dou , making impossibl attempy ean establiso t e hth exact line of the walls. A flat slab or two and a little cobbling formed the only recognisable flooring. sideE Oalignmene n ,na th earth-fasf to t boulders suggeste dprolongatioa side th e f walno l beyond the limit of the hut-circle by about 8 ft (2-44 m). A somewhat similar alignment may have occurred whole th t e bu structur , ruinouo W to certaine s b e th o est wa n .o The peripheral wall was not unlike those of Hut-Circles I and II though it was more ruinous, somewhat more roughly constructed and much rather loose rubble was involved, especially in the SE. Nor could a central earth core be clearly detected. In general, the width averaged ;about (1-5t f increasin ) t 5 stooabov) I 2 m . cm d e1 NE abou groune (6 abou go t t th f (2-1t n t2 f i t) d7 3m t sugges debrie no th leve d originad n sdi a t an l l heigh morf o t e tha n cm)centr0 e abou(9 Th .t f e3 t of Hut-Circle III was sunk about 1 ft 6 in (46 cm) below the ground level at the periphery of the floor (see section fig 10), which would allow more headroom within this relatively small dwelling. In the NE sector, where the walling was wider than usual and very well preserved, the method of construction was clearly indicated; a double row of boulders (see pi 14a) both at the inner and outer face had been laid down, much as in the case of the western walling of Hut-Circle I. In all, there are a number of minor features which appear to differentiate Hut-Circle III from otherse th t thesbu , e peculiaritie explainee b n sca d simpl adaptatios ya unusuae th o nt l settinn gi a convenient natural hollow. There can be no doubt that this circle structurally is of the same type as Hut-Circles I and II.

Pottery Four wall sherds \\-1\ in (38-57 mm) across were recovered, together with another eleven fragments. The caml yal e fro mlarga e vesse across) r vesselcm lo 9 , (1 t witleas sa n i ht7 walls 0-3- 0-4 in (8-10 mm) thick. There was no decoration but the pot had been wet-smoothed before baking. The form and the coarse gritted fabric are identical with those of the sherds from Hut-Circle II and with some sherde oth f s from Hut-Circl. eI Pounder The only other find consisted of two parts of a large, smooth pebble which had been cracked into at least three parts. It measured about 4 in (10 cm) across and was If in (4-5 cm) thick. It had been abraded along the rounded edges, presumably in usage as a pounder.

Hut-Circle ) (fiV I 11 g This lies between Hut-Circles II and V on a rather steep southerly slope upon which two of lone th g bank boulderf so s converge. p Neithema SutherlanS e O rth n i 6 w d ne Inventory e th r no (Sheet NC 91 NE) indicate the site and in fact it was traceable only as a slight bank with occas- sional stones diameteaboun s i i (6-t f ) t I .10 m t2 r internall maie th nd purposyan trence th f eho cut across the structure was to confirm its existence. relativelA y deep quarr deptn i ybeed ) fachha cm n e 1 mad nearl(6 t f e y2 int risine oth g groun accommodatdo t flooe eth r which even then sloped markedly down toward entrance sth e th en i encircline Th . SE morgo bann s e kthawa nstona e foundation abou (1-t coult f widi ) 5 t m edbu be traced very clearly round the upper side. A flat stone near the centre might have marked the heartsockee theron d posr s hto an ewa t hole visibl trenche th t 3-n ei (90-124se ,t f fro) e 0mcm th wall at the back. No small finds were recovered but the presence of flecks of carbon in the floor and the general form suggested a dwelling in miniature of the type of Hut-Circles I and II. 82 | PROCEEDINGS OF THE SOCIETY, 1970-71

KILPHEDIR HUT CIRCLN E

Approximate Floor

FIG 11

Hut-Circle ; pis13 , 14b (figsV 12 , 15a c , ) Hut-Circle V differed from the others in having such a substantial encircling bank that, by analogy with similar hut circles in the vicinity (Sutherland Inventory, Nos 318, 327 and 328, discussed in more detail later) a souterrain might be expected within it. There was evidence, too, olona f g entrance passage wit counterparo hn Hut-Circlen i t s I-IV. Moreover fifte th ,h circle seemed to be more intimately related to the boulder banks. The site was in a relatively sheltered position in what must have been a slight valley before the peat formed, with a trickle of water at the bottom. A clearly marked alignment of boulders came from away to the NW, curved round in front of the doorway and abutted onto the encircling wall just beyond the entrance in the SE (se 12)g eseconfi .A d alignment paralle firse th t o agailt shora nd camtan trenceW froN he mth FA1RHURST AND TAYLOR: A HUT-CIRCLE SETTLEMENT AT KILPHEDIR, SUTHERLAND | 83

KILPHEDIR T CIRCLHU V E

lNili||il,/,//>. '/A

FIG 12 (actuall searcn i souterrain)a t f ycu ho , indicate dimpingeo tha to encircline t i tth n do g waln o l sideW e . Betweeth twoe nth , tonguthera s egrounwa f eo d whic slightls hwa y roug appearanchn i e seemedirecd a e an b to dt continuatio mooe th f rno (se 14) g visuae efi .Th l evidenc slights ei t bu , this could represent a route for stock from the open grazings on the hill into the round house. Finally, another boulder alignment came towards the circle from the N. Ther smala s ei l pla Hut-Circlf no Sutherlane th n i eV d Inventory (nowhicn i ) . 325h40 g fi , the low boulder alignment in front of the doorway is indicated; this alignment is shown slightly 84 | PROCEEDINGS OF THE SOCIETY, 1970-71 positiof o t seemd ou nan stoo st p wal e shorth lf neato entrancee rth . Insid circlee eth Y-shapea , d partition wall of boulders appears on the plan with the arms bonded into the back wall and the leg of the Y ending in the centre of the floor. A clearly denned entrance passage is also shown. By 1965 passage th , beed eha n almost obliterated with larg eY-shapee stoneth d san d walling seemed havo t e been disturbed sincInventorye time th eth f eo constructioe plath y nb renewar no f o l lambing pens, and also, we suspect, by stone pulling on the part of enthusiastic antiquaries (pi 14c). Similar loose wallinw ,lo showgs Sutherlane i th y nb d Inventory othen 16i r circle preciself so y the same type as Hut-Circle V, notably at Brae (129), Aberscross (279), Caen (318) and Knock- arthur (499 and 500). In the examples we were able to inspect these internal dividing walls were of very loose stone and so wide that they could be regarded rather as platforms. With these other examples in mind, the Y-shaped walling was left intact as long as possible in case it was original, but in the end it proved to be modern. The arms of the Y were not bonded into the revetment: loose th e boulders were clearly embedde pean werdd i t beneatan t wholee no th : heit overlae yth debrilase th t f occupationso , includin hearthe g th difficulte .Th tha s bouldere ytth wa s were found to overlie, with disconcerting precision, a pre-existing kerb-like structure which ran outwards from the back wall to enclose the central hearth (fig 12). This co-incidence arose perhaps because a low bank, markin kere line th bgf th e belowo suggested ,ha dpattera stocf no k shelter somo st e herd who had then pillaged the inner facing wall of the circle for building stones. Apart from the Y-shaped accretion, the central area of Hut-Circle V was covered with grass and bracken, not with heather as in the other circles. The peat, too, was very thin and the soil below was loose. There was scarcely a trace of a secondary iron-pan though the lower, regional iron-pan could be detected above a well-developed indurated layer. The difference in the soil profile as compared with Hut-Circles I and II was sufficiently strong to suggest to us a less mature development, but Mr Romans considers that better drainage on this particular site could provide adequate explanation. After excavation, the entrance was such a prominent feature that it almost suggested the vere passagth y d inty stron an osmala ewa n gdu lencirclin g bank reinforced this impression. Perhaps Hut-Circl Gaelie 'rathe th th n s i ci ' trench-like nameV th e f eo th , N e e valleth o yt Clais Rath Fhinn. The extraordinary length of the passage, some 15 ft (4-6 m), was produced by a sudden expansion of the encircling wall of the house from its average width elsewhere of about 9 ft (2-7 m). This expansion was a secondary feature produced by adding about 7 ft (2-1 m) of wallin originae th o gt l outer face, part f whicso h could stil discernee b l d continuind ol e th t ga width finas it ln (69-10outline I .n s in-form i 3 6 t passag wa e widf ) 3t f d th 2 , y 6 dcm s eb an ewa large slabs on edge with smaller stones packed in the interstices. The SW side appears ragged on plae th n (fibecaus) g12 slabo falled etw sha n inwards als d innee o,an th r corneE N r e stonth n eo sidslipped placeedgef s eo ha t wa dt ;i ou back into positio outline th d dotteplans e ni an th n .do floo e paves Th rwa d with large flat stone whicn si hdecidea d brea t abouka t 1 (3-t 1 f outwards) 4m , again suggests a late addition. There was an upstanding sill at the inner end, while outside cobbling curvine wath s sa tracegr banfa s dbouldera f k o entrancee fronn sth i f o t . The wall of Hut-Circle V was built of stone and earth, much overgrown with grass and heather. In general, it stood 3-4 ft (90-120 cm) high and measured 8-10 ft (2-4-3-1 m) across. The material was kept in place by large slabs or boulders at the outer foot, as shown in section E-F (fig 13), but inside there had been a revetment of large slabs or boulders placed on end, which was apparently a secondary feature with additional stonework above. In places, the stones had fallen inwards and the bank had slumped, suggesting that the revetment had not been securely bonded into the main structure. Originally the wall must have risen at least 5 ft (1-5 m) above the adjacent floor inside, allowin gmuca h larger are usefuf ao l space than wit low-see hth te rooth f fo 71 KILPHEDIR HUT CIRCLE V X SECTIONS

