Trial Excavations at the Old Keig Stone Circle. 87 Iv. Trial

TRIAL EXCAVATIONKEID OL G E STONTH T ESA CIRCLE7 8 .

IV. TRIAL EXCAVATION KEID GOL STONE TH T ESA CIRCLE, ABERDEEN- SHIRE PROFESSOY B . . GORDORV N CHILDE, B.LITT., P.S.A.ScoT.

Stone circles wit a hRecumben t Stone o highesbetweetw e tth n pillars represent a type peculiar to Aberdeenshire and the adjoining countie f Banf o sd Kincardine an f . Despit e surveyth e f sixty-fouo s r of these monuments and the excavation of one by Mr Coles,1 the age of the type is still a legitimate subject for discussion 2 that can only be settle y mordb e productive excavation e StonTh .e e Circlth n o e farm of Old Keig seemed a suitable site for a trial dig aiming primarily at a solution of the chronological problem; the monument is so ruinous as to be of small value in solving structural problems, so that investiga- tions on the latter would not be prejudiced by a partial excavation if unproductive t hasI . , moreover, been previously disturbed lettea s a , r writte n 169i ny Professo2b r Garde f Aberdeeo n o Joht n n Aubrey demonstrates.8 The initiatio e fruitfuth f no l operation e describe b e sit o th t e t a sd here was due to Mr Mansfield D. Forbes, of Clare College, Cambridge. Lord Forbe Messrd an s s Mortimer s tenantKeigd ,hi Ol ,t gavsa e ready permission and much assistance, for which the excavators wish here recoro t d their gratitude Forber M . s aske writee dth tako rt e command of operation d preparan s e report th er W . VarleJ M . . f Liverpooo y l University kindly undertook the surveying, and assisted as well in e worth f excavationo k s DoriMr . s Dingwall . ,Crichto B Mis . M s n s Varle . Layard. FranklinMr W MitchellW y. d . J A an als ,r r oM ,M , lent invaluable assistance hireo N .d labou s employewa r d excepr fo t filling in at the end. The excavation in the occupation levels was carried out with trowel and penknife, which accounts for a relatively high percentage of relics. Operations were begun on 1st September and lasted fourteen days. The circle4 stands on a slight crest on a ridge running up to the Hil f Airlieo l , whic crownes hi fora y dtb terme Barmekie dth n (929 feet 1 Proceedings, vol. xxxir . 140p . ; vol. xxxv . 188p . ; vol. xxxvi . 489p . ; vol. xxxvii . 91p . ; vol. xxxviii ; vol 7 .25 xxxix. p . . 193p . . 2 Hadrian Allcroft Cross,e Circlee th d Th , an chap . viii. 3 Archceologia, . vol320p . i .. 4 Earlier reference : Jame e sar s Garden (1692), Archceologia, 320vol. p . Loga. J .;i n (1827), ibid., vol. xxii. p. 201 and pi. xxiii; F. K. Coles (1901), Proceedings, vol. xxxv. p. 211; and Bishop G. F. Browne (1921) Somen O , Antiquities e Neighbourhoodth n i f Dwnechto House, Aberdeenshire, existine Th . g78 wind-brea . p showt no Logan's kn i n o earln a s figurn yo estatr eno e plan (undated, but probably older than 1827). 8 3 PROCEEDING E SOCIETYTH F O S , DECEMBE , 193212 R . O.D.). The circle now lies in a narrow windbelt planted with firs and spruces. This plantation is probably subsequent to 1827, though it includes older trees. A farm road, banked on either side, crosses the plantation immediately to the south of the circle. Of the monument a Recumbent Stone, two flanking Pillars, and a third undoubted orthostat to the east survive. Five other earthfast stones in the vicinity look as if they had been disturbed. The Recumbent and Pillars stand in a more r leso s circular bank, which merges inte dyketh o s boundine th g plantation on either side but crosses the belt again on the north. Here two large blocks projected above the turf, but were found to be loose. Three trees were still growing within the circle, and-the stumps of others are visible on its bank. Our excavations were limited to digging a trench 7 feet 9 inches wide across the circle at right angles to the Recumbent and equal widtn i hal o ht s length it f(southe Th . ) easter nexcavationr edgou f eo , hereafter called line A (the opposite edge being referred to as line B), accordingly coincides wit diametee hth circl y an whic f ef o ro h Recumbent may form a chord, and so is likely to represent the true diameter of our circle (fig. 1). The bearing of line A is 43 degrees W. of S., magnetic, r conveniencfo t bu n descriptioi e t i wilne treateb f li runnin s a d g north and south. Distances are reckoned in feet along line A, the inner fac f Recumbeno e t turf-levea t l (691'62 O.D.) being . take0 s a n The section was extended southwest of Recumbent on the same line fee4 fo2 r t (but onl yfee2 e last th widt r 12)efo . Here distancee ar s given negative values with referenc e samth eo t ezer o point. From sectioe th widenes 9 - feetn 0 wa exposure 1 alloo o t t th ,o dt 3 f w o f eo base th f Weseo t Pillarextendes wa t .i 9 feeBetweed2 3 d t easan 5 nt3 of line A, where a scraggy larch-tree was removed. The section was cleared down to virgin soil over its whole width, save between 24 and 32 and beyond 66; in these sections a strip 2 feet 6 inches wide from line s cleareAwa thio dt s depth e strith . p4 5 Betweecleare d s an 2 dwa n3 wider, but varied in breadth owing to the presence of large stones. Immediately unde e turfth r a looss e buff loas encounteredmwa . Stones projected throug e loath h m sout f Recumbeno h t between - 6 and - 9, between 0 and 9, between 24 and 54 (save for a strip 2 feet 6 inches wide from line A between 32 and 53), and on the northern segment of the bank from 65 to 77. Between 10 and 24, and again from stra w onl, fe 65 yy a 5 o loos4t e stones occurre loame th n di . In these stoneless areas the loose buff loam passes over almost insensibly, 6 to 9 inches below the turf, into a darker, more compact soil sometimes resembling clay in consistency. This "clayey layer" s normallwa y abou inche5 t s thic d containean k d everywhere small TRIAL EXCAVATIONS AT THE OLD KEIG STONE CIRCLE. 39

EXCAVATION OF A STONE CIRCLE AT OLD KEIG. ^°+ <&

STONEY AREA EXPOSED BUT I STONES OF BANK B LEF SITN TI U I (vEXPOSEB BUT — O<7 I % LEFT IN 00 !__%&_ _SITU

| • .;|40 50 60 7'0

REFERENCE. > STONES OF CIRCLE ^NATIVE ROCK XT FISSURE W HEARTH ^RELIC-LAYER

SCAL F FEEEO T 20 10 0 0 0 5 7 3 e*o 0 2 10

SECTION ALONG A-A 693.5 O.D + <693-50D fll TURF T E (sa-^—MODERN JLINE/- ^ ~———^g°^g^ ^fWWF P.I ar.K Mm n n Keid FigOl g . Ston1 . e Circle: Pla Sectiond nan . 40 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1932.

fragments of charcoal. Most of the relics were recovered from it. It rests between 54 and 66 upon a fine yellow sandy gravel. This is virgi founs nwa dsoild alsan , o outsid circle eth e nort e banth d o hkan t sout f Eecumbentho placesn I . , notably between -2-18d 2an , between e underlyin th d nea, an 45 r, 13 g d bedrocan 9 k outcrop n ridgei s s throug . Withihit e circlth n e area (save unde e stonth r y patch near the centre) and under the bank on the north, the surface of virgin soil lies almost leve t 689'40±'15a l . Somewhere abou t begini 0 t o slopt s e perceptibl e south-westth o t y o thas , t betwee e outeth n r facf o e

Keigd FigOl . Ston.2 e Circl e: Segmen banf o t north-eastn ko . foot 1 thifal e 2n - I s 2 i Recumbenth ls. d stri s an fale 6 pwa th l - t a t observed to be quite continuous (apart from minor unevennesses due o roct k fissures) o thas , t her t leasa e fosso t n ther s e rounewa e dth circle. North of 70 a similar but more gentle slope seems to begin. The clayey laye generan i r l follow e contou e gravelth s th f t o I r . passes continuously under the bank on the north and under the westerf Recumbento d en nt i disappear t bu , s completely outside th e bank at 77 and beyond Recumbent at -9£. The interpretation of this deposit must be postponed till the structural features of the monument have been described. The Bank to the north of the section is composed of boulders with loose loam between them (fig. 2). The innermost were noted at TRIAL EXCAVATIONKEID OL G E STONTH T EA S CIRCLE1 4 . 65 x 5/7 and 66 x 2/4, but these were quite loose and rested on the clayey f i layefalle s a rn forward. Beyond cam a emas f bouldero s s from mor1 7 t betwee ed Bu 6 8an closel . 8 t77 o6 n y juxtaposed slabs4 1 , to 18 inches high and 4 to 6 inches thick, stood on edge or on end on the clayey layer.1 Some were embedded in it to a depth of 1 to 3 inches, but it was in every instance interposed between thei e gravelr th base t 68d A a sher^. an s s foun dwa n thii d s layer unde stonee bane onld th th r kan f incy1 o s h abov e graveleth . The Recumbent is a block of sillimanite gneiss, 16 feet long on top, 6 feet thick (at 691'50 on line A), and 6| feet high (at -3x2|). The s nearli p to y d lieflaan st approximately horizontal e outeTh r .o r south-west face is slightly convex largA . e slic bees eha n spli alonf of t g a plane of cleavage from the inner face near the western end; a portio e tur f s thilyinth no f wa sn againsgo e stone e bodth th tf o y. base vers Th ei y uneven longitudinan I . l profil t slopei e towardp u s s the western end. Viewed transversely, it is seen to be keeled. The lowest point t f fee slin2 n 2- ou , reacheo (689'20)JtA e3 t - a e t a li d, (688-84), and at - 5f, 3£ feet from A (68910). The flatness and horizontal taperine th wese featuree d th ar to an gt alignmenp sto e thath I tf o t have observe othen di r Aberdeenshire circle f thio s s type. s Acentrit t e e turalon th r linf ou A gecam o withit e n almost 4 feet of the top of the Recumbent on the inside and to within 4J fee then o t - outside ture eacn i fTh .h case rested upo nstona y bank which slope towardp du westere sth n stume end fir-trea Th f . po e that d growha n upon this bank right agains e outeth t r fac f Recumbeno e t stoo t fee3 dfro ou e banktm Th lin. s A ewer e compose f smalo d l stones and medium-sized boulders resting in loose loam. The fragment spli f Recumbenof t e innet e banth reste th n ro k n sideo d , projecting above the turf. On the inside the bank seems to slope up from between 6 and 9 to 0, though disturbed by a fir growing close to the edge of our section at 8. Further excavations may show this bank to have been a platforpar f o t m suc s occurha n othei s r similar monuments. The large boulders at 9 might mark the edge of such a platform. On the outside the bank extends from - 6 to - 9 only. On both sides bane th ke claye restth n yo s layer ,e innewhilth n r o e sid t covereei d also the larger boulders immediately in front of and under Recumbent. The cavities under Recumbent on either side of its keel were filled wit a verh y loose black mould relico N . s were recovered from this deposits texturit d an e, show d sbee ha than t i tforme d naturalln yi spaces protected from downward pressure. Boulders and broken slabs of rock lay in this mould, but only a few of these were tightly wedged Compare Aikey Brae, Old Deer, Proceedings, vol. xxxviii. p. 266. 1 42 PROCEEDING E SOCIETYTH F O S , DECEMBE , 193212 R .

in betwee e underside nativth n th e ston th d e f ean roco e r virgio k n soil. Unde e upwarth r d tapering f westerRecumbeno d en n n a t irregular row of large boulders lay in and under the loose mould (fig. 3). largese a pris Th s f locamo wa t l rock (garnet biotite hornfels fee2 ) t long,by 1 foot by 1 foot. None of these boulders, when fully examined, seemed to be giving effective support to the Recumbent itself. All rester wero n edo quite superficially embedde e clayeth n di y layer. Recumben e s restinfounb wa to t n bedrocdgo k (biotite-cordierite

Fig. 3. Old Keig Stone Circle : End of Recumbent Stone after excavation from south-west. hornfels) only alon a glin e abou f t feefro2 tou tm lin betweeA e n -2f and -4J. At these points the lower face of the stone was in contact with ridges of bedrock at 689'04 and 688'83 respectively. From here the keel of Recumbent slopes upwards in the direction of line A, d fir-rootsan , presumably fro e tremth e whose stump e standth n o s bank outside linn , O paseA . s3 - righ s a t n i unde r e stonfa th rs a e itself a boulder was firmly wedged in between the base of Recumbent and bedrock. e graveTh l runs continuously unde e holloth r w f westero d en n Recumbent, dipping from 689-90 at -1 to 689'50 at - 5. There was, however n irregulaa , r depression a stone f o , ,d perhap aboube e tth s d beee ha n th d scoopen be 4i t incheA linn ou d. o B es5 - dee t pa TRIAL EXCAVATIONS AT THE OLD KEIG STONE CIRCLE. 43 gravel, at least for the lowest part of the Recumbent. Its lowest point fulls i inchey4 se grave beloth f average o wlth p surfacto ee leveth e f o l t -.6a , and remarkeds a , , that surface slope ove p spacde u occu w th r eno - e tunne th e fac g unde e f stoneth o du eth lpie n I ry .b d Recumben t along line A the edge of the excavation was actually visible and has been faithfully reproduce Varley'r M n do s section. The "clayey continuousln layerra o "to y unde e holloth r w western e Recumbentth en f do , underlyin d coverine bouldean th g w e th ro gr sloping gravel, but filling up the shallow depression at - 5 B. It also runs a couple of feet in under the Recumbent from its inner side, but is definitely interrupted by the keel and is conspicuously absent d bee latterha e n th d t scooper followwherI be .fo e t th edou s either thae layerth t t n questiothrougi cu d s hollowinbe n i nhwa e th t ou g for Recumbent, or that its formation was hindered by the presence firse e stoneth oth t f n caseI . , which momene seemth t a smore th t e probable, the layer must have been in existence before Recumbent was place positionn di . Rim sherds wer e clayee base th foun f th eo yt a dlaye r againse th t foo f Weso t e pockee graveth tth Pillan i n i t l d (Nor an (No) . 36)59 . . Anothe sherm rri founs d wa (No fee) 2 d 60 .t under Recumben foo1 t n i t fro same mt parth f bu elin , o t vesseA e e clayes founth wa l n di y layer under the stones of the bank at 1| x 3. West Pillar is 8 feet 10 inches high (from its true base), 3 feet wide, and 2J feet thick along the edge nearest the Recumbent. Its summit rises 5} t considerefeeno ts abovs basewa it e turt dI t eth a .f saf o t e expos base eth f thi eo s heavy stone except alon edge gth e next Recum- bent. Her e pilla s supporteth e e moswa r th r t fo dpar y packingb t - stones resting on the clayey layer (fig. 4). From its south-west corner, however, there projects downwards fro e maimth ne ston masth f eo s a sort of tongue about 1 foot 1 inch wide and only 2 to 3 inches thick. This spur penetrates through the clayey layer into the under- lying gravel to a depth of 4£ inches. It cuts clean into the gravel, which is undisturbed all round, and is surrounded with a thin envelope of darker soil fittin t liki g a eskin . Charcoa s foun wa ln thi i d s soil right unde e stonepotshera th rd an ,depressioe edge th d th f (35 eo n o ) n e gravele tonguith nth d beet f I looki ha e. n s a sforce d inte th o gravel rathe weighe pillae th th y f rb o ttha deliberaty nb e preparation of the ground. The Stony Area presumabl4 5 betwee d an 4 2 ny mark e sitf th so e the central ring cairn characteristi f thesco e Aberdeenshire circlest I . has unfortunately been disturbed, presumably by the operations mentioned by Garden in 1692, and by subsequent tree-planting; a 44 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1932. young larch was factn i , , growin t 37ga Je apparen closth o t e t centre of the circle. The stones are boulders and rock-fragments of substantial size, often ove foo1 r n diametei t d closelan r y packed together. All seemed to rest upon the clayey layer, but it must be remembered that here excavations were carried dow o virgit n n soil only alon a gstri p 2 fee inche6 t s wide fro e cairmTh nlin . materiaA e l actually crossed this strip onl , thougy 32 betwee d han projectin4 2 n g int t oi betwee n e last-nameth n I . 53 d regiod 50an nlayera f charcoao s observewa l d immediately unde storiese th r . numerou6 Betwee4 d an t minut6 snbu 3 e

Keid FigOl g . Ston.4 e Circle: Boulders under Recumbent Stone.

fragments of cremated human bone and fragments of at least seven pots, one nearly complete, turned up in the loam and disturbed soil from 2 to 15 inches below the modern turf surface, immediately above virgin soil. Between 38 and 39 there was a layer of particularly black soil, and adjacent to it, but 2 to 3 feet out from line A, a patch of soil 18 inches square baked red by fire. Between 33 and 54 the average leve f virgio l n soi fros i l m 12o £t inches lower than elsewhere in the section, and the material is distinctly paler in colour. The surfac grave e interruptes i th f o eo lto projectiny db g ridge f bedrocko s . Between thes e fissures ear betweee On . n 44d 43Jan | attaine ddepta h of 18 inches, but was filled with rather dirty loose gravel quite devoid of relics. ther2 anothes 3 Betwee d ewa an 1 rn 3 hol inchee8 s deep filled TRIAL EXCAVATIONS AT THE OLD KEIG STONE CIRCLE. 45 with loose gravel equally sterile. I believe both depressions to have been natura havo t d e an lbee n sealed over befor depositioe th e e th f no bone . e kindlinpotterd 39 th e firan s t d th a e f yan go We may now turn back to the so-called clayey layer. In texture and appearance this resembled the muddied soil that all too soon covered our nicely cleaned section gravef strampleo e w f li d upon them afte showerra , and that can be seen upon farm paths in the vicinity. It is, in fact, n occupationaa e builder e th monumenl th deposio t f o se du t t and/or subsequent visitors thereto. A connection between this deposit and the monument is established by its restriction to the area of the circle and a narro'w strip in front of Recumbent. That it is also coeval with the circle might be a plausible inference from the following facts. In the first place, it is older than the bank at the north of our section, the irregular boulder row under the hollow end of Recumbent, the banks on bot hstonee side th somd f so an e boulder,t leasea th e centra f th o t f o s l cairn, sinc thesl eal e rest upo layere nth t mightI . , however arguee ,b d that the banks are secondary, while the central area was admittedly disturbed. r excavatioou Secondlys a r fa n s wenta , Wese th , t Pillar seeme reso dt t supportee b d oa packingnan y db , itself e beddeclayeth n yo d layer, though a tongue of the stone admittedly penetrates the layer. Thirdly, t seemei d r Recumbenlikelfo d ybe tha d beee ha th ttn scoopen i t ou d and through the deposit. If this be right, the relics embedded in the clayey layer must be accepted as dating the monument. Of course, the formation of the deposit may have continued after e erectio e th stonesth t t as i exhibitf Bu o n. o stratificationn s s it , formation presumably went on without interruption during the period whicn i monumene hth use n theri d meano s ,an n es twa i distinguishinf so g suc perioha d from e actua thath f o tl erection. The relics upon which we must rely in dating the monument consist of potsherds thesn I . e ther realle ear y onlclassessecone o th ytw t dbu , class can be subdivided into three varieties. Class I., fine red ware, 6 to 7 mm. thick and fairly homogeneous though including visible grits, exceptionally attaining a length of . n botWhilo mm 4hd re esurfaces a dar, k cor s visibli e n sectioni e . e quitsurfacw Th no e s i erough t mighbu , t once have been covered with a slip of which no trace survives. This ware is very 'rare and was represented exclusively by minute fragments. All without exception were foun smala n i d l y arela ad closan 0 lino et 6 betwee eB d an 6 n5 either right on the gravel or not more than 2 inches above it. One sherd bears a herring-bone pattern executed with a square-toothed comb. It is thus marke belongins a t l likelihoodou al n gi beakera o dt . a quit f Claso e . s (figi II sdifferen ) 5 . t character t I comprise. l al s 46 PROCEEDINGS OF THE SOCIETY, DECEMBER .12, 1932.

Fig. 5. Old Keig Stone Circle : Sections of Pottery. (}.) TRIAL EXCAVATIONS AT THE OLD KEIG STONE CIRCLE. 47 e potterth y found wit e cremateth h d human e disturbeboneth n i s d central area and the sherds from under Recumbent and below the banks shere On nearldy . la occurre t i inchey 3 t bu , ds betwee60 d an 6 n5 higheclayee th n i ry layer tha beakee nth r sherds sherdl Al .Clasf so . sII are comparatively hard-burned, but are gritty and unpolished. We distinguish the following varieties, though the differences are probably of no historical significance :— . thick . Coarsmm 1 ,4 include1 wared ere o t 6 , sgooa d dea f largo l e grit; exterior reddis slightld bufo ht an f y smoothed over; core interiod an r black; rather soft. 2. Coarse black ware, agreeing generally with 2 but black throughout. 3. Brown ware, 6 to 11 mm. thick; paste comparatively even; brown on both surfaces but darker towards the core. Most sherds seeme belono dt rougo gt h cooking-pots with either quite straight wall sligha r so t bulg . e Nonth up e f y eo two-third wa e th f so sherds suggested the presence of shoulders or necks. One or two sherds belonge o shallot d w dishes with inverted rime rim l showeTh al .s d careful treatment e mosTh .t distinctive, e flattenedgrouar , A pe th , edges being left relatively sharp. Such rims may overhang slightly, y, inward , outwardsz , or se potte s Th produce. ha r e overhanth d y b g runnin thume gth b along jus whilt m undefirse ri ee th t rth finge r pressed down upon it; the imprints of the digits can just be felt in some cases. A rare simpls ,i B,m typri y f roundedeo . Everted rim absolutele sar y missing. The bases are all flat and tend to be slightly splayed out. Sherds often clustered in groups. The more important specimens are as follows :— 2Jx ,6 fragmentBetwee4 d an t leasa 0 n4 f o st four vessels : warf 29o , wit , n portioneA3 m ur ;h ri n perhap a whicn f sur o e hsth had contained the cremated human bones found lying hereabouts. 27. ware 1, rim A y, discontinuous row of very shallow finger-tip impressions jus toutside n beloo m wri . , witB 28 .h m warri doubl , 1 e e groove immediately belom wri produced by a blunt-pointed implement. From 34 to 39: 45. rim A y, ware 1. 46.' rim A, ware 2; below rim a wide shallow groove. 43, rim A, ware 3. .No sherds were found under the stones of the cairn between 24 and t unde32 bu nead , an rr Recumben collectee w t d : t rathewar, z bu , A e1 r60 m lighri , t blackenebufd an f d externally 48 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1932.

just below rim. Parfouns wa t d unded e banan th , rt 3 1ka Jx part right under Recumbent at - 2 x 1. t , rathewarB bu , e2 36m rri ,find unusuallan e y hard-burntt I . belong inturnes i bow a d o st an l d slightly. Founpockee th n di t gravee inth inche4 l linn . o eB s5 dee- t pa 35, rim 'of ware 2, with shallow groove below it from top of gravel right against bas f Weseo t Pillar. Beyond Recumbent several coarse sherds of ware 1, including pieces basoa d perhapf an e l belonginal s vessele on o t g, turne betweep du n . 8 - d an 7 - Besides potsherds smalo tw , l scraper f flinso t turnee righp th du n o t gravel, one at 58 x 4 and the other at 65 x 3. The excavators and the Society have to thank Lord Forbes for most courteously agreein preseno gt Nationae relice th th t o t s l Museum. Amon e relics th ge potterth , y assigne o Clast d . needII s s special mention, since it seems to date the monument. The following points may help to fix its chronological position:— Flattene seee d manb n rimno y ysma late cinerary urns from Scottish urnfields d typologicall migh9 an 2 , e regarden b t ur a finar s you a d l degeneration of the overhanging-rim-cordoned urn series.1 At the same tim e relativelth e y hard qualite ware th e finger-and-thum f th ,o y b treatmene associatio th e rims d th f an ,o tf fla no t with rounded an d inverted rims differentiates our ware from any recognised group of Bronze Age pottery in Scotland. On the other hand, it does to some extent approximate in technique to much of the pottery o>f the Roman period, generally termed here Early Iron Age. Eve e rim-flatteninth n g distinctiv r grouou s f i po e occasionally found on sites of that period, though the edges are seldom so sharp.2 Nevertheless, the complete absence of the eversion that was trule th y characteristic mod Roman i treatinf eo m ri n e timegth s would preclud eKeid sucOl g hdatina e wares th f go . These ought ratheo t r occup intermediatn ya e position. NowSculptor'e th n ,i s Cav t Coveseaea 3 Miss Bento find ndi d pottery absolutely identical with ours, attributable to a Late Bronze Age (or Hallstatt) occupation. More recently Mr W. serie1A Museue urnf th so e Unite n sth i f mo d Colleg t Sea t Andrews illustrates this very well. Cf. Elgee, Early Man in N.E. Yorkshire, fig. 30, 6. 2 rime r centAbouth pe f s o ,n fro te t m Traprai show nLa w some sor f flatteningo t . Pro-Cf . ceedings, vol. xlix. 12-13fig., 12 ; more remote parallel Museue th n si m come from Irositee nAg s on Lewis (Bragar . 601)HR , ; Por Nesf to s (HR. 757); Rudh Duin a ' n (GT. 67)Cold an ;l (H.D. 324). 3 Proceedings, vol. Ixv . 190metae p . Th , lfig. find 11 .bone-word an s k suffic provo t e e Late Bronze Age and Roman period occupations. The pottery belonging to the latter is easily dis- tinguishable from that here mentioned, which might thus be connected with the Late Bronze Age types even had it not been stratigraphically associated therewith. TRIAL EXCAVATIONS AT THE OLD KEIG STONE CIRCLE. 49

Thorneycroft has shown me rims, identical with those from Old Keig, from two hut-circles in Glenshee, which he will describe at a subsequent meeting. Other sherds fro e sammth e circles hav internan a e l bevel, apparently produced by the same finger-and-thumb treatment as used Keig d r showinbelonge ja Ol on ;a t a o st gdistinca t shoulder exactly sherdw fe e s th fro f o m like Heathereon y Burn Cav Countn ei y Durham. A small fragment of iron was found in one circle, but no Roman pottery, and only a saddle quern. Still more recently Mr A. O. Curie has collected, in the course of his fruitful excavations at Sumburgh Head, Shetland, flattened rims from a deposit which was at least pre-broch. These sites sugges isolatioe basia tth r ceramia s fo f no c group intermediate between e nativth e cordone e de domestiLat th th urn ef d o sBronzan c e Ag e pottery of Roman times.1 Such a group has long been recognised in England; indeed, several such groups are now distinguished. Rims of our type A would in Southern England be assigned to the Late Hallstatt phase and the culture terme y Hawkes.b dA t ther e IroBu 2 eAg n such sherde ar s often associated with vessels exhibitin e well-knowth g n "Hallstatt" profile, and even polished Hallstatt ware. Farther north the settlement on Castle Hill, at Scarborough, offers similar rims from urns with less pronounced profile d unmixean s d with burnished wares t oftebu , n showing finger-print ornament either on applied strips or on the body of the pot itself.3 While this group is probably no later than Iron e nort outlasy th e flattene th Ag n hma i e Hallstat, eA th tm ri d t phase (to which it is confined in Southern England) and is, in fact, seen on a very crude pot of La Tene II. Age from Danes' Graves in Yorkshire, East Riding.4 The English evidence seem justifo t s connectioe yth f potterno y like ours with intrusive movements fro e Continenmth n Lati t e Bronze Age-Hallstatt times. In drawing attention for the first time to such movements, O. G. S. Crawford5 proposed treating them all as one and identifying them with that responsible for the flood of foreign types with which the Late Bronze Age in Great Britain opens. Such a simple formula is no longer accepted. At least three phases of intrusion are t (EEpo . A 100) , from Sordale Hill, Caithnesssherm ri da frod an ,m Tents Muir, Fife (BN. belon y samIlle th )1ma eo gt group . 2 Ant. J., vol. iv. p. 355, fig. 14; vol. vii. p. 483; ' St Catherine's Hill,' Proc. Hampshire Field Club, xi., fig. 12, PI, A131, Q4, X3 (for type A), and Mis. 14, E64 (for type B); Arch., vol. Ixxvi. , figsp15 . . 4-6; Proc. Bristol Spelceo. Soc., vol . fig(1931), , 2-7iv .3 27 . . . p , 3 Arch., vol , . 31. Wheelecf 23 Ixxvii; , 18 . 187 p n Row, .i r ,16 figs, n tree'15 . s Historyf o Scarborough, pp. 20 and 404. 4 Arch., vol. Ix 283. p . havI . thano et Hawker kM drawinr sfo attentioy g m origina e th n no o lt this feature, whic invisibls hi Greenwell'n ei s figure. 5 Ant. J., vol. ii. pp. 27-35. VOL. LXVII. 4 A 50 PROCEEDINGS OF THE SOCIETY, DECEMBER 12, 1932.

represente pottery:e th y firse db Th t J brings Deverel-Rimbury urno t s Southern England and originates ultimately on the Lower Rhine in Hallstatt B-C, abouy a triflB.C0 sa 2 r 70 o .te earlier e seconth ; d from the same quarter, but pronouncedly influenced by the "Harpstedt" type from beyond the Rhine,3 is best illustrated at Scarborough; the third introduce o Southert s n Englan e potterth d f Finayo l Hallstatt affinitie welo s s l known fro l CanningmAl s Cross additionn I . , Hawkes seems ready to postulate at least two intrusions to explain the diffusion respectively of the group of foreign bronze types associated with leaf- shaped swords with U hilt-plates and that associated with "carp's tongue" swords. No potter s beeyha n hitherto directly associated wit e foreigth h n bronze types, but it would be tempting to identify the movements responsibl r theiefo r introduction with som f thoso e e demonstratey db intrusive pottery. That is, however, chronologically impossible as far as the first group (U swords) is concerned, if the bronzes dredged up from Huelva Harbour4 be taken as an unitary hoard; for there a Sicilian fibula otypa f e whic fashiof gono d t hha eou n befor foundatioe eth Syracusf no e (734 B.C.) was associated with a native British spear-head of the Late Bronze Age (as at Denhead). Perhaps Mr Curie will produce the relevant pottery from Shetland. Or perhaps the new bronzes merely reflect the journeys of merchant-smiths and tinkers trading new goods goldr foou r , tin scrap-metald ,an t settlingno t t bu leas,a , distincs a t t communities. At the same time the foreign bronzes were being used by the immigrant t Scarborougsa occupante th d han f Heatherso y Burn Cave. And the latter are connected by their pottery (through Glenshee) and bone-work (through Sculptor's Cave) with the folk cremated and buried Keigd Ol .t a Neithe t Scarborouga r r Heatherno h y Bur e truar ne Hallstatt forms included in the metal-work, though the former site is probably Hallstatt in time.5 On the other hand, in the Braes of Gight hoard, armlets like those from the Sculptor's Cave are associated with neck-rina g apparently derived fro Late mth e Hallstatt type belonging to the Selz-Dangstetter culture,6 while the bracelets themselves are knowRhine th n nei same valleth et y a period. 7 Eve e associateth n d metal types would thu compatible b s e witha

1 Fo thesl ral Kendric e ese Hawkesd kan , Archceology Englandn i 1914-1931. x , chapsd an x i , 2 P.Z., vol. xxi. p. 169. 3 Mannus, vol. xvii . 292p . . Ebert's Reallexikon, s.v. Huelva.

54 Gf. Hawkes and Kendrick, toe. cit., pp. 137 ff. 6 Cf. P.Z., vol. xi. p. 174; Schaeffer, Tertres funeraires . . . de Haguenau, vol. ii. p. 217. ' Schaeffer, op. cit., fig. 175,15. TRIAL EXCAVATIONKEID OL G E STONTH T E A SCIRCLE 1 5 . Late Hallstatt date for the Old Keig pottery. But jxist as in Scotland (as elsewher e Highlanth n i e d Zone e Hallstatth ) t typese b ten o t d accompanie producte th y db cerami w f nativo sne e ecth smithso to o s , features may be assimilated to native traditions. The absorption of intrusive culture whicn so h Foxl insists would explai e comparativnth e rarity of such innovating pottery and perhaps even the similarity already note oldeo dt r native formsindeed An e occurrenc.dth f suceo h a ceramic group in connection with a megalithic monument is itself an instance of the assimilation of exotic by autochthonous elements. Nevertheless importane on , t excavatioe resulth e Keif d o tth Ol s gi t na clearer definitio ceramia f no c class first recognised onlyearo ytw s ago. No less important is the suggestion it affords that Recumbent Stone Circle havy ma se been erecte Latn di e Hallstatt times. That conclusion is not so revolutionary as might appear; for the evidence in favour of a pure Bronze Age date is very slender. Against the bracer fronj Candle Hill, Old Rayne, might be set the stone ladle from under the causeway leadin circla o gt t Crookmore ea , near Alford. But, after allshere th , d obeaker-war? f e froKeigd mOl , thoug t absolutelhno y certain, suggests reserv attributinn ei monumene Irodatn e th ga o net Ag a whole s a t . hope completinW y eb excavatioe gth site th e f nexno t yea determino rt e monumente th e rese relatio f e th beakerth o tth o t f , n o f suc i .,be t hi The excavators wis thano ht k Professor Elliot Smit f Universitho y College, London; Dr Robert Campbell, Reade Petrologyn i r , University of Edinburgh ; Mr M. Y. Orr of the Royal Botanical Garden, Edinburgh ; e Macaulath f o g y Og Institute r D d an , Aberdeen r examininfo , d gan reporting on material collected during the excavation.

Report on the Bone Fragments. By Professor G. ELLIOT SMITH, F.R.S. The bone o fragmentsmalto o permit e l ar e st m mucsen o t t h information to be revealed. It is possible to say, however, that— (a) The bones are human; (b) They have been cremated—se e characteristieth c fissurine th n gi enclosed X-ray photograph (not reproduced here); (c) They belonged to a young adult (B is the fragment of a fully grow t stilnbu l separate cranial bone).

Petrological . CAMPBELLR r Report.D y B , F.R.S.E. Recumbent Stone : Sillimanite gneiss. Prism : Garnet-biotite hornfels. Bedrock under Recumbent : Biotite-cordierite hornfels. Personalitye 1Th of Britain,. 31 . p 52 PROCEEDING E SOCIETYTH F O S , DECEMBER 12, 1932. Recumbene Th f locao t t Stonl no origin s ei . Similar rocks occun i r the Don valley between Kemnay and Tillifourie, and there are also area f gneisseso e th so t e 1-incwesshowe botd p th th an tn o h ma t no north prise Th mprobabls .i y local.

Report on the Charcoal from the Clayey Layer. By Mr M. Y. ORB, by courtesy of the Regius Keeper. The charcoa o mucto s hi l carbonise o t admid f microscopio t c examination e identificationth d an , s given beloe e th basew ar n o d appearanc e broketh f o en surfac e thereforear only d an , e subjeco t t correction. It is not possible, for example, to distinguish Willow from Poplar by macroscopic characters alone. Nor is it easy to separate either from Birch. Material of other trees represented does not present e samth e difficulties e distinctivth r fo , e feature f Aldero s , Oakd ,an Hazel are obvious, even without microscopic examination. Willow or Poplar: 35 pieces. 4 3 „ Hazel: Alder: 25 „ 3 1 „ Birch?: Oak: 6

Report . GrW . n GravelOGGr o D d Clay. y an .B The amoun f materiao t s insufficienwa l mako t t e anything lika e complete investigation. . . , All we can do at present is to give you some idea of the possible nature of the two samples from their appearance. Before doing much in the way of analysis I should require to take much larger sample f eaco s h clas f materiao s l (includin e surfacegth ) frod an m several different spots. The sample marked " gravel" whic mentioneu yo h youn di r letter a glacia s a l sand d whic u yourselan , yo h f questioned, appeare b o t s part of the soil profile and not the unaltered parent material of the soil. It obviously contains an accumulation of oxides of iron, and is probably from wha tere "e "w t mB horizoth e soith l f (thno e layef o r accumulation). It appears to be slightly more gravelly than your sample marked " clay," but if the parent material is a glacial sand this variation in texture is quite usual. Your " clay" sample has all the appearance of the surface or subsurface layenormaa f o r l soil t containI . s organic matter—remainf so roots, pieces of charcoal, etc.—and had it not occurred below the level of the surrounding surface soil, I should have said it was the surface THREE SEVENTEENTH-CENTURY SHETLAND TOMBSTONES. 53 or " A" horizon. Afte rn appearanc a ignitio s ha t ni e very similao t r that of your sample marked " gravel," and I suspect that within the circle the original surface soil may have been removed, and during occupation the subsurface material (similar to your sample marked "gravel") had organic matter incorporated with it. The organic matter rathee giveth t i sr " clayey" feel. I think, therefore, that both sample y havma se beee samth n e originally, and that the difference is due to organic matter worked intso-calle"e e gravel.th oth d dHa " " clay " been originall surface yth e layer, it would probably have given a greyer appearance after ignition. A miiieralogical examination of the coarser material from both samplet sho y differencno wan d di s n origini e .