178 PROCEEDINGS OF THE SOCIETY, APRIL 9, 1917.

I. EARTH-HOUSES AND THEIR OCCUPANTS. BY DAVID MACRITCHIE, F.S.A. SCOT. The term " earth-house s heri " e s commouseit n di n acceptations a , denotin undergrounn ga d structure almost invariably buil f stoneo t d ,an used as a dwelling. Yird-hoose, or eird-hoose, is the form of the word in vernaculae th Lowlanf o r d iord-husd an , Scandinavian i n speech. Tigh thalaimho f (" undergroun e Gaelith d f o hous ce equivalentson e "s )i , but a more usual Gaelic name is uam or uamh, " a cave," which word is corrupted into weem or oo-ag by non-Gaelic speakers. As is well known antiquarieso t , such structure foune sar varioun di s Britise partth f so h Isles and in foreign countries. The present paper relates chiefly to those of Scotland, with an occasional reference to other examples. Althoug e correctnesth h e terth m f o s"house generallw no s i " y admitted, it may be useful, for the sake of those who have not hitherto turned their attentio thio nt s subject stato t , e briefl reasone yth r sfo regarding earth-house s dwellings a stime on e t ther A s consider. ewa - able dissent from this conclusion pointes wa t tha t I dou . t underground structures were use varioun di s countrie s storehousesa s s tombsa , d an , s templesa . Undoubtedl ye statemen thith s truei sd an , t appliet no s only to past ages but also to the present. Moreover, there are under- ground structures, of the ordinary " earth-house" plan, "which were in some cases primarily and solely used as tombs, in other cases as dwellings, and in others as tombs after being used as dwellings. Attention was draw thio nt s circumstanc Svey eb n Nilsso t earlier.1867n ni no f ,i 1 e "W may rest assured, sayse "h , " that befor e savagforese th eth f to e plains of Scani Wesand a t Gothland bega builto n d gallery-chamberthe for s dead, he had already constructed similar ones for the living."2 Nilsson therefore divides these galleried structures into classes—gallerytw o - huts and gallery-tombs. Sir Bertram Windle writes to the same effect s LifeEarlyn ii nhi Britain (London observee H , . 1897)55 . s p ,tha r Si t Arthur Evans "points out that the early barrows of the North are in fact a copy of a primitive kind of mound-dwelling, such as is still represente e gamme Lapp.th th y f b do "e This idea find s fullesit s t development in the huge burial mounds of Japan, which were con- structed as sepulchres and not as dwellings, but with an entrance 1 Perhaps as early as 1844, but my knowledge of Nilsson's work is limited to the enlarged edition published in English in 1868 (London), and edited by Sir John Lubbock, under the title of The Primitive Inhabitants of Scandinavia. tit.,. . 142-3Op pp 2 . EARTH-HOUSES AND THEIR OCCUPANTS. 179 passage whic s verhi y unsuitabl e formeth r fo er purpose althougf ho e latterth r muc.fo e Amonhus e Eskimo gth e Aleutia th f o s n Islands a custom formerly obtained which resulted in their semi-subterranean buildings serving both purpose e samth t ea s time; for whe nmembea r of their large communal dwellings died, his corpse was walled up in the compartment which he had inhabited, and his kinsfolk continued to live on as before in the tribal home.1 It is unnecessary to enlarge further upon this detail, whic s beeha h n introduced herr thefo e . purpos f showineo g tha e earth-housth t e sepulchrth d an e e cannoe b t altogether dissociated from one another. s regarda r Sofa s Scotland, however e earth-houseth , a dwellin s i g before anything else. The pronouncement made by Dr John Stuart, Secretary of this Society, on 14th December 1868, holds good to-day. " It has been doubted," he remarks, " if these houses were ever really used as place f abodeso purposa , whicr efo h they degreseeo suitedn e b mn i o et . But as to this there can be no real doubt. The substances found in many of them have been the accumulated debris of food used by man, and indicat s presenchi e s e surelkindrea eth s ya d kitchen middens which have recently attracted so much attention, while their occurrence in groups mark e gregariouth s s e habitearlth yf o speople bonee Th f .o s .th, deer othed eox an , r like creatures have been founde th wels a ,s a l shell f fishso , mixed with fatty eart d charrehan d wood. Ornamentf so bronze have been foun fe a f themwo n i d d bead ,an f streakeo s d glass. In some cases the articles found would indicate that the occupation of these house comd ha s e dow comparativelo nt y recen case t timesth es i s a , of the Irish crannogs, where objects of the rudest times are found along- sid e seventeentf thoseo th f eo h century t tracee foundationhu Th f .s o s on the surface in connection with some of these underground chambers e alsar o conclusiv s place t a leastf theia e o e , us of sr, occasional retreat of man."2 The fact sr Stuar citeD y db t will prove convincin o most g t people. f furthei t Bu r evidenc requirede eb t wili , foune b l thosn i d e instances where obstructions have been deliberately buil placer entrance o t th n di e passage, rendering accese interioth o t s r difficult bese Th t .example f so this peculiarit foune yar Irelandnaturn e di th d f theseo an , e obstacles t onca wil e eb l understoo examination a y db e sectionth f plannd o san s of the earth-house at Rathmullen, County Down. There is no room for doubting that this structurr fo buils refugt a s ewa no r mana t d fo e ,an the purposes of a storehouse, a tomb, or a temple.

1 W. H Remainse . Ballth n O , of Later Pre-Historic obtainedn Ma from ... Cavese th . e oth f Aleutian Islands, Smithsonian Contributions to Knowledge, 318, Washington City, 1878, p. 7. * Proceedings, vol. viii. p. 25. 180 PROCEEDING E SOCIETYTH F O S , APRI , 1917L9 . The accompanying illustrations explain the Rathmullen earth-house. fee0 inche6 t12 s i st I in length s generait ; le heighth feetd 6 s an i t;

BL..JP«H!iSfc—-"

Fig. 1. Sections and Ground-plan, of Earth-house at Cove Hill Rathmullen• . DownCo , . doorway barriere th fee2 t inche6 se a t sar heightn si . These barriers are in some cases built from the ground upward, and in others from the

Fig. 2. Enlargement of Section and Ground- Fig. 3. Figure of Ulsterman at Barriers in pla Earth-housf no t Rathmullenea Down. Co , . Earth-house at Rathmullen, Co. Down. roof downward. They for integran ma e lorigina parth f o t l structure, and it is obvious that th'eir purpose is to impede the advance of an intruder e thirTh d. illustration showsn imaginara y b , y figura f o e modern Ulsterman e risth ,k attending intrusio e occupanth e f ni th f o t earth-hous e n enemintrudera th e s b f o yn passingwa ey I . ma t i , EARTH-HOUSE THEID AN S R OCCUPANTS. 181 observed that this human figur s drawi e scaleo t n , thereby indicating the general dimensions of the passage.1

ARDTOLE SOUTERRAIM ARDCLASS Co DOWN.

Fig. 4. Another Ulster earth-house, tha t Ardtolea t , County Down, exhibits the same idea of a deliberate impediment placed in the way of an

Fig. 5. Interior of Earth-house at Ardtole, Ardglass, Co. Down. intruder. The total length of this souterrain is 118 feet 3 inches; most of it consisting of a long, cur-ved gallery, the height of which throughout is apparentl feey chambero 3 e widtinches6 tw tth e feetd h3 Th an , s. rise I am indebted to the late William Gray, M.B.I.A., Belfast, for these and other Ulster illustrations1 . 182 PROCEEDINGS OF THE SOCIETY, APRIL 9, 1917. to a height of 5 feet 3 inches. The only obstacle occurs at a distance of 60 feet from the entrance. It is formed by a " step" down to the level e inneoth f r chamber, which necessitate visitoe th s r lowering himself through a contracted ingress ,e floowhere onlth ar re yd roo th ean f about 18 inches apart. The photograph here reproduce e passage th Ardtolth f o dn i e e earth-house (fig) show5 : s tha , likit t e other Irish earth-housesf o s i , the same style of architecture as those of Scotland. The obstructive "step agais "i n see souterraia n i Head g Bo t ,na Coxinty Antrim.1 In Scotland there do not seem obstacley an e b so t correspond- ing exactly with those of Ireland. But the'same idea is not wholly absent from our,Scottish weems. A notable instance is in the mound - dwelling at Milton of Whitehouse, in the Braes of district of Aberdeen- s firsshirewa t t describeI . n di Scottish Notes and Queries, Aber- deen (vol , Marc10 .. ix.hNo , 1896), by Mr George Gauld, who had "recently" discovered it.2 The Fig. 6. Ground-plan of Mound-dwelling at Milton ground-plan shows that its pass- of Whitehouse, Cromar, Aberdeenshire, show- ing obstructions at H and I. e is ag intentionally blockedt ,a o place tw y largb s e upright stones. The larger of these, stands at the point where the passage turns abruptly to the right, forming an acute angle. It is " a granite slab, 3 feet 11 inches high. This slab almost fills the passage, which is narrowes t thia t s point. e passag"e entranc th th "n I t a e e standa s flat hornblende slab (2 feet 4 inches in height)." Mr Gauld adds that " this probably formed the door to the house," but it is difficult to see. how it could be so used.- More likely it was intended as a stumbling-block unwelcomn a f o y wa e e intruderth n i ; lik tallee eth bendre slath t b.a With regard to this earth-house, it may be explained that it stands, or stood, in a pre-existing knoll; and Mr Gauld concludes, with good reason, 1 This souterrain ia described by Mr W. J. Fennell in the Ulster Journal of Archceology, July 1896, pp. 272-3. A supplementary notice by myself, more fully illustrated, appeared under the title of " An Aberdeenshir2 e Mound-Dwelling" in The Antiquary, London, vol. xxxiii., May 1897. It was also notice Hone th y .d b Joh n Abercromb Proceedings,r ou n yi vol. xxxviii . 121p . . EARTH-HOUSE THEID AN S R OCCUPANTS3 18 . tha buils f roofet "knoll i e ti suiwa slopo e t th d th tf dean ,o with wood, as it most likely was, would present the appearance of one-half the roof ordinarn oa f y house." South Uist furnishes another exampl f whao e - t in appear e b o t s tentional blocking e cas f th Uainh eo n i , Sgalabhad—the Wee r mOo-ao g of Skalavat.1 Here ther slaa f s transportebei o d rock which narrows mako passage t th s t a almosei o es t impassable buildere th d t no sHa . wishe n obstruction a utiliso dt s a t i e , they could easily have buile th t opposite a greatewal t a l r distanc ef I thifro . s it massumptio e b n correct entrance th , thio et s weem would e left-hanprobablth t a e dyb pictureegrese righe e th sidth th t f d tesa o hanan , d migh concealee b t d from outside view by bushes or otherwise. However, this is merely a detail. The constriction of passages at certain points ought probably to be regarded als intendes oa imposo dt temporarea y advancchece th o kt f eo an intruder. l sucal hn I cases, especiall e notablth n yi e instanc t Rathmullena e , Ulster manifess i t i , t tha earth-house th t othe o buils n r ewa r tfo purpose tha afforo nt retreaa d menoccupatioe r tfo Th . havy nma e been casual and intermittent, or the earth-houses may have been occupied (as I am incline believeo dt ) throughou wintee th t r months abandonee b o ,t a r dfo nomadic life during the milder seasons, but that they were human habita- tions is evident. It is hardly conceivable that anyone would maintain that only those which were built with intentional obstructions were human abodes. Not unlikely these represent an earlier type, just as a fortified house belong olden a o srt period tha undefenden na d mansion. Of the date of these buildings we can speak with certainty in some cases, namely those which have stones of undoubtedly Roman workmanship built into their walls cleaA . r instanc weee f thith e o t m Crichtons sa i , Mid Lothian (fig wher 7), . Romaethe n origi mornof e than forty stones bees ha n established face f theio tTh . r situatio s integrana l e partth f so structure prove s buils thawa tt i tafte e advene Romanth th r f o tn i s Mid Lothian, and after their abandonment of some station in that neighbourhood; perhaps even subsequent to their departure from Scot- land. Of similar character was the weem at Newstead (figs. 8 and 9), now obliterated, which stood near the famous Roman post in that locality. It was discovered in 1845, and is described in vol. i. of our Proceedings by Dr John Alexander Smith. It was built of hewn stones, many of them bevelled on one edge, and two of them presenting " a rope-moulding of distinctively Roman character." f theso e ee hav On twow 2 es a , been

1 Describe Captaiy db n Thomas, Proceedings, vol. vii. 167-8pp . . * Dr Joseph Anderson, Scotland in Pagan Times, 1883, p. 301. 184 PROCEEDINGS OF THE SOCIETY, APRIL 9, 1917. lately informed by Mr Alexander Curie, is now preserved at Mertoun House. T E E F 0 4. 3 3 0 3. 3 2 0 2 5 . IS 10

Fig. 7. Plan and Wall Elevations of Earth-house at Crichton. The Crichto Newstead nan d weems prov theiy eb r construction that t exisno the td e arriva prioydi th f o tht o rl e Roman Scotlandn i s . Other weems show that they \vere occupied contemporaneously with the EARTH-HOUSES AND THEIR OCCUPANTS. 185 presence of Romans in their neighbourhood, or very soon after the invaders had left. This conclusion is arrived at from the fact that they have been found to contain specimens of the red glazed ware known as t reall" Samian.no d y di com t I "e from Samos mory an , e than Dresden

Pla, Fig1 f Earth-hous. no .8 t Newsteadea , Stone3 ; , ornamented with rope moulding, therefrom. china conies from China t thes termo bu , e convenienttw ar s , although t strictlno y accurate gooA . d exampl f thieo s Samian bowe warth s lei e earth-houseother th o f e founo r on n di t Pitcursa , Forfarshiref o e On . these two earth-houses, that of Ballo-fleld, was discovered in 1863, and in it were some fragments of Samian ware which were presented to our Societ Stewarr M y yb t Hood 13tn o h April 1863. That earth-house seems

to have been quitJ e demolished. The other, situated on the farm of Pitcur, was not dis- covered until 1878. It is a little uncertain out of which of these earth-houses the Samian bow preservew lno (fig) 10 .t Hallyburtoda n House was obtained. Formerly I assumed discoveree on e thath f 1878camt tn di o i t ,eou Pig. 9. Section of Newstead but subsequently there seemed reason to be- Earth-house. lieve that it was found at Ballo-field. It is said to have been intact at the moment of its discovery, but got broken by rough handling. It was put together at the Museum under the supervision of Dr Joseph Anderson. An interesting poin connectioin t n with this bow thais lpresenc its t e intact seems to show that the earth-house in which it was found must hav ef occupatio o gon t eou n soon afte e Romanth r s left Inchtuthilf I . earth-house th e continue occupiee b thousana o d t r dfo d year r morso e after that dateinconceivabls i t i , e that suc hfragila e vessel could have 1 Mr Hood's contribution from Ballo-fleld consisted of " two small portions of embossed red Samian war e; portion rustef so d iron; teet boned han cattlef so , founundergrounn a n di d build- ing or ' Pict's house,' at Pitcur, near Cupar-Angus." (Proceedings, vol. v. p. 82.) 186 PROCEEDING E SOCIETYTH F O S , APRI , 19179 L .

survived presumptione Th . , therefore s thai e ,occupant th t d lefha ts their home one day, perhaps to begin their summer wandering, and had never returned, being kille r captureddo .s likely i Ifs a , , the closed yha d e entrancth r entranceeo e purposth r f sfo concealment eo e Samiath , n bowhavy ma el rested inside undisturbed for some fifteen centuriest I . ma e addeb y d tha t leasa t t two other Forfarshire weems contained Samian ware—those at Fithie.and Tealing.

Fig . Fragmen10 . f Samian-waro t e Bowl preserve t Hallyburtoda n House. The larger of the Pitcur earth-houses may be noticed here. It appears largese th e b t o specimet Britise th n ni h Isles totas ,it l length being about 190 feet. Judging from the portion still roofed, we may estimate its height throughou fee 6 r more wide o t ta t Th e. annexe that forme th s western portion of the structure must have been roofed with wood. The reason for this assumption is that its great width forbids the idea s spannethawa t i t d with flagstones. Moreover y traceo n an f s o s ha t i , kind of stone roof.1 While noticing this important, specime Forfarshira f no e earth-house,

f placo givt o et ou passine a t no s i g t i glanc t anotheea r earth-housn ei "Descriptioy m e Se Earth-housn a 1 f no t Pitcura e , Forfarshire," Proceedings, vol. xxxiv. . 202-214pp . EARTH-HOUSE THEID AN S R OCCUPANTS. 187 the same county, which is remarkable for being in a complete state of preservation forethoughe th o t . e formea Thi du f o s ti s r Ear f Airlielo . The discovery was made in or about the year 1794, and Lord Airlie, who had realised thaotheo tw t r earth-house se sam founth en o dfar m (the Barn f Airlieo s beed ha )n utterly destroye d theian d r stones user fo d building material, had a clause inserted in the lease of this farm by

f vttan.sAeu'tn^ the £astfide,andtfiegreatly inclined ffaliery by which it is enfered'.

Plan of the first and second courses ofilfie roof stones. Plan Section& of subterranean Nets Souse situated t/ir£inin

Fig. 11. whic successors tenane hi h th d an t s were boun proteco dt e remainth t - g specimenin d interestinfulA an .l g descriptio f thio n s weey mb W.I . Jervise1A , Brechin e Societ s reath wa o ,d13tt n yo h June 1864.1 It has been stated that the wide chamber at Pitcur must have been roofed with wood. But, although wooden roofs were probably not infrequent,2 there was another method employed in roofing spaces which 1 Proceedings, vol. v. pp. 352-355; plan and section opposite p. 302. 1 Ther a definit s i e e referenc e tenth-centurth n i e y Sag f Thorgil o ae woode th o t s n balks supportin earth-housn a rooe f gth o f Irelann ei d into whic forcen hme Thorgilds hi thei d rsan way and encountered the inhabitants. Sir Daniel Wilson also refers to wooden roofs in earth- house Prehistorics hi n si Annals of Scotland, ed.d 2n , 1863, vol. 121p . i .. 188 PROCEEDINGS OF THE SOCIETY, APRIL 9, 1917. wer broao eve r eto longese dTo nth t flagstone e erectioth .y b This n swa of pillar pierd an st interval a s e floorth n . o s These formed supportr sfo flagstones of ordinary length, whose other end rested on the side walls. This method will be clearly understood by an examination of three specimens and one in South Uist. The Orkney specimen e thosar s t Pierowalea l (fig e islan. th 11) n di , of Westray; at Saverock, near Kirkwall; and at Grain, near Kirkwall. Illustration f thes o se her ar e e shown. Orkney furnishe a fourts h specimen at Yinstay, on the estate of Tankerness. The earth-house at

Fig . Ground-pla12 . f Earth-houso n t Saverockea , near Kirkwall.

Fig. 18. Section of Earth-house at Saverock, near Kirkwall.

Grain appear havo st e been opened abou yeae th t r 1827, after whice hth entrance became choke covered dan reopenes d overwa t I .d in-185y 7b Mr Farrer and Mr Petrie, and again reopened in 1901 by Mr James W. Cursiter contributeo wh , ddetailea d Scotsmane descriptioTh o t f t i o f no 19th October 1901e coursth n f ,whici o e sayse h :mose "Th t remark- able feature of the chamber is the fact that the roofing is supported by four massive pillars, each pillar consisting of a single water-worn block of stone place endn do , supplemented when require smaly db l ones placed on the top of them. Resting on these, with their other ends built into the nearest side wall, are other stones forming cap-brackets. Over these cap-bracket e placear s d heavy lintels lengthwis d crosswisean e th f o e building, these lintels servin supporo gt e roofingth t , which consistf so EARTH-HOUSES AND THEIR OCCUPANTS. 189 heavy water-worn stones, laid chiefly crosswise, some being laid flat and others on their edge."

. Interio14 Fig. f Pillareo r d Earth-hous t Grainea , Kirkwall. [From photographa Jamesby Tudor Cundall, B.Sv. The South Uist specimen occurs at Usinish, the supports in this instance being piers, not pillars. This structure has been fully described

HO.US1 E

CROW) PLM o, SUBTERRANEAN" STWJCXURE « CRESS. Fig. 15. by Captain Thomas (Proceedings, vol. vii. pp. 166-7). A similar structure existe t dGrresse easa th tn coaso , f Lewiso t , nort f Stornowayho , having a central pier to support the roof of a circular building. An account of it 190 . PROCEEDINGS OF THE SOCIETY, APRIL 9, 1917.

by Peter Liddle, Gress, appeared in the Proceedings, vol. x. p. 741. The adjoining galler papea describes n i yrwa e ream y db befor e Societeth y on 10th December 1894, but the paper was not printed in the Proceedings. It is interesting to add that the Third Report and Inventory of Monuments d Constructionsan e Countyth f n Caithnessi o (issuey b d Royathe l Commissio Ancienthe Historicaon nand t l Monumentand s Constructions of Scotland) contains the statement that in the parish of Latheron, in the south of , there are several " galleried dwell- ings " which have their roofing-space extende meany db f pillarsso e th n ,i s thosa sam f y Pierowallo e ewa , Saverock, Grain d Yinstayan , . This information, which is one of the many results of Mr A. O. Curie's labours than i t district, show architecturan a s l kinship between Caithnesd an s the Northern and Western Isles. The same method of supporting by means of- pillars is found in the Balearic Isles n connectioi , n wite samth h e kin f structureso d e Se . f Cartailhac'o 0 4 d illustrationan 8 s 1 Monuments . pp t a s primitifss de Ties BalSares, Toulouse, 1892. Other instructive varietie e earth-housth f so foune ear t Taransaya d * Harris; Seffister, Shetland; Briboll, ; Belleville, near King- ussie; Broomhouse, Berwickshire; and Meall na h-Uamh, South Uist. Dr John Stuart's observation, already quoted, that " in some cases the articles found would indicate thae occupatioth t f thesno e housed ha s come down to comparatively recent times," raises an interesting question. On the one hand, the existence of a Samian bowl, intact, demonstrates pretty clearly tha e housth t s founwhic n ei beed wa ha t dhi n closer dfo about fifteen hundre de othe th years n r O hand. e kin th f ,article o d s r StuarD referre y b t- shoto d wa lat e occupation. "Speakine th f o g above-ground structures of this natur n Irelandi e r ThomaM , . J s Westropp, M.R.I.A., remarks e caherscontinue:e th " Th f o , e lissesdus , and raths is very .interesting," and he enlarges upon this theme in the Transactions Royale th f o Irish Academy, vol. xxxi . 624-632pp . e Th . subterranean retreats connected with such overground buildinge ar s naturally included, and Mr Westropp cites the statement that in the year 1317 "even every man in a caher's souterrain"• was summoned by Prince Donoug e figh th f Corconiro o to ht e (ibid., vol. xxxiie . 158).p Th . inference therefore is that the weems of that part of Ireland (County Clare) were in regular occupation at the time of the battle of Bannock- burn. Casual occupation in more modern times may also be inferred, and there are definite cases of this by the Chouans of La Vendee, and by the Highland Jacobites in 1745-6.' A very remarkable instance of an inhabited dwelling closely resembling, if not identical with, an earth-house was • that visited by Dr M'Culloch in the early part of the nineteenth century, EARTH-HOUSE THEID AN S R OCCUPANTS1 19 . islane th f Nortn do i h Roiia. M'Culloch's accoun quotes i t t lengtda y hb s Ecclesioloyical . MuiThi S . n i r Notes (Edinburgh I hav ,. 1885)e86 . p , myself seen traces of tinker occupation of such places in recent years. Captain Thomas gives similar evidence (Proceedings, vol. vii . 189).p . It is impossible to understand the kind of life led in those sub- terranean abodes unles e stud w se similayth e Arctirth liff o ce races, European, Asiatic, and American, whose dwellings have been so often compared with our earth-houses. The comparison was made with great elaboration some fift yy Sve b years beeno ha ag Nilssonnst i d an , repeated since by many others. The most recent of these is, I think, Mr Erskine Beveridge, who observes:l—" The ancient Earth-Dwelling or 'Erd-House,'—often e Hebridesfounth n i d. woul . . , d see o havmt e had widely extended use. A recent visitor to Ungava Bay, Labrador, describes the Eskimo dwellings as of three kinds, the third, or Igloshuak, being 'simply an underground cellar or tilt. It is impossible to stand upright in such a residence, but the natives being small in stature experience no difficulty on this point. A subterranean passage about twelve feet long forms the principal entrance to the house, and it requires no small amount of gymnastic agility to wend one's way through squarA . e roo th hol f n i eserve a liks e purpose.' This Igloshuak [remarks Mr Beveridge] appears strongly to resemble an Earth-House, both in size and arrangement." Captain Scoresby describe sa deserte d Eskimo village visitem hi y db in 1822, in similar terms.2 The village was situated at the southern extremit f Jameson'yo s Land, Scoresby Sound, East Greenlands wa d an , composed, says Scoresby n ,hut te f "ninn closo i sr o ee combination, besides many others scattered abou e e margie flatth th tTh .. f . o n e huteithed th ha sl roofal r f beeo s n remove falled ha r no din ; what remained consisted of an excavation in the ground at the brow of the bank, about 4 feet in depth, 15 in length, and 6 to 9 in width. The sides of each hut were sustained by a wall of rough stones, and the bottom appeare gravele b o dt , clay accesmossd e ,an Th theso .s t e huts, aftee th r mannee Esquimauxth f o a horizontar s wa , l tunnel perforatine th g ground, abou fee 5 lengthn 1 ti t , openin extremite on side t gf a th e o n yo the bank, intexternae oth l air , othere andth t ,a , communicating with interioe hute th th .f ro This lowtunneo s s , thalwa persota n must creep hands o nhi t intknee d dwellinge ge san o th roofeo s st wa t di : with slabs of ston sodsd ean . Thi st bein kinhu f gdo deeply earth e sunth d n ki an , being accessible onla subterranea y yb n passage generalls i , y considered as formed altogether under ground. As, indeed, it rises very little above the surface, and as the roof, when entire, is generally covered with sods, Coll and Tiree, Edinburgh, 1903, pp. x-xi. • Journal, Edinburgh, 1823, pp. 208-214. 1 192 PROCEEDINGS OF THE SOCIETY, APRIL 9, 1917. and clothed with rnoss or grass, it partakes so much of the appearance of the rest of the ground, that it can scarcely be distinguished from it." A later writer, Captain Hall, gives a similar description of the older Eskimo dwellings :: " Formerly they built up an earth embankment, or a wall of stone about 5 feet high, and over this laid skeleton bones of the whale or spars of drift-wood, then on top of that placed skins of the seal

Fig. 16. Eskimo Earth-house, St Lawrence Island, Aleutian Isles. s seeni y rightn Behino , bo e mound-dwelling th e e figur, th dth f o e . On left is the shelter above entrance to tunnel.

Fig. 17. Eskimo Earth-house, St Lawrence Island, Aleutian Isles. Section: viewed from the opposite side. or walrus. . . . The entrances were serpentine tunnels under ground, with side walls roofed ,an d with slab f stoneso paso T .s through them on'all-fours.'n o s obligeei o g o dt " Commentin sucn go h descriptions, Nilsson i n t observesno s i t "I : Greenland only thae meew t t with dwellings constructe s her- a d de e scribed ; we find them amongst all Esquimaux tribes, wherever they are domiciled. They are invariably and everywhere characterised by the long, narrow, straigh r curvedo t , covered side gallery, pointine th o gt south or east, and by the chamber about five feet high."2 1 Life with tlie Esquimaux, London, 1864, i. 130-1; ii. 111. 2 Op. cit., p. 136. EARTH-HOUSES AND THEIR OCCUPANTS. 193 e Aleutia r wesAth fa ss a t n Islands e Nortth , h American Eskimos followe same dth e custom, although with architectural variations. e "Th natives, especially in their winter villages, were used to construct large, half underground habitations, often of extraordinary size. These were so arrange internay db l partition afforo t s sa d shelte eveo rt mans na s ya one hundred families. No fires were built in the central undivided portion, whic enteres hwa d throug e roof hholth a ,n ei provide d witha e f smalladdero th y n lI wa .compartmentnotche y b g dlo s each family had its own oil lamp, which, with the closely-fitting door of skins, and the hea f numerouo t s bodie a ver n i ys small space, suffice keeo t d p them •warm."l n morI e modern time e oldeth s r fashions have been modified an d improved upon. Examples of Eskimo dwellings in the Aleutian Isles

"w"' _A

-\ Fig. Ground-pla18 . Sectiod nan Norta f no h Greenland Winter Hut.

Nortn i d han Greenland, here shown, demonstrat close eth e resemblance betwee e Eskimth n o mode dwellerr th f d thalifou o ef an eo t n i s earth-houses. Kane Arctie ,th c explorer, writin 1857n gi , describe specialla s y small variety of Eskimo house. " The most astonishing feature," he says, " is e presencth f somo e e little out-hutsI firs s a t , thoughor , t them, dog- kennels. Thes e aboun ground-plani ar e 3 fee 4 ty b t somd an , feee3 t high." They were stone-built, of bee-hive shape, covered over with turf, and havin gstona e doora sla r spitn I b fo .f thei eo r small dimensions, they were used as dormitories in cases of emergency. It is clear that t regar s possiblKanno a comrade d s t di edi hi d e thaan se h tcoul d have accommodated themselves within such narrow limits, and it is equally obvious that the natives of that region must have been small people. Eskimos have certainl e knacyth f packinko g themselves within very small compass, but a bedroom 4 feet long by 3 feet broad, with a height feeto3 f , demand n occupana s r occupanto t f distinctlo s y smaller size than that of ordinary Europeans. W. H. Ball, op. cit., pp. 6-7. Nilsson quotes a similar description relating to Greenland, at p. 1341 : " separate cells, likstalle stahle,ea th n si " divided ofhidy fb e partitions. VOL. LI. 13 194 PROCEEDING E SOCIETYTH F O S , APRI , 1917L9 . Thos havo ewh e visite undergroune dth d house Holmth n f Papeo o , a Westray, in Orkney, have seen little rooms there whose dimensions coincide almost absolutely with those of the Eskimo sleeping-places describe Kaney db . This fact serve emphasiso t s e another link between the earth-houses of North America and of Scotland — their frequently

' -^il3?l?I:* 1 £%•'•:•..:?""-,> ' •: '••• " 'I..-:-...',-,;- , s,;; -/^^fe.Ji5*^PS^pr^;*p'>t* . -*?%p^^"' -*^%.;r S^.^ff '- .- :• '.:: ' > *&:.:«!-: v'-.^5.«*i-k:Jl:H8A«'i-_^«*Miai^*::;4^^A*S.&^

Fig. 19. Ground-plan of Earth-house at Holm of Papa Westray.

Fig. 20. Interior of Earth-house at Holm of Papa Westray.

small dimensions. This characteristi earth-houser ou f co s oftesha n been remarked upon. Of modern explorers of these archaic buildings one of e earliesth James twa s Farre f Ingleborougho r a s lefs , u ha M.P.t o ,wh description of a specimen which he visited in 1855. It is situated in the islet known as the Calf, or Holm, of Eday, in the Orkney group. " The entire length of the building," says Farrer, " was found to be 16£ feet; e entrancth vers wa ey narrow a larg d ean , ston e s placeth ewa t a d EARTH-HOUSES AND THEIR OCCUPANTS. 195

mouth. There were four chambers, the largest being at the end of the building, and measuring 6 feet 2 inches long, 4 feet 6 inches in height, fee2 inche6 td an s widedoorwae Th . r entrancyo thio et s chambes rwa 1 foot 11 inches wide." After giving further measurements, Farrer observes:—" Whilst the size of the stones used in its construction is evidenc f greao e t persona l e buildersstrengt th e par f th o e tn th o h, small and narrow rooms seem to indicate a diminutive race."1 Other investigators have been impressed in the same way as Farrer. " What size coul e peoplth d e have beecrawleo nwh n througi d h such rabbit holes as the passages of this eirde house are?" asks J. R. Tudor, after

describing in detail a similar mound-dwelling on the north-west sid2 e of Wideford Hill, near Kirkwall. Captain Thomas, R.N., who examined many of the Orkney and Hebridean souterrains, had previously (1867) asked the question that Tudor put independently in 1883. " What are we to think," asks Thomas, " when the single passage is so small that only a child could crawl through it ? "3 In" An Account of some Souterrains in Ulster," read before the British Association at Leicester in 1907, an Ulster archaeologist, Mrs Mary Hobson, writes to the same effect: " The entrances are small, but the tiny doorways between one chamber and another are even of more diminutive dimensions—great numbers bein o smalto go admit l e th t average-sized man—a person having to lie down flat in order to get through eved ,an n the widte nth h wilt allono l w other tha shouldere nth s of a woman or a boy to pass through." Captain Thomas, again, in his description of the Weem of Skalavat (Uamh Sgalabhad), in South Uist, remarks tha transportee tth d rock already referre mado dt passage eth e so narrow that, although some of his brother officers squeezed through, " I contented myself by looking through it." He adds:," This incompre- hensible narrownes featur a building e s si th n i e f thiso s period." There

is nothing at all " incomprehensible " 4 in this circumstance, the reality of whic s testifiei hy thes b o et d observer n Ulsteri s , e Orkneyth d an , Hebrides; not to bring in my own experiences. It is quite obvious that no sane race would construct buildings of this nature into which they coulcommonsense t enterth dno d an , e deductio tha s ni e dimension th t s buildere oth f s were proportionate wit dimensione hth buildingse th f so . Those who have read the descriptions of the semi-subterranean houses inhabited by modern Eskimos must be struck by the recurrence of similar observations writee Th . r quoter ErskinM y db e Beveridge (ante, . 191p ) says of the earth-houses of Ungava, Labrador: " It is impossible to stand upright in such a residence, but the natives being small in stature 1 Proceedings, . 156volp . ..ii Orkneyse Th * Shetland,d an London, 1883 . 285p , . 1 Proceedings, Proceedings,* vol. vii . 186p . . vol. vii . 168.p . 196 PROCEEDING E SOCIETYTH F O S , APRI , 19179 L . experience no difficulty on this point." In passing, it may be noticed that this writer illustrate e attitudth s f otheo e r observer Europen i s e H . begins by saying that " it is impossible to' stand upright in such a residence. e nexth t momenBu " e contradicth t s himsel y addingb f : "but the natives being small in stature experience no difficulty on this point." Thao say t s perfectl i s i t,t i y possibl o stant e d uprighn a n i t Ungava earth-house, provide e belong e racon dth e o t stha t buile th t earth-house. The "impossible" of that writer is a pendant to the "in- comprehensible" of Captain Thomas. In both cases it is artlessly assumed that the average modern European represents mankind, in the present and in the past. When the fact is once grasped that there are and were races of distinctly smaller dimensions than the average European, then the " impossible " and " incomprehensible " features vanish altogether. Kindred evidence from Greenlan furnishes di Nansenr D y tello db s,wh of a fat Danish storekeeper who stuck fast at a difficult point in the entrance-gallery of an Eskimo house. He was, of course, crawling along on all-fours. "There he stuck, struggling and roaring, but could not advance stild fout an ,l ge lesr o t smal s d retreatende ha l th boye ,n h I .s o helt p him o shovin,tw g behind, d twoan , from withi e houseth n , draggin fronn i they b m t "ghi arms . They laboure e toiled th dan n di a n i jammes d sweawa wa f n a theis fas o dta s ma a tr e browsth t bu , gun-barrel, and there was some thought of pulling down the walls of the passag orden ei liberato t r e him, befort lasa e t h emanage squeezo dt e through. If I remember rightly," adds Dr Nansen, " a window had to be torn down in order to let him out of the house again." There is another story of a missionary who, in his domiciliary visits to his Eskimo flock, had-to be dragged along the entrance passage in a sack, in which he lay straight arid rigid. From the evidence cited, it appears that Ireland and Scotland once contained a race of people whose habitations, in character and dimensions, were similar to those of Eskimos. Nilsson, who exerts himself to show the same analogy in Scandinavia, is strongly opposed to the idea that this denote raciay san l affinity. Buffone otheth n r o ,hand , believesn o , thes othed an e r grounds, thae Arctith t c race f Americao s , Asiad an , Europ ef course o wer , l akinis eal t , I understoo. d that th,e connectios ni t f Siberiaacroso no y e Nortbd yth swa an , h Atlantic late eTh .Charle s H. Chambers, a member of the Anthropological Institute of Great Britain and Ireland, concisely states the conclusion which he had arrived at, in this matter, as follows: " I believe the race which inhabited the northern shores of Europe to have been akin to the Laps, Fins and Esquimaux, and Pickte th r Pechtso f Scotlandso havo t d e an ,give n ris mane o et th f yo EARTH-HOUSES AND THEIR OCCUPANTS. 197 dwarf, troll, and fairy stories extant among the Sagas and elsewhere." Chambers makes this statemen e Anthropologicalth n ti Review r 1864fo , but he does not there state the reasons which lead him to this deduction. It is probable that he was influenced to some extent by the discovery islane madth f Burrayn edi o , Orkney previoue th n i , s year groua f o , f po twenty-seven human skulls, some of which were pronounced to be " of e Esquimauth x type, shor broad."d an tt unlikel familias No 1wa e yh r with the fact that, as recently as the seventeenth century, the Church of Burray possesse dskia n kayak belongin e rac th f Finns eo o gt r Finno , - Men who formerly frequented the waters of Orkney and Shetland.2 Like the small earth-houses, these skin-canoes are constructed for a small people, and the specimen preserved in the Anthropological Museum in Aberdeen Universit e average-size yth usee b couly b dt dno d Britof no presene th t day.3 Much migh e sai b tconnection di n with Chambers's closing words t tha bu ,don te b coul t e dwithouno t extending this paper extravagann a o t t length.4

1 Wilson's Prehistoric Annals of Scotland, ed.d ,2n 1863, . 120volp . .i . 2 For fuller accounts of these people I would refer to my " Notes on a Finnish Boat preserved in Edinburgh," Proceedings, vol .o "Th t xxiv. d ean , Aberdeen s Congeners,Kayait d an k " Proceedings, vol. xlvi. 3 This detai articln a s explaine li n "Dwar n i eo e m y fdb Skeleton Grean i s t Britaine Th n "i Scotsman 14tf o h July 1914. e addeb y d ma tha t I 4 t Chamber s probablwa s y acquainted wit e statementh h t madn i e Tulloch' Orcadibuse D s Insulis 1443n i subsequentld an , y quote modery db n writers from Barry onwards, to the effect that the Orkneys contained a population of " Peti" prior to the Norwegian colonisation thaand , t these "Peti, wer ver"who eof y small stature, were accustome takdto e refuge in little underground houses.