z a

Cobbles o Alignment of Boulders as Remnantf so c HEARTH o P CB

SECTION ACROSS WEST WALL f OUTER REVETMENT Kerb W FACE I

1 H

a CB O

10 15 2O _iFeet I r1 a FIG 13 86 | PROCEEDINGS OF THE SOCIETY, 1970-71 'simple' typ f round-houseo e represente Hut-Circlen di . DiscountinII d an sI g slip diametee th , r of the house after the probable reconditioning was 32-35 ft (9-8-10-7 m). In the western sector of the revetment just to the south of the junction with the Y-shaped boulder shelter ) apart verticao cm tw ,3 , indicate(5 l n corneri 9 entrance t f dth 1 sintern a o et - mural cell or souterrain. It had been blocked with rounded stones which had spilled out on to the adjacen thousee flooth f o r. Walling coul tracee db d inwards fro cornere mth s onl r abouyfo t 1 ft (30 cm). A trench was cut across the bank (E-F, fig 13); it revealed no abnormality on the N Se , confuseth a ther s n o side ewa t bu d, mas loosf so e stone. Certainly some structure n suca s ha intermural cell had existed or had been partly built, but had subsequently been dismantled. It was impossible in the time available to strip the wall down to foundations in an attempt to establish structuree plae th th f no souterrai.A uncommot no s ni n with this typ rounf eo d house (Sutherland Inventory 318s no , 328 sunked )an n chambers seem als occuo t r (nos 344-5). The floor of Hut-Circle V had been excavated into rising ground to the NW to a depth of materiae th d beed an entranc e lha n) bouldee th spreath cm n i s 0 t aboua e(9 dr r ou aret f fa s t3 a a alignment outside slighA . t natural hollo beed wha n house selecte th site thaf o th e s o r th t dfo regional iron-pan, when it had not been worn away by trampling, and the induration layer below it, dipped towards the centre of the house and the entrance where the occupation deposit was much deeper. Indications of occupation were found immediately below about 3 in (8 cm) of peaty sod, in the form of quantities of charcoal from branches of birchwood, some of it unbarked and 3- (8-10cm4n i ) across. Apparentl roofine yth g materia collapsed still-smoulderina ha l o t n do g hearth. At the end of the occupation, the interior arrangement of this round-house was markedly different from tha f Hut-Circlewell-laiA o t . II d d an pavemen sI t continue floorine e dth th f go entrance passage towards t stoppebu , d centrae shorth f o t l hearth (fi, heartg12 h II). Thiy sla withi curiouna s alignmen slabsof t , carefull curvinyand placegend evenldon y aroun centradthe l area of the house, delimiting a tongue of flooring at a slightly higher level. The hearth itself was a circular settin flagstonef go s abou t (1-5f across) 5 t ker2m w smalf lo b o ;a coulld slaben e dn b so t thibeed bu sha napparentlE destroyeW detecteN S e e th th stono y dt n ydb i e robbinge Th . hearth was covered with ash which extended outwards over some flagging behind. The ash was in three distinct layers uppen a ,lowe d ran r deposi blacf o t k woodas interveninn a d han g spreaf do friable orange brown material which was taken to be peat ash. On the flagstones beneath this ash and obviously contemporary with the hearth, there were two distinctive potsherds, one with an everted rim, differing from the wares found in Hut-Circles I and II (see below). A saddle quern la position yi n nea hearte rth jus d htan outsid kere eth b stones. purpose Th e behind this alignmen stonef to sobscures i comin t hu frot e bacge .ou th m th f ko Unfortunately linbeee d th e,ha n somewhat destroyed especiall constructioe th n yi lambine th f no g shelters in the NE. In the NW, however, on approaching the peripheral wall of the house near the blocked entrance to the presumed inter-mural structure, the stone alignment changed character and there seemed to be a double row of rounded stones as shown on the plan (fig 12). Betwee pavemene hearte centre th n th th d t ha ean t running inwards fro entrancee mth , excavatio s continuewa n d downwards int a olowe r stratu f occupatiomo n materiae th d an l rectangular corner of a low kerb was found. This was clearly recognisable as part of a hearth as it was accompanied by a distinct ash-spread. Although we did not attempt to dismantle either the central upper hearth flagstonee th r o , s leading toward entrance sth e passage doub,o n ther s tewa earlien oa f r occupation represente Hearthy db I (fi, 13) g12 . This would see mconfiro t m other evidence that the house had been extensively reconditioned for the last occupation. The earlier house presumably resemble typn di e thos Hut-Circlef eo . II d an sI FAIRHURST AND TAYLOR: A HUT-CIRCLE SETTLEMENT AT KILPHEDIR, SUTHERLAND | 87 A number of post-holes or stone-sockets had been noted and again there were suggestions of alteren a d arrangement, thoug pattere hth incompletes nwa . Fro possibls plae mwa th nt i (fi) e g12 to think in terms of an older ring, diameter about 24 ft (7-3 m), set at a distance of 3-5 ft (90- fro) 15 wale 2mcm th l faccomprisind ean g originally about 11 post-holes; these were traceable mainly in the E half of the floor at the horizon of the indurated layer. A later ring, diameter about 20 ft (6-1 m) and possibly associated with Hearth II, seemed to involve about 6 post-holes of which coul5 recognisede db . Eve , somnso e examples remain unplace interpretatioy an d dan n muse b t doubtful. summarn I woulyt i d appear tha houstypta e subsequentlHut-Circleth s f eo f eo wa I I d an sI y modifie giv o dt round-house eth Hut-Circlf eo wit, e V insid n ha e revetmen bouldersf o t highea , r encircling ban d thereforkan e with more utilisable floor space finallyd an , , wit hvera y much extended entrance passage souterraia f I . r intra-murano l cell existed t belongei , earlien a o dt r phase than the final occupation. Pottery No more than 8 potsherds were recovered. Two closely resembled the coarse sand-gritted ware from probabl d Hut-Circlean , II yd belon horizoan e sI th o g t Heartf no . hI flagginOe nth g withi alignmene nth f stoneo t s around Heart , therhII e wer sherdo etw s which, while resembling some of the sand gritted pots of Hut Circle I in texture and colour, had significant differences; four other sherds seem to go with them. The clay was mixed with grits of a rather larger size, up to i in (6 mm) across and the pot had been wet-smoothed. One sherd formed part of a slightly everted rim (fig 7 no. 10), diameter about 4i in (11 mm) with depressions immediately below the out-turned lip made wit hbluna t stic r boneko t wit finger-tipe no , hth s . Thi(p8) i s coul parallelee db d fro Roae mth d Broc Keisf ho s (National Museu Antiquities)mf o Crosskire ,th k Broc Forsef o als d g hoan 1 fro7Wa e mth thoug t closel lattee hno th rn yi case designe Th . , however vers i , y rea e simplth l d significancean thif eo s sher contrase d th lie n si t wit flat-rimmee hth d ware Hut-Circlef so . II d an sI Saddle quern Found in association with Hearth II, this was a granite slab sub-rectangular in form, measuring 17 by 13^ by 5^ in (43 by 34 by 14 cm) and worn smooth by rubbing on the upper surface to a depth of about i in (12 mm). Pounders, rubbing stones and used pebbles Elongated pebble, 3 by 2 in (7-6 by 5-1 cm), faceted in place rubbingy b s . Elongated water-worn pebble in-22 , 4 y H(10-y ib n ^ i b 5-y 81b cm-6-) 3-y 4b 8cm rubbe chipped dan t botda h ends, expeciall broadere yth . Rounded pebble, 52 y (13-J|n b i in-32 1 y £b by 7-7 cm-8-9 by 5-1 cm), with the broad end chipped and narrow end rubbed. Elongated pebble, 4^ by 2 by 1 in (10-8 by 5-1 by 2-5 cm), rubbed at one end. Disc, flat and irregular, apparently smoothed, diameter 2^- (6-4-7-n £y i 3 b 1-y 6b 3 cm). Several rounded quart z(2-n i ) pebbles o 5-t 52 1cm o t 1 , in diameter, selecte slinr fo gs d a stones . Flint Five pieces of flint, mostly flakes but there was one yellow brown example, 2 in (5-1 cm) long, whic scraperpara s f ho twa . Bone Eleven well-burnt fragments, white in colour, with a striking resemblance to cremated bone but they cannot be identified as human. The Improved Land (figs 2, 14; pis 12b, 15b) Withi wholepossibla sits e s nth ea wa t i , distinguiso et separato htw e group lowf so , over- grown heaps of stones of the type which the Sutherland Inventory describes as 'mounds'. At least 10 occurred on the higher ground in the NE near Hut-Circles I and II, while a second group of 8 were scattered nea southere th r nsitee limith .f Moro t e almost certainly existed belo peae wth t facn i s t locatewa e on d d whilan e investigatin f Huto nature g-W th sligha f eo WS t e banth o kt Circle III. The 'mounds' varied in size between 8 ft (2-4 m) and 25 ft (7-6 m) in diameter but they rarely rose more than about 2 ft (60 cm) above the peat. Although the stones were neatly piled together, they exhibite boundarsiga o dn f no yexternaf o ker r bno l walling. Two were sectioned; both vicinite wereth n f i Hut-Circl yo t otherwisbu , eII e they were chosen at random (pi 15b). They consisted of boulders and stones thrown together indiscrimin- botn atelyi d h cases,an , rested directl soin yo l withou suggestioy tan layea f npeao f ro t underneath. PROCEEDINGS OF THE SOCIETY, 1970-71

IRON AGE SETTLEMENT-KILPHEDIR CULTIVATION

- •"~ *- -•.! > j. i- ,1- , * * a/ ,,- * ;, % * i' i" * * ,„-„.-*" * * " y

i -•••J-' --u - - ^ ^ •± ± iL ,i, ,., i \\\\\

Moor 7\' Clearance Cairn / -"; Boulders ;:: Stones 0 5O 1OO 15O 200 ' I \ feet '/ Approximate Arable Area

FlO 14 There was no sign of a grave deposit nor did the sloping site seem at all appropriate for a burial mound reasonablo n . e Therb n eca e doubt tha primare th t y purpos freo t es patcheewa lanf so d for cultivation by piling the surface stones and boulders together into clearance cairns. A single flint scraper about £ in (13 mm) across was found near one of the cairns. e alignmentTh f boulderbankw o s lo sd predominatean s e centrath n i dl area around Hut-Circl t firsA . t esightV , they suggested ruinous field walls t thebu ,y wer facirregularn eo i to t . Closer study showed that they were made up of unsorted collections of stones, heaped along the edges of what appear to have been cultivation plots. A trench across a characteristic stretch near Hut-Circle III seemed to confirm this interpretation. There was brown soil on the flatter, down- side of the alignment, i.e. on the 'field side', and a stone-spread above with no sign of a built wall. relevans i t I noto t e tha continuouo tn s ban apparens kwa t alon mooe edge th g th f reo above eth site, correspondin heae th do g t dyk latef eo r times. FAIRHURST AND TAYLOR: A HUT-CIRCLE SETTLEMENT AT KILPHEDIR, SUTHERLAND | 89 Both between the boulder banks and also in the vicinity of the clearance cairns, the ground was smoother and devoid of stones as compared with the undeveloped moorland. It was not possible to delimit very precisely the old cultivation plots, partly because of the mantle of peat, and no high degree of accuracy can be claimed for the differentiation of the arable ground as , however 14)showg is fi t .I n n(o , clear tha totae tth l are improvef ao d land cannot have exceeded fou fivr o r e acre hectares 2 probabls ( wa d )an y less, bearin minn gi d that minor patche rougf so h ground suc thoss ha e upon whic clearance hth e cairns rested must have interrupte arable dth e ground generan I . lansmoote s th ld wa rolling d aspecthan W gradientS e r , o Th wit. S ha s varied whe; 6 grounmuce : s na unde1 th s froo s ht a d wa mp ru cultivationabou2 1 : 1 t , soil wash must have been a difficulty. Far more serious, however, was the problem of preserving even a low level of fertility, for the soil was thin, stony and acid, judging at least from its present state. It seemed almost impossible to grow crops on such land, but the visible evidence of the clearance cairns and boulder banks could not be doubted. Presumably under these conditions, spade cultivation would e onlth ye b possible method aroun betweed dan clearance nth e cairns. This leave more th s e elongated stretches define bouldee th y db r bank problema s sa ; under more auspicious conditions, it might have been tempting to think in terms of plough cultivation with an ard. The nature of the ground rendered any attempt to discover plough marks quite out of the question.

Soils A detailed report on soils and pollen analysis of the peat is contained in Appendix 2 by Dr Durno and Mr Romans of the Macaulay Institute of Soil Science; we wish to acknowledge our appreciation of their work and also to thank Mr D Futty who undertook the sampling. While the Appendi matterxo speaktw r itselfr o s se fo requir on , e discussio t thina s stage. Our first indication of a date for the settlement came when Mr Romans suggested on very general grounds that the peat might have started to accumulate on the site of the earlier hut- circle t lonsno g afte Sub-Boreal/Sub-Atlantie th r c transition (Godwin's Zones Vllb/VIII). Durno and Romans18 have indicated the period as about 2500 BP. This development could perhaps be regarde extension a s da n downhil generae th f o l l peat blanke describes a t Grampians.e th r dfo 19 It is noteworthy that there are only very slight signs of cultivation in the soil profile - a mere scratching of the surface must be envisaged. We have maintained that two periods of cultivation appea involvede b o rt associatee on , d wit clearance hth wite boulde e cairnon hth d san r banks; this depend t r investigationmerelsno ou n yo t Kilphedisa evidence th t als n bu ro o e from other Sutherland sites as discussed in the Introduction to the Inventory. Nevertheless, the pollen analysis show burse son Plantagof o t pollen which indicates cultivation nearby seconda t bu , , much less marked peak would result only from cultivation withi regioe nth n generally r RomanM . s sug- possibilite th geste s u o dt prehistorif yo c fields downhill fro sitee mth , where greater supplief so wood-ash might be available. No trace of such plots can be found before the old fields of the early nineteenth-century clearance settlement reachede sar . Nevertheless loate ar abandoo he t w , e nth hypothesis of two separate phases of cultivation of the fields on our site, especially as the Carbon 14 datings indicat distinco etw t period f occupationso .

Carbon 14 Determinations The excavations of Hut-Circles I, II and IV produced charcoal only in minute quantities, while in the case of the lower horizon in Hut-Circle V unmixed samples could not be obtained with certainty. Hut-Circle III, whic specialls hwa y excavate obtaio dt n charcoal, yielde dsmala l amount from twigs mainly of birch and hazel but with the possibility of some alder (identifications Donalr byM d timBrette th et a lecture, Botane th n ri y Department, Universit Glasgow)f yo . This | PROCEEDING 0 9 SOCIETYE TH F SO , 1970-71 latter sample represents the only dating material for the earlier phase of the settlement and was obtained from occupation material on the hut floor and immediately beneath peat. A determina- r mads AlatioD nwa n ei Walton's laborator e Chemistrth n yi y Department, Universitf yo Glasgow followss a , : GU 299 2370±40BP This date would accord wit idee hth a thapeae th t t bega accumulato nt site eth shortln eo y after the presumed date of the Zones Vllb/VIII transition, i.e. very broadly about 500 BC using half-lifew ne C14r e fo th . Unfortunately, this firse singl datth 4 t r occupatioeeC1 fo t Kilphedina r is otherwise unsupported. Charcoal was available in quantity from the reconstructed Hut-Circle V. It occurred in block mucs sa across) h cm a isn5 2 ( , from roofing pole hearthe s whicth fira n falleed o o t . hha n no The charcoal lay immediately below turf and specimens were prepared by cutting away the outside surface avoio st d root contamination sample theTh .s enwa divided intportionso otw y B . special arrangement witcalibratins r Waltohwa D o nwh g equipmen e tim th n 1966 i et a t e ,on portio retaines nwa Glasgo n di othee th rd wsen an r compariso fo t Lamone th o nt t Geological Observatory, Columbia University respective Th . e determinations fro same mth e sample wers ea follows: GU 10 1908±60BP GU 11 2064±55BP 7 1922±60B6 U G P 106L 1 2100±8P 0B t anotheYe r determination fro same mth e charcoa bees lha n provide laboraw d frone e -mth tory at the East Kilbride Centre, Glasgow, which is closely in accord with L 1061 as follows: SRR 3 2100±50BP averagn combinee A th r efo d figures would centr 202n eo 0 BP (approximately 13l 0al BC t )bu the results must be regarded as significant and indicative of the lack of absolute precision inherent methode inth .

GENERAL DISCUSSION firse Th t occupatio placee b site n dth eca f provisionallno perioe th n yi d aroun B0 n Co d50 both the C14 and pollen evidence. The five round-houses of a relatively simple type, with a roughly rectangular heartcentraa d han l rin postf go supporo st roofe th t , nee t necessarildno y have been in use simultaneously but they can hardly be regarded as a successive series. The floor construction involved considerable effort and, once one of these level platforms had been com- pleted, it seems reasonable to suppose that it would have been re-occupied had a succession of single family units occupied the site as a whole. The general similarity of construction of the houses and strong resemblance of the pottery seems also to point to a group occupation. The very small are arablf ao e ground indicates eithe considerablra e relianc stockn eo f mucr o ,o e hus fish, game and wild produce generally. Perhaps the location of the clearance cairns amid the arabl spada ef o indicatlandiggin r y ee o dus ma e geth stick rather tha ploughna associatioe .Th n of these simple round-houses with the cairns cannot in fact be proved from the site, but the close correspondenc mano s n yeo other site Sutherlann si Inventorynotee s da th n di vera s 20i y strong argument. FAIRHURST AND TAYLOR HUT-CIRCLA : E SETTLEMENT AT KILPHEDIR, SUTHERLAN1 9 | D

KILPHEDIR DISTRICT

DISTRIBUTIO T CIRCLEHU F O NS BASED ON O.S. 6" MAP 1964

FIG 15

An important point whic emerges hha tha s dblankei e tth peaf to t bega forno t m soon after the occupation of the hut-circles at Kilphedir. This has had the long term effect of precluding later agricultural exploitatio fossilisee th d nan d landscape survives beneat peae hth t mantl stils i ld ean recognisabl surfacee th t ea . locae Th l distributio showinp nma g site hut-circlef so s (fig bee15s )ha n compiled froe mth new OS 6 in map (Sheets NC 91 NE and NC 92 SE) with slight modifications from personal 92 |; PROCEEDING SOCIETYE TH F SO , 1970-71 knowledge. It will be noted that the haughlands along the river are devoid of sites and that very few circles occur below 200 ft (62 m) which Mr Romans suggests was the approximate upper limit f continuouo s woodland. Again sitew fe ,s occu t somr t abovbu (12f ) e0 4m nea e40 heade rth - streams of the Torrish Burn in the NW corner occur at over 800 ft (250 m) so that no simple altitudina drawne l b limi n ca t. totae Th l numbe f hut-circleo r Kilphedie th n si r largo ares s aei tha t cannoi t t readile yb explaine relativterme n di th f so e attractio smala f no l pocke goof o t d otherwisn lana n di e poor region. It suggested comparison with the population at the time of the Clearances in the early nineteenth century whe densite nth f populatioyo alss onwa very high plaA . n showine th l gal building existencn si 181n e i bee s 0ha n preserved amongs munimente tth Sutherlane th f so d estate alreads ha d yan been describe previoua n di s paper.21 Without going into detailst onca s i e t i , clea hut-circlee rth thatl al f bee,i d sha n timee occupie populatioe on ,th y an are e t da th af nwoulo d have been muc excesn hi f tha so f 1810 o welts i t lI . known that just befor Clearancese eth e th , faminestandard an w slivinf lo d o wers gwa e common, although deepee cultivatioth n o r s nwa shelteree soilth n so dStrathe flooth f o r . deterioratioA climate th n ni e might perhaps invokee b accouno dt acceleratioe th r tfo f no peat formation afte perioe th r f settlemendo extreme t th eve , bu t nso e povert thesf yo e upland soils make inherentlt si y unlikel hut-circleye thath l Kilphedital e th n si r district were occupation i n simultaneously. Shifting cultivation most probabl involves yi d with fresh land-winnin t shorga t intervals. It must be noted, however, that within the altitudinal limits of the hut-circles, there is no suggestion that the ground had been developed to the maximum in terms of cultivation. In fact, far from dealing with a distribution showing a fine adjustment of the sites to local conditions of slope, aspect, drainage and soils which could only be worked out over generations, we may well seeine b resulte g th vera f yso incomplete development pioneea , r distributio termn ni f smalso l patche f cultivatioso n occurring very irregularly. In other words the large number of hut-circles and clearance cairns, which are such a feature of centra easterd an l n Sutherland, woul resule dth seee tb neithemo t densa f ro e populatior nno o lonfa g perio settlementf do visible Th . e remains must have survive mosa o dt t unusual extent becaus developmene th f eo peae th tf whicto itselhavn hy i t shor fema periocu e tth cultivatiof do n of these upland areas at the altitudinal zone involved. Too little evidence is available to say who these colonists could have been. The havy yema bee descendante nth Bronza f spopulatioo e eAg n whic turned hmorha a o dt e sedentar liff practises o e a y ywa morn di e southerly area Britaif so n fo thousana r d year r moreso . Thehavy yma e been struggling wit pooe hth ropee soilth nf so uplands before they had adequate techniques to clear and drain the lower straths. Whether these colonists had technically progressed to the Iron Age is questionable; effectively they would seem to have been still largely in a Stone Age. five th esitd f hut-circleeol o e clearance th Th d san e cairn re-occupies swa d some threr eo four hundred years later by the occupants of the reconditioned Hut-Circle V perhaps in the later second century BC difficulties e spitn .I th f eo bearint bu , minn gi evidence dth e from other sites, we associate with this re-settlement the boulder banks which are so characteristic a feature at Kilphedir almose .Th t dun-lik Hut-Circlw ne e beee s forth ha nf me o V mentione d above. Socially, one family group seems to have replaced the old loosely clustered settlement of up to five families. A change in potting techniques is also indicated. similaA r hut-circle-which seem havo st esouterraia n occurs acros Claie sth s Rath Fhinn (Inventory no. 327); the builders had again utilised a pre-existing hut-platform and the traces of the older circle can be detected outside. The same elongated entrance occurs, and the familiar boulder alignments are to be seen in the vicinity. Across the Kilphedir burn, another dun-like FAIRHURST AND TAYLOR: A HUT-CIRCLE SETTLEMENT AT KILPHEDIR, SUTHERLAND | 93 circle (Inventory shor. a 328 s no ) tha souterrai ndoorwa w entere insidlo e a th y dn yb ei e walth o lt left of the entrance. Boulder alignments occur nearby but are less obvious, and once more, one oldee th f ro circleo tw sr o appea closn ri e proximity. Further afield similaa , r typ hut-circlf eo e occurs in the neighbouring glen of Caen; it is again set within what appears to be an earlier circle and is characterised by a short souterrain (no. 318). Two others are located in the upper Strath Bror Knockarthut aa r (nos 498-9 whicf )o formee hth agait rye n seem havo st e been placed within a pre-existing circle. In two examples at Kildonan (nos 344-5) a sunken circular chamber seems to have been entered through the wall of the hut from the interior, as though to replace the souterrain twoe maie th firsth e ,f nth o t entrancn I . provides ewa d with check doora e r sfo Th . writer Sutherlane th f so d Inventory were fully awar thif eo s distinctive typ hut-circlf eo e but, from descriptione th frod san m personal inspectio numbea f no f themo r , some variety occura d san precise differentiatio t possibleno s i nr distributio Ou . (fip ) includeg1 nma s som whicn ei e hth chief distinguishing featur expansion a s ei walline th f neitheo n go entrance r th sidt f a eo s e(a Hut-Circle V) though it is not clear that this trait in itself should be diagnostic. These strongly built hut-circles, usually with the expansion at the entrance, often occur singly and the distribution is strictly confined to eastern Sutherland. The area includes Strath Brora, Strath Fleet and the Strath of Kildonan but so little is known of the archaeology of the neighbouring county of Ross and Cromarty that this statement may be misleading. It is interesting perhapd an s relevan noto t t e that another distinctive typ f dwellingo e e so-calleth , d 'wags' describe Inventorye th n d i r Caithness fo , hav similarlea y limited distribution, this time being confined almost exclusivel southero yt n Caithness.22 They have been include distributioe th n do n map (fig 1) for comparison, though very little is known of their age. Most appear to have had encirclinn a g wal f stoneo l within which stone endn slabso t , werse , e use supporo dt roofe th t , instea innen a f dro rin postf Kilphedirt go a s sa . This highly local phenomenon representee th y db wag perhapy sma explicable sb gooterme n i th df s o buildin g stone availabl southern ei n Caithness. occupationo tw e HoKilphedie th r th w t fa s a fittee rb sitdn eintca pattere Latoe th th ef no Bronze Age/Early Iron Age as currently envisaged, is a very difficult problem, and has already been briefly discussed elsewhere.23 We can see no need to invoke mass invasion to account for a type of land exploitation and of house form which had been current at least in southern Britain for centuries. Moreover, we would make the point that some features of the type of dwelling represented by Hut-Circle V might perhaps be regarded as vaguely foreshadowing some of the characteristic feature brochse th f so cultura .A l continuum fro Late mth e Bronz tentativels i e eAg y postulated.

REFERENCES . Feachem1 , 'UnencloseW R , d Platform Settlements', Proc Antiqc So Scot, xciv (1960-1), 79-85. ScotlandM A C ,R Inventory. County2 e th r fo of Roxburgh, For7 59 t . Eildo no vo , ii l n Hill North. ScotlandM A C ,R Inventory . County3 e th r fo of Sutherland (1911) xxvp , . ScotlandM A C ,R Inventory . 4 Countye th r fo of Caithness (1911) xxxviip , . 5. R C A M Scotland, Inventory for Orkney and Shetland (1946). Sutherland.M A C 6R . 1. R C A M Sutherland, pp xxix-xxx. Sutherland,M A C R xxixp . 8 . 9. Graham, Angus, 'Cairnfield Scotland'n si , Proc Antiqc So Scot, (1956-7)c x . 7 , 10. Abercromby, John, 'Exploration of Circular Enclosures and an underground House near Dinnet, on Deeside, Aberdeenshire', Proc Antiqc So Scot, xxxvin (1903-4), 102. Ogston, Sir A, 'Prehistoric Antiquities of the Howe of Cromar', 3rd Spalding Club, 1931, 22. 11. Thorneycroft, Wallace, 'Observations on Hut-circles near the Eastern Border of Perthshire, North of Blairgowrie', Proc Soc Antiq Scot, LXVII (1932-3), 189. 94 | PROCEEDINGS OF THE SOCIETY, 1970-71 12. Fenton, Alexander, 'Early and Traditional Cultivating Implements in Scotland', Proc Soc Antiq Scot, xcvi (1962-3), 264, (see page 274). 13. Curie, Ale , 'OExaminatioe xO n th Hut-Circleo Tw f Stratne o th Kildonanf n shi o , Sutherlandshire, one of which has an Earth House annexed', Proc Soc Antiq Scot, XLV (1910-11), 18. 14. Hamilton , ExcavationsC R J , t Jarlshofa (1956). 15 Phemister. d Wilsonan , W , G (revise ,J d Anderson C1946)G ,J . 'Tal othed can r Magnesium Minerals Chromite'd an . Dept f Scientifio . Industriad can l Research, Geological Surve f Greayo t Britain, Wartime Pamphle. 9 . No t Sutherland.M A C 16 R . 17. Curie, Alexander O, 'An Account of the Partial Excavation of a "Wag" or Galleried Building at Forse, in the parish of Latheron, Caithness', Proc Soc Antiq Scot, LXXV (1940-1), 23. 18. Durno , 'EvidencE S , e regardin Rate g th Peaf eo t Growth' Ecol,J , (1961)9 4 , 347. 'Pollen Analysi Peaf so t Deposit Eastern si n Sutherlan Caithness'd dan , Scot Geog Mag, (1958)4 7 , 127. 19 .Romans d Durno , 'Evidencan C , E C S ,J , Variationr efo Altitudinae th n si l Zonatio Climatf no n ei Scotland and Northern England since the Boreal Period', Scot Geog Mag, 85 (1969), 31. Sutherland.M A C R . 20 21. Fairhurst, Horace, 'The Surveys for the Sutherland Clearances 1813-20', Scottish Studies, 8 (1964), 1. Caithness,M A C 22 xxxixR p . . 23. Fairhurst, Horace, 'Kilphedir and Hut-Circle Sites in Northern Scotland', Scottish Archaeol Forum, 3 (1971), 1-10.

APPENDIX I Sutherlane Th d Field School The investigations at Kilphedir were made possible through the organisation of an adult field- schoo archaeologyf o l . This Sutherland Field School represent experimenn sa t whicf believo e he w b o et more than local interest in these days of high labour costs and a widespread amateur interest in field work and especially excavation which can so easily become destructive through excess of enthusiasm and absence of technical skill. 1961n I conferenca , hels Dunden de wa i Scottis e th y eb h Regional Grou Britisr pfo h Archaeology, on the place of the local society in archaeology. Delegates requested a course of instruction in field work and simple methods of survey with lectures and the practical planning of ancient monuments. Kilphedir seemed to be the ideal site for such a venture. The monuments, of considerable variety, were contained withi nwell-definea d area regio e attractivn whola th ;a students s o nt a e wa e woul o eon swh d havo et devote parlimitea f to d summer holida sucyo t h instruction; above McLellan J all r ,M , Directo Educaf ro - Educatioe th tiond an , n Committe f Sutherlaneo d gav invaluabls eu e suppor placinn i t McLeoe gth d Hostel at Golspie and transport at our disposal. The School was held in the summer of 1962; Mr Gordon Petrie of the Geography Department of the University of Glasgow gave instruction in survey. While at Kilphedir we became interested in the problem presented by the hut-circles. When the student Schooe th t sa l requeste dcontinuatioa developmend nan projece th f o t t wit excavatioe hth a f no selected site, the investigation of some of the hut-circles seemed a worthwhile subject for their enthusiasm. Four seasons' work at Kilphedir followed with from 12-30 members taking part. Many discussions on various aspect excavatiof so problemgenerae n i th d lan s arisin t Kilphediga particulan ri r providew dne line thoughf so stimulua d an t s whic pleasura s i t hi acknowledgeo et woule w d d ,an than k membere th f so Schoo their lfo r hel sustained pan d enthusiasm. When required, labour, mainl supplies fillinr ywa fo , gin d by Sutherland County Council, who also provided the tools and an indispensable 14-foot ladder. Mrs Dolores McBride also deserves our special thanks for her care in tracing the plans. We have also to offer Durn E sincerr cam o S Romans ou C r o wh eFuttD C eD fro J d r thank Macaulae r yman M ,th M o st y Institut Soif eo l Researc tako ht e soipead lan t samples abov d discuso t an , l evarioual e sth s horizonf so soie th l profile which dominate ofteo ds day-to-dae nth y excavation. FAIRHURST AND TAYLOR: A HUT-CIRCLE SETTLEMENT AT KILPHEDIR, SUTHERLAND | 95

APPENDIX 2 Kilphedir - Hut-Circle Excavation Site . Durn . RomanE C . . S oC d b . yJ san The site at Kilphedir is situated at an altitude between 400 and 500 ft (120 and 150 m) on the N Strate sidth f Kildonaf eho o n about 3 Helmsdalef i o mile W sN ) (5-6.km more Th e freely drained soil thin so s site range from peaty podzols wit hstrona g thin iron-pan (which are found on the hut-circle site and above it), to podzols without this iron-pan (which occur between the site and the A879 road below). Patches of relatively smooth ground surrounded by small stone cairns, which have been describe 'possibls da e area formef so r cultivation' generalle ar , y founn di near transitioo e rth n area between thes soio eltw types microtopographe .Th site generalls eth i f yo w ylo moundy and the soil parent material is a morainic till derived from granite. Lower down the valley slope ther) increasinbelon m a s e0 i w(9 t abouf g0 admixtur30 t acif eo d schist material wit granitee hth . Thin peaty or peat gley soils occupy the wetter patches and hollows between mounds. presene Ath tbot y hut-circlee da thth surroundine th d san g moorlan coveree dar substantia a y db l peatthicknessn i (30- n i ) y 2 5layecm 1 o t ,r indicatin) rangincm 0 g(2 gtime fron i th tha f me8 o t a t occupation the development of a peaty layer can scarcely have begun, though the regional thin iron-pan alreads wa y well developed presens i t i somewhaa s n a i ,t t dilapidated conditio surface e th th n f no eo indurated B3 horizon, where this forms the floor of Hut-Circle I. The presence of a secondary thin iron-pan forme debrie th hut-circlee n di sth lefn o t , afte deserteds site eth rwa , indicates that conditions favourable to the formation of thin iron-pan persisted for a while after the abandonment of the site. This suggests that formatioe th thif no n iron-pa thin no s sitbees eha n associated wit initiae hth l stageaccumulatioe th f so n of a peaty surface layer. establishmene Th thif to s settlemen that whata , s tton timewa soi,a l transitional between podzod lan podzol with thin iron-pan suggest probabls thawa t i t y nea uppee rth rvalle e edgth f eyo side woodland, wher tree e thinned eth sha intt dou o open birch pine woodland wherd settlemen,an e eth f o t t ou itsel s fwa the direct line of sight from the valley bottom. Chemical analyses are available from a control profile (Kilphedir no. 3 peaty podzol) 100 yards (90 m) uphill from Hut-Circle I, from inside Hut-Circle I (Kilphedir no. 1), and from a profile (Kilphedir no. 2) 4 yards (3-7m) outside the same hut-circle. Phosphate analyses only are at present available from a profile (Kilphedi take) patchee 4 th . nf rno o possiblf fro so e mon y cultivated ground associated wite hth hut-circles. A general consideration of the chemical analyses indicates that the figures for profiles 3 and 4 are well withi limite nth s which have been recorde r peatdfo y podzols develope granitn do e parent materian i l Aberdeenshire (Charr Countesswellseriee th f so s Association1). When the figures are considered in detail it is noticeable that the total phosphate contents of the three uppermost mineral horizons of the Kilphedir control profile no. 3 (1-4 in, 5-7 in, 7-8 in) (2-5-10 cm, 13-18 cm, 18-20 cm) are somewhat lower than those recorded on Aberdeenshire profiles.1 The total phosphate contents of the two uppermost mineral horizons of the possibly cultivated field no. 4 (0-1 in, (0-2-) slightl e in 7-5-1, ar 5 5) cm 3- y3cm higher than controthose th n ei l profilthoug, 3 . t higheeno h no r than might be expected from an Aberdeenshire seminatural profile. The figure of 197 mg P2O5/100 gm soil recorde t d17-2 a 'fielde th 0(43-5n n i i slightl 's i profil) 4 1 . cm ye no highe r than would have been expected in a seminatural profile and is in the position within the profile where an archaeological accumulatio f phosphatno e migh expectee b t locatede b difficulo s dwhethet i y t sa I . o phosphate t t th r e pattern in the field profile no. 4 is significant or not, but if it is, then the depth at which the 197 mg P2O5/100 gm soil occurs would suggest that cultivation of this field was associated only with the very earliest years during which Hut-Circl e I itseloccupied s wa f . Thin sections made fro msampla f soieo l taken fro uppee mth r minera 'field4 . l horizono ' profile th f no e confirm tha soie th tl parent materian i s i l fact derived essentially from granitic material suggesd ,an t tha t beet therno ngreaa s eha t dea mechanf lo - ical disturbance, and that any 'cultivation' that may have taken place was in the nature of 'scratching' surface oth f e rather than t generallploughingno d di yd disturan , soie bth l more) thacm n0 abou(1 n i t4 belo minerae wth l surface. The phosphate pattern in the no. 1 profile taken inside Hut-Circle I indicates that there has been a significant archaeological accumulation of phosphate. Comparison (fig 16) with a hut-circle previously examined at Dalnaglar in Perthshire (Dalnaglar no. 1 profile2) shows that while the post occupation | PROCEEDING 6 9 S OF,'THE SOCIETY,: 1970-71 ; t: ,;:>,;•: <, ' ^ ;.;; J V. •;.> T*;' 'iH;;n

0 30 0 20 0 10 0 40 SCO 0 30 0 20 0 10 0 0 0 20 ttO O M 0 10 0 0 10 200 0 0 20 0 10 B

KILPHEOIR N?4. KILPHEDIR N?3. KILPHEOIR N!2. DALNAGLAR N?5. KILPHEDIR, N«l. DALNAGLAR N?l. FIELD PROFIL ' CONTROE L PROFILE PROFILE OUTSIDE PROFILE OUTSIDE PROFILE'INSIDE HUT CIRCLE H°. ' I' . PROFI'LE'INSIDE' HUT CIRCLE N*I; :HUT CIRCLE »°l.' HUT CIRCLE N«2.: - : ' ' v: " ' ; Fig 16 Comparison of total phosphate distribution in soil profiles from Kilphedir and Dalnaglar. Total P 6 in ' ' m-' e ue''•'••'r • lOO • ••'-•''• • e " • soi l '••• ' •'•'"•'"">••: • •'"••: '•"•' ' '•••/•• , ••"-:••<• • -,-.•. ••.,.•••.• a 6

.10-

ISO-1

0 20 40 60 80 90 0 ZO 40 50 0 20 30 0 10 0 10 0 20 40 0 20 40 60 70 BETULA PINUS ALNUS ULMUS QUERCUS CORYLOID % ARBOREAL POLLEN

-( 175 7. A. P.)

0 200 400 600 800 1000 1200 0 20 40 60 BO 100 0 0 4 20 0 2 40 0 0 10 ERICACEAE (REDUCED SCALE) GRAMINEAE CYPERACEAE FILICALES PLANTAGO FIG 17 Kilphedir peat profile expresse f totao % l tres da e pollen (recent) accumulation of phosphate at Kilphedir has been less than at Dalnaglar, the archaeological accumulation has been considerably greater at Kilphedir than at Dalnaglar; suggesting that the period of occupatio t Kilphedina eithes rwa r longe r moro r e intensive tha t Dalnaglarna . Pollen analyses have bee nshalloa carrien o t wd ou pea t monolith taken nea hut-circle rth e siten o , peate th y layer overlyin peate controth e ygn th o layer d l an profil s3 whic . formeed no hha d ove hute rth - circle profile no. 1 and the nearby no. 2 profile. FAIRHURS TAYLORD TAN HUT-CIRCLA : E SETTLEMENT AT KILPHEDIR, SUTHERLAND

PEATY TOP.

0 4 60 0 2 0 0 10 0 5 0 0 1 100 2 2 00 ID 0 0 4 0 6 2600

BETUL, A PINUS ALNUS ULMUS qUERCUS CORYLOIO % ARBOREAL POLLEN JS

0 20 40 CO 60 100 0 20 40 60' SO 100 0 20 40 60 0 10 0 10 ERICACEAE' GRAMINEA ' ' E CYPERACEAE FILICALES PLANTA60

MINERAL SOIL I LL I

0 10 0 20 M 0 10 PINUS ALNUS ULMUS

80 0 10 0 20 40 CYPERACEAE FILICALES FI 8 GKilphedi1 r profil; value2 . eno s expresse totaf o l% tres da e pollen

pollee Th n zonin peatf go s from these norther welt ye lnt defined thisitesno n i s si d contexan , e th t age of the base of the peat monolith is 'probabl3 y post Atlantic' (post 5000 years BP) - that is either 'Sub Boreal 'Sur o ' b Atlantic' (bas t 250ea 0 year hut-circlse basBP)e th f Th e. o e peaty layer (Kilphedis i ) 1 . no r somewhat earlie datn ri e thapeate base th n th f y eo laye yard4 r s (3- outsid) 7m ee (KilphediTh . 2) . no r base of the peat profile is again somewhat older than both (differences involved being probably of the hundreo tw r o de orderatheon f ro r than thousand years)f so . Howeve comparisora pollee th f nno patter weee grasnd th lefdan y b speate florth t n ai Monolit h (Plantago Graminead an particularn ei ) from Kilphedir (figs 17-20) with that previously establishet da Dalnagla Perthshiren ri 2 indicate versa y close correlation hut-circl e profile th th n n .I i e d e an profil 1 . eno 4 yards (3-7 m) outside it, no. 2, Plantago pollen is absent from the uppermost mineral horizon and present inchew fe sa further down uppermose Th . t mineral horizo botn ni h profiles probably represents 'post occupation debris' indicating that Plantago polle distributes nwa d ove site erth whilst this hut-circls ewa inhabited. A second small accumulation of Plantago pollen is present at the base of the peaty layer of peate hut-circl base th yf th e o t surfac controa e d th e f an profil eo 1 l . profil- thi 3 e no s. seconeno d occur- renc absens ei t peatfroe base th m th f y eo surfac yard4 eprofil, e 2 layeth s. f (3- eno o r) outsid 7m e eth hut-circle and instead there is a rise in Gramineae (grass) pollen and a marked drop in Ericoid pollen. This could be interpreted as localised burning of the surface vegetation around the hut-circle - the remnant wall-bank actin fira es g a break. Both accumulation f Plantagoo s polle recordee nar peae th t n di monolit h from Kilphedid an r appear to match the pattern at Dalnaglar in east Perthshire.2 notabls i t I e tha earliese th t t accumulation of Plantago pollen, quote percentaga s da totae th f leo arboreal pollen in the peat monolith at Kilphedir reaches the very high figure of 175 % as compared with 98 PROCEEDING SOCIETYE TH F SO , 1970-71

0 20 40 CO 80 90 0 20 40 50 0 20 0 10 0 012 0 0 0 6 0 0 4 20 % ARBOREAL 8ETULA PINUS ALNUS ULMUS QUERCUS CORYLOID POLLEN

0 200 400 600 WO 1000 1200 0 20 40 eo so 100 120 0 20 30 0 20 0 20 'ERICACEAE BRAMINEAE CYPEMCEAE FILICALES PLANTAGO

MINERAL SOIL

I [

0 M0 2 10 0 0 9 0 O 4 S (0 0 2 O S 0 4 0 2 I ID0 0 7 0 6 0 0 4 20 BETULA PINUS ULMUS QUERCUS CORYLOID '/.ARBOREAL POLLEN

0 2 0 0 2 0 0 O 2 B 120 0 0 164 0 0 0 10 0 6 0 6 0 4 0 2 0 ERICACEAE ^ "SRAMINEAE CYPERACEAE FIUCALES PLANTAGO * REDUCED SCALE FIG 19 Kilphedir profile no. 1; values expressed as % of total tree pollen

10 to 12% at Dalnaglar. This indicates a local source of higher than regional intensity, and in this connec- mentionee b y tioma t ni d that Plantago polle virtualls ni y absent from peat profiles previously examinen di Caithness and Sutherland.3 Plantago (plantain) is one of the weeds often associated with periods of cultivation.4 It seems probable therefore that the earliest and most local of the possible periods of cultivation correlated wit accumulatioe hth n of Plantago contemporaneous wa s wit occupatioe hth Hut-Circlf no eI tha d othee an tth r post-date destructioe dth thif no s hut-circl relativela y eb y short time longet ,no r thana few hundred years. Additional pollen analyse peate sth carrieyn o tur t fdou overlyin fiele gth d profile t sho presencno e (Kilphediw d accumulatioth y di an ) 4 f e o . rno n of Plantago pollen baseithee th f eo t ra uppermose th n peae i r th o t t minera lsoil e laye thith n I .f sro absence of Plantago pollen coupled witha high percentage of Gramineae pollen the field profile is comparable with profile no. 2 taken 4 yards (3- outsid) 7m e Hut-Circl ericoie th t dbu figure, eI relativels i y hig compares ha d wite hTh profil. 2 . eno patter f phosphatno e accumulatio e fielth d n ni profil e does however suppor correlatioe th t y an f no 'cultivation' on the site of the field profile with the earliest of two phases of occupation at Kilphedir. FAIRHURST AND TAYLOR: A HUT-CIRCLE SETTLEMENT AT KILPHEDIR, SUTHERLAND | 99

PtATY TUP 0' so 100-| 150- 200- J50- 300- lit

0 0 4 20 0 20 40 0 S2 O 0 0 2 00 4 0 0 2 10 0 0 10 BETULA PINUS ALNUS ULMUS QUEMUS CORYLOI. ARBOREA'/ D L POLLEN

0 200 400 600 1000 0 20 40 60 80 100 0 20 40 60 80 100 0 20 0 10 "ERICACEAE BRAMINEAE CYPERACEAE fILICALES PLANTASO

MINERAL SOIL

0 10 0 7 0 ( 0 4 0 2 0 0 20 40 0 0 0 4 0 23 0 20 02 0 BETULA PINUS ALNUS ULMUS CORYLOID % ARBOREAL POLLEN

0 40 60 120 160 200 0 40 60 120 160 200 0 ill 0 20 40 "ERICACEAE *GRAMINEAE CYPERACEAE PLANTAGO * REDUCED SCALE FI 0 GKilphedi2 r control profilvalue; 3 . eno s expresse totaf o l% tres da e pollen

ACKNOWLEDGMENT The chemical analyses were carried out at the Macaulay Institute for Soil Research by the Analytical Section of the Pedology Department under the supervision of Dr H G M Hardie.

NOTES soile Th s . roun1 d Aberdeen, Inveruri Fraserburghd ean , HMSO 1963. . 'Th2 e excavatio circulao tw f no r enclosure Dalnagart sa , Perthshire' Stewart,C E Scot,t M , An Proc c So (1961-2)v xc , 134. . 'Polle3 n Analysi peaf so t deposit Eastern si n Sutherlan Caithness'd dan DurnoE S , , Scot Geog Mag, 74 (1958)1 , , 127-35. . 'Fores4 t clearanc Stone th n ei Age', Johannes Iverson, Scientific American, 194 (Marc3 , h 1956).