I.
THREE METAL-WORK HOARDS OF THE ROMAN PERIOD FROM SOUTHERN SCOTLAND.
By STUART PIGGOTT, D.LITT. HUM., F.B.A., F.S.A., F.S.A.Scoi., Professo Prehistorif o r c Archaeolog Universite th n yi y of Edinburgh. Read February 11, 1952.
CONTENTS.
PAGE INTRODUCTION ... 1 THE CHARACTER AND CONTENT THE CIRCUMETH SITED -AN S HOARDE OPTH S ..8 . STANCE2 . FINDSE OTH F S. THE SIGNIFICANCE OF THE THE DATE AND NATURE OP THE . .1 . HOARD6 . S HOARDS ...4 . THE CATALOGUE ..0 2 .
INTRODUCTION. Some twenty-five years ago the late James Curie published in these Proceedingse classith f o c e paperon Romano-Britisn so h archaeology.n I 1 the course of this study, devoted to objects of Roman origin found in native site Scotlandn si discussee h , t dsoma e length witd an ,h characteristically illuminating comment, three metal-work hoards mad largelp eu f iroo y n objects, found respectivel n Kirkcudbrightshirei y , Roxburghshird an e Berwickshire. The first and last of these, from Carlingwark Loch, and Blackburn Mill in the parish of Cockburnspath, had been briefly published in early volumes of the Proceedings; the third find, that from Eckford, was published by Curie for the first time, though found in the 1880's. It is in no spirit of criticism of Curie's treatment of these finds that a reassessmen f theo t mundertakes i presene th n ni t paper. Full thougs hhi study was e threth , e hoardtako t ed theisha r widee placth n rei survey of Roma nativd nan Scotlandn i e ; sinc e 1930'e th r knowledgou s e th f eo Earl Britainn yi e Iroespecialld Ag nan , Nortn yi h Britain s greatlha , y increased. Furthermore, the publication of an analogous (if more spectacular) hoard from newa t Anglese se 194n . i r Cyri s x Si 6ha y Fo ly b standard for the treatment of such material, and has shown what can be
> "An Inventory of Objects of Roman and Provincial Eoman Origin . . .," P.S.A.S., LXVI (1931-2), 277-397. .1 1 VOL. LXXXVII. 2 PROCEEDINGS OF THE SOCIETY, 1951-53.
wrested from seemingly unpromising scrap f ironworo s y detaileb k d examination and comparison.1 Sir Cyril's Llyn Cerrig report was the direct occasion of the study of the south Scottish metal-work hoards presente thin di se papeth s a rr andfa o s , differing character of the hoards permits, the work has been organised along the same lines as he set out. The problems presented by the Scottish finds consideree ar d under various thein headsi d r widean , r setting firse th t n i , part of the paper, the second part consisting of a detailed Catalogue of the objectthree th en i s finds with their relevant comparative materiale Th . existenc thif eo s documented Catalogue render t unnecessarysi generae th n ,i l argumen papere th f givo o tt , e reference parallelo st othen si r than general terms, sinc entre eth y under appropriate th e number will give such detail aavailables si . Each hoar separatels di y numbered unde initian ra l letter— E. for Eckford, C. for Carlingwark Loch and B. for Blackburn Mill. The line illustrations, which form an essential part of the paper, were originally prepared by members of the Honours Class in Prehistoric Archaeo- logy in the University of Edinburgh under the supervision of the writer and with his eventual checking and revision. In their present form they are almost wholl wore yth f Misko s Audrey Henshal e Nationath f o l l Museum of Antiquities whoo t , e writer'mth s grateful thank l three duear sAl .e hoards are in the National Museum, and their publication here is made possible by the courtesy of the Keeper, Mr R. B. K. Stevenson, who has given constant help during the preparation of the material.
THE SITES AND THE CIRCUMSTANCES OF THE FINDS. e discover Th e eachoardn i th s hf yo wa instancs e accidentalo n d an , precis obtainede eb informatio w e detaile findinno th th .n o f t o s ca g s na Eckforde Th mads fin d wa 188workman a ei y 3b n while diggin littla n gi e plot of land known as Toddle Bigs on the farm of Easter Wooden in the parish of Eckford, Roxburghshire. Toddle Bigs is an oval patch of ground which dried-ua e b y pma loch lying withi Kale nth f 100eo Wate . northers 0ft it n ro n side. Some 25 pieces of metal were found, according to Curie, who first published the find: 19 are catalogued here, and among the missing fragments Curie identified pieces of a hanging-lamp or lamp-stand (cf. Blackburn Mill, gridiroa f o d nan (cf.) 26 Carlingwark. B 71). C , . The Carlingwark Loch find was made about 1866 by two fishermen near r Islane LochFi th e outskirte n di ,th th whicn o f Castls o shi e Douglas, Kirkcudbrightshire objecte Th . s comprisin e hoargth d were containen di larga e bronze cauldron which appear havo st e been dredge frop de u mth shallo believe o referren w havwh o me t lak, t o di edto e tw bottoe th y mb i A Find of the Early Iron Age from Llyn Cerrig Bach, Anglesey (National Museum of Wales, 1946), referred to in this paper as Llyn Cerrig. (See also note on abbreviations, p. 20 below.) THREE METAL-WORK HOARDS OF THE ROMAN PERIOD. 3 belonge Edwardo t . dI More islandthae oartificialocf e e fth o nth on h s i n si l crannog type or is revetted with wooden piles, but no association between islands and hoard can be established.1 Blackburne Th Mill r Cockburnspato h fin s madd wa 2 e before 185y 2b labourers cutting a drain " in a haugh adjoining the Water of Eye, at a depth of abou ins5 1 t. belo e surfacewth , lyin blun go e clay, belo peate wth f "o "a moss on the farm of Blackburn Mill, parish of Cockburnspath, Berwick- shire." s Her e objecteth s were containe bronzo tw n di e cauldrons: "the one appeare havo dt e been inverte othere th n d,o wit articlee hth s within
l^\JM Ap=K ? METAL- WOR. K H0AUD
Pig. 1. Location-map for the hoards. them." This, the earliest and most precise of the three accounts of dis-
covery,3 certainly implies tha e beefind th tha d n lakedeposited ol ,n a n di late becomo rt epeat-mossa . e deduceb Littl n ca ed from these meagre reports save commoon e n factor. The Carlingwark hoard came from an existing lake, Blackburn Mill almost certainly from an ancient one, and Eckford not improbably from a simila rbearine siteTh . f thigo s circumstance upo e originanth l naturf eo deposite th consideres si d . belo7) . w(p As we have seen, 19 identifiable objects or fragments survive from the Eckford find, which appears to have been larger: we are not sure, either, that the finder recovered the whole of the deposit. The Carlingwark find, inclusive of the cauldron which contained it, numbers about 100 pieces
1 P.S.A.S. VII (1866-8), 7; X (1872-4), 286; Munro, Anc. Scot. Lake-Dwellings (1882), 28, with further refs. 2 For the desirability of using Blackburn Mill as the name of the find, cf. Bosanquet in P.S.A.8., LXII (1927-8), 250. > P.S.A.S., I (1851-4), 42-43. 4 PROCEEDINGS OF THE SOCIETY, 1952-53.
exclusive of indeterminate scraps; Blackburn Mill, with two cauldrons, totals some 65 pieces. We have then over 180 objects to consider, and, as wil seee b l n below, over 16 thesf 0o e have features which enabl mako t s eu e some sor classificatiof o t themf no .
