PREHISTORIC SETTLEMEN DURNESN TI S by R. W. K. REID, F.S.A.SCOT., G. DAVID, F.S.A.SCOT., and A. AITKEN SUMMARY THIS paper is devoted to a study of the correlation between ancient monuments of the Neolithic period, Bronz d Iro ean physicae n th Age d san l backgroune th n di Durness Peninsula, Sutherland. The study is divided into two sections. The first describes the physical background of the area. The second attempts to analyse the distributio e archaeologicath f o n l material quantitativA . e summary appearn i s Appendix B. A schedule of ancient monuments is included under Appendix A, but, e paucitn vieth i f wo f detaile o y d information studo n ,s beeyha n e madth f o e monuments themselves. INTRODUCTION firse Ath s 193ts a n editioPersonalityr i Cyrilon e o 2Si x Th gag Fo lf no of Britain^ demonstrate relationshie dth p between early settlemen chale limestond th kd an an t e Lowlane hillth f so d Zone mais Hi .n thesis distinctiorestee th n do n between porous soils and non-porous soils, the former based on limestones, sands and gravels and the latte heavon r y clays. Porous soil equateshe d with primary settlemen clayand t s with secondary settlement. For some time now archaeologists have tended to think termn i 'claf so y soil mixed san forestk doa , inimica earlo t l y settlement'. Thia s si t unreasonablno e generalisation generalisatioa f i , f thio n s kin desireds di . When applied to small areas it may, however, be misleading, if not plainly erroneous. This is especiall e casyth e when s oftea , n happens, cla s equateyi d with boulder clay. The term 'boulder clay' is doubly unfortunate, in that ground moraine of the kind referred to is often not clay and frequently does not contain boulders. For this reason the term 'till' is to be preferred. Grimes shows 2ha n tha Anglesen i t y Neolithic, Bronze Age Irogroupd e an ,n Ag s settle arean do til f preferencsn o i l localitieo et s free from tillalse H o. demonstrated the increased value of the detailed soil map in distributional studies compared with either the solid geology or drift map. There can be no doubt that when published soil maps are available the results obtained are frequently more precise, a fact amply demonstrate wore th Cramptof ky o d b Webled nan Soutn yi h Wales.3 Maps showing soil serie soir so l association t obtainablno aree e sar th ar discusseefo thidn i s paper. Although withi y geologicaan n l classification ther e soiar el differences resulting from the character of the superficial deposits, in the area discussed a close correlation between the solid geology and soil type may nevertheless be demonstrated. studAe bassa th earlr f yo e fo y settlemen geologe tth drifd yan t maps therefore have enhanced value. r Cyri1Si l Fox Personalitye ,Th of Britain (1932), 53ff. . GrimesF . 2W , Antiquity, xrx (1945), 169-74. 3 C. B. Crampton and D. Webley, BBCS, (1960), 387-96; (1963), 336-37. 21 22 PROCEEDING E SOCIETYTH F O S , 1966-67 DURNESS 23 The area covered by the survey is the Durness Peninsula, Sutherland (fig. la). It first came to the notice of the writers during the excavation of a souterrain in 1964 and IQ65. In the course of examining Durness Parish for comparative material

two facts becam1 e apparent. First, there appeare numbea e b hithertf o dt o r - oun recorded ancient monuments. Second parise th , h coul dividee db d into well-defined, contrasting geological regions accounn O combinatioe . th f o t thesf nfactoro o etw s decides wa t i attempo dt archaeologican a t l surve Durnese th f yo s Peninsula, treating it as a pilot area on the basis of which, if successful, other areas in the NW. Highlands of Scotland migh investigatede b t . possibls wa t I begio et n compilin distributioe gth beforp nma e field work com- menced. Archaeological bibliographies yielde dsmala l amoun informatiof o t d nan Archaeologicae th l Divisio Ordnance th f no e Surve Edinburgn yi h allowee us e dth of their card index. A party of undergraduates from the University of Glasgow directed by the writers undertook the field work, which was completed in a period of two weeks in July 1966. Basic information was plotted directly on Ordnance Survey 6-inch sheets by field parties separately responsible for the archaeology or geology of a given section. In practice this resulted in any section of the total area being

visited at least twice2 . A certain degree of overlapping of sections was intentionally introduce givo dt e some accurace checth fiele n ko th df yworko mape Th . s are, there- fore complets a , possibles ea . They show monument apparenf so t Neolithic, Bronze Age, and Iron Age date plotted alongside a background of solid geology and superfi- cial deposits. GEOLOGY The ages of the rocks of the Durness Peninsula range from Lewisian to Lower Palaeozoic, specifically Cambro-Ordovician. Distribution is defined more by faulting than any other single factor. Faults have brought the limestone against the Lower Cambria d Pre-Cambriannan thao s , t ther e contrastinar e g adjacent rock types. Ther closa s ei e correlation between rock typsoid ean l character t musi t e b tbu , remembered thae elementth t f climateo s , vegetatio relied factore nan ar f soin i s l formation thesn I . e cool, northerly latitudes Nort e closth o eht Atlantic storm tracks higa h proportio dayf no s wit hcorrespondingla raid nan insolatiow e ylo b y nma expected e PeninsulTh . exposes ai reached dan elevation sa ovef no r 250 . O.D.0ft , so that a high percentage of the area, all the surface above some 400 ft. is commonly subject to wetting by low cloud. Above 400 ft. ground conditions are only inter- mittently e resuldryTh .s tha i uppee t th t r layer f soie subjeco s ar l leachingo t t . Vegetatio nutrienw lo f o n te expecteb statu n ca sn suc o d h soils. Calluna, Erica (heathers) and the poorer grasses are dominant, plants which leave acid humus residues and contribute to further leaching and to peat formation. e DurnesTh s limeston f dolomites mado i e p u ed limestonesan s , with chert occurring fairly commonly. The significance of the limestone as a parent material is t onca e apparent. Ther sucs ei abundancn ha calciuf eo m carbonate that leachins gi

1 Discovery and Excavation Scotland 1364, 51; 1365, 39. . NW SE 6 5 4 4 ,, SW 6 4 , SE 6 3 , 2 O.SNE 6 4 . , loo-kilometrNW 6 4 , NE e 6 squar3 , : sheetsSE e NC 7 3 , PROCEEDINGS OF THE SOCIETY, 1966-67 DURNESS 25 insufficient to remove it entirely from the soil. Moreover, the maximum elevation of this roc . Figureft k6 typ28 s ei availabl soilr eDurnesfo n so s limestone indicatH ep values of 6-5 to 6-8.1 The Lewisian comprises mainly biotitic and hornblendic gneisses with frequent migmatisation. It is resistant to weathering and chemically very varied figureo N .available ar s soilr efo s develope Lewisian do n H gneisp t sbu values of less than 6-5 might be expected. Peaty podsols, common on the Lewisian, should exhibit pH values of under 5-0 in the top horizons. Chemically, quartzite is at oppositthe e extrem Lewisiato e n gneiss. These arenaceous Cambrian rockare s liabl physicao et l weathering, especially frost shattering d largan , e area f bareo s , shattered quartzite pavement occur. Becaus e extremth f o e e impoverishmenf o t nutrients vegetation tends to be poor in type and often scanty, particularly at higher elevations. Having demonstrate line dth k between solid geolog soilsd e yb an t musi , w no t emphasised tha solie th t d geolog source parene th th y e f eb neeo tt materialdno n I . Durnese th s Peninsula superficial deposits often constitut parene eth t material. Where this is the case the nature of the superficial deposits may materially alter conditions from those normally associated with the rock types already discussed. In general, the till has a tendency to coarseness, although its composition is varied t leasa s i ta t spartialli y derived from rocks groun situ.n di Deposits thin northwards. maximue Th m thicknes aree sth a appears n centrei . some ft b 0 Nato n e5 t d o . Grid Ref. 425612 limestone th n I . e zone where ther hollowe ear basinr so s containing till, drainag sufficientls ei y poo havo t r e engendered peat accumulatio growte th d hnan of sphagnum. With the exception of occasional patches, ground moraine does not, othee th n r o hand, occur abov e. O.D ft som 0 . e50 Erratic foune b t y aboveda sma 1000 ft., and it is not improbable that a, possibly thin, covering of till has been removed fro hile mth l slope erosiony sb . Moraines occur locallPeninsulae th n yi . . sid particularle E eThosar e th n eo y hummocky, creating small areas totally unsuit- able for settlement. Deposits of sand are confined to Faraid Head and to the small peninsula and its south-eastern margin lying betwee Kyle n th BalnakeiDurnessf ee o d th an n y O .Ba l distinctioa p ma madns i e between fossi 'modernd an l ' dunes modere Th . n dunee sar mobile, while fossil dunes, although sand-blows occur in them, may be considered stabl thes ea y hav ethia n coverin finef go , springy turfexposurn A . sand-bloa n ei w at 39206674 afforded evidence of two periods of soil formation below present ground- level (fig. 2). Near Loch Lanlish at 38436811 there was further indication of three soil surfaces lowese th whicf n o to , hcirculabeea d ha n remainint e builtth rhu n I . g sand-blows, where a sufficient depth of material has been removed, a soil below the present one is easily identifiable. Wind erosion has often been markedly arrested at secon e levee th th f lo d soil surface thin O s. soi t 37766 lt 39066a a d numeroue 7an 3 ar s example dry-stonf so e dykes, cairn loosf so eexampl e t circlestoneson hu d a .f an e,o The available evidence suggests a period, or periods, of sand movement followed by a perio stabilitf do y sufficiently long botsoia for r o settlemenlt d h fo m an tako t e place,

Those writers are indebted to Mr D. Futty of the Macaulay Soil Institute, Aberdeen, for information on soils1 . 26 PROCEEDINGS OF THE SOCIETY, 1966-67

inches 0 Sand with humic material

11 High sand content Sand with humic material 18 Sand 26 30 Peat: probably localised 33 Bleached white

Sand

51

SOIL PROFILE: SAND BLOW AT '39206674. Fig.2,

but it is problematical whether or not these soils may be correlated on a regional phenomenoe basisth f associatee I .b y regionaa ma s nt i di e witon l h some external factor, perhaps climatic change n presenO . t evidenc s equalli t ei y possible that localised movement of sand of the kind demonstrated by the current sand-blows is involved biotia , thif so I c. s si factor , suc overgrazins ha summery dr a n gi , coule db critical. Pea bess i t t develope Lewisian do n gneis r wherso e ther covea s ei f til n o ro l relatively flat or undulating land. Much of the peat is hagged. Where this is the case oftes i t i n possibl obtaio et nprofila e pea e basdowe th whic n th f ti e o o n t h tree stumps and large pieces of wood occur. Examples were observed on the Lewisian, and even quartzite localitien i , t elevationa d san s where tree longeo sn r grow. While y thisma or may not be evidence of climatic change, there can be no doubt that forest clearance of lower hill slopes and the depredations of sheep have modified the vegetation. One result has been a dislocation of the drainage pattern, particularly on the E. side of e Peninsulth a wher a braidee d stream patter s commonni . Under conditionf o s heavy rainfall water drains fro hillsidee mth s mor lesr eo sheetsn si , emphasisind gan illustratin essentialle gth y poor drainage. GENERAL ARCHAEOLOGICAL INTERPRETATION The most prominent feature of the distribution map is the concentration of sites on the limestone. Of the ancient monuments shown some 9% occur on the Lewisian and quartzite which constitute some 80% of the total area. Prehistoric societies evidently favoure lightere dth , freely-drained, less acid soils which developee th n do limestone. This accords wit knowe hth n distributio medievaf no earld an l y modern settlement. Those pre-Clearance townships whice . th sid E hf o ee existe th n o d Peninsula, and not on limestone, seem to have been small, late, and were probably DURNESS 27 associated with fishing e present-daTh . y townshi f Laipo artificialls di y sited dan situated, in that it belongs to the Clearance period. Crofting in the immediate neighbourhood has never been successful and the township can be considered a fail- ing community. The gneiss and quartzite, much of it above 400 ft. O.D. and with a high proportion of steep slopes, has always been more or less ignored. Four of the seven prehistoric sites in this zone appear to be funerary. Of the remaining three, site Seanachaisteaf eth o l (40566943:82), promontora y fort, was coursef o , , chosen

with a view to defence but is1 , nevertheless, backed by cultivable or pasture land. The souterrain at Portnacon (425o6i2g:8i) is associated with a small area of 2 flatte coasre lanth n td partiallo y forme streay db m deposition. Abov souterraie eth n are slopes suitable for grazing. Finally, whereas all the other sites in the Peninsula examplee ar f monumentso . mainlansNW typicae Scotlandf th do f o l rounde th , - house (40496101:80)3 is exotic. It lies at a high elevation in an area of broken, poorly-drained land. ARCHAEOLOGICAL INTERPRETATION BY PERIODS Without a precise chronology for the prehistoric and early historic periods in . ScotlandNW impossibls i t ,i offeo et r meaningfu lperiod e dateth r sfo s involveds i t I . also often difficult to relate a particular monument to a particular period on the basis of the available evidence. Consideration of the distribution of finds period by period is nevertheless desirable, for, despite the inherent difficulties, other facts emerge when this is attempted. Evidence of Neolithic settlement is scanty. It is almost certainly represented by a much ruined cairn (38976609:25). A structure at 38846556:14 may possibly be the remains of a chambered cairn, but precise identification must await excavation. more Th e monument o northerltw e well-draineth n f o y o s si d limestone wit hthia n coverin tilf go l merging into sitesand y othee dr , Th a althougals.s ri n oo h thers ei evidenc mucf eo h peat accumulatio t leasna t sinc adjacenn s ea wa t thu Iroe nAg constructed limestonA . e outcro whicn po structure hth buils ei t overlooks gently sloping land wit hthia n coverin loamf go yfacte tillTh s. accord wit conditione hth s which might be considered desirable by early agriculturalists. Soils are light and freely drained and may be placed in an intermediate group between the very light, inherently less productive, sands, and the heavy clays. It is surely significant that although funerary monuments account for the greater Peninsulae evidence parth th Bronze f n i o t th e weighf e eo eth ,Ag distributiof o t n remains on the Durness limestone. Six certain Bronze Age cairns and one probable cairn marke standiny db g stones stretc lina en hi from 3689615 Keoldalef o . E o 1t . There can be no doubt that good pasture and well-drained soils attracted the builders cairnse ofth directl;s i eac e hon y associated with good grazin patca r g o cultivabl f ho e lande cairn Th oftee . sar nshouldee builth n ridgea o tf t necessarilo r no , e th n yo highes tpositio a point in appea and ,be n whicto r prominenhis t relativ onceto e

1 Digits appearing afte grie th rd referenc eschedule refeth o t r ancienf eo t monuments. (Appendi) xA 2 R. J. Buxton, PSAS, LXIX (1934-5), 431-3. . MathiesonJ , PSAS, LIX (1924-5), 221-3.

3 C 28 PROCEEDINGS OF THE SOCIETY, 1966-67 settled localities caire Th t 38876641:3.n a probabls 9wa y intende visible b o dt e from Keoldale arabln a , e area whicrecorn o s hi havins da g been cultivate r somdfo e considerable time. At Ach' na h-Anaite the tomb is close to the shore between the . contourft positions 0 it 5 place-namt e d 2th sbu 5 an s a , e suggests relates i , gooo dt d land. The cairn at Sarsgrum conforms closely to this pattern, contrasting with the finding CramWebley.f d so Valan e n th Glamorgaf p eo to n I 1 n they suggested that burial sites had been pushed to the edge of the settled areas, perhaps to minimise the possibilit deae th df y o affectin livinge gth . quartzite e cairnth Th n so e for mloosa e agglomeration. With other, smaller, cairns (77; 78), the nature of which cannot readily be determined without excava- tion, forthey myBronzma a cemeterye monumente eAg th l Al . constructee sar f do frost-shattered quartzite boulders inclusioe Th . occasionaf no l lump quartzitf so n ei Neolithic and Bronze Age funerary monuments is a well-known occurrence, but it is probably altogethe fancifuo to r explaino t l , even partially cairne sitee th th f , so under discussion in these terms. The locality has a high proportion of bare, quartzite pavement with little evidence of soil formation. Vegetation is sparse and poor. t impossiblWhilno s i t ei e that ther eBronze wasth n i ,e Age ,thia n coverin tillf go , since stripped off by erosion, there is neither evidence of this nor necessity for such a hypothesis. The cairns were not necessarily associated directly with any form of settlement t eveno , n with temporary summer settlement, suc mighs ha usee b t y db shepherds or cowherds. There may have been totally different motives for the choice cairne ofth siteabovl e sli Al e 100. e th . contou 0ft r with inspiring views eastward across Loch Eriboll to Ben Hope and beyond. Cam an Righ is situated above 2000 ft. O.D. On clear days it is possible to see far to the N., E., and W., and the structure itsel vers i f y conspicuous visibls i t I .e from almos positioy e an tth n no Durness Peninsula, and from a considerable area beyond. If the position of Carn an Righ does not belie its importance it is possible that other cairns were attracted to the area on account of its 'sanctity'. It must, however, be admitted that a relative chronology for the cairns has not been established. They may, perhaps, be associated with Bealac easiehe th Loc f ro Seilga hn routee on , s acros Peninsulae sth e th n I . recent past, Bealach Mor use s N.e littla conjunctio,n d th wa ,i o et n wit hferra y yb travellers journeying to Tongue. Before modern roads round the sea lochs were built, a more direct route which crossed the lochs and peninsulas appears to have been generall usen yi . Monument st onl attributablno y e morar Iroe e eth n numerouAg o et s than those of earlier periods but also include well-attested dwellings. Apart from funerary monuments, the only definite evidence for settlement in the Bronze Age may lie in numerous small heaps of loose stones, possibly field-clearance heaps, centred on 38806525 and 38886578. From one of these mounds came a fragment of Late Bronz potterye eAg . three th f e O Irositee nsAg outsid limestone eth e belt positione th ,pro e th -f so montory fort and souterrain have already been discussed. The site and situation of the round-house must remain something of an enigma. Reference to the distribution 1 Crampto Webled nan y (1960), 391; (1963), 335. DURNESS 30 PROCEEDINGS OF THE SOCIETY, 1966-67 map (fig. 3) shows that a concentration of monuments related to this structure occurs thren i e separate regions Outee th n rI .Hebride Northerd san n Isles some thirty-five round-houses are known with an outlier on Tiree, several on Islay, and two reported in Caithness ; some sixteen wags are known in Caithness. The round-house at 1 2 Durness is thus well outside the focal zones of the distribution pattern. It must, however, be remembered that no matter how carefully a distribution map of these monuments is compiled, the result is what may be called an immature distribution map. Such a map not only reflects the survival rate of a particular class of monu- ment, but also the state of archaeological knowledge at a given moment in time, a reminder whic particularls hi y . Scotland appositNW d relation ean i . . TherN o nt e stily lma remai ncasa regardin r efo round-house gth t Durnesea exotics sa . Nearly relatee th l al d structure Westere th n si Orknenn i Isled Shetland san yan beloe dli w . contourft 0 5 Tign e A .h th Clach elevation Barra n o feett 0 a s a95 i . f no Eve n no 3 the E. coast, where generally higher elevations may be expected, the highest recorded elevatio waga r nfo , Allt Preas Bhealaich s alread soms i ,ha feet. 0 s e80 A y4 been remarked, ground condition s Durnese . O.Dabovft th 0 n i . e40 s Peninsul onle aar y intermittently dry, and at 950 ft. there is a high incidence of hill fog. Agriculture in the proximity of the round-house would be impossible though animals could be grazed and fish obtained from the hill lochans. There must have been a telling reaso induco nt constructioe eth mose th f tdwellina o remot f ne o on leasd n gei an t hospitable placePeninsulae th n si therevidencs o n A . s ei climatif eo c amelioration, neither her r elsewhereno Scotlandn ei , durin perioe gth d whe round-housee nth s were in use, an explanation in economic terms would be difficult to sustain. The thought tha reasone tth s were social, tha people tth e were strangers perhaps, is , t no , unreasonable t sucbu , hsuggestioa t susceptiblno s ni proofo et . Both duns (38466539:22; 38396753:46 associatee ar ) d with well drained, gently sloping land. The availability of cultivable land backed by good grazing seems to have been the critical factor, for, although both structures are built on small pro- montories, neither is well defended on two sides. mose th r t numeroufa t circley b Hu e sar s clas monumenf so t representee th n do (figp ma . ic) t musI . rememberede b t , however, that although the here yar e being treated as a group this is a reflection only of lack of detailed evidence from NW. Scotland singla t eNo . exampl bees eha n excavate Peninsulae th n di . Froe mth evidence availabl estandara dt regioe circl typth hu f n ei o n would measure from diametern i . ft 0 4 2o ;5t walls constructe doubla f do e rin boulderf go s infilled with entrancrubblee th d an , e facing approximatel . AlthougySE h some two-thirdt hu f so circles measure from 25 to 40 ft. in external diameter the extreme range varies from ft.0 4 , o perhapt . in 6 s . 1suggestin3ft g functional differences. Whethet no r o r variation in form, examples of which are shown in fig. 4, implies differences in function or date, it is almost certain from evidence elsewhere in Scotland that hut circles were used over a long period; it is among this group of monuments that 1 Discovery and Excavation Scotland 1960, 44. 2 A schedule of round-houses and wags appears in Appendix G. 3 Discovery Excavationand Scotland 1965,so. * RCAM (Caithness), . 249No . ssot- o 0,

MAS OEEKI FROM 'X* AND CUMPCD/C' T'Y TO OETTAlKl A. FL-AT

H oA *a c- 3 s 3> c Ig fs a* 7 o c

\ ^ tto

I 5 !7 > » i 5? ?5

O3 32 PROCEEDINGS OF THE SOCIETY, 1966-67 Bronze Age as well as Iron Age living sites must occur. In this respect, the results excavatioe oth f t Wesna t Plean, Stirlingshire relevant.e ar , excavatoe Th r found 1 tha housa t e typica fora f mo l Late useth e been d i Bronz ha ne replaceeAg n a y db Iron Age farmstead. It was calculated that the house timbers might have lasted for thirt r fortyo y years thad totae an ,th t l duratio occupatiof no havy nma e extended over some eighty years. This raises another point. If house timbers had to be replaced every thre r foueo r decadee b o st house therreasoo th d n y eha s ei nwh rebuilt on exactly the same site and a single family may well have occupied two separate hut circles at different periods fairly close in time. Apart from the question of survival of evidence there are thus several other pertinent reasons why any attempt to postulate population densities based on numbers of huts would be, to say the least, premature. A cursor sufficiens i p y ma emphasis glanco t te th t ea importance eth limee th f -eo distributioe stonth n ei t circlessingla hu t f no eNo . example occurgneise th r n so so quartzite. With only three exceptions the huts shown fall into four more or less well denned groups. Grou comprisee pon s those structure peripherar o s , whicon e , har lto the area of fossilised dunes. A second is situated near bodies of water, either Loch Borralaid Kyle th Durnessf er o h o e remaindeTh . locatee ar r d along r closo ,, eto the limestone outcrops which trend in a SSW. direction from Loch Caladail to Ach' a' Chorrain, wit hsmala l concentratio t eithena r extremity. Discussioe th f no elevation of these structures is irrelevant, for the limestone attains over 250 ft. only in isolated bare outcrops, and its maximum elevation is a mere 286 feet. There is thus the propitious combinatio calcareouf no s rocks proportioelevationw w lo lo , a d an n, of steep slopes. Within the limestone zone it is nevertheless possible to distinguish between used apparentld an y unused areas t circleHu . s hav t beeeno n foun lann do d whict a s hi present under cultivation. This is, of course, negative evidence of the most unreliable maio tw n kinde areaTh .cultivationf so t Keoldala , Sangomored ean , have been i use for some considerable time and there is no reason why the soils could not have been worked by prehistoric peoples. The corridor running between Allt Smoo and a line drawn from Sarsgru . edgE Sangdevoif s tracee ei y o th y an m earlo f f ot dBa s o o y settlement. This corridor comprise basia s n wit hcoverina tilf whicn go o l h con- siderable peat formation has taken place. The marshy valley which debouches on Loch Croispol must als regardeobe negativedas . Overlooking Loch Borralaidhis aren a roughf ao , broken ground compose limestonef do , patche whicf so bare har e frettedd an singlA t .circl hu e e (38646686:16) overlook a smalls , gently-sloping hollow which could have provided a patch of arable. This is one of the three excep- tions outside the four main distribution groups. remainino tw e Th g exception generae th o st l patterstructuree th e nar s overlook- Balnakeig in l Bay e exampl.Th t 38266867:5ea n externai . s onl1i in y6 l . 14ft diameter and that at 38446880:52 measures only 18 by 15 ft. Both structures are resting on bedrock for the sand cover is minimal with many small pieces of chert on e surfaceth . Vegetatio s scantyni , consisting principall f patcheyo f thio s n grass. Steer. A . 1,K PSAS, LXXXIX (1955-6), 227-51. DURNESS 33 The sites are exposed, open to winds from the sea, so that even in summer there are often strong breezes. Possibly these small huts were occupied only seasonallyn i , summer, and might be regarded as shielings. Only four structures occur within the zone of fossilised dunes, two in sand-blows (37736673:7) 38436811:49) whero tw e dept th ed f sanho an , d accumulatios ni considerably less than normal for the area (37636732:10, 11). The presence of three soil surfaceevidence th d san e for land-use seconbasee th n do d soibees lha n discussed elsewhere. With wall corbellef o s d construction structure th , lowese e th (49 n o t) 1 soil is unusual. No dimensions are recorded as the monument was only partly consideres wa visible t i d dan , undesirabl disturo et bsita e which hopeds i t y i , ma , excavatee b futuree aree th n dTh ai . occupie sand-blowy db relation si totae th lo nt are fossilisef ao d dune t seemsmali s s i d san l reasonabl assumo et e tha numbeta f o r structure buriede sar t alsI . o seems likely thaknowe th t n sites other tha core nth - belle lated d an structurIrore e Bronzth nf Ag o e eSub-Boreaar n Agei r fo , l conditions e sandth y areas would doubto n , , suffer from seasonal drough summen i t r which would limit their full usefulnes pastures a s sprino t s d autumn wettee gan th n I r. conditions of the Sub-Atlantic the pastures would presumably be improved. It is unlikely that the stabilised sand would be used as ploughland for the top-soil would be blown away very quickly. In this respect, the site of the corbelled structure is significant t impossiblno s i t I . e sites thai tilln t i dto , small patche whicf so - hap peared to be present in the very lowest levels exposed in the sand-blow. This would accord with what is known about soil conditions immediately to the N. of the stable sand zone, where coarse til coveres i l thia y ndb laye sand.f ro large Th e numbe t circlehu f ro s arean peripherao bettef l s o al san e e th r dar o t l soil. Soils often hav admixturn ea sanf et o hav dbu emuca h higher proportiof no humic material. Adjacen groue th o pt t centre 36767n d o numbe a e 8ar probablf ro e clearance cairns, although thes contemporaryee b nee t dno evidens i t I . t that this periphera ladvantage grouth s pha cultivablf eo e land backe excelleny db t grazing. sitee thosf Th s o t circleehu s included within secone th d group exhibit interesting features. The huts at 38546730:40 and 3860673114.1 are sited on the narrow zone betwee shore nth Locf eo h Borralaid steed han p limestone cliffs. Cultivable lans di negligible although a small patch of such land is associated with the two other huts on the loch shore (38406696:17, 18). At the present time many sea-birds nest and roost on the island in the loch suggesting that at least the former pair of hut circles may have been seasonally occupied at a time when eggs, or even young birds, were available. Similarly, the huts on the shore of the Kyle of Durness, particularly those on the raised beach (37286349:56, 57 and 37196327:58, 59), suggest fishing, possibly with stake nets lowes s whes it ,a larg tidt b e a nth teb s e i e expanse sanf so e dar left temporarily dry. Again, seasonal use of these sites, during the period of salmon runs, is not impossible, especially as the River Dionard is today rich in salmon. The final group, between Loch Caladail and Ach' a' Chorrain, is in three diffuse cluster sfivt circled oftenehu an x ssi ,wit hfurthea r four strun t alongou d gan at the foot of the outcrop S. of Cnoc na Moine. A typical site for these huts is a 1 p. 25, supra. 4 3 PROCEEDING SOCIETYE TH F O S , 1966-67 structural bench overlooking a small area of flat or gently sloping land, always at a premiu Highlandse th mn i excellenn A . t exampl groue th s ei p centre 392660n do . Those farmsteads on the bench had the advantage of a dry site; those below the benc nowadaye har s covered with peasphagnumd an t . Ther thus eha s been con- siderable peat accumulation on the natter land with which they are associated and which was, presumably, farmed in Iron Age times. The tendency to build on the E. side of the ridge is marked in its northern half. Experience of conditions has shown why this is so. Even in summer the E. side has markedly lower wind velocities and there is some protection from the rain. Farther S., Ach' a' Chorrain is sheltered by the hills to the W. so that a cluster of six hut circles occurs on the W. side of the ridge. CONCLUSION Consideratio f agriculturano l potential indicatee limestonth y wh s e zons wa e favoured by prehistoric peoples. When the digging stick and spade, or cas-chrom, wer usen ei , befor introductioe eth doub o ploughe n s th f tn o wa possibl t ,i cultivato et e slopes of steeper gradient. Slopes on the Lewisian and quartzite are, however, often excessively steep n thesI . e northerly latitudes soils develope thesn do e rocke ar s leached and the upper limit of cultivation is low. Soils developed on sand were also avoided unti exceptIroe e th ln Ag , perhaps, seasonally eved an ,n during this period their use must have been for grazing. It seems to have been realised that inherent fertilit sandf yo y soil lows si tha d ,an t ploughing would lea wino dt d erosion. Sandy soils, on the other hand, produce excellent pasture and, as their thermal conductivity is high, warm up early in spring. At a time when the keeping of young animals through the winter was difficult the early growth of young grass must have been a considerable advantage. realises i t I d that this pape poses ha r d many more problem ssolveds thaha t ni , and has indicated the limitations of the techniques employed. It has, nevertheless, at least shown the validity of the physical background as a basis for the study of pre- historic settlemen givea n i t n region. Unquestionably mora s i e p soie usefu,th ma l l base than the geology map, but detailed soil maps are available for only a very limited t likelavailablnumbee b no o e yt ar areaf mucr o rd e fo Scotlanf san ho r dfo some time. Many of the problems posed can be solved only by excavation, but it is almost an axiom that field work is an essential preliminary to excavation and no detailed field wor befors kha e been attempte Durnese th n di s region. Apart fros mit usefulnes n plannini s n excavationga , this approach would see o offemt r other opportunities. Much archaeological researc attemptes hi basie th correlatinf so n do g information from often widely dispersed sites. This approach has, of course, shown itself to be a valid one. To date there has been no attempt, at least in Scotland, to stud ysmala l are depthn ai . Possibly suc hstudya , involving field seriea wor d skan of planned excavations, would give results differen qualitn i t y from those obtained using more usual methods. ACKNOWLEDGMENTS writere Th s. Corcora P woul . W advicdr . nX likfo thanmano . n et J eo r kyD 5 3 DURNESS s SandemaoccasionsMr d an Durnesf nr o D ; locar sfo l informatio succoud nan n i r time need. f HesselgreaveJ s o d ;an assistancr sfo fielde th .n e i Undergraduatee th f so University of Glasgow assisted in a number of ways. In particular, thanks are due to those member Glasgoe th f so w University Exploration Societ undertooo ywh e kth sometimes arduous fiel . GibbA d o workt . d an ,

REFERENCES Buxton, R. J., 'The Earth-House at Portnacon, Sutherland', PSAS, LXIX (1934-5), 431-3- Crampton, C. B. and Webley, D., 'The Correlation of Prehistoric Settlement and Soils in the Vale of Glamorgan', Bulletin Boarde oth f of Celtic Studies, xviii (1960), 387-96. Crampton Webleyd an . B . , C ,D. , 'The Correlation Prehistorif o c Settlemen Soilsd an t : Gowed ran e Soutth h Wales Coalfield', Bulletin of the Board of Celtic Studies, x (1963)x , 326-37. Fox, Sir Cyril, The Personality of Britain, National Museum of Wales. Grimes, W. F., 'Early Man and the Soils of Anglesey', Antiquity, xix (1945), 169-74. Mathieson, J., 'Earth-Hous r Gallerieeo d Building near Durness, Sutherland', PSAS, LIX (1924-5), 221-3. . Steer'ThA . e K ,Earl y IroHomesteae nAg t Wesda t Plean', PSAS, LXXXIX (1955-6), 227-51. 36 PROCEEDINGS OF THE SOCIETY, 1966-67 APPENDIXA Schedule of Ancient Monuments in the Durness Peninsula 1 No. Type N.G.R. Comments References

1 Hut circle Centred Partly overlain by modern sheepfold. Over- O.S. Card index on all diameter 38 ft. Elevation 75-100 ft. O.D. 38796789

t circlHu e2 38796789 Adjacen . i. No Overal o t t l diamete . ft O.S 3 3 r . Card index Elevation 75-10 . O.D0ft .

t circlHu e . Elevatio3 in 38796786 . y 19ft b . nft 9 3 O.S2 Ova t . Carhu l d index 75-1 oo ft. O.D.

t circlHu 4e 38796789 Scooped int. sideS o s .slopit Overaln o e l O.S. Card index diamete . Elevatioin r6 . 19ft n 75-10. 0ft O.D.

. Elevatio ft t circl6 2 Hu ey b 5 n 3879678. in O.S 6 . . ft Car9 2 3 dOva t indehu l x 75-100 ft. O.D.

6 Hut circle 38796789 Overall diameter 26 ft. Elevation 100-1256;. O.S. Card index O.D.

7 Hut circle 37736673 In a sand-blow and built on soil surface two. Dyke clearancd san e cairns also e occuth n i r blow. Partially destroyed. Overall diameter some 34 ft. Elevation 100-125 ft. O.D.

8 Hut circle 37736661 Well preserved hut 27 ft. in overall diameter. Elevation 75-10 . O.D0ft .

t circlHu e 9 3773666 . 1Overal8 . EighteeNo f lo dia. n- NE fee o t t mete. Elevatioft 3 r3 n 75-10 . O.D0ft .

t circlHu 10e 3763673 sligha n O t 2ris e where ther vers ei y little sand accumulation. Overall diameter 30 ft. Elevation 125-150 ft. O.D.

11 Hut circle 37636732 125 ft. to NE. of No. 10. Much ruined hut n 29-3overali . ft 0 l diameter. Elevation 125-15 clearancA . ft 0 e cair d dykan n e nearby.

12 Hut circle 38826548 Elevation 150-175 ft. O.D. Recorded on O.S. Card index O.S. 6 in. sheet.

13 Hut circle 38826548 Some 180 ft. to N. of No. 12. External dia- mete . Elevatioin r 6 1. 8ft n 150-17 . O.D5ft .

1 Sites 1-52 are situated in O.S. I oo-kilometre square NC 36 NE, 53-73 in square NC 36 SE, 74-79 in square NC 35 NE, 80-81 in square NC 46 SW, and 82 in square NC 46 NW. DURNESS 37 No. Type N.G.R. Comments References

14 Chambered 3884655 e arrangemenTh 6 f stoneo t s markee th y db cair) n(? O.S. does not appear to be a ruined hut. The arrangement is somewhat confused, but there are indications that the remains may be those ochamberefa d cairn.

15 Sub- 38926594 Built agains steea t p outcro enclosurn a s pi e circular wit hdoubla ekine walth df lassociateo d with t circles animalr probabls hu i fo t I n . pe senclosur ya e and could have been built within the past century. Not shown on map (fig. ic). t circlHu 16e 3864668 thin O n6 soil amidst serie f limestono s e out- crops. Overlooks small area of flatter land somewhaw whicno s i h t damp. Overall diameter 28 ft. Elevation 150-175 ft. O.D.

t circlHu 17e 38406696 Compose D-shapeo tw f do d enclosures back backo t . Overall diamete . Elevatioft 8 2 r n 50-7 . O.D5ft .

t circlHu 18e 3840669 . 17No .6 f Overalo Som . yds0 NE 8 e l o .t diameter 13 ft. Elevation 50-75 ft. O.D.

19 Hut circle 37876695 Overall diameter 18 ft. Elevation 100-125 ft. O.S. Card index O.D.

t circlHu 2e0 37936716 Partially scooped into ground. Overall dia- O.S. Card index meter . 2Elevatio1ft . O.Dft 0 n5 .

21 Cairn 38426566 Cairn with cist known as Cnoc nan Ceannan. RCAM (Sutherland) Diamete ft8 . 4 . Humarc n bonebronza d san e No. 16 'elliptical cockade' found about 1835. No 4 9 NSA, , xv evidence of cist now. Elevation 25-50 ft. O.D. n Du 22 38466539 Planned by O.S. Very ruined. On a slight O.S. Card index promontory within 100 yds f hig.o h water mark. Elevation 75 ft. O.D.

23 Hut circle 38726530 Overall diameter 27 ft. 6 in. Elevation O.S. Card index 125-15 . 0O.Dft .

24 Hut circle 38776511 Much mutilated. Elevation 125-150 ft. O.S. Card index O.D. Between Nos. 23 and 24 lie a number of presumed clearanc whicf eo cairne hon n si fragmena f Lato t e Bronz pottere eAg s ywa foun 1958n di .

25 Chambered 38976609 Badly mutilated. Elevation 175—200 ft. O.D. O.S. Card index cairn D & E ig6o, 44 38 PROCEEDINGS OF THE SOCIETY, 1966-67 No. Type N.G.R. Comments References

26 Hut circle 39046614 Badly ruined. Overall dimensions 24 ft. Elevation 175-20 . O.D0ft .

27 Hut circle 39186593 Overall 29 ft. by 37 ft. 6 in. Built against a steep bank. Extended entrance. Cf. fig. 4, . Elevatio5 . No n 150-17 . O.D5ft . t circlHu e 8 2 39186597 Overall dimension . Elevatioft 6 3 s n 15. 0ft O.D.

t circlHu 2 . 9eOveral 28 . 3918659No lf dimeno . S 7 o t - Som. ft 0 e5 sion. Entrancft 2 s3 e sheltere boomerangy db - shaped arrangement of stones, thought to have protecte a dcooking-fire . Elevation 150ft. O.D. Cf. fig. 4, No. 4.

30 Hut circle 39066603 Well-defined example; 175-200 ft. O.D. O.S. Card index

31 Hut circle 39066603 Some 50 metres from 30. Elevation 175- O.S. Card index 200 ft. O.D.

32 Cairn 39066613 Show n O.S o . nshee in s 'standin.a 6 t g O.S. Card index stones'. Six upright stones form a circle 18 ft. in diameter. Probabl retainine yth g sida f eo cairn. Elevatio . O.Dft 0 .n20

33 Hut circle 39226614 Overall diameter 32 ft. Elevation 150 ft. O.D.

34 Hut circle 39226613 Elevation 150 ft. O.D. On bench overlook- ing basi t heana Locf do h Caladaile ar s a , Nos. 33 and 35. t circlHu e 5 3 39206609 External diamete . Elevatioft 0 4 r n 15. 0ft O.D.

36 Hut circle 39226605 Overall dimensions 36 ft. Elevation 125- 150 ft. O.D.

37 Cairn 39486613 Low, heather-covered mounn i O.S . in .d 6 Car . 31ft d index . highin 6 .diamete. ft Ther 2 foue d ear ran large kerb stones in situ and two recumbent in the SE. segment. Others may have been removed as there is a slight ditch round the cairn. e Threkerth bf eo stone situe n i sar only 4-8 in. apart, so there may have been a mor r leseo s continuous kerb. t circlHu e? 3960659 8 3 2 Possibl t circl verea hu n ei y ruined condition. Overall diameter some 28 ft. 6 in. Elevation 150-17 . O.D t 5showft No .fign no . ic. DURNESS 39 JVb. Type N.G.R. Comments References

39 Cairn 38876641 Much mutilated, 40 ft. 6 in. in diameter. O.S. Card index Elevation 175-20 . O.D0ft .

40 Hut circle 38546730 Remains of hut with butt or lambing-pen built e secondarinsidth , it e y structur- in e corporating material from the hut circle. Overall diamete . Elevatioin r6 1. 9ft n under 50ft. O.D.

41 Hut circle 38606731 Hut circle with later sub-rectangular struc- ture built partiall partialld an n yo y insid. eit Overall diameter 29 ft. Elevation under 50 ft. O.D.

42 Hut circle 38706765 Very small; overall diameter only 13 ft. 6 in. Elevation 75 ft. O.D.

43 Hut circle 38576769 Partially built into rock outcrop. Overall O.S. Card index dimension y . 1b Elevatio4ft . in s6 . 1n7ft 50 ft. O.D. t circlHu e 4 4 38506778 Nea f Locro sprin. h BorralaidhNW g . Overall diameter 39 ft. Elevation under . O.Dft 0 5 .

45 Hut circle 38546772 Overall diameter 30 ft. Elevation under . O.Dft 0 5 .

46 Dun 38396753 On a rocky eminence above Loch Borralaidh. O.S. Card index Planne O.Sy db . Elevation 75-10 1960,E & 4 O.D. 04 ft D .

47 Enclosure 3857676 limestonA 8 e outcro bees pha n utilise fordo t m n enclosesideo a tw f o s d area, y . 21b 14ft e otheo side Th e formetw rar s f largo d e boulders. Date and function uncertain. Not shown on fig. ic. t circlHu e 8 4 38816788 Very small structure, overall diameter . 16ft Elevation 75-100 ft. O.D.

49 Circular 38436811 In a sand-blow. Well preserved stone dykes e alsar o presenstructurt belo ee curren th w t soil. Elevation 12 . O.D5ft .

t circlHu e 0 5 38806817 Overall diameter . 1Elevatio8ft n 100-12. 5ft O.D.

51 Hut circle 38266867 Small hut, overall diameter 14 ft. 6 in. Elevation 100-125 ft- O.D. 0 4 PROCEEDING SOCIETYE TH F O S , 1966-67 7V0. Type N.G.R. Comments References

52 Hut circle 38446880 Overall dimensions 18 by 15 ft. Elevation 100-12 . O.D5ft .

3 5 Cairn 379°6435 Cair t Sarsgrumna diametern i , . somft 0 e5 . RCAM (Sutherland), Stone cist in centre with covering slab sup- No. 165 ported on three upright slabs. Elevation O.S. Card index unde. O.Dft 0 . 5 r

54 Hut circle 37866414 External diameter 36 ft. Elevation 50-75 ft. O.D. t circlHu e . 5 Overal5 378664154 . No lf diameteo 4 . S . Somft 0 re8 2. Elevatio9ft n 50-7 . O.D5ft .

56 Hut circle 37286349 On raised beach. Elevation under 50 ft. O.S. Card index O.D.

57 Hut circle 37286349 Some i6ft. to S. of No. 56. Very much ruined. O.S. Card index

t circlHu 58e 3719632 n raiseO 7 d beach. Overall diamete . ft O.S 2 .3 r Card index Elevation under 50 ft. O.D.

59 Hut circle 37196327 Some 43 ft. NE. of No. 58. Only the western O.S. Card index segment remains. Elevation . undeft 0 5 r O.D.

60 Cairn 37336314 Round cairn with central chamber. Dia- RCAM (Sutherland), meter some 43 ft. 6 in. with a maximum No. 164 heigh . Remainin 2 . t oft cis a f 5 centre f n sti o . O.S. Card index Fragmentary remain tracee b kera n f so db ca roun e easterth d d south-easteran n n edge. Elevation 75-100 ft. O.D.

61 Cairn 37166234 Round cair diamete n . i ft . somRCA 7 ft 4 d e4 Mran (Sutherland), high. Elevatio . 163No n 100-12 . O.D5ft . O.S. Card index

62 Hut circle 36996208 Overall diameter 30 ft. Elevation 125-150 ft. O.D. t circlHu e 3 6 36976206 Material scooped from upslop dumped ean d downslop o makt e a eleve l site. Overall diameter 28 ft. Elevation 100-125 ft. O.D. Cf. 3-. No fig , 4 . t circlHu e 4 6 36896202 External diamete . Elevatioft 0 2 r n 125- 150 ft. O.D.

65 Hut circle 36896202 Overall diameter 36 ft. Elevation 125-150 ft. O.D. DURNESS 41 jVb. Type N.G.R. Comments References 66 Hut circle 36856201 In places, stonework three or four courses high probably constitutes a fank built directly t wall t somhu A t e fro. e th ou m th . 1n 1ft o outeirregularlt n wala hu s i rle edgth f eyo spaced circle of large, upright boulders. Dimensions 29 by 26 ft. Elevation 100-125 ft. O.D. Cf. fig. 4, No. 2.

67 Hut circle 37366200 Built against limestone outcrop. Series of sub-oval enclosures associated. External diamete . Elevatioft 3 2 r n 125-15 . O.D0ft . Cf. fig. 4, No. 6.

68 Hut circle 36776198 Overall diameter 30 ft. Elevation 75-100 ft. O.D.

69 Circular 37366190 Insid n outeea r circl f irregularleo y spaced hut stones of diameter 28 ft. is a circular hut with wall f typicao s l double constructione Th . outer diameter of the hut is 16 ft. Elevation 125-150 ft. O.D. Cf. fig. 4, No. i.

70 Hut circle 37366190 Some 70 yds. to N. of No. 68. Overall dia- . meteElevatioft 0 3 r n 125-15 . O.D0ft . t circlHu 71e 37236185 Overall diamete . Elevatioft . 0 ft 3 r 0 20 n O.D.

72 Hut circle 37296180 Overall diamete. ft 4 2 r Elevation 175 ft. O.D.

73 Cairn 36896151 Badly mutilated. Large slabs amidse th t cairn material probably forme - cistda De . limited by a shallow ditch. Diameter of cairn some 54 ft. Elevation 200-225 ft. O.D.

74 Carn an 37355797 There is some doubt as to whether or not this O.S. Card index High structur antiquityn a s ei t measureI . y b 0 s3 e souther. 2overall6ft th n O n. side large, thin slabs have been use construco dt drya t - stone wall. At least three courses survive. The interior is made up of an indeterminate mas f looso s e boulders. Elevation 205. 0ft O.D.

75 Cairn 38285798 Oval cairn 35 by 26 ft. Elevation 1225 ft. O.D.

76 Cairn 38775866 Diameter 21 ft. and maximum height of 12 ft. O.S. Card index remaine e th th a cis e n b f Theri t o s y ema centre. Elevation 105 . O.D0ft . 42 PROCEEDINGS OF THE SOCIETY, 1966-67 No. Type N.G.R. Comments References

77 Cair ovan nA l. cair75 . n3877586No f o . 6 NW . Somft 0 e8 16 by 13 ft. It is difficult to be certain whether antiquitn thia s i s t showr notyo No n . no fig. id.

78 Cairn 38775866 . ConstructeSom76 . eNo 10f o 0 . dydsNW . of thin slabs placed on edge with their long axes tangentia e 'circle'th o t l , . y 11b 47ft Elevation 107 showt O.D. 5ft No fig.n n o . id.

79 Cairn 38775866 Some 80 yds. N. of No. 77. Cairn 25 by 16 ft. wit uprighn ha t pointe end. .dSW stone th t ea Elevation 107 . O.D5ft .

80 Round- 40496101 An account appears in PSAS, LIX (1924-5), O.S. Card index house 221-3. Elevation 950 ft. O.D.

81 Souterrain 42506129 Elevatio . O.D brieft A 5 . 7 fn accoun - O.Sap t . Card index pears in PSAS, LXIX (1934-5), 431-3-

82 Promontory 40566943 Know s Seanachaistealna . Elevation under RCAM (Sutherland), . 158No fort soft. O.D. O.S. Card index DURNESS 43 APPENDIX B Quantitative Summary In the past, archaeological field studies have generally been attempted on a qualitative basis. In an attempt to give a more accurate and objective assessment numerical data may be used to render a quantitative summary methoe Th . d used her areas ei l samplin randoy gb m numbers modifies a o ds includ o ancient e th l eal t monuments show fign i . I. Random sampling give representativsa e cross- section of the whole range of conditions, the choice of sample being made by reference to a table of random sampling number kine th df so availabl Cambridgen ei Elementary Statistical Tables.1 On the i: 63360 O.S. sheet of the Durness Peninsula, a grid was drawn based on the kilometre tablesquaresy ke 2e showTh . s that 100 units were use thao ds t total figure tablen so s 1-6 also represent percentages. This give almosn sa t complete coverag Durnese th f eo s Peninsula. Referenc tabla o et e of random sampling numbers gave 100 values, the first digit of each value being used as an easting and the second as a northing. At the various points arrived at in this way, elevation, superficial deposits solid an ,d geology coul reae db d straight fro relevane mth t maps. Slop assesses ewa y db reading the horizontal distance between contours and taking an average of the limits shown. Arch- aeological sites were treate groupso tw s da , funerar non-funerard yan y monuments. Numbers referring to archaeological sites on the key table give the actual number of remains within the kilometre square. In this respect, the sample is not, therefore, strictly random. Contingency table 7 illustrates that the introduction of this modification has led to certain anomalies. When dealing with archaeological data it must be remembered that the data is almost inevitably only partially extant. The total available information is a sample of the total information and is, at least partly, a biased sample. This sets limits on using such data as samples from which facts concerning the total population can be calcu- lated. Quantitative methods suc this ha s should, nevertheless t wors,a t obviate conclusion doubtfuf so l validity t besa d t shoul,an d enable mor lesr eo s firm deduction drawne b o s t preliminar A . y numerical assessment may also permit the full study to be made more effectively. The method adopted has the advantage onle thasimples i th t y i t d abilitan , y require tha s dcompilei e th t r shoul able d b rea o et d maps. Moreover, the data can be presented in easily understood tables with or without an accom- panying text. Applicatio morf no e sophisticated techniques would increas informatioe eth n which could be derived from the sample area, particularly as far as items 1-4 on the table are concerned. t beeno Thi n s merel s i attemptedsha m indicato ai yt e kine Th .eapproacth f do h which mighe b t adopte solutioe th n d i f thi no s typ problef eo mconscioua o andhopedt s i y t i , poinwa o st , e th t attempt to provide a more quantitative approach to the problems of the archaeological field worker.

. MillerP . Lindle. C V . , J . Cambridge 1d D yan Elementary Statistical Tables (Cambridge, 1953). . 44 . p . cf 2 44 PROCEEDING E SOCIETYTH F O S , 1966-67

1-2 D G 1-7 D G G 2-C 1 i-i D G G 0-E 4 0-QD 5 . 0-4 C L L 1-F 4 1-6 E L L 0-C 7 400 400 400 400 400 2 0 I 400 404 200 200 i 2 3 4 5 6 7 8 9 10 L 1-E 2 o-i F L L 1-E 3 L 1-C 9 L 1-D 7 L 0-D 9 0-4 D L i-i E L 2-2 D L 2-1 E L 40 3o 13 4 200 200 I O O I O O 4 o 4 424 227 too 11 12 13 14 15 16 17 18 19 20 i'QC 9 . 0-9 D G L 0-C 9 2-4 D L 2-3 D Q.G 3-C 7 4'B Q5 . G 3-B 1 1-9 C G L 0-D 9 I O O I O 0 225 103 I O O I O O I 0 0 I O O I O 0 I I 2 21 22 23 24 25 26 27 28 29 30 2-6 D L 3-4 C G G 3-C 8 G 9-C 1 7-1 B G 4-1 B G 1-4 D G L 1-D 3 2-1 D L G 3-G 2 I O O I O O I O O I O 0 I O 0 I O 0 I O 0 i i 4 I O O I O O 31 32 33 34 35 36 37 38 38 40 G 6-D 2 6-4 C G G 8-C 1 3'B Q2 . i'9 D Q. 0- D Q9 , 0-7 D L G 0-D 9 3-8 D G 6-4 G G I 0 O I O O I O O I O O 200 I O O 3 ° 5 3 i r I O O I O O 41 42 43 44 45 46 47 48 49 so 1 1-4 C G n-i C Q. 4-7 G G G 1-D 9 0-3 C Q. 0-3 D L 1-9 D G G 5-C 6 G io- C i 8-B Q9 , I O O I O O I 0 O 200 200 3 i i 104 I O 0 I O O I O I 51 52 53 54 55 56 57 68 59 go 9-2 B Q. 2-6 C G 0-6 D Q.L i-D i 7-8 B G 5'4 C G 7-C Q7 . 8-B Q7 . 3-C Q2 . 2-2 D Q. I O 0 2 O I 200 300 I O O I O O I O O I O O I O O 2 O O 61 62 63 64 65 06 67 68 69 70 4-7 A G G 8-B 9 r4'9 A Q. 10-6 C Q. 9-3 D Q. 3'3 C Q. 0-4 C Q_ 10-4 C G G 13- A 1 13-9 D G 100 I O O I 0 0 too I O O 200 200 I O O I O O I 0 0 71 72 73 74 75 76 77 78 79 80 10- QD 9 . 8-C Q2 . 2-7 C Q. 0-4 C Q. G 19- A 9 23- B Q9 . •4-7 C Q Q "• 7C 6-2 C Q. o-7 E Q. I 2 O I O O 200 200 I O 0 I O O I I 0 I I O 200 200 81 82 83 84 85 86 87 88 89 90 G 18- B 3 22- B Q1 . •7 QB 7 . 6-7 C Q. 1-2 C Q. 18-7 B G 13-7 B G 9'QB 4 . 0-7 E G 0-4 E Q. 100 I O 0 I 0 0 200 200 I O O I 0 0 I O 0 200 200 91 92 93 94 95 96 97 98 99 100 Key Table of Numerical Data obtained from random sampling. (1) Heights—1-2 = 120ft. O.D. Height Slope Solid Geology (2) Slope =A — sever ) e(o steeB= p (50) (Figures in brackets i i 3 C = moderate (100) horizontae refeth o t r l Funerary Dmediu= m (200) distance between con- Superficial Monu- Other Deposits ments Sites E = gentle (400tours) ) sligh= F t (800)J 4 5 6 (3) Solid Geology— Q Quartzit= , e G = Gneiss LimestonL= e (4) Superficial Deposits— 1 = hill peat 2 = till 3 = river and marine gravels 4 = sand (5) Funerary monuments (6) Sites other than funerary monuments DURNESS 45 Height Height (ft.) Q. Total (ft.) A BCD F Total 2400 2400 2 2 2 2 2200 2 20O 2OOO 2OOO 3 3 i 2 3 1800 1800 I I I i 1600 1600 2 2 i i 2 1400 1400 3 3 i I i 3 1200 I2OO 4 3 7 6 i 7 IOOO IOOO 6 2 8 4 3 i 8 800 800 3 5 8 2 5 i 8 600 600 6 6 i 2 3 6 400 400 6 8 6 20 2 10 7 i 20 2OO 200 ii 12 17 40 10 20 8 2 40 o O Geo- Total 42 23 Total 4 16 38 31 9 2 Slope 35 logy TABLE i TABLE 2 Correlation Heightof with Slope. Correlation of Height with Solid Geology. The area is between o and 2400 ft. in Allimestone th l d an . belos eft i 0 w40 elevatio e surfacth f o ne wit% 40 h most is less than 200 ft. O.D. Quartz- lying below 200 ft. and a further 20% apparentle it y form greatee th s r part between 200 and 400 ft. The indica- of the 700-1100 ft. erosion surface. tions are of a low erosion surface be- tween 400 ft. and sea level, a further surface between 700 and 1100 ft., and isolated peaks risin 240o gt . Wit0ft h only 11 % of slopes in the E and F categories a scarcity of flat land is indicated. Slope 1234 Total Slope Q. Total A 4 4 A I 3 4 B 16 16

B 10 8 18 C 24 12 2 38 C 16 17 4 37 D 17 6 4 4 3i

D 6 12 12 3° E i 5 3 9

E 2 2 5 9 F 2 2

F 2 2 total Superl 62 23 4 II deposi Total 35 42 23 Geolog TABLE 4 TABLE 3 Correlation of Slope with Superficial Correlation of Slope with Geology. Deposits. Ther highea s ei r proportio gentlf no y Sand produces the gentlest slopes with sloping land under limestone. Slopes the emphasis on category D. It is developed on quartzite are steepest evident that much till on flat land will wit hhiga C hd proportioan B e th n i be covered with peat, a fact brought categories. out extremely wel thiy b l s table. PROCEEDING E SOCIETYTH F O S , 1966-67

Height Geology 123 4. Total (ft.) 3 4 Total 2400 Q 23 12 35 2 2 2200 G 32 4 i 5 42 2000 L 7 7 3 6 23 3 3 1800 Total 62 23 4 ii I I 1600 TABLE 6 2 2 1400 Correlation of Solid Geology with Super- 3 3 ficial Deposits. 1200 7 7 Peat predominates on both gneiss and IOOO quartzite. On the limestone there is 8 8 relativela y even distributio peatf no , 800 sandd tilan l . 8 8 600 6 6 400 '3 6 i 20 200 9 '7 4 10 40 o Total 62 23 4 ii TABLE5 Correlation of Elevation with Superficial Deposits.

Height Height (ft.) (ft.) 22OO TABLE 7 22OO TABLE 8

2OOO Correlation of Height with 2OOO Correlation of Height with other Funerary Monuments. Monuments. 1800 table Th e doe brint st no gou 1800 Again the trend is clear and l6oO the actual very high eleva- l6oO the anomalous location of tion of Carn an Righ but the e wheelhousth t onca s i e 1400 genera lt mistrenno s -i d 1400 evident. 1200 represented. I20O

IOOO IOOO i 800 800

600 60O

400 40O 1 1 2OO 20O 5i O O Total Total 63 DURNESS 47 TABLE 9 TABLE 10 A A Correlation Slopeof with Funer- Correlation Slopeof with other B ary Monuments. B i Monuments. The high proportion on G 4 C 6 slopes of category D may re- D 8 D 32 flect the information on equallmay Tably but 3 e E E 2 7 reflect a desire for drainage. F F 17 The anomalous position of e wheelhousth s i agaie n Total 14 Total 63 evident.

TABLE ii TABLE 12 Q. 4 a 2 Correlation of Solid Geology with Correlation of Solid Geolotjy G i Funerary Monuments G 6 with other Monuments. The inclusion of one funer- L 9 L 55 ary monumenx si d an t Total '4 Total 63 other monuments on gneiss is inaccurate, but the tables bring out the general trend. proportiow lo e Th sitef no s on gneiss and quartzite would indicat necessite eth y of checking these results.

TABLE 13 TABLE 14 I 21 Correlation of Superficial De- Correlation of Superficial De- posits with Funerary Monu- 2 14 posits with other Monuments. ments. The high proportion of 3 o sitetw s e tabulateTh d monuments on sand is an 4 undeedge th ern o san e dar 4 28 inaccurac e tablth t e bu y of the area of blown sand emphasises the large num- Total 14 Total and the soil is composed of 63 ber of sites which are peri- till wit n admixtura h f o e pheral to the sand area. sand merging into till. 48 PROCEEDING E SOCIETYTH F O S , 1966-67 APPENDIX C Schedule Round-Housesof ' Wags'and No. on Sketch Name1 N.G.R. References* Map North Uist group Machair Leathann 17/8075 PPS, xiv (1948) ,71-72 PSAS, LXVI (1931-2), 33 RCAM2 27 , Eilean Maleit 17/773742 PPS, xiv (1948), 72-73 (A) PSAS, LXVI (1931-2)3 3 , RCAM0 27 , Cnoc a Comhdhalach '7/76974! PPS, xiv (1948), 73 (A) PSAS, LXVI (1931-2), 33 RCAM, 269 Carry lochdrach 17/772743 PPS, (1948)v xr , 73-74 (A) PSAS, LXVI (1931-2), 32, 63

Foshigarry i? 7/ PPS, xiv (1948), 74-75 (r mor3o e houses) PSAS, LXV (1930-1), 299-357 a-(A) PSAS, LXVI (1931-2), 63-64 PSAS, LXIII (1928-9), 319

Bac Mhic Connain i7/ ? PPS, xiv (1948), 75 PSAS, LXVI (1931-2), 42-66 RCAM, 271 7 Dun Thomaidh 17/757758 PPS, xiv (1948), 75 PSAS, LXV (1930-1), 299-357 (AlVallaabove n i th le ear y Strand.) 8 Glettraval 17/752711 PPS, xiv (1948), 46-68 (A) PSAS, LXIX (1934-5), 483-4 9 A Cheardach Ruadh 1963,E !7/7756i & 31-3D 72 South Uist group 10 Usinish 23/848340 PSAS, vii (1866-8), 165 (A) RCAM, 395 11 Kilpheder 23/735205 PPS, xvin (1952), 176-93 (A) Feachem, 98-99 CheardacA 12 h Mhor 23/ ? PSAS, xcm (1959-60), 135-73 (Drimore Smiddy) 1956,E D3 & 3 I2A Drimore Farm 23/768403 D & E 1956, 33, 34, 38 (N.B. 'At least seven sites' between Isle Ornsay and Daliburgh—vide Lethbridge, PPS, xvin (1952), I93-) followine 1Th g abbreviation usee thisar n d i s column: (A) = Aisled Round-house. (WG) =Wag. All others are wheelhouses or presumed to be so. 2 The following abbreviations are used in this column: 'Feachem' —Guid prehistorico et Scotland (1963). 'Hamilton' —Jarlshof (1956). RCAM —Appropriate inventory Roya e ofth l Commissio Ancienn no Historicad tan l Monuments. DURNESS 49

Situation Outer walls outbuildingsor

sana n dI dune, overlookin beache gth . 'Outer courtyard'

On a small islet, on the beach. Several chambers adjoining. machaire th n O , overlookin beache gth . Several outbuildings. machaire th n O , overlookin beache gth . Oval outer enclosure.

sana n dI dune, overlookin beache gth . Souterrain: 'harbour'.

On the machair, overlooking the beach. Several outbuildings.

hillocka n O , overlookin machaire gth .

On the South slope of Clettraval Hill (438 ft. Byres, etc, and outer wall. O.D.) Just above High Water Mark, West coasf o t ? Baleshare Island. lowee th n r O slope Heclf o s a (198 . O.D.)8ft , Souterrain. and half a mile from Usinish Bay. Among the dunes, on the machair. ?

On the machair, 300 yds. from the shore. 'Forecourt.' machaire th n O . Souterrain. 50 PROCEEDINGS OF THE SOCIETY, 1966-67 No. on Sketch Name N.G.R. References Map Barm group 13 Tigh Talamhanta 32/ ? PSAS, LXXXVII (1952-3), 80-105 (A) RGAM, 459 ? 13A Allasdale 1962,32/65702E 2 & 3 D 9 Tign A h14 Glac h 32/667997 D & E 1965, 20 Tiree 15 Vaur DunMo l 44/042492 Antiquity, xxxix (1965), 272 (Phase 4—post-broch) (A)?

Islay group 16 Loch Gorm 57/ 1 D & E 1956, 10 (Five or more)

Benbecula group 17 Bruthach a Tuath 23/785567 PSAS, L (1915-16), 12 PSAS, LXIV (1929-30), 14 RCAM4 35 , 18 Bruach Ban 1956,E 2 & 3 D 23? / (Two or more) PPS, xxin (1957), 227 Lewis 19 Berie Sands i? a/ PSAS, LXVI (1931-2), 2855 39 , (since destroyed) RCAM8 9 ,

North West Sutherland 20 Eriboll 9/405610 PSAS, LIX (1924-5), 221 Caithness Group (a) Latheron Water 21 Wagmore Rigg 15/003262 RGAM, 248 (WG) 22 Allt Preas Bhealaich i5? / RGAM9 24 , (WG) 23 Borgue Langwell 16/102218 PSAS, XLVI (1911-12), 77-89 (WG) RCAM0 25 , Feachem, 95 (b) Berriedale Water 24 Dail a Chairn 15/028294 RCAM, 255 (WG) 25 Achnihavish i6? / RCAM, 251 (WG) 26 Carn Tighe Chreagaich 16/089294 RCAM, 252 (WG) RCAM, 253 27 East Achnihavish i6/ ? RGAM, 254 (WG) DURNESS

Situation Outer walls outbuildingsor a rock n O y outcro e machair th e th n i pn Souterraino , , kiln, steading, barn, byre, etc.a , South slope of Cora Bheinn. within an outer wall. ? Amon sane gth d dunes. On the machair, at an elevation of 500 ft. ? 3 miles S. of No. 13.

On a rocky eminence, next to the sea. ?

gentln O y sloping groun hundrew fe da d yards N. of Loch Gorm.

On the edge of the machair.

On the edge of the machair.

sana n dI dune, just above high water mark.

rocka n O y outcrop betwee seto lochansf sno tw . Crescent-shaped enclosur South-Wesn eo t side. Remote situation (950 ft. O.D.) Outer walls to NE.

Upper part of valley, I mile S. of Morven, at an Secondary structure- en se (?fronn on i )f o t elevation of 750 ft. O.D. trance. As abov Morvenf t Jo mile. bu eSE . Elevation Oblong structure to the SE. 800 ft. S. side of Langwell Water. Elevation 350 ft. O.D.

South ban f Berriedalo k e Water. Elevation . O.Dft 0 .70 . BanN Berriedalf ko e Water. Elevatio. ft 0 n45 O.D. S. bank of Berriedale Water. Elevation 500 ft. 'Various indeterminate structures.' O.D. . BanN Berriedalf ko e Water. Elevation about 500 ft. O.D. 52 PROCEEDINGS OF THE SOCIETY, 1966-67 JVb. on Sketch Name N.G.R. References Map (c) Dunbeath Water 28 The Wags (3 or more) 16/106330 RCAM, 257 (WG) Burg ) Wa n(2 2 9 RCAM, 258 (WG) 30 Wag Mor (a or 3) RCAM, 256 (WG) 31 Achorn (Bad na Glac) 16/119304 1960,E D& 4 4 (WG-3 Wheel-houses-2) 32 AchorI nI RCAM, 259 (WG-4) 3 3 Bridg Rhemullef eo n RCAM, 261 (WG) (d) Burn of Houstry 34 Cor Tulloch 16/152356 RCAM, 262 (WG) (e) .Forse Wae Th g35 16/204352 PS AS, LXXV (1940-1), 23-39 (WG) PSAS, LXXX (1945-6), 11-25 , PSLXXXIAS I (1947-8), 275-85 Feachem, 95 RCAM3 26 , (f) Possible outlier 36 Yarrows 16/308434 RCAM0 26 , (? WG) Shetland group

37 Clickhimin 4/464408 PPS, XXIH (1957), 227 D & E 1956, 34 38 Jarlshof 4/398096 PPS, xrv (1948), 84-89 (A) and Wheel-houses (4) Hamilton 39 Hestensgot 1964,E 4/38412& 9 4 D 2 Note: Nineteenth-century reports suggest further examples of round-houses in association with brochs. Orkney group 40 Howmae Brae 5/758524 RCAM, Vol. II, 48-50 PSAS, xix (1884-5), 23-32 PS AS, xxiv (1889-90), 451-61 PPS, xiv (1948), 82-83 41 CalfofEday 5/578386 RCAM, 24) 5(3 PPS, xm (1947)5 2 , PPS, xiv (1948), 81-82 42 Little Howe of Hoxa 7/423940 PPS, xrv (1948), 83-84 DURNESS 53

Situation Outer walls or outbuildings

Burng . banWa N .f ko Elevatio . O.Dft 0 .n 40 Possible outbuilding(s).

? . banN Dunbeatf ko h Water. Elevatio . ft 0 n35 O.D. E. side of Raffin Burn. Elevation 500 ft. O.D. ?

Head of Allt Bhrailteach and half a mile from ? this burn. Elevatio . O.Dft 0 .n55 ? . banN Achorf ko n Burn. Elevatio O.D. ft 0 .n 40

? 0 yds48 . NW f Bridge o .n elevatio a t a , f o n 250 ft. O.D.

? Near head of glenelevation a t ,a O.D. ft n0 o.f 40

South slope of Ben-a-Chielt, at an elevation of Traces of turf wall and ditch on one side— ? . O.Dft 0 . 40 farmyard wall.

Outside broch n edgo ,f Loch o e . Elevation 300 ft. O.D.

Within broch tower edgn o ,loch f eo . Several outbuildings.

On edge of bay.

sana n bay dI e edge th dun th f .eo n eo Several huts, conjoine vera n di y irregular plan.

On the shore of Calf Sound.

Just abov beache eth , overlookin gbaya .