IV.

ON THE RELATION BETWEEN A RAISED BEACH AND AN IRON AGE MIDDEN ON THE ISLAND OF LEWIS, DONALY B . D BADEN-POWELD LAN CHARLES ELTON. . INTRODUCTION1 . The problem of the geological date of the separation of certain parts e Britisoth f h Isles from each other has occupie r attentiodou r somnfo e time, and we decided that the Outer Hebrides possessed certain features which made this an important area in which to start investigations. Accordingly n 193i d 193, undertooe an 35w k field wor n capacitiei k s geologiss a animad an t l ecologist, wit e generahth l objec f discoverino t g as much as possible about the physiographical and biological history of islandse th , especiall Lewisf yo . Three papers have already been published (Elton, 1934 and 1936, and Hora, 1934) dealing with the parasites of certain Hebridean mammals, which are believed to be important as indicators of the Pleistocene histor f theso y e islands d anothean , n freshwateo r r animals (Elton, 1936) n thaI . e generat th par f o t l work which deals wit e morth h e recent history of Lewis, it became important to find out the con- nection between the periods of human occupation and the natural events, such as the formation of peat, river deposits, raised beaches, and ancient sand dunes, and during this part of the field work a raised beach was discovered and its relation to a kitchen-midden was proved. archaeologicae Th l objects fro e middemth n have been depositee th n di National Museum of Antiquities of at Edinburgh, the animal bones in the Royal Scottish Museum, and the shells from the midden in the University Museum at Oxford. The descriptions of the animal bones in the present paper are quoted from manuscript reports by Miss D. M. A. Bate and Miss M. Platt, and copies of these reports are kept at the Bureau of Animal Population at Oxford. 348 PROCEEDINGS OF THE SOCIETY, APRIL 12, 1937.

We are greatly indebted to the Ancient Monuments Board for Scotland and to the officials of the Galson Estate for allowing us to carry out excavations alse W o .wis thano ht r GrahakD m Callande workinr fo r g out the pottery and implements; Mr A. J. H. Edwards for supply- s witu ha lisg f bone o in t s foun y himidentified b d an , y Professob d r James Ritchie; Miss D. M. A. Bate and Miss M. Platt for their careful investigation of the fragmentary animal bones found by us; and . Crai M r advic . gfo R r e D about condition Lewisn si .

2. THE GALSON SECTION. The site showing the connection between the raised beach and the kitchen-midde wese th tn coaso ns i Lewif to s opposite Galson Farm House, about 7 miles south-west of the . This is the site from which occupation layers, includin n earth-housega , were describey b d . EdwardH A. J . s s positio(1924)it d s alsan ,nha o been recordey b d r CallandeD e Reportth f n theo i r Royal Commission Ancientn o d an Historic Monuments f Scotlando e middeTh . (1928n9) laye. p , s i r exposed at intervals near the top of a high talus slope of sand, which occupie e coasth s t between moute TeampulNaomo th Cr d hn an h na l of the South Galson Burn, a distance of about 200 yards; the vertical height from the modern beach to the top of the slope is about 35 feet. On the western part of this slope—that is, towards the track from Galson' Farm to the sea—fine raised-beach shingle crops out about 25 feet above present high-water mark. It should be mentioned here that all heights were measured from high-water mark, which was taken as the highest poin t whica t h seaweeds were growine rocksth n o ;g measurement was mad meany eb smala f so l hand-level generae Th . l site statth e f eo before excavation consiste f thido s talus slop f sando e , wit n interha - mittent outcrop of the midden near the top of the slope, an outcrop of raised-beach shingle rather over half-way up the slope towards the west end. Four small excavations wero speciatw e s witu made l th hy b e objects of— (a) Discoverin e stratigraphicath g l relatioe middee th th f o no t n raised-beach shingle. (b) Collecting as many animal bones and archaeological objects as possible froe middeth m n e shorlayerth r tn i sou tim t a e disposal. Excavation A.—In the midden layer only, towards the eastern end e exposureoth f . Depth fee4 , inches2 t . RAISED BEAC IROD E MIDDENHAN NAG , LEWIS9 34 .

Excavation B.—A. trial hole was put down in the talus slightly farther east than Excavation A, starting from the level of the base of that section. Thi suns depta fee7 swa o k t f tho without strikin raisede gth - beach shingle. It is possible that the shingle is missing, both here and vertically under Excavatio. nA Excavation C.—A vertical section was cut through the midden at a point much nearer the western end of the exposure, and was continued down through underlying dune sand unti e shinglth l founds a d ewa an ,

HEI6HTS SAND ABOVE MIDULN H.W.M. RAISED BEACH TALUS •yi 10"

28' 2'

25' 8"

2or 4 FEET1 2 1 0 1 8 2 6 4 Fig . Sectio1 . t Galsona n raised beach.

volume of midden layer, 3 feet 8 inches thick, was removed over a horizontal area measurin feeg4 inche 6 tfeet 3 y .sb Excavation D.—A trench was also cut a few yards east of Excava- tion C through the talus and sand at right angles to the direction of e exposures taketh wa n d e surface raised-beacdowan ,th th o f t no e h shingle at its base. A general section through Excavation shows i D fign . ni d 1 . an sC The raised beach at the base consists of fine, well-rounded beach shingle, with practicall interstitiao yn l sand therdoubd o n an ,s ei t that this deposit is continued under undisturbed dune sand as well as under the talus. Unfortunately, no fauna was found in the raised beach; but this is not surprising, considering how rare shells are on a pure shingle beach at the present day. No artefacts or pottery were found. 350 PROCEEDINGS OF THE SOCIETY, APRIL 12, 1937.

The dune sand consists largel f comminuteyo d shell fragmentsd an , no stones or whole shells or any archaeological objects were seen in this deposit. In this it differs from the sand making up the talus, which contains large angular stones, animal bones, shells, fragments of pottery, and in one case a small worked piece of flint. All of these objects have slipped down from the midden above. The midden layers consis f hearto t he sand th site n ,i s mixed with refuse wil d describee b an l, greaten di r detail below t seemI . s thae th t human occupation of the site began in the middle of the period of dune formation proves botd wa t an Excavation ,ha t di C tha d t an ther A s e were at least 2 or 3 feet more dune sand 'which accumulated after the site had been evacuated.

3. THE GALSON MIDDENS. The two sections, A and C, which were cut through kitchen-midden materia t connecteno e continuoua ar l y b d s outcropseparatee ar t bu , d wid a covere p y eb ga sandy db y talus alreads . wa Thi p existencn yi s ga e Edwards'r e timM th f eo t a s excavation suggeste h d an , s (1924 . 186p , ) that either the removal of the sand was "for some utilitarian purpose or the effect of a severe storm." As the archaeological objects found by us were not quite identical in .sections A and C, they will be described separately, but in Dr Callander's opinion there is no great difference in age, either between the two sections or between the top and bottom of each section. Excavation A. layere Th s expose Excavation di 1933n (i nA ) wer follows:s ea —

Thicknes f Layersso .

Feet. Inches.

Blown sand, with impersistent lin sea-shellf eo s 3 0 Layer of large stones ...... Hearth layer, brown earthy sand towards the top, black toward bottoe sth m .... 0 8 Blown sand ...... 0 6 Thin black laver ...... Blown sand, base not seen ..... RAISED BEAC IROD MIDDENE HAN NAG , LEWIS1 35 .

Metal Objects.—Three small rusted piece f iroo s n were e founth n i d top part of the hearth below the layer of large stones. One of these is describe r CallandeD y db smala s a r l iron knife iroo N n . objects were blace loweseee th th kn nhearte i n rparsectioni e th parr f th o tf h no o t . One bronze pin inche3 , s long s founwa , d abov e hearteth h layer (fig. 2). Bone Implements.—One borer made from a splinter of bone, length just unde well-polishee inches2 ron d an , d bone pin, just unde inche5 r s long, were both e hearthfounth n di layer. In the same layer there was also found an object described by Dr Callander as a "bone whorl, of domical shape, made froe epiphysimth n animal'a f o sg bone le s , 1-| inc diameten hi jd f an rinc height.n hi " Miss Bats eha also examined this objechumaa d considere b an tn o t t i s bone: "There is one human bone in the collection from the hearth; this is the head of a femur which had not yet become completely fused to the shaft of the bone. The depression for the insertion of the ligamentum teres has been artificially enlarged to form a hole through the bone." No other human bones were found either in Excavation A or C. Stone Objects.—One flat stone whorl was found in the thin black layer 6 inches below the hearth. A few flakes of flint occurre browe th n di nhearthe parth f t cannoo t bu , e b t

recognised as implements. These must have been brought j>j 2. Bronze -

by human agency from some distance, as no flint was seen Pig n fr°m Gai-

occuo t s u r y naturallb Lewisn yi ,e glacia eveth n ni l drift" ' . ewls

piecee Th f flin o sthin i t s t deriveheartno e har d fro e flintmth y crus son h roc f Easterko n Lewis, with which they have been compared carefully by us. Pottery.—Small piece f darso k grey pottery were e heartfounth n di h layer. Non f f thesthe o eo s decorated mo ewa tw wer fragt m ri ebu , - ments. There were not enough pieces to fit together so as to get any e originaideth f ao l siz r shaputensilse eo th f eo . e pieceSomth f f eo so decorated pottery which were collected later fro e talumth s (fig wer) 4 . e awar fa t y frono m Excavatio. nA

ExcavationC. The position of this excavation is towards the west end of the outcrop of the midden, near where the track from Galson Farm reache s situatei e shore th st I . d abov e outcroe raiseth eth f do p 352 PROCEEDINGS OF THE SOCIETY, APRIL 12, 1937.

e maith beach nd objecan , f undertakino t a gfres h excavation here (in 1935) was to find out the relation between the midden and the raised beach; also better pottery was found here tha n Excavatioi n . A n Excavation C showed that the midden layers are 3 feet 8 inches thick at this point, and that there are conspicuously large stones between the 2 fee d t an dowdepth 1 nf o s from the top of the midden. The lower part of the midden con- tains oblique layers of burnt red earth and pellets of greenish clay, e basth ane t da ther prea s i e- pared platform of the same clay, which occurred als between oi n large th e stones. Metal Objects.—No pieces of iron were foun thin i d s excava- tion, and the only finished metal object found was a bronze rivet, from the very top of the midden. Dr Callander reports that many similar rivets have been found from the Glenluce Sands, Wig- I | ^f^^^f 1 o townshire thad an ,t such rivets are known to have been used for fixing patches on an Iron Age Pig . Potter3 . y from Galson. bronze cauldron. Pottery.—Fragments were found at all levels throughout the midden e depthth layer d t whica san , h they were obtaine measuree dar d from the top of the layer. The pieces of greatest chronological value were: At 1 foot 2 inches, a rim and wall fragment of dark brown ware, rim projectind an p to flan go toutward s very slightly, showin rowo gf tw o s deep, almost vertical incised lines under the brim (fig. 3). At 1 foot 4 inches, rim fragment of dark ware, the rim hollow on the top, and projecting outwards and inwards, with finger-made depression s(no m underi t e figured)rth . fee2 inches3 t t A piece on , e with applied (wavy) decoration (fig, 4 . bottom left). At 2 feet 6 inches, rim and wall fragment of thin red ware, rim everted, decorated an d with short vertical incised lines unde same rth t rime A . leve l smala ther s lewa fragmen a differen f (o t t pot) with incised lines (curved) (figright)p , centrto 4 . d .ean RAISED BEACH AND IRON AGE MIDDEN, LEWIS. 353

Several decorated fragments were found on the talus between fragmentm ri e o On Excavation,tw . everte d C lipe th an d , t an da sA wall fragments, all of thin red ware, the first wall fragment bearing

J?CM * Inches.

Fig. 4. Pottery from Galson.

double zigzags and the second a lattice design, all incised. Also three decorated fragments with applied wavy line (fig. 4). It -was noticed in the field that the thinner red ware tended to be e lowefounth n rdi e midde parth f o te coarse n th laye d ran r grey ware toward uppee larga sth n excavatioe ro e th partt enougt no bu ,s hnwa scale to show whether this observation has any great significance. In any case, the pottery as a whole is considered by Dr Callander to belong to one period—the early part of the Christian Era. Similar ware has 3 2 VOL. LXXI. 354 PROCEEDINGS OF THE SOCIETY, APRIL 12, 1937. been found in many of the kitchen-middens and earth-houses in the Outer Hebrides, and although hand-made pottery continued to be pro- duced until well inte nineteentoth h century r CallandeD , r considers these pieces from Ctalso comparativele b o nt y early. The large stones- consist of angular pieces of the local Lewisian gneiss, some of which had been artificially broken. The dimensions of the largest measured wer feee2 inche 3 tfoo 1 inchey 7 tinches sb 6 y n sb I . both Excavations A and C well-rounded beach pebbles were found, which measured up to. about 5 inches long. These are bigger than the pebbles seen in the raised beach in the immediate locality, which are only 1 inch in grade, although pebbles up to 4 inches grade occur in the raised beac poina t ha t about one-thir mila f deo farther west alone gth coast. If, therefore, these pebbles are from the raised beach, they may have been brough shora t t distancemors i t i e t likelbu , y that they were collected from the contemporaneous shore by the men who inhabited the Galson site. The petrology of these beach pebbles in the middens hat beesno n worke detailn i t t thedou bu , y consist mostl acif yo d gneiss, hornblende gneiss, and pink granite gneiss, all rocks which occur in the neighbourhood. Pebble Torridoniaf so n sandstone were also foundt bu , these nee t havdno e been brough distancey an t s thea , y occur botn hi the raised beach and in less rounded form in the boulder clay farther north alon e coast r gth Callande D . r foun sigo n dn thae pebbleth t s submitted to him had been used as hammer-stones, but as the natural matrix of the midden deposit is dune sand, they were presumably brought hane th greemanf e e sit do y th eb Th n.o t clay found associated with hearthe th liks si egreenisa h boulder clay soute founth Gralsonf ho o dt , and the pottery may have been made from this boulder clay. Angular fragment veif so n quartz were hearthsfoune th n di , which almost certainly came from this bouldet possibl whethey no s sa i ro t et i clay rt thesbu , e sho artificiay wan l flaking. Two pieces of slaggy material were found in Excavation C, at differen e indebte r ar F .tM e Parke levelsw o t d d r makinfo an r, a g chemical examinatio thesf no e specimens. The founs firs depta inchewa 6 e middet da f th h o n si n layerd an , it appears to be a piece of iron ore, but as it is slightly vesicular it has been subjected to heat. It was found by qualitative analysis to contain the following:— Iron, 64-580 r cent8pe . Silicon, abundant. Magnesium, calcium, and aluminium, a trace. RAISED BEACH AND IRON AGE MIDDEN, LEWIS. 355

Under the microscope this material is opaque, except in very thin fragments, which are reddish brown, and isotropic, as they show no interference figure. This substance does not give the normal streak for limonite. The second of these two objects "was found at a depth of 3 feet 6 inches in the midden layer—that is to say, near its base. It is much more vesicula d slaggan r n appearancyi e tha e firsnth t piece, and fact,n i , appears to be a true iron slag. Chemical examination proved that it contains: Iron, 31-441 per cent. Silicon, abundant. Magnesiu phosphorusd man , fairly common. Calcium tracea , . Carbonate tracea , . Unde e microscopeth r whita , e mineral whic presens hi proves wa t d bopticas yit l propertie darkequartze e b th o t st rspecimee bu , parth f o t n wat mineralogicallno s y determinable. This specimen gav streako n e . The quartz was included in the above analysis, mixed with the darker fraction. s quiti t I e certain that neithe f thes o specimeno r etw naturae ar s l volcanic rocks, as their iron content is much too high. Considering the rarity of iron ore in Lewis (see Jehu and Craig, 1934, p. 847), it is unlikely that these stones were introduced accidentally. It seems, therefore, that their presence in the Galson midden, as ferruginous material which has been subjecte greao dt t heat, suggests that iron-smeltin practises gwa d awayr fa t .no 4. ANIMAL REMAINS. One of our objects in excavating at Galson was to obtain dated information about the past faunal conditions on Lewis. Except for deer and a wild cat, most of the remains of mammals belong to domesticated species. The list which follows includes the animal remains found both by Mr Edwards durin excavations ghi earth-housee th f so ourselvesy b d an , .

[TABLE. 356 PROCEEDINGS OF THE SOCIETY, APRIL 12, 1937.

Excavations.

A. C, Earth- Below Upper Lower IlOUSG Hearth. Hearth . part. part.

Mammals : Felis sylvestris Schreber (Wild Cat) .... - - - X — Oryctolagus cuniculus (L.) (Babbit) - X X - - Cervus elaphus . (ReL d Deer) X X X - X Bos longifrons Owen (Celtic Shorthorn) X X X - X Ovis sp. (Sheep) X X X X X Sus sp. (Pig) ? X X X X

Whale .... - - X X X Birds : Larus marinus . L (Grea t Black-backed Gull) X • - X - - Larus argentaius Pontopp. (Herring Gull) X - X - X Numenius phceopus L. (Whimbrel) . X - - - - Fratercula arctica grabce (Brehm) (Puffin) X - X — —

Sula bassana (L.) (Gannet) . X X X _ Phalacrocorax carbo carbo (L.) (Cormorant) . X - - - Calidris alpina schinzii (Brehm) (Dunlin. ) X - - - Turdus merula merula L. (Blackbird) . X - - - ?Pyrrhocoraxpyrrhocorax (L.) X - - - (Chough)

? Some kind of Eagle — _ X - _ "Duck" .... — — — — X RAISED BEACH AND IRON AGE MIDDEN, LEWIS. 357

Excavations.

A. C. Earth- house. Below Upper Lower Hearth. Hearth. part. part.

Fish: Gadus callarias . (CodL ) X X X X X Gadus pollachius L. (Lythe or Pollack) X - - - - Gadus virens . (SaitheL ) - X - - - Molva molva (L.) (Ling) — X — — -

Crustacea : ? Cancer pagurus (Auct.) - - X - X

Mollusca : Mytilus cdulis L. X - X X X Pecten maximus L. X - - - X Patella vulgata L. X - X X X Gibbula . cinerariacf (L.) - - X - - Littorina littorea (L.) . X - X X - Littorina littoralis (L.) — — X " —

Mammals. Miss Plat s identifieha t complete on d e smal frob ri l m ExcavationC belongins a Wile th do g t Cat animan a , l whic longeo n s hi r found living Lewisn o . Red Deer were represented by an upper cheek tooth row, ribs, a small hyoid bone, fragments of pelvis, tibia, radius, ulna, humerus, and tarsals. The three associated upper molars hav etotaa l antero-posterior length . Thesmm 9 oe5 f remain Deerf so , which were more commo Excavan i - tion A than in C, belong to a small race. This was considered by Miss e specimenth Bate onlo w ysa wh , s from Excavatio o resemblt , nA a e founkinw no dNorwayn i d t Misbu ;s Platt identified them wite hth true Scottis Deerd hRe . remaine Th f Rabbio s probable ar t y intrusive e bone s nonth a , f eso 358 PROCEEDINGS OF THE SOCIETY, APRIL 12, 1937.

belongin o thigt s animal were calcined, whereas those f somth o e f o e other animal shod sdi w sign havinf so onle g firee beeTh yth .othe n ni r bone a wil f o ds mammal which "were found f -wi , e omi e possibilitth t y of pigs having been wild, belong to some species of whale, but were not further identifiable. e domestiTh c animals from Galso e cattlear n , sheep d pigsf an O , . thesmose cattle th eth te interestingear belond smale th an , o lgt Celtic Shorthorn type. The e representear y y ribsb d , scapula, carpalsa , fore phalange d manan , y teeth entiro N . e skulfounds t wa lpart bu , s e skula youno th f o l g animal were represente n Excavatiodi y b nA the frontal, nasal bones, and an ear bulla, also by a complete lower jaw e milth wit kl hal molar s e firspresenth t d permanenan t t molar just showing. Altogether three horn cores were f whica foundo s e hha on , length of 15 cm., taken along the curve, and a circumference at its base of about 15-5 cm. Harvey-Brown and Buckley (1888, p. 43) record that remains of Bos longifrons have been found in a Picts' house in Harris. remaine Th f sheeo s p represent both adul verd an ty young animals, d consisan f scapulao t , rib, cannon bones, small vertebrae, humerus, tibia, and a lower jaw which has both the permanent molars and the third milk tooth present. As no horn cores of sheep were found, the exact breed of these animals is not known. Excep r fragmentfo t f ribd limo s an sb bones e remainth , f pigo s s consist of fragments of skulls, while deciduous molars were found at more than one level. These remains were more numerous in Excavation C than in A, and one canine tooth from below the hearth in Excavation A belonged a ver o ,t y youn g e othe mand pigth an f ,r yo remain f thio s s species belonged eithe youno t r r smalgo l individuals. remaino N horsef so s were found.

Birds. e mosth r t fa Sea-birdcommo y b e nar s among those represented, although remains of the chough and the blackbird -were also found. Bones of birds occurred more frequently in the upper than in the lower part of each excavation. sea-birde Th s were identifie follows:s da — Great Black-backed Gull, various limb quadratbonea d san e bone. Herring Gull, limb bones. Whimbrel g bonle ee froon , m Excavatio . MisA n s Bate observes that this species is said to breed at the present time on the island of (which is rather over 50 miles north-east of Galson). RAISED BEACH AND IRON AGE MIDDEN, LEWIS. 359

Puffin fea , w bones, including tarso-metatarsus. Grannet, only foun Excavation di win a tibi a , d gnC a an bone . Bone f Cormoranso Dunlid an t n were only foun Excavation di . nA Among non-marine birds, Miss Platt recorded from Excavation A "a wing bone, which has not been fully identified, resembling in size that of a Chough (Pyrrhocorax pyrrhocorax (L.))." From Excavation C, "the vertebr a bird f ao , some kin f eagldo e e (Falconidai)th f o e On ." most interestin remaine g th Blackbire finds th f wa so d (Excavatio, nA) whic essentialls hi ywoodlana garded dan n species thid an s, recory dma hel throo pt w som ee biologica lighth n o t l histor f Lewisyo , whico s s hi conspicuously devoid of natural woodland at the present day.

FishCrustacea.and Fish remains were extremely numerous, the most common being cod. Both vertebrae and fragments of the head were found, and in additio therd co e o nwert e also pollack, materiasaithee th ling d n I an , . l from Excavatio MisnC s Platt recorded e larg"parth f e o tpince a f o r crab, possibl edible yth e variety (Cancer pagurus)."

Marine Shells. By far the most common shells in the occupation layers were limpets (Patella vulgata) and winkles (Littorina littorea), but other shellfish which might have been usebaits foor a dr fo ,do suc mussels ha s (Mytilus edulis), were also found. These three species occurre t botd a excavations e hth ; Excavation i t bu winklenA s were definitely more common than limpets, and another species found at this part of the site were fragments of a large edible scollop (Pecten maximus). t ExcavatioA , howeverC n , limpets were more abundant than winkles throughou e layerth t f o s s especialloccupationwa t i d y an notice, d that they were collecten i d heaps at the foot of the large stones. Two other species, Gibbula cineraria and Littorina littoralis, were also found in Excavation C. We can see from the contents of the Gralson middens that fishing was practised, cod being a favourite article of food. Cattle, sheep, and possibly pigs were tended t evidenco thern bu , s i e e that horses were kept; -wild deer were hunted diee f theso Th t . e people -was also varied by crab, limpets s interestin i winklesd t i an , d an , g tha winklee th t f o s this coas stile ar t l highly esteeme croftere neighbourhoode th th y df b o s , were w ed tolan d that these shellfis t presena e har t exported from here Londono t alss i ot I interestin. comparo gt inferree eth d habit thesf so e 360 PROCEEDINGS OF THE SOCIETY, APRIL 12, 1937.

peopleIroe occupieo nAg wh , e Galsoth d n site, wit e e tastehth th f o s recent St Kildans. According to Heathcote (1900, pp. 154-7), the most valuable sea-birdt Kildan visis S tim e hi e th f tth e o t "wer sa o s t e fulmars, second an importancn di theseo et , gannets .• No remain fulmarf so s were discovered in the Galson excavations. animae Th l remain sr Edward founM y b d s durin s excavatioghi f no the earth-house at Galson were identified by Dr 'Ritchie, but have not previously been publishe detaildn i . The almose yar same th tthos s ea e found by us in the midden layers; the most common domestic animal bein e Celtigth c Shorthorn e nex th ,orden i t f frequenco r y being sheep, and thirdly the pig. Of wild animals which probably had been used also Deed s fooda Re r a ,wa s represente smala y db l fragmen antlern a f o t . The only other mammal found was some kind of whale, represented by bonw fe e a fragment intervertebran a d san l disc. Birds were represented a fragmentar y b y wing bone, apparentl e Herrinth f o y g Gulld an , apparentl onle th ys fis wa duckshy yd b identified Co . onle th ;y inverte- brates found were mussele on :limpets o ,tw scolloe ,on p (Pecten maximus), edibln e cla a pard th f wo f an e o t crab. Edwards (1924 . 196p , ) states tha pottere tth y fro earth-house mth same th tha s es a i t fro middene mth . Thprobablo et tw diffe no o ry d greatl agen yi .

5. THE GALSON RAISED BEACH. Before describing the extent of the raised beach at Galson, we must poin t thaou t t very little wor bees kha n .distributioe donth n eo raisef no d beache e Outeth n ro s Hebrides, becaus geologiste eth havo swh e visited these islands have usually mad e pre-Cambriae studth th e f yo n rocks their main objective. Even James Geikie was mainly concerned with the glaciation, although He hinted at the presence of a late "post-glacial" raised beach (1894 . 160)p , . Jeffreys gives " Raised Beach" as a locality for Trochus cinerarius and Fusus antiquus (vol. iv. p. 325, and vol. v. p. 203), but it is possible that the reference is to the interglacial marine beds of Northern Lewis, becaus a especie s allie Fususo dt s commoi thesn ni e deposits. These marine beds are much older than the Galson raised beach, and do not come within the scope of the present paper. Jehu and Craig (1934, . 842)p , after remarkin e generath n go l scarcit f raiseyo d beaches along coaste th f Lewisso , drew attentio face th t tha o nt peninsule th t e th ao t eas f Stornowao t y whic he Peninsul uniteEy e th s a wit e mainlanhth d of Lewis consists "partl f blowyo n sand partlan d f beacyo h gravel." Durin fielr gou d wor were kw e abl confiro et suggestioe mth n that marine RAISED BEAC IROD E MIDDENHAN NAG , LEWIS1 36 .

material form substratua s f mthio s nect certai t ii f no landsk o t nbu , whethe describee rb thin trua sca s dea raised beach. Gregory reported raisea d beac t Lochmaddha heigha fee2 1 t y a tf r abou o o abovt 0 1 t e O.D. (1928 mention. e 119)p h , t ,bu s tha tonle thith ys si exampl e seen by him in the Outer Hebrides. In a paper on the surface features of Lewis, Panzer (1928, p. 195) hints that low-level raised beaches may exist at Tarbert (Harris), but he gives no details. Galson.—The coast at Galson is completely exposed to the open Atlantic d storman , s here extremelar e y severe, especiall n winteri y . On s therefori e e force o makt d e sure thae apparenth t t raised beach along this coast is not built of shingle thrown up by storms at the present t Galsoa t time t leasBu na . t thi disproves si severay db l considerations. Firstly pebblee raiseth e , th n sdi finf beaco ee gradehar t Galsoa , n itself a size of over 1 inch being exceptional, whereas the beach now forming e samth t ea place consist f vero s y coarse material e averagth , e pebbles being 6 inches long. Secondly, this difference of grade cannot be due sortiny an o t g e existinactioth y nb g waves, becaus e raiseth e d beach outcrop occure smalth n lo s capes alone tendenc e th coast gth d an f , yo storm action at present is to drive the shingle up into the bays. Thirdly, we hav explaio et undoubtee nth d fact tha blowe tth nmiddee santh t da n site at Galson was formed later than the raised beach, and was laid . it f o p dowto n no As regard s laterasit l distribution raisee th , d beac traces n a hwa n di easterly directio moute e th th n s f middee a hfro o r th site mf fa th e o s na North Galson Burn—tha distanca r fo , f abouis eo t t hal milea f . (The names used here wile 1-ince founth b l n ho d Ordnance Map, Popular Edition e coas, Th Sheet ) immediatel8. t y north-eas e Nortth f o ht Galsot beeye nt n e whol examinedno Burth s s ea nha coast bu , t from this e cappoin th t Cobh ea o t t a Sgeir opposit e Deleth l Rock consists of high ground unlikels i t i , y that extensive outcrop raisee th f sdo beach will be met with in this direction. Beyond Cobha Sgeir the whole coast -was examine Fithichn da frod mout e Delme Ba th , th s la f ho River fa s ra but no sign of a low-level raised beach was seen. To the south-west of the site of the midden the raised beach was traced for about half a mile along the coast, as far as the place, easily visible on the 1-inch Ordnance Map, wher e 50-footh e t contour first bend t towardou s e th s coast. So far no attempt has been made to trace the raised beach farthe thin i r s direction. No marine shells were found in the raised beach at Galson, except for Patellaw fe a vulgata, f whico e provenanchth doubtfuls ei . Where eth t e beachiddee storbasth no th f s mhey i o n b proces n i beac w f hno o s 362 PROCEEDING E SOCIETYTH F O S , APRI , 193712 L .

formation eroden t resti a , n so d platfor Archseaf mo e th n t rocksa t bu ; farthest point to which it was traced towards the south-west it appears o rest n bouldeo t r clay, thoug e sectioth h s obscurni d woulan ee b d very difficult to clear. There are two types of boulder clay in Northern Lewis, one of which is brown in colour, and contains erratics of Torri- donian arkose and sandstone, while in the other, greenish boulder clay, these erratics are extremely rare, or altogether absent. The boulder clay on which the raised beach appears to rest to the south-west of the Galson site is that -which does not contain erratics from the Torridonian. These erratics, however raisee occuo th d , n di r beach havd an , e therefore been derived from the north-east, beyond the North Galson Burn. We should mention that the remarks about the small size of the shingle in the raised beac t Galsoha n only e immediatapplth o yt e neighbourhoof do the midden site, and that to the south-west there is a development of alternate layers of coarse and fine shingle. As no fauna has been found in the raised beach near Galson, it -would be rash at present to attempt correlation with raised beaches elsewhere, especially outside Lewisalss i ot I difficul. estimato t t e amouneth f o t submergence represented by this raised beach, because it consists of a shingle bank facies along its entire length, as far as it was seen by us. e beace heighe basth onls th Th f eho f feewa y5 o t t above high-water mark toward moute Norte sth th f ho Galson Burn alreads A . y noted, e middeth t a n e sit th t seene t no e beacitsel bu s e basth , th f fhwa o e t e shinglaboua th s feef 5 o 2 tewa t p abovto e sameth e datume Th . height at the point to the south-west where it rested on boulder clay wa fee6 1 s t above present high-water markcoase Th t. from this point to the mouth of the North Galson Burn is backed by an unusually flat stri f landpo j behind whic e grounhth d rises abruptly fro altitudn ma e of 30 feet O.D., and, considering these facts, we can say that the sub- mergence represented by the raised beach is somewhere between 10 feet and 30 feet. As regards the age of this beach, it seems impossible that it is in any sense "Pre-Glacial," because even if we ignore the obscure section in which it appears to rest on boulder clay, we still have to account for the Torridonian erratics, which \vere almost certainly derived from "Torridoniae th n Boulder Claynorth-easte th o "t . Stornoway.—During r fielou d wor f 193o k e foun3w d that Loch Branahuie mile3 , s eas Stornowayf o t ancieno tw y b hels ti , marinp du e beaches, respectively north and south of the loch. These beaches can only doubtfull e calleb y d true raised beaches e onlth ys a ,section e w s sa thewn i m wer t wouli e t poordBu .see m tha sand-dunee th t s -which cover most of the surface of the ground here have been formed on top RAISED BEACH AND IRON AGE MIDDEN, LEWIS. 363

marine oth f e feature, thoug t s cleaGralsone evidenca a h th s t a r no .s i e At their base they were seen to rest on the Stornoway Conglomerate (probably of Torridonian Age), and although boulder clay occurs quite near, below , the junction of the boulder clay and beach was not show n sectioni n e heighTh f thes.o t e ancient beache s founwa sy b d measuremen feet0 1 norte ,e b botth o n he peninsult tho th sid f o e t a Melbost and at the western end of the south side. Since 1933 the sectioe soutth n hno sids beeha e n covere de buildin th ove n y a b rf go embankment to protect the road. Where there are no sea-walls, the section muce sar h obscure modery db n beach sand. Eight species of Mollusca were found in the marine deposit between norte th h sid Locf eo h Branahui Melbostd ean : Donax vittatus da Costa. ? Solen marginata Montagu. (Solen vagina Lamarck.) Patella vulgata Linne. Gibbula cineraria Linne. Littorina littorea Linne. Littorina littoralis Linne. (L. obtusata Linne.) Trivia europcea Montagu. Purpura lapillus Linne. These shells represent intertidal conditions, excep Donaxr fo t vittatus. Accordin Stepheo gt n (1930 . 532)p , , this species r beloliveo t wa s low- wate Outee r th presene marn ri th Hebridest y ka da t t emptbu , y valves are washed up by wave action to high-water mark. Elton found empty valves of Donax vittatus similarly washed up on the shore of Pabbay (South Harris) Baden-Poweld an , s alsha lo found valve f thio s s species in a like situation at Studland in Dorset and at St Andrews (Fife). The presence of this shell as a fossil is therefore no indication of the depth of deposition, and the assemblage as a whole represents the intertidal zone. As regards climatic indications speciee th founl e thib n al , si o sdt lise ar t living round the Outer Hebrides at the present day. If this deposit at Loch Branahuie is a raised beach, the amount of submergence is between 5 and 15 feet. It is therefore possible that we dealine ar g here wit hraisea dthas a sam e beac t Galsone a tth f eag ho , t mucbu h evidenc needes ei d befor consideree eb thin sca proveds da t I . is also tempting to correlate with Gregory's raised beach at Lochmaddy, at a height of 10 to 12 feet, but one would like to see more faunal or archffiological evidence before coming to a decision about this. There does, however, appear to be a bench cut in the Stornoway Conglomerate on the Eye Peninsula, which may be a continuation of the Branahuie 364 PROCEEDINGS OF THE SOCIETY, APRIL 12, 1937.

raised beach. This featur mentiones ei figured dan Jehy db Craid uan g (1934 . 865 thid p ,pid . figan ii s .an , , sit. 2) t Dunea , near , was investigated in 1935; the height of the bench was only 5 feet above present high-water markt certainli appearance d th an , s yha marinea f eo - t platform difficuls i cu t i t decido t t bu , e whethe t representri s differential movement of sea-level. 6. CONCLUSIONS. The facts which have been establishe excavatione th y db t Galsosa n are that— (1) Ther raisea s ei d wese beacth tn e Islanho coas th f f Lewio dto s fee5 2 t o abov t heigh a 0 t 1 a ef o texistin g high-water mark. (2) This raised beac formes hwa d earlier tha overlyinn na g kitchen- midden, whic bees hha n s archseologicadateit y db l conten earle th yo t t part of the Christian Era. The raised beach is also probably earlier than earth-housIroe n a nAg same th t eea site. (3) This raised beac believehs i youngee b do t r tha ngreenish-colourea d boulder clay at Galson, but verification is needed of this relation. e anima(4Th ) l remain e middenth n i s , besides givin ga pictur f o e food-habite th peoplee th f o s , show tha faune th t a about 1500 yearo sag was very similar to that of the present day. One species, however— the wild cat—was not previously known to have lived in the Outer Hebrides. There is a suggestion that the fauna of Lewis is derived partly from an earlier woodland fauna. This wild cat, the recently extinct pine-marten, the occurrence of long-tailed field mice (Apodemus), whose nearest relative e mainlanth n o se predominantl ar d y woodland forms, the pigmy shrew (partly a woodland form), an island race of song- thrush, and also the presence of blackbird in the midden, all support this suggestion e woodlanTh . d e faunfacieth f s especialli ao s y strikinn gi view of the almost entire absence of any trees in the Outer Hebrides except those planted within the last hundred years or so, or buried in older peat deposits.

REFERENCES. CALLANDER, J. GRAHAM, 1928. Royal Commission on Ancient and Historic Monuments of Scotland: Outer Hebrides, Skye, and the Small Isles. 230 pp. EDWABDS , 1924H. . J ,. Proc. Soc. Ant. Scot., vol. Iviii . 185-203pp . . ELTON , 1934C. , . "Metazoan Parasites from Isle Micf Lewisth eo n ei , Outer Hebrides," Parasitology, vol. xxvi. pp. 107-11. ELTON, C., 1936. "Parasites from Rodent Shrewd Outee san th n sri Hebrides," Ent. Mon. Mag., vol. Ixxii. pp. 31-4. RAISED BEACH AND IRON AGE MIDDEN, LEWIS. 365

ELTON , 1936C. , . "Note Freshwaten so r Animal e Isl f th Lewise o n si , Outer Hebrides," Proc. Zool. Soc. Lond., . 939-44pp . GEIKIE, J., 1894. The Great Ice Age, 3rd edition, 832 pp. GREGORY, J. W., 1928. "A Raised Beach on the Outer Hebrides," Inter- nationa] Geographical Union, First ReportCommissione th f o Pliocenen o and Pleistocene Terraces, p. 119. HEATHCOTE, N., 1900. St Kilda, 229 pp. HORA, A. M., 1934. "Notes on Mites Collected from the , Outer Hebrides," Parasitology, vol. xxvi. pp. 361-5. JEFFREYS . GWYNJ , , 1867-1869. British Conchology, London. v d , volsan . iv . JEHU, T. J., and CRAIG, R. M., 1934. "Geology of the Outer Hebrides. Part V. North Harris and Lewis," Trans. Roy. Soc. Edin., vol. Ivii. . 839-74pp . PANZER, W., 1928. "Zur Oberfliichengestal r Ausserde t n Hebriden," Zeits. f. Geomorphologie, vol. iii. pp. 169—203. STEPHEN, A. C., 1930. "Studies on the Scottish Marine Fauna. Additional Observation e Faune Sandth th Muddd n f ao o syan y e TidaAreath f lo s Zone," Trans. Roy. Soc. Edin., vol. Ivi. pp. 521-35.

MONDAY, y Wth1937Ma . SIR GEORGE MACDONALD, K.C.B., President, in the Chair.

A Ballot having been taken, the following were elected Fellows:— JOHN CARTWRIGHT, Architect, Ancient Monuments Department, H.M. Office of Works Carric0 4 , k Knowe Hill, Edinburgh. 12 , FRANK GRAY, Balnagowan School, Downfield, Dundee. JOHN MclNTOSH, M.A., Glendelvine Schoolhouse, Murthly, Perthshire. CHARLE . SMCLELLANDJ Waldee D 2 1 , n Terrace, Kilmarnock. Rev. JAMES PATRICK STEVENSON, B.A. t Mary'S , s Vicarage, Prince's Road, London, S.E. 11. WILLIAM Ross YOUNG, "Drumveigh," Barnton Avenue, Edinburgh. 4 , There were exhibited:— (1) By the Rev. E. B. A. SOMERSET, The Vicarage, Mexborough, near Rotherham. A Stone Moul r castinfo d ga crescenti r horse-shoe-shapeo c d object , foun (figheaa 1) .n f e stoneroadsidi dpo th n o s e abou mile2 t s south of Beattock, Dumfriesshire.