THE DATE AND NATURE OP THE HOARDS. The dates of specific objects in the hoards are commented upon in greater
detai1 l in the Catalogue, and at various points in the commentary which precedes it. But in general it is clear that all three finds belong to some phase or phases of the Roman occupation of southern Scotland, and are the products of a native population able to acquire a considerable complement of Roman tools and other objects in addition to the products of native Early Iron Age craftsmanship. Most of these Roman objects cannot be dated except within broa de Blackburlimitsth t bu , n Mill hoard includea d broken bronze pater f Antonino a e ) hoarwhiceth s 52 typ f . i wa dh (B e deposite circumstancee unita th s ds a ,a findinf so g suggest, would providea terminus a quo for this event. The essentially peaceable character of the vast majorit objecte three th th f yo en si hoards, commente Curiey b n do , would sugges data t e afte initiae th r l Roman impact campaignine th n i ; g phase e occupatiooth f e acquisitionth f Romano n object f husbandro s e th y yb native populatiod 2n e th n o woult t unlikelye e wholedb 1s th e n th ,O . century A.D. represents the most likely covering date for all three finds. questioe naturTh e th f n theso f eo e deposits, however, canno answeree tb d so easily. Hoards of metal objects are well-known features of prehistoric archaeology therd generaa an , s ei l tendenc regaro yt d the representinms a g property buried in times of insecurity and not recovered, on analogy .with coin hoards. Distinctio s furthenha r been made between small groupf o s objects likely to represent the possessions of an individual, and larger finds of metal objects whic includy hma e scrap-meta broker o l n pieces lattee th , r being regarded as the stock-in-trade of smiths or traders. On suc analogyn ha three th , e hoards under discussion, with their broken and useless objects, would fall under the heading of smiths' hoards—mis- cellaneous irod bronzan n e objects collected togetherd ol , storen a n i d cauldro r conveniencenfo thed ,an n burie r droppedo laka r safetyn dei fo . Curie, while recognisin possibilite gth e smith'th f yo s hoarexplanan a s da - e finds howeves tioth f wa n,o t e certainoriginath no r s wa l: ownee th f o r Blackburn Mill hoard, he asked, "a husbandman or a travelling tinker"? Furthermore discussinn i , cauldronse gth , tripod gridirond san s associated with the finds, he laid emphasis on the ritual quality of brewing herbal decoctions or beer in cauldrons, illustrating his point with reference to the Filzen and Rheinzabern finds from Germany. He also drew attention to the British find most closely comparable with the Scottish hoards, that from THREE METAL-WORK HOARDROMAE TH F NSO PERIOD5 . Santo Norfolkn ni , wher emetal-wora k hoar containes dwa cauldroa n di n of Carlingwark type.1 The questio nature th f nthesf o e o similad ean r hoard raises n a s wa n di acut greae th for lake y tm b th f Llyfin eo n i dn Cerrig Bac Angleseyn hi . Her circumstancee utilitariaea th s a regar o t t deposix no e t dnth i Fo f sd o tle hoard of peasants or metal-workers, but as a votive deposit made in accord- ance with the known practice of the ancient Celts, attested by the classical writers f makino , g such offering pooln i slaked san s held sacre certaio dt n deities .interpretatioe th Suc s hwa Scottise th f no h finds half hintey b t da Curie, with his stress on ritual possibilities inherent in some of the components of the hoards, and we must now examine the matter in more detail. The deposition of objects in lakes or bogs for presumed votive purposes is attested throughout Northern European prehistory, with the Final Palaeolithic votive deposit of a drowned reindeer in the Meiendorf lake as s earliesit on f o e t manifestations. Becke s recentlha r y studied Neolithic finds in the Danish bogs which can be best interpreted as votive; in
Scandinavia generally the custom seems to have continued through 2 the Bronze Age and is well represented in later finds mentioned below. In Britain ther soms ei e evidenc likelf eo y votive depositd en e t leasth sa y b t Bronze oth f bronz x si et e a Age eg Th shieldedg .n circlbo a o a n t e i n esse i Beith in Ayrshire, and probably the similar shields found at Tetholm in Roxburghshire, Scottise ar h example sucf so famoue h th finds d san , hoard
from Llyn Faw Soutn 3 i r h Wales, which include cauldronso dtw , coule b d interpreted in this manner.4 suren o Wit e rEarle ar hgroundth e yw Iroe , nAg sinc ambiguoue eth s evidenc archaeologf eo supplementee b n yca commenty db Celtin so c religious practices by the classical writers. Fox has drawn attention to Caesar's referenc votivo t e e deposit se Gaul madth y seb afte victorya r , citine gth Swiss find at Tiefenau near Berne as an instance. This hoard of metal-work, dating from La Tene II-III times, is of some interest in showing the difficul- ties of interpretation of such finds: it includes unfinished blanks or rough- outs for iron tools and weapons, and is regarded by Tschumi as representing stock-in-trade th •workshop.a remarkablf sita e eo s th ei t Bu 6 e a one n o , peninsula formed by a loop of the River Aare, on which have also been found a La Tene settlement and graves, a Romano-Celtic temple and another Roman temple, and Roman graves. One could then put up a case for
regardin peninsule gth Celtia s aa c sanctuary persisting into Roman times, 1 E. A. Smith in Comma. Camb. A.S., xm (1909), 146, and refs. under C.I below; for "Sauton" rather than the older "Santon Downham," R. B. Clarke in Arch. J., xcvi (1939), 71. E.g. Mosefundne Lerkar pa Yngre Stenalder (1948). 2Sprockhoff, Handelsgeschichte der Germ. Bronzezeit (1930), 12, with full refs.
3 * Archaiologia, LXXI (1921), 133; Ant. J., xix (1939), 369. s JahresbericM Schweiz.r de Gesell. r Urgesch.,fu xxi (1929)° , 131. 6 PROCEEDINGS OF THE SOCIETY, 1952-53.
and see the hoard (or rather hoards: there were two finds made in the last century dones ha votives x .)a Fo s a , A find very comparable with those under discussio thas ni t from Diirnau in Wurttemburg a peat-bo n I . g nea e Federse th ra hoar s f irodo wa e n objects, includin hammerso gtw sicklea , chisela , pairo tongsf tw , so , tire fragment othed an s r pieces belongin mountinge th o gt chariotf so cartsr so , para bull-headea f o t tripod-stoola f o g dle fire-doge th . d Witan , iroe hth n objects were two bronze vessels, one a situla and the other a flask. Reinerth regarded the find as belonging to the middle of the 1st century B.C.1 famoue Th s awrum Tolosanum however, .the Celtic treasur t Toulouseea , also quoted by Fox, was specifically a series of votive deposits, "some placed in sacred precincts, and some in sacred pools," 2 and in this context fine Vemmerl0f th do Swedenn vi animaf o , humad an l n bone whan si d ha t been artificial pools, believe t imprecisele Bronzbu b e o dt eAg y dateds i , apposite.3 By the Roman Iron Age of Scandinavia, votive bog-finds are well recognised, ranging from Hjortsprincenturd 3r e yth B.Cn g i suc o t . h great hoard thas s a Thorsbjergf o t , ranging over four centuries ending about A.D. 400.4 Among these finds, the ritual carts of Djebjerg are of outstanding interest in their relation to the account of the rites accompanying the procession of the sacred car of the goddess Nerthus as related by Tacitus.5 custoe Lighth n makinf mo t o g votive offering laken i s Roman si n Gaul is throw passaga y nb whicho et I believe, , attentio t beforno s e nha bee n draw thin i s connection stort a Hilar S n f I y o Arief . yo s give Gregory nb y of Tours is a description of ceremonies carried out at a lake identified as the Lac de St Andeol, near Aubac in the Cevennes.6 At a recognised e yeatimth rf eo peopl e would e lakegatheth ,y b rcomin cartn gi s laden with food and drink for three days' feasting, and throw in as offerings bread, cheese and beeswax, as well as rags and clothing, and other objects accordin theio gt r standing. mentioe Th textilef no s recall Thorsbjere sth g
1 Reinerth, Das Federseemoor7 als Siedlungsland des Vorzeitmenschen (1929), 162. 2 Llyn Cerrig, . 69-70pp , with commen . RichmonA . I quotatione y th tb n do s from Strab Justid oan n involved. > 3 Real., xiv, 112; Hilda Ellis' statement tha poole th t s were enclose sharpeney db d stakes seems hardly justifie e evidencth n do e (Road Helo t (1943), 14-15). 4 Sheteli Falkd gan , Scand. Arch, (trans. Gordon, 1937), 186, 204, with refs.; Jankuhn, Forschungen u. Forschritte, xn (1936), 202, 365-7; Antiquity, xxvi (1952), 14. I am indebted to Professor Jankuhu for photostat earlies hi f so r papers. 6 Germania, XL. 6 The identification is that of Dumezil, Festin d'Immortality (1924), 245, citing Revel, Relig. de la
Gaule avant le Christianisme. 7 Gregor f Toursyo n GloriaI , Confessorum, Cap (M.G.H.2 . Script. Merov., I (1885) , e 749)Th . passage runs as follows: "Mons enim erat in Gabalitano territurio cognomento Helarius, lacus habens magnum. Ad quem certo tempore multitudo rusticorum, quasi libamina lacui illi exhibens, lenteamina proieciebat ac pannos, qui ad usum vestimenti virili praebentur; nonnulli lanee vellera, plurimi etiam formas casei ac ceree vel panis diversasque species, unusquisque iuxta vires suas, quse dinumerare per- longum puto. Veniebant aute plaustrim mcu s potum cibumque deferentes, mactantes animalia et per triduum aepulantes." Cf. Olwen Brogau, Roman Gaul (1953), 203. THREE METAL-WORK HOARDS OF THE ROMAN PERIOD. 7
garments e Santoth d nan , find include s wela s dleathe a x l piecewa f r o s fragments, and one must reckon with perishable elements having vanished in all finds suspected to be of a votive nature. A suggestion has facn i t been made thae depositth t f "bog-buttero s welo s " l known i n Highland Scotlan relatee Irelann b i y d votivo ddt an dma e practices,d 1an in this connectio bronzfina e f nth do e cauldro typa f no e comparable to.that from e Blackburth , n Mill hoard (B.I n clos"i ) e juxtaposition kego "t f o s bog-butte t Kyleakia r Skyn f somni o s ei e interest : cauldro butted nan r
were foun thef d i peat indeee e 1\. deed yth ar ft an n ,pi d contemporary2 t i , •would give us not only another likely votive deposit of the Iron Age in Scotland, but would take back the practice of depositing bog-butter to a respectable antiquity. There is, then, a case for regarding the three Scottish hoards under discussion as representing some form of votive deposits made in lakes or pool accordancn si e wit hwell-establishea d Celtic custom presence Th . n ei the hoards of scrap-metal, broken objects, or rough-outs would not neces- sarily militate against such an interpretation, for their presence •would be symbolic t AndeolS e ,thed d lik c shrede an ,ye th La ragd e clotf san so th t ha would in fact constitute a form of wealth to the metal-worker. Unlike the Llyn Cerrig deposit, spread over a couple of centuries, or that at Thorsbjerg, over four actuae th , l depositio Carlingware th f no Blackburd kan n Mill finds at least must have been the result of a single act whereby a group of objects, themselves perhaps collected over a period of time and deposited in their container, were together committed to the "indwelling spirit of the pool," in Fox's phrase. The Santon cauldron and its contents would represent a similar set of circumstances. ' Bu t thita s poin muse tw t conside large rth e ironwork hoards which have been found on Roman sites. At Silchester two such finds were made, one in othee disusea th n d ri centurh an dt 4t well pi e bot d th a y e f an ,hA.D o Th . first hoard Insuln i , N , froconsistea I t m Pi pieces6 6 f do , including iron axes, hammers, gouges, chisels, plough-coulters, tongs, files, a carpenter's plane hippo-sandala , gridiron.a lama , d secone pan founs Th 3 dwa n i d Well no. 2 in Insula XXIII, where the mass of ironwork accounted for 7 ft. of the filling: over 100 objects were present, again including tools such additiohoardn i N t d Pi an , n e thoss a scythesth f eo , field-anvils, bucket- handles, linch-pins, door strap-work, a padlock and cauldron-chains, and bronze vessels.4 A very similar hoard, also century h probabl4t foune s a th wa ,f n ydi o e Roma . th deeft n 6 o pn t town-sitpi f Greao e t Chesterford, containing
1 Estyn Evan Ulstern si J.A., (1947)x , 59-62. potCfe th . s containing buttee animath d n i ran t lfa Thorsbjerg find (Jankuhn loc. cit.). » P.S.A.S., xix (1884-5), 309. s Archceologia, m (1890), 742; ibid., ixv (1895), 139. « Ibid., LVII (1901), 246. 8 PROCEEDING E SOCIETYTH F O S , 1952-53.
96 objects, includin a gdoze n scythes, field-anvils, coulters, linch-pins, nave-hand f wheelso s , pole-sheathing cauldron-chain.a d an s againd An 1, at Newstead, the 1st century Pit XYI contained 94 whole or fragmentary
metal objects, almost all of (iron, and including hammers, tongs, chisels, gouges, nave-bands, scythes fragmena d an probabl,a f o t e cauldron-chain.2 Similar finds are known from Roman forts in Germany.
Ther markea s ei d uniformity between these hoards betweed an , n the3 m and the three south Scottish finds under consideration. It is noticeable that eve t Newsteana make-ue d th hoare th f peaceabls dpi o charactern ei , and reflects the farmers and craftsmen of the Romanised British population rather tha armourerse nth ' workshop e militarth f o s y forces. These town ford an t hoards sugges t firsa t t sight accumulated blacksmith's e scrath r po stock-in-trad craftsmana f eo , hiddegroune th n ni d with inten recoveo t r in times more favourable than those at the time of deposition, but in view knowe oth f n Belgic practic depositinf eo g votive offering welln i s shaftsr so , it is possible that the utilitarian explanation may not wholly apply in all examples.4
E CHARACTETH CONTEND E HOARDSRAN TH F O T . As wil demonstratee b l d below hoarde th , s contain objects e whicb n hca divided into three main classes—those of Roman origin, those of native Early Iron Age type, and those of types common to Iron Age and Roman material culture 16f O 5 .object other s o relate whice these f b ro n on hedo ca t classes, 82 are Roman, 34 are native and 49 Romano-British: they are typicall e productyth f communitieo s f Romaniseo s d Britons drawinn go sourcee th craftsmanshif so p available durin earliee gth roccupae parth f o t - tiondifferenco N . f contendetectee o eb n ca t d betwee three nth e groups, and many types are shared in common between them. They can then be treated as a whole, representative of Romano-British culture in southern Scotland from Carlingwark Loch in the west to Blackburn Mill, some 90 miles to the north-east. Perhaps the most striking feature of the assemblage, as Curie was quick to point out, is that it represents a peaceable population engaged in husbandry. At Carlingwark alone the tips of eight swords and fragments of chain-mail indicate the presence of warriors, for the probable fragment of an iron shield-boss from Blackburn Mill is slender evidence enough. All else point farmerso st , shepherds, peasante villagth d es an blacksmith r fo , 1 Arch. , xmJ. (1856). 1 , 2 Newstead, pp. 119-20. Saalbure th t a s gA (Jacobi Bomerkastells Da , Saalburg (1897) pi, ,II . xlvi , Zugmante1) , l (Germania 8 (1932), xvi, 159), and the Heidenburg, Kaiserlauten (Alt. uns. Heid. Vorzeit, v (1911), 255, pi. 46). 4 Cf. Burchell in Arch. N.L., Jan. 1949, 13. THREE METAL-WORK HOARDS OF THE ROMAN PERIOD. 9 e enamelleth d cheek-piece oa bronza bridlf d ean e terret from Eckford could as well or better have decked the harness of a chariot-pony used in peaceful pursuits than one drawing a battle-car. Iron tires and a pole-tip of rough workmanship could as well belong to a farm cart as a chariot, and the utilitarian iron bits again are more appropriate to the farm than the parade inseparable from Celtic warfare. The contrast with the pre-Roman martial assemblage from Llyn Cerri complets gi strikingd ean hoarde th ; s are the archa3ological counterparts of the hut-villages sprawling over the defence obsoletf so e hill-fort Scottise th n si h Lowland Romae th n si n period, and like them show tha phrase emptn a tth Romanax t eypa no ones wa . Whether Roma nativer no materiae th , l fro dividehoarde e mb th n sca d reasonably enough into categorie agriculturalf so , domesti technologicad can l significance farme craftsman'e housth th e : th d , ean come s th shop -n I . mentary which follows, the Roman tools and objects are dealt with first,
followed by the native material and that which is represented by types1 commo boto nt h cultures.
(a) The Roman Element in the Hoards. As detailed commen specifin o t c type e Catalogues madsi th n ei , there is no need here to do more than touch on certain points. An important innovation in husbandry which we must attribute to the Romans in Britain scythe,e isth represente fragmenty db t Carlingwarsa e th y 21-23. b k(C d )an necessary concomitan field-anvile th f Blackburo d t an t Eckfora ) 19 . n(E d Mill (B. 38), used with a hammer to repair injury to the edge while in use. Scythe representee sar t Newsteada mann o s dya Roman sites althougd ,an h known on the Continent in late La Tone times, their appearance in Britain, as in Scandinavia, must be attributed to the Romans. Sickles are common boto t h nativ Romad ean n agricultural equipment example th t bu ,e from Blackburn Mill (B. 34), retaining its original wooden handle carved into e forphallusa th f mo , mus Romane b t , associated with classical imagery Priapie th d can aspect harvest-homee th f so . Roman farriers' practic s reflecteei thn i de characteristic "buttress"r o farriers' tool of the Roman world (E. 17) and by the hippo-sandals (B. 20, 21). These latter clumsy horseshoes hav adden ea d interes than e i t on t preserve e "sandal,e treath th sf do " whic s studdehi d with stout spikes, evidently to guard against slipping in icy weather: no small risk in a Berwickshire winter.
shoule On d not passinn ei g tha adequato n t e treatmen Bomaf o t n tool similad san r metal objects 1 in Britain exists, and the evidence has to be gathered from those excavation reports of Boman sites in whic iroe hth n objects have been considered worth detailef yo d publication. Fortunately, Curie's great publicatio Newsteaf no d provide ricsa h serie typef so specifif so c applicatio Scottise th o nt h Lowlands 10 PROCEEDING E SOCIETYTH F O S , 1952-53. On the whole, the tool-kits of Roman craftsmen differed in degree rather tha kinn i d from Gatdise thosth f e o Britisr ho h carpente blacksmithd ran . An increased command over iron technolog classicae th n yi l world enabled various form f heavo s y hammers, picks, axesproducee d adzesb an o t n di forms superior to those of the native Celtic world, and all these types in their characteristic Roman form are represented in our hoards. The heavy draw-knife or spokeshave (C. 60) appears to be specifically Roman in type, but in the normal run of chisels, gouges and saws native and Roman forms appear to be indistinguishable, and they are touched on below. The Roman contribution' to the British house as reflected in these hoard s convincini s e gridironth f d 71. o d an significant,(C e g an us e Th . perhaps Eckford) as an alternative to the open hearth and traditional cauldron meant a new culinary amenity; so too the tripod (C. 73). The hanging-lamps' or lamp-stands (B. 26, and again perhaps at Eckford) introduc n elemenea f refinemeno t eved an tn perhap increasef so d privacy within the chieftain's house, and the key, spring padlock and pieces of hinges (B. 9; C. 13; E. 12; C. 84, 85; B. 28) aU imply improved house construction and greater security than that afforded by the primitive locks opened by the latch-lifter of native type (cf. C. 14 below). e bronzTh e patera from Blackburn f courso Mil s i l a echaracteristi c Roman type, one of three finds in the eastern Lowlands assigned by Bosanquet to a mid or late 2nd-century date.1 From the Carlingwark finds came fragment greenisf o s h glass (now lost); "opiecee non ins3 , . lony gb 2 ins. in breadth, is in relief the letter A and I, which may be a portion of M or some other letter." 2 These pieces of glass must have been Roman in origin: the relief inscription suggests the well-known "gladiator" cups of t centur1s e Carlingware th e th siz yf th o eA.D. t bu , k fragments impliea s larger vessel tha norma e thesf nth o n e lru smal l bowls.
Alforegoinge lth likela d yan , proportio undifferentiatee th f no d3 Romano- British types of tool, represent of course completely novel and intrusive equipment brought by and in the train of the Roman armies. They are products of mass-production and standardisation, and an indistinguishable series could be produced from any part of Britain and from most of the Continent during the first couple of centuries A.D. But the native element is of greater potential interest. How far does it reflect a population of Early Iron Age culture in southern Scotland, and of distinctively local character, or how much of the material is common to wider areas of Britain? Can affiliations betwee e nativnth e culture e nort th d thos f han f o othes o e r region establishede sb ? 1 P.S.A.S., LXII (1927-8), 246-54. 2 Ibid., vii (1866-8), 9. 3 Cf. Harden in Camulodunum (1947), 300; Kisa, Das Glas im Altertume (1908), II, 471 ff.; . if 3 Ill,92 THREE METAL-WORK HOARDE ROMATH F NSO PERIOD1 1 .
(b) The Native Element in the Hoards. thin I s sectio shale nw l conside t onlno r y types specifically non-Roman in character t als bu ,ocertai a n numbe itemf o r s which could equally reflect Roma nativr no e traditions t wilI conveniene . lb deao t l with this material as before in groups of related objects. Once again the Catalogue must be referred to for detailed documentation. Three type objecf so t represent warfare sword-tipse th : from Carlingwark (C. 26-33) e pieceth , f chain-mailo s fro e sammth e e hoarth . 74) d d(C an , probable fragment of a round iron shield-boss from Blackburn Mill (B. 30), perhapd an e Carlingwarsth k fragment . 38) f these(C sO .e sword-tipth , s with their absence of mid-rib, their outline and their variation in width, seem on the whole to be best matched in the Llyn Cerrig series of native sword-types. This series comprises at least two main types—a narrow- bladed weapo a typ f o ne know d centurn3r fro e myth B.C. onwardn i s Britain comprisind an , e writer' d (Grougth a an ) , sV p IY Group o t I I s broader blade (and a scabbard-mount of comparable proportions)— •whic Britise h th appea e b h o representativet r e Continentath f so Tena L l e I seriesII likeld an ,hav o yt e been introduce Belgse.e Carlinge th Th y db -
wark swords, like those at1 Llyn Cerrig, indicate two traditions, the latter beginning not earlier than the 1st century B.C. The chain-mai mighe on l t firsa t t sight assig Romaa o nt n origin, citing Newstea locaa s da l findt whatevebu ; ultimate th r e origi thif no s typf eo armour beins wa g t ,i acquire Celtiy db c warrior Britain si n befor Romae eth n Conquest, if presumably from Roman sources. The locus classicus is the Lexden burial, which bold spirits have conjecture havo t d e been thaf o t Cunobelin himself: 2 this contained corroded fragments of such a corselet, and whatever may have been the name of its owner, he was a Belgic chieftain and was buried early in the 1st century A.D. The second find in a native contex metal-wore Britain i th t n i s ni k hoard from Stanwic Norte th n khi Ridin f Yorkshirego e dateb o t d, around A.D. 50—7 d representin0an e gth e heighnativth f o et kingdo f Brigantiamo e presencTh . f chain-maio e l t Carlingwarka e susceptibl,b theny ma , o interpretatiot e n termi n f o s Brigantian relationships in native princely houses, as well as in those assuming a later acquisition from local Roman sources in Scotland. Wheeler, commentin e Hjortsprinth n go g (Denmark) fin f chain-maio d f o l the 3rd century B.C., has hinted that Asiatic sources may have been drawn
Celte othed th v osy an nb r barbarian peoples.3 The shield-bosses (if the fragments indeed represent these) are similarly
1 P.P.S., xvi (1950), 1—28; Group V swords listed and mapped (Carlingwark not here included), fig. d 13 an . 8 2 . p Hawke. F. C .C 2 Camulodunum,n si . 13 Borne Beyond the Imperial Frontiers (1954), 56, 59. 3 12 PROCEEDINGS OF THE SOCIETY, 1952-53.
ambiguous. Circular bronze bosses are known in the Early Iron Age of Britain, and round iron bosses at Hunsbury, some with curious spikes
through1 them, may well be from shields, and on the Continent the circular iron shield-boss appears to be a La Tene III development, Dechelette quoting examples from Alise, Carniola and East Prussia. On the other hand equalls i t i , yRomaa n type occurring locall t Newsteaya elsed dan - where in Britain, so that it is impossible to decide the precise status, Roman or native, of the Blackburn Mill piece. Objects connected with chariot consideree b r carty so sma d e nextTh . enamelled e specificallbronzar ) e 2 terret . th e (E cheek-pieced f yo an ) 1 . (E native types, the former sporadically but widely distributed in England from Somerset to East Anglia and northwards to Yorkshire. The enamelled design belongs to a school with Belgic affiliations, and comparable pieces are harness-mount Westhale th n si l hoar Boudiccaf do n date and Scotlandn i , , Auchendolle th y terret undoubten a , d import fro Belgie mth c are southf ao - east Englan f thi o s Leeds da sp typma ' e shows. Eckforde Th 2 . terre, is t however, not of southern English type, but a member of a class which has a distribution almost wholly Romabetweeo tw e nnth walls, with dated examples ranging from Trajanic to Antonine times and (in a modified form) centurh 4t r o yd A.D3r int e .oth e iroTh n uniqus i pole-tip ) 3 . itseln ei t wit (E bu f h clear analoguen i s nativ Romano-Britied ean h contexts—the Great Chesterford pole-sheathings particularle ar y interestin thin gi s connection single-piece Th . e tires must again be reckoned a native contribution, in view of the evidence brought d otheran o shox t sforwar e distinctioFo wth y b d n betwee e Celtinth c practic shrinking-of eo nsingle-pieca Mediterraneae eth tired an , n tradition e owheel e ffelloe linch-pinsth th nailinf Th o eo .t t gi fro e hoardmth s compris typeso etw firste th , , represente e pai bronzf th o r y db e caps from Blackburn Mill (B. 6), belonging to a native group with their best parallels Belgin i c examples from East Angli mid-lsf ao t century remainin e A.DTh . g linch-pins, of iron (E. 4; C. 7, 8; B. 4), are of Romano-British types, but their full distributio neves nha r been worked out: ther severae ear l Scottish examples, English specimens have been noted E.R. Yorks., Northantd san Gloucestershire, and they occur in Roman sites in Germany.3 The horse-bits (C. 4,5; B. 7, 8) are of a simple form frequent in Romano-British contexts but with good La Tene prototypes, and they differ completely from the specifically Roman type sees sa t Newsteadna frod an m, London. The furnishings of the hearth, represented by the cauldrons and fragments y cauldron-chainsb , . 14) 12 ,d B an ,; 18 , r otheif17 fo other, r1 . (C s
1 Fox in Arch. Camb. (1945), 200. - Celtic Ornament, fig. 33, terrets of Leeds' Type 2. 3 Several from Zugmantel: O.R.L., xxxn ; (1909)66 , 65 , , 10063 , , ,Taf 19036 XV . . , no Abb , 44 . c/. Saalburg, Jacobi, Riimerkast. Saalbwrg, pi. xxxxii. 7 , 1 , THREE METAL-WORK HOARDROMAE TH F NSO PERIOD3 1 . suspensio . 6-9(E n; B. 17; probabl f considerablo . 10. 18) B C ye ; ,ar e interest. Hawkes has discussed the Carlingwark and Blackburn Mill cauldron showd thean w s ynho represen Tena L I t typeseII e Britisth , h examples being not earlier, on available evidence, than the 1st century A.D.
The Santon cauldron, containing a hoard closely comparable with those1 under discussion, is a case in point; it is in a Belgic context and is of the same typ thas ea t from Carlingwark smalA . l exampl same th f eeo typf eo cauldron comes fro mwela t Silchestera l , wher t cannoei earliee b t r than c. 100 A.D., and a fine specimen, almost identical with that from Carlingwark, turned up as a stray find in the Black Moss, High Grains, Bewcastle. The o Iristw h examples f Alieno g , Bo ,sho e e fro. Antriwth th mCo d man widespread distribution of the type in the north and west. A remarkable little model of such a cauldron, cast in one piece, comes from a cremation buria t Tarlanda l , Aberdeenshire centurd , probabl2n e yth f A.D.d yo an , Hawkes has drawn attention to a parallel custom of burying miniature cauldrons amon e Gallo-Romagth n Nervie firsth t n i threi e centuries A.D. It may be worth while noting that such Scandinavian cauldrons as that from Rynkeby in Denmark, or the iron-bound types such as those from Brokaer in Jutland and Bjergelide near Horsens, come close to our Santon-Carlingwark type probabld an , y shar ecommoa n origin.2 The globular for sees ma t Blackburna n Mil less i l s dowe easti o yn nt i term f dato sr cidtureo e , excepe broadesth n i t t terms d comparablan , e examples have Hawkes d beean n x Scotlan n notlistey I Fo . ema y db e dw Kyleakie th n (Skye) cauldron alread yvera referred yan , abovo t d7) . e(p fine unpublished exampl t Abercairney a e Perteg th fro bo n hi m a w no , Museum. Further t i shoul, e pointeb d t thae cauldroou dth t n from Kincardine Moss, Stirling, is a particularly fine example of Fox's globular, composite type, and must surely be of Iron Age date and not, as sometimes
inferred rather 3 than stated a membe, e Latth ef o rBronz groue Ag ef po globular cauldrons.4 e Y-shapeTh d e suspensiochainth r fo s f cauldrono n s ove n opea r n hearth, represented at Eckford (E. 6-9) and Blackburn Mill (B. 17), and probabl a fragmen y yb t from Carlingwar f grea o . 18) B e t . ,10ar . cf ;(C k interest. In addition to the examples just mentioned, there are some 16 finds of cauldron-chains or fragments from Britain, which can be divided
1 "Bronze-Workers, Cauldrons, and Bucket-Animals in Iron Age and Roman Britain," Aspects of Archceology (1951), 172-99, esp. 179-83. 2 Rynkeby, Brensted, Danmarks Oldtid, in, 97; Brokaer and Bjergelide, Muller in Aarbager (1881), 110; cf. Stenberger, bland under Yngre Jernaldern (1933), fig. 1; Shetelig and Palk, Scand. Arch, (trans. Gordon), pi. 30a. Cf. also Eggers, Der Bomische Import in Freien Germanien (1951), 40, Taf. 2 and Karter distributiofo an, 0 n1 d11 f sucno h cauldron Northern si n Europe; Wheeler, Rome Beyonde th Imperial Frontiers (1954), 72-90 for summary. " Llyn Cerrig,; P.S.A.S.,88 . p xix (1884-5), 313. ' Cf. Ohilde, Prehist. Scot. (1935), fig. 44. The Kincardine Moss cauldron was not, however, included by Leeds in his classic study of Late Bronze Age cauldrons (Archceologia, LXXX (1930), 1-36). 14 . ' PROCEEDING E SOCIETYTH F O S , 1952-53. into two types. The first, or Stanfordbury Type, is that which is also known fro mnumbea sitef ro Hallstat f so Tena L r eContinente o t datth n eo e on : English fin whaf do probabls ti r yo fragmena h suce 4t on he f th chaio tf o ns i 3rd century B.C., but the remainder form a compact group in time and space, confined to the Belgic area of south-eastern England and ranging in date from the 1st century B.C. into the earlier Romano-British period. The Eckford fragments are from a chain of this Stanfordbury type. At Blackburn Mill, however, and probably also at Carlingwark if the broken fragmen s correctli 0 irof 1 . o t nC y interpreted have w , e representa- tivea secon f o s d typ f chaineo e Greatth , Chesterford Type, whicn i e hth looped rods joined with rings which are characteristic of the Stanfordbury type are replaced by relatively elaborate chain-work. In three examples at least ther complea s ei x openwork decorativ chaine th f o , p e to featur e th t ea e Y-junctioth d an elaborates ni decorativn di e fashion t BlackburA . n Mill this feature incorporates ornamenta e Grealth spiralsn i t Chesterford an , d exampl e junctioeth s formeni a fantasti f o d c tour forcee blacksmithd f o - ing in which two iron rods have been tied into a reef-knot which forms the central feature of a reduplicated triple lyre pattern, the outermost members in the lower hah7 of which are returned as tight spirals in precisely the manner of Blackburn Mill.1 The elaborate double-looped links of the Blackburn Mill chai alse nar o e Greafounth n tdo Chesterford specimenn o , Stanfordbura y type chain from Ove (non .Fe rSaalburg e II)th n t o a , d an , slave-chain Romano-Belgif so c type from Lagore which, with other evidence from that indicaty sitema , occupation ea n earlier tha ne Dar thath kn i t Age whicy sb crannoe hth bess gi t known. The known examples of Great Chesterford type chains come from Romano-Belgic contexts in southern England, and there is a probable fragmentary example from Newstead. Their distribution patter s conni - cordant with the Stanfordbury type chains, and both of course represent, with the cauldrons they supported, a feature of the native household wholly e Romaalieth o nt n tradition e greaTh t. e lengtcompletth f ho e .Great Chesterford chain, slightly over 7 ft., gives an indication of the proportions of house th whicn ei h suc hchaia n could han gcauldroa n abov fireea e th : larg euppee chai rine oth d f th t gren a n show heltransversa s than do wa t ti e pole, along which it could be shifted to bring it into the desired position over or to one side of the hearth.2 The distribution in the British Isles of cauldron-chain foregoine th f so g . 18) type(p . 3 gives g si fi n i Celtie Th c househol furthes di r represente objecto tw comparativy f dsb o e luxury e iroth :n e bronzmirror-handleth d ean tankard-handle) 3 . (C , 2) . (C
Cf. Fox in Ant. J., xxvm (1948), 129; P.P.S., xvin (1952), 50. The spirals are of course present in
classicae th 1 sub-classicar lo l prototype lyre th e f patternso . 2 I owe this suggestion to Sir Cyril Pox, who has also pointed out to me a medieval counterpart of the pole-hung cauldron-chain. THREE METAL-WORK HOARDS OF THE ROMAN PERIOD. 15
Fig . Iro2 . n centre-piec cauldron-chainf eo , Great Chesterford. 16 PROCEEDINGS OF THE SOCIETY, 1952-53.
both from Carlingwark mirror-handle shows th ha w x nho Fo nort.a s ei h English type, with outlier south-wese th n i s t Stamfora t d Hill, Plymouth, and to the north-west in the Lambay Island hoard off Dublin.1 The tankard- handle, on the other hand, comes into Type II of Corcoran's classification, which he regards as a Belgic form,2 so that •while the mirror-handle may be a more or less local product with probable Brigantian origins, the tankard- handl likels ei havo yt origin ea southern ni n England, withi Belgie nth c province. In the sphere of agriculture, the hoards have produced two ploughshares of different type (E. 10 and B. 31) but both within Payne's group of Belgic or Romano-British shares.3 Comparable shares come from Trapraiw nLa Oxnad an msouthern i n Scotland otherwise ar t bu , e confine south-easo dt t Englan Belgin di Romar co n contexts. difficuls i t I commeno t t t upo craftsmen'e nth s hoarde toolth n si s which cannot be assigned to a Roman origin, or which are more likely to be native than not 20). tangee . (C th , Sucw dsa h sicklee form 37). th e adzes (C th s a , (B. 39), or the files (C. 68, 69), gouges (B. 40, 50), etc., can be seen to have a general origin in the later La Tene cultures of the Continent, and in Britain are the common stock of Iron Age B and C at least. Good parallels exist in southern England, but this may reflect more extensive excavation rather than a real concentration. The striking feature which does emerge from this analysis however is that, wher precisea e attributio madee b almoss i n t i ,n ca t invariably withie nth Belgi r Romano-Belgico c culture south-easterf so n England. Distinctively North British types are rare—the Eckford terret and the Carlingwark mirror-handle stand out as exceptions. Certain well-known native types of metal-work ,e develope sucth s ha d form f horse-bito s s characteristif co North Britain in the 1st and 2nd centuries A.D., are noticeably absent from the hoards. Our final task, then, must be to seek for a reason for the presence of these alien types in southern Scotland.
E SIGNIFICANCTH HOARDSE TH P EO . The mixtur f Romaeo nativd nan e meta l e hoardtypeth n i ss could reasonably be supposed to reflect one of three possibilities: 1. The native types are not sufficiently distinctive to be attributable to phasy e latean r areth eo f r ao cultureIro e nAg f Britaino s t bu , represen e commoth t n stoc blacksmithinf ko d time an th f eo t ga shortly after the Roman Conquest. They would therefore indicate
a local contribution from south Scottish e culturesIroAg n , 1 Arch. Camb. (1948), 24-44. P.P.8., xvin (1952), 85-102. 2 3 Arch. , ciJ. v (1948), 82-111. THREE METAL-WORK HOARDS OF THE ROMAN PERIOD. 17
combined with Roman types imported into the region in the Agricola lated nan r campaigns. 2. If weight is attached to the Belgic character of much of the native metal-work presencs it , Scotlann ei d resule coulth immigraf e do tb - tions into the Lowlands from south-east England on the eve of, or following immediately upon, the Roman Conquest of A.D. 43. The Roman element in the hoards would indicate subsequent contact with military forces stationed at Newstead and elsewhere southern i n Scotland. 3. The hoards could be regarded as representing the importation into Scotland of a mixed Romano-British or Romano-Belgic culture from south-eastern r earl Englancenturo d e latt y2n th 1s e n ydi A.D., in which the fusion of native and Roman elements had already taken place before the move north. Such a movement within the period of the Roman occupation could hardly be other than an expressio deliberatf no e Imperial policy. The first of these three views depends upon assumptions easy neither to confirm or refute. In our admittedly incomplete knowledge of Early Iron Age ironwork in Britain, it might be urged that to define regional types as likelBelgie b forco t evidenc o e yt s c i eth e beyon propes dit r boundst Bu . cauldron-chaine th ploughshared san s seem distinctive forms enough, absent for instance in the large metal-work series from Grlastonbury, and the differenc composition i e r Scottisou f no h hoards from tha f Llyo t n Cerrig suggests that those responsibl r theifo e r deposition drew upon sources neither pre-Roman in date nor characteristic of the Jurassic Zone or the south-west virtuae th : l absenc then ei metal-worf mo k types knowe b o nt Brigantia distribution ni n would again argue against derivation from that area. (Fig) .3. native Ith f e hoarde typeth n regardee si s b likels da derivo yt e from south-eas t e e BelgiwidesEnglandb th o t n ci t d sensean , muse w , t then conside dat e manned rth ean theif ro r acquisitio communitiey nb southern si n Scotland. It is on the whole unlikely that they represent objects acquired by trade from Belgic sources by an existing iron-using population in the north, muse w td supposan e that they represent actual Belgic elements amone gth mixed refugee population from southern England which seems to have been responsible for the establishment of Early Iron Age culture in the Scottish Lowlands. These immigrants appear to have entered largely by way of
the Tweed1 mouth, in the 1st century B.C. and later, and are mainly repre- sente theiy db r hill-fort e Chevioth n si t foothills Belgae Th south-easn .3i t Englan t appeano o havdo d rt e made fortification type th ef o srepresente d on a small scale by the southern Scottish sites, and the metal-work objects 0. M. Piggott in P.S.A.S., MEXXIV (1949-50), 129-135; S. Piggott in The Problem of the Picts 1 (forthcoming), ch. ii . VOL. LXXXVII. 2 18 PROCEEDINGS OF THE SOCIETY, 1952-53.
CAULDRONS, k CAULDRON-CHAINS CAULDRONS CHAINS « 'ONST AN ' 3'BATTEIUEA'
BELC U COIN ACE
Pig. 3. THREE METAL-WORK HOARDS OF THE ROMAN PERIOD. 19
under discussion would be the only archaeological evidence of their establish- northe menth f sucI n i t. h communities were brought into being shortly befor Romae eth n Conques f Scotlando t , they would provid requisite eth e native element in the hoards, to which would be added Roman objects from local military sources. otheOe nth r hand objecte th , suspectef o s d Belgic origihoarde th n ni s are those known in southern England in contexts after A.D. 43—they are in fact to be regarded as Romano-Belgic rather than representative of the pre-Roman e nativphasth f o ee culture e wholTh .e compositioe th f no have hoardw es a seen , sis , strikingly simila thoso rt e from such Roman sites Silchestes a Grear ro t Chesterford lase thref th to d e an possibilities, , thae tth metal-work under consideration represents communitie peoplf so Romanof eo - Belgic origin, newly established in Scotland after the Agricolan campaign, is in many ways the most attractive.. In a recent study of native agriculture at the time of the Roman Conquest of Britain, the writer has pointed out that north-west of the Jurassic Ridge there are likely to have been large areas of pastoralism, with subsistence agricultur Bronza t levelea e thad eScotlanAg n an ,ti d even more primitive conditions may "well have prevailed.1 The ploughshares of Romano-Belgic type in southern Scotland suggest a deliberate innovation in agricultural techniques, eithe resula s rnativa f o t e migratio arounr o t ntime a dth f eo the Roman Conquest of southern Britain, or, as is now suggested, as part of a Roman scheme for increasing the corn production of northern Britain to meet some at least of the requirements of local garrisons. We are re- minde corveee th f do s later employe Hadrian'n do s Walln me , f mado p eu from southern England including member tribae th f so l e Dumareath f so - nonii Durotrigee th , Belgie th d c san Catuvellauni, Reginaf o d an 2gir a , f o l the same tribe, who married a Palmyrene and was buried at South Shields. The deliberate f necessari d an , y forcible, transferenc t onl eno f militar yo y f wholo unit t esbu tribal groups froregioe mon anotheo nt r was, after all, by no means unknown as an instrument of Roman policy.3 The alternative suggestion could therefor t forwarpu e eb d tha hoarde tth s (and other evidence of Belgic contacts in southern Scotland) might be interpreted as indicating deliberata e transferenc civif eo l population fro . Englanm ScottisSE e th o dt h Lowlands at the end of the 1st century A.D» Romann I Natived an Northn i Britain (forthcoming), oh. i . 1 Bruce, Handbook Romane th o t Wall . (edRichmondA . .I , 1947), 160. 2 3 Cf. Collingwood, Roman Britain (1936), 146; Macdouald, Roman Wall in Scotland (2nd ed., 1934), 48. 20 PROCEEDINGS OF THE SOCIETY, 1952-53.
CATALOGUE. e cataloguTh objectf eo l thre al n ei s finds which follows her s arrangeei n di concordance with the illustrations in figs. 4 to 13. Reference letters indicate the appropriate find, B. representing Eckford, C. Carlingwark, B. Blackburn Mill. Material, dimension d detailsan f eacso h objec e givear t n first, followee th y b d Catalogue Numbe e Nationath f o r l Museu f Antiquitiemo bracketsn i s . There then follow general and comparative comments. In addition to abbreviations normally in use, certain works, referred to on several occasions citee ar ,shortene n di d for follows:s ma — Camulodunum . Hawkes , Camulodunum, d . Hull R. . C. F . an ,. M ,C , Soc. Ant. Lond. Research Committee Report (1947)V XI . . no Celtic Ornament Leeds, B. T., Celtic Ornament in the British Isles down A.D.o t 700, 1933. Dechelette, ill, IV . Dechelette, J., Mdnuel d'Archeologie . . ., 2nd Ed., m (1927) (1927)v i , . Excav. Gran. Chase, I, II Pitt-Rivers, A., Excavations n i Cranborne Chase, i (1887), ii (1888). Glastonbury, II Bulleid, A., and Gray, H. St G., The Glastonbury Lake-Village, II (1917). Lagore . 'Hencken, H., "Lagore Crannog," P.R.I.A., LH (Sect (1950)) C . , 1-247. Llyn Cerrig Fox, C., A. e Early Findth f o Ironfrome Ag Llyn Cerrig Bach, Anglesey, 1946. Lydney . Wheeler . M.E , . Excavatione ReportR ,th n o . . . in Lydney Park, Gloucestershire, Soc. Ant. Lond. Research Committee Repor(1932)X I . no t. Maiden Castle Ibid., Maiden Castle, Dorset, ibid.,I ReporXI . no t (1943). Newstead Curie, J., A Roman Frontier Post and its People: The Fort of Netvstead, 1911. Richborough, II, IV Bushe-Fox , SecondP. . J , ExcavationReporte th n o f o the Roman Fort t Richborough,a Kent, Soc. Ant. Lond. Research Committee I (1928)ReporVI . no t; ibid., Fourth Report . . ., ibid., Report no. XVI (1949). Verulamium Wheeler, R. E. M., Verulamium: A Belgic and Two Roman Cities, Soc. Ant. Lond. Research Corn- - mittee ReportI (1936)X . no .
I. THE ECKFORD FIND. E 1. BRIDLE CHEEK-PIECE, bronze, 4-25 ins. long, maximum breadth 0-. 9in The cheek-piec circulas ei sectionn ri , wit hflat-sidea d central expansioo nt accommodate the leather strap of the bridle within a slot 1-2 by 0-2 ins. The slightly expanded terminals are decorated each with three incised lines. e centra th e fac Of on o ne l expansio a desig s i nn champlevni d re e enamel, in good condition. (DWA 1.) The symmetrical "fold-over" desig s basei na repeate n o d d paif o r voided acorn-shaped nodes unite a "broken-back y b d a " s i curve d an , THREE METAL-WORK HOARDE ROMATH F NSO PERIOD1 2 .
simplified, thoug t debasedhno , versioe enamelleth f no d harness-mount from Westhall, Suffolk (Leeds, Celtic Ornament, pi. i, 6), or the terrets from that sitfrod eman Lakenheath (ibid., pi , 3.i ; Lakenheath identification by R. R. Clarke loc. cit. inf., 70) and Auchendolly, Kirkcudbrightshire (P.S.A.S., (1886)x x , 396). There goos i d reaso regardinr nfo Westhale gth l find as of Boudiccan date (R. R. Clarke in Arch. J., xcvi (1940), 68-70) mid-lsa d an t century A.D l thes. datal r eefo enamels seems reasonable. Bronze cheek-piece t commobridler sno fo Britis e e th s ar n i h Iron Age, and seem unknown in comparable Continental cultures. They may repre- sent a local translation into metal of the horn and bone cheek-pieces of Iron Age "A" and Late Bronze Age type.
BRONZE BRIDLE CHEEK-PIECES.
Site. Reference. Comment.
Birrens, Dumfriesshire. P.S.A.S., LXII (1937-8), Massive bronze cheek-piece with 335-7, fig. 38. terminal knobs and panel of double-trumpet relief mould- ing, associated wit terreto htw s with enamelled knobs in a deposi . A.Dc f o t. 157. Bowerchalke, Wilts. Wilts. A.M., XLIII (1925 blud - an eEname d squarere n i l s 37), 352. and discs. Cambridge. V.C.H., Cambridge, i Small bronze cheek-piece with (1938), 292, fig. 25, ring-and-dot ornament ( ? sunk no. 8. for enamel). Eckford, Roxburghshire. Present paper. High Cross, Leicestershire. Leicester Mus., unpub- Bronze. Famt traces of incised lished. design probably similar to that Eckfore onth d specimeno N . enamel traces. Found in as- sociation with bronze chariot horn-cap (Llyn Cerrig, pp. 17, 77); probably second half o f 1st cent. A.D. Polden Hill, Somersetshire. Brit. Mus. E.I. A. Guide Five bronze and enamel cheek- (1925), 143-4 Henry. ;F , f iroo no wertw piece d e an s Prehistoire, n (1933), 99. found in this hoard; probably of 1st cent. A.D. Santon Downham, Norfolk. F. Henry, Prehistoire, u Enamelled bronze cheek-piece (1933). 98 , with lateral expansion t slosa t simila Eckfordo rt . Silchester, Hants. Ibid. Simple form similar to High Cross, bronze with shakings for enamel. Stanwick, N.R., Yorks. Archceologia, LX (1906), Small exampl bronzee— - un , 289; York vol. Arch. decorated. From the large Inst. n (1848) , pi. iii, metal-work hoard, probably p. 37, no. 3. mainly mid-lst cent. A.D. 22 PROCEEDINGS OF THE SOCIETY, 1952-53. E. 2. TERRET, bronze. Loo f bronzeo p , broke d distortean n t approxibu d - mately 1-7 ins. across internall originan yi l state, terminatin moulden gi d collars linke flay db t platelooe decorates Th pi . threy db e groups eacf ho three pellets or knobs spaced symmetrically. (DWA 2.)
Fig. 4. Bronze objects from Eckford (E. 1; E. 2} and Carlingwark (0. 2). (i.)
Knobbed terret thif so s typ Leeds'e ear s clas (Celtics7 Ornament, . 125)p , and have a predominantly Lowland and Scottish distribution. To Leeds's list (Kirkby Thore, Yorks; Muircleugh, Lauder; Eckford; Middlebie, Dumfriesshire; Traprain Law; but not Ardoch) add: Castledykes, Lanark (Hunterian Mus.); Nether Denton (Chesters Mus.); Fremington Hagg, Swaledale (York Mus.); Poltross Bur . Cumb.n(T West.d an A.S., N.S., xi, 13). The distribution is almost entirely between the two Roman Walls. The Nether Denton terre s almosi t t certainly Trajanic, though that from Castledykes is Antonine, and Poltross Burn (a modified form) 3rd to 4th centur ytype A.DTh e .obviousl lona Northe d g th yha lif n ei . E. 3. CHARIOT or CART POLE-TIP, iron. A massive iron ferrule or shoe, circular in cross-section and of truncated conical form. It is broken at its larger end, which is open, and its maximum length is now 4-6 ins. The THREE METAL-WORK HOARDE ROMATH F NSO PERIOD3 2 .
external diamete e smallerth f o r , s 2-e internai close 8th d ins.d en d an l, diameter approximately 2-3 s insperforatee i close d Th . en d d centrally b ya hol e 0-. diameter5in a secon d an ,d hole, . 0-diameterin 7 , perforates the side 2-2 ins. from the closed end. (DWA 14.) The most probable use of this object seems to be a pole-tip for the central shaft of a two-horse vehicle, secured by a broad-headed nail driven in lengthwise fro frone mth t wit e transvershth e perforatio alloo nt a f wo vertical pin to which the yoke would be lashed. No precise parallels can be quoted, though the relevant material is discussed by Fox in describing analogoue th t preciselno t sbu y similar iron sheathin charioa f go t pole-tip from Llyn Cerrig (op. cit., 92), and the bronze pole-tip from the famous Charioteer's Barrow at Arras, E.R., Yorks (Ant. J., xxix (1949), 81-3). In the hoard of iron objects from Great Chesterford in Essex, of Roman date, two iron pole-tips were, however, found, open at the end and with a square e yoke-pinth hol r fo e , intermediate between Arra d Eckforan s n typi d e (Arch. J., bronze xmTh (1856) pied . i) .mountingan , 4 , e pole-tith f so p and other Djebjere partth f so g ritual alscary notee ob ma t d (H. Petersen, Vognfundene i Djebjerg Prcestergaardmose (Copenhagen, 1888), pi, nosv . .6 and 7). Although lighter and of more elaborate design, they present the same fundamental elements of design as the Eckford specimen, of truncated conical shape, -with a (decorative) boss or large-headed nail at the smaller, close hold transvers the end efor ,and e pin. Precise datin s impossiblgi e lonvien ei th gf w o persistenc chariotf eo s in Iron Age Europe and during Roman times. Arras is probably late 3rd century B.C., Djebjerg about 100 B.C., and Llyn Cerrig from the 1st century B.C. to c. A.D. 50. Great Chesterford is within the Roman period but with native elements (e.g. cauldron-chains). The Eckford specimen may be anywhere withi range nth othee datf th eo f reo associated objects. E. 4. LMCH-PDJ, iron. A plain square-section iron bar 0-5 in. thick with a flat spatulate head. Total length 5-6 ins. (DWA 22.) Parallel thio t s s linch-pin come fro ma Roma n contex Hadrian'n o t s wall at Chesters (two examples in Chesters Mus.); another comes from a Roman site at Elslack, Yorks (Skipton Mus.). There seems little doubt that this decorate hande type th on related,s e o i t , e th dn o spatulate-heade d linch-pin of enamelled bronze from King's Langley, Herts (Ant. J., xx (1940), 358), and, on the other hand, to the large series of iron linch-pins with looped spatulate head normally foun Roman di n contexts [Ant. J., xxi (1941); 67 , c/. Newstead (Newstead, pi. Ixx ; Blackbur6) , , I,3 n below)4 Mil . . (B Cf l . alsprobablothe e example from Carlingwar 7)]k(C. . The Kings Langley pin is dated to the first half of the 1st century A.D., anpine dth Roman si n context centurieh 4t e s th rang so t A.D. t e 1s fro , e mth nativf o e b e y originma t .bu . TIRE5 . E , x ironfragmentSi . s ranging from 12- o t 4-5 4 ins. longl al , probabl e tireon . f o yMuc h rusted t widtbu , h estimate t a 1-d 3 ins., maximum thickness 5 mm. Apparently Fox's type A/C, worn rather thin (Llyn Cerrig, p. 75). (DWA 19.) nail-holeso Theren e ar d thes an , e fragments presumably represena t single-piece iron tire shrunk on to the felloe of the wheel, a type of tire and techniqu f fixineo g that seems non-Roma a TenL f eo origid nan n (op. cit., 12). Comparable tires, complet fitted e an thei o dt r wheels knowe ar , n from 24 PROCEEDING E SOCIETYTH F O S , 1952-53.
Newstead (Newstead, pi. Ixix) and Bar Hill (Macdonald and Park, Bar Hill (1906), 92-9). Tire fragments also Carlingware occuth n ri Blackburd kan n Mill finds described in this paper (C. 6; B. 3). . TWISTE6 . E D LOOPED EOD, iron square-sectioneA . d rod, twiste o givt d e a decorative effect, with flattened loops at each end. Overall length 11-0 ins., thickness 0-. 3in (DW A 15.) . POT-HOO7 . E K WITH TWISTED SHAFT, iron. Broken fragment 8-8 ins. long, made from a twisted rod 0-3 in. thick, expanding to a flattened hook 0-45 in. wide at the unbroken end. (DWA 16.) E. 8. POT-HOOK WITH TWISTED SHAFT, iron. Smaller fragment 5-5 ins. long, matchin . 6 (DW . gno A 17.) . HOO9 . E K WITH TWISTED SHAFT, iron. Fragment 3-9 ins. long, . 0-in 2 thick twistef o , witd dro h terminal hook (not flattened). (DWA 18.) Nos. 6—9 should be taken together as elements of a double-hooked cauldron chain. British cauldro n dividee t uncommochainb no y e dma sar intd o noan tw types, both represente Soute th n dhi Scottish hoards under discussione Th . Eckford fragments represent a chain of the Stanfordbury Type, in which looped rods joined by rings form the main components, with lengths of chain adjusted to the height above the fire required. This type follows closely the Continental pattern, where it goes back to Hallstatt times. English find e mainlar s Belgin yi c contexts (Bigbury, Stanfordbury), thouge hth Bledlow hook indeef i , cauldroa f do n chain, centurwould 3r e db y e B.CTh . second r Greato , Chesterford Type f chaio s representeni t Blackburda n Mill and is discussed later (B. 17 below); it is a Romano-Belgic type.
DOUBLE-HOOKED CAULDRON CHAINS. . EUEOPEA . (The following provisional lis certainls i t y imperfec t indicatebu t widee sth range of the type.)
Site. Beference. Comment.
Anderlecht, Barbant, Bel- De Loe, Belgique Anc.,, II Twisted rod chaind san , from pre- gium. 162, fig. .64 Boman level under Bomano- Belgic house. Camp de Chateau, Salins, Dechelette, m, 292, fig. Chain, rods and hooks. Not France. 323, I. twisted. Hallstatt. Duhren, Baden. Dechelette , 293m , . a TenL I graveI en I , datey db Jacobsthal to late 2nd cent. B.C. (Early Celtic Art, i, 157). Emmendingen, Baden. Dechelette, in, 926, fig. a SimilaTenL eo t r example, 636, I. found with a cauldron of Car- lingwark type. Late La Tene. Frogg, Bosegg, Carinthia. P.8.A.S., xxm (1889), Chaid hooan n k fragments. 244, pi. iv, 8. Twisted. Hallstatt. THREE METAL-WORK HOARDROMATHE NOF S PERIOD25 .
E1Z ^ E 10
Pig. 5. Iron objects from Bckford. (J.) . Europe—A (continued).
Site. Reference. Comment.
Tenea L , Switzerland. Vouga, La Tene,.y\. xxvii, Of La Tene II date. 4:4; Dechelette, in, 293, fig. 323, 2. Saalburg, Germany. Jacobi, Ronikas. Saalburg From Well no. 6 in 1st cent. A.D. (1897), i, -157; n, pi. Roman fort. XIV, vii. Tielle Inferieure, Switzerland. Dechelette, in, 293. Vertault, C6te-d'Or, France. Dechelette, m, 295, fig. Similar to the La Tene example. 323, 3. From a Gallic Wall fort of late La Tene date with occupation into Roman times. Zarybrick, Bohemia. Dechelette, 292, in,85 . Fragments of chain and hooks with twisted rods. Hallstatt.
. CAULDROB N CHAIN TYPF O S EI (STANFORDBUR Y TYPE BRITAINN I ) .
Site. Reference. Comment.
Bigbury Camp, Kent. Arch. J., LIX (1902), 215, Several examples, with fragments pi. ii, 5; ibid., LXXXIX of tripod f o Stanfordburs y (1933), 107. type; 1st cent. B.C. /early 1st cent. A.D. Bledlow, Bucks. Records f o Bucks., xiv Twisted, hooked and looped rod (1944), 197, fig. l,no. 2. in Iroe "AnAg " contexf o t 4th-3rd cent. B.C. Eckford, Roxburghshire. Present paper. Hunsbury, Northants. Arch. J., xcm (1936). 67 , Fragment f moro s e thae on n example; 2nd cent. B.C. /early 1st cent. A.D. Kingsdown Camp, Somerset. Archceologia, LXXX (1930), Twisted, looped rod, probably 85, fig. 7. from cauldron chain . A.Dc ; . 40-50. Over Fen, Cambs I (prob. - Cambridge Mus. of Arch, Complete example, no associa- ably in or near). and Ethn., no. Z.I 1494. tions. Over Fen, Cambs. II (prob- Cambridge Mus. of Arch, Complete example o associan , - near)r ablo n yi . Ethn.d an Z. . 11495,no . tions. Silchester, Hants. Archceologia, LVI (1899), Tw t leasto a fro e mon , Hous, eI 242, fig. .3 Insula XIX and another from Insula XXIII associated with three of Great Chesterford type. 4th cent. A.D. Stanfordbury, Beds. Fox, Arch. Camb. Reg., Chains suspended from iron tri- 100, pi. xvii; Arch. J., pod, in grave of mid-lst cent. i,xxxvn (1930), 260. A.D. THREE METAL-WORK HOARDS OF THE ROMAN PERIOD. 27 E. 10.. PLOUGHSHARE, iron, 7-8 ins. long, maximum breadth 4-7 ins., made of a 0-2-in. thick flat plate folded in at the upper end to form an oval socket 3-6 by 1-4 ins. to receive the thinned-down tip of the share-beam. (DWA 9.) This typ widf eo e flat shar bees eha n discusse Payn. Fy d.G b e (Arch., J. civ (1948), 82-111), and seems to be of Belgic and Roman date; the Eckford share is the widest of its type from Britain, though comparable examples come from Traprai Oxnad nan msouthern i n Scotland frod an m, Blackburn perhapd an 1 . 3 33a sB . B .Mill . no , E. 11. SOCKETED EEAFIMG-KMPE or SICKLE, iron, 9 ins. long, with upper end missing; split socket . 0-internain 6 l diamete t a maximumr . Blade 1-6 ins. wid t baseea . (DWA 11-12.) It is difficult to find close parallels for this knife with its distinctively long socketexampln a t bu , e from Newstead (Newstead, pi. probabl) Ixi10 , y represents the type. It is likely to be Roman rather than native in origin. E. 12. MASSIVE DOOR-BUHGE, iron. Overall length 7 ins., but broken off at end; perforation 1 in. in diameter. (DWA 21.) Thi presumabls si y Roman •worked iron an , a nn do L-shape d staple such as that from the Carlingwark hoard (C. 52). E. 13. SMALL PICK, iron, 7-3 ins. long overall, 1-8 ins. wide at centre, with oval flashaft-hole ton hammer-heads ha et I 0-9 0-. y 5b 8in , burred from use, narroe on d w an pick blade. (DW) A3. . SMAL14 . E L PICK, iron, 4-4 ins. long overall, 1-9 ins. wid t centrea e , with shaft-hole 0-7 . diameter5in . (DW) A4. Both these tools are representatives of common Roman types which were either masons minersr o ' ' tools (cf. Blackburn Mill 37). B , . Wheeler, Lydney (1932), 92, gives references to ironminers' picks (cf. Newstead, pi. Iviii, 5,12) nativA . e type, with characteristic narrow oval shaft-holes, occur n pre-Romai s n context e Continentth n o s ' (Dechelette . 879p , IV ,, fig. 607), in England (Bredon Hill, Arch. J., xcv (1938), 73-4), and (as a hammer t Carlingwarka ) . 42 . C , E. 15. ADZE-HAMMER, iron, 8-4 ins. long overall, hammer end 1-1 in. square, blade originally c. 2-5 ins. wide. Slightly oval shaft-hole, 1-2 by 0-9 ins. (DW) A5. This massive tool appearRomaa e b o snt type (cf. Collingwood, Arch. Rom. Brit. (1930), fig. 65, p, though with tubular shaft-hole). . MATTOCK16 . E , iron, 9-5 ins. long overall, both blade . 1-sc 4 ins. wide. Oval shaft-hole 1-7 by 1-0 ins. (DWA 6.) This axe-adze or mattock seems to be related to the Roman military dolabra (cf. Newstead, pi especially).1 mattoce Ivii. th no o ,t kd whican , h normally has a broad spade-like end (Collingwood, op. cit., fig. 65, q). It lacks, however e "clipse socketh ,th characteristio n s o "t f Romao c n shaft-hole axes, adze similad san r tools. . FARRIER'17 . E S "BUTTRESS," iron, 11-3 ins. long overall, 2-0 ins. wide t lowea r end t uppera , 1-. 1in . (DWA 13.) 28 PROCEEDINGS OF THE SOCIETY,- 1952-53.
An essentially Roman tool, commented on by Curie (P.S.A.S.,. LXVI (1931—2), 317), with references to examples from Silchester (Archceologia, LVII (1901), 248) and the Continent. The "buttress" is used for paring horses' hoofs.
E. 18. COLD CHISEL, iron, 6-8 ins. 'long, square-sectioned, 0-5 in. across. Head slightly burred. (DW) A8. Blacksmiths' chisels of this type are known from Continental sites of La Tene date (La Tene, Gross, La Tine (1887), pi. ix; Dechelette, iv, 873, fig. 601), but in this country the type is probably more likely to be Roman (cf. Newstead, pi. Ixiii. 7) ,
. 19E . FIELD ANVIL, iron, 5-9 ins. long, with octagonal head . 1-across1in . Remains of the rusted lateral ground-rests still in central perforation. (DW) 7. A mower'e Th s anvi s usei l d wit ha hamme r straighteninfo r g scythes, which, although known on the Continent in late La Tene sites (Dechelette, IV, 888, fig. 613), do not seem to have been introduced to this country until Roman times. (As also in Scandinavia: Steensberg, Anc. Harvesting Imps. (1943), 101; cf. scythe fragments from Carlingwark, C. 21-23 above). Roman examples of field anvils may be cited from Silchester, Newstead (Newstead, p. 285, pi. Ixii) and elsewhere.
. THII E CARLINGWABK FIND. C. 1. CAULDRON, bronze, 24 ins. diameter at mouth, 26 ins. at maximum bulge, 18 ins. deep. The vessel is made of three pieces of thin sheet bronze of whic e piecon h e form e lowe th ss bee rha npartd beateo an t p u n form a shallow carinated bowl 9-6 ins. high. To this two bands 8-5 ins. wide are riveted to form the upper part of the cauldron; these are each rather more than hal a circumferencf e vessel'th f o e s mouth d overlaan , p dia- metricall abouy yb ins8 t eacn .o h side, where the hele threyy ar d b e vertical rows of rivets. This provides double-thickness bronze at the points for handle-attachments e torth , n rivet-hole f whico s stiln e seenn hca b l O . analogy the rim would also have been strengthened by an iron hoop now missing! The rivets are small and round-headed, and the joints are of fine craftsmanship. (DW 1.) The vessel has been extensively repaired in antiquity by means of sheet- bronze patches riveted to the interior. Two styles of repair can be seen, (a) with small neat rivets simila thoso t r ee origina useth n di l jointsd an , (b) bent pieces of thin bronze strip pushed through and flattened in the manner omodera f n paper-clip . (cf.18)C , .nos17 . .C Cauldrons of this "Santon" type have been discussed by Hawkes in Aspects of Archaeology (1951), 172-99), who shows that this two-piece form goes back to the 1st century B.C. on the Continent (e.g. Emmendigen, Baden; Dechelette, Manuel, iv, fig. 636). The Austwick cauldron (P.P.S., in (1937), 164) laterf appearo e b , Romano st , date (Hawkes, loc. cit., 185-6). THREE METAL-WORK HOARDS OF THE ROMAN PERIOD. 29
EI9 E 16
Pig. 6. Iron objects from Eckford. (J.) 30 PROCEEDING E SOCIETYTH F O S , 1952-53.
CAULDRONS OF SANTON TYPE
Site. Reference. Comment.
Ballymoney, Co. Antrim. J.R.S.A.I., LIII (1923), Pound in a bog; 26-5 ins. dia- 25, fig. 13; P.R.I. A. (c), mete moutht ra , 19-75 ins. high. LI (1947), 59. Bewcastle, Cumberland. Tullie House Mus., Car- Foun n i Blacd k Moss, High lisle (no , 1947)51 . . Grains, about 1907. Rim diameter 24 ins., height 19J ins. Bog of Alien, TJrlingford, Co. As for Ballymoney. Foun boga ins4 n di 2 ; . diameter Kilkenny. at mouth, 18-75 ins. high. Carlingwark Loch. Present paper. Santon, Norfolk. P. Camb. A.S., xm (1909), Containing metal- work hoard ; 147; Arch. , xcvJ. i probably Boudicecan. 18£ins. (1939), 71 ff. diameter d ringan , -irom ri n handles. Silchester, Hants. Archceologia, LVII (1901), From Wel , Insul1 l a XXIII0 1 . 246. ins. diameter, beaten from single piece ; iron rim. Tarland, Aberdeenshire. P.8.A.8., XLIX (1914-5), Miniature version in cast bronze, 203.- with burial probably of 2nd cent. A.D.
C. TANKARD-HANDLE2 . , bronze, 4-5 iris, long 2 ins, . wid t a endse . Ring ornament at the extremities and on the central member presumably held enamel with, at the end, central pins holding the handle to a wooden body (DW 80.) Early Iron Age tankard-handles have been discussed by Corcoran (P.P.S., xvm (1952), 85-102). The Carlingwark handle (his no. 22) he relate tho dt e series t centurbeginnin1s e th y n B.Cgi . with that froe mth Aylesford cemetery t emphasisebu , s degeneratit s e quality e ringTh -. ornament, if it did in fact hold enamel, would relate it to such other North British pieces with bichrome enamel ring-and-dot ornament (e.g. a terret Middlebie inth e Plac e hoardth d e Felan ) l (Ullswater) bit. These presum- ably show Roman influence in the use of polychrome enamelling. C. MIRROR-HANDLE3 . , iron, 3-8 ins. long. Baluster-moulded with rectangular- section ring terminal 77.W )(D . Thi representativa s si Fox'f eo s Typ eI (Ba r Handles classificatios hi n )i n of mirror-handles (Arch. Camb. 1948, 24-44). Such bar-handles appear to derive from the Arras type of the 3rd century B.C., but Carlingwark need earliebo en r than mid-lst century A.D. C. TWO-LIN4 . K BIT ,rin irone gOn .missing t origina,bu l overall lengt 10-. hc 2 ins., surviving ring 2-8 ins. external diameter. (DW 76.) THREE METAL-WORK HOARDS OF THE ROMAN PERIOD. 31
Cl
Pig . Bronz7 . e cauldrons from Carlingward Blackburan ) 2 . B ; nk1 . Mil. Loc1) . (B l h(C 32 PROCEEDINGS OF THE SOCIETY, 1952-53.
. CTWO-LIN5 . K BIT, iron, fragmentary d witan ,h ring, crushe f shapeo t dou , fragmen 2-w 2-y 7b 0 no t ins. overall. (DW 74.) Fragment f similao s r bits occurre e Blackburth n di n Mill , hoar7 . d(B B. 8) and at Newstead (Nat. Mus. Ant., nos. FRA 420, 450, not published as such). Broadly similar bits occur in Roman contexts in Britain [e.g. Wroxeter, Corbet Anderson, Roman City f Uriconiumo (1867), pi. xii; Colliton Park, Dorchester (Dorchester Mus.); Silchester (Reading Mus.); Woodcuts, Dorset, probably in Roman context (Excav. Cran. Chase, I, pi. xxv, 3)]. But they are in contrast to the obviously Roman military types as at Newstead (Newstead, pi. Ixxi, nos. 1-3) or London (Wheeler, London Romann i Times (1930), 149), and e exactl n vie,th i f wo y similar bitt a s Tena L e (Gross Tenea L , (1887), pi. xii possibls , 12)i ,8 t i , e tha Britise th t h example realle sar y native types. Fragment similaf so r bits were founa n di Belgic contex t Maidea t n Castle (Maiden Castle, . 274-5)pp . Cf. iron bits t Llya n Cerrig, nos , 12885 .f whic,o approache5 8 h r typeou s . They survive int Dare oth k t AgeLagora s sa e (Lagore, . 103)p . C. 6. TIKE, iron, fragment, 4-3 ins. long, 1-5 ins. wide. Fox's Type A, unworn. notee Se s under Bckfor cfd .d an Blackbur , find 5 85o. . W E , )(D . n 3) Mil . l(B C. 7. LINCH-PIN (?), iron, fragmentary, 2-6 by 1-8 ins. It is suggested that this object may be the top of a linch-pin having a spatulate head and hook, in this instance wit hooe hth k hammere flan di t agains heade certaith tA . n example of this type occurs in the Blackburn Mill find (B. 4) and discussion wiU be found there. Cf. also the unlooped Eckford pin (E. 4). (DW 856.) C. LINCH-PI8 . N (?), iron, fragmentary, 1-y 1-b 76 ins. This hooked fragment may also be a linch-pin head of the type described under C. 7. (DW 63.) C. CLEA9 . r STAPLETo , iron, 4-y 1-b 34 ins. Thia large s i s r versioe th f no common Roman boot-cleat (e.g. foun1 d wit hobnail6 h3 t fee sa f buria o t l at Rotherley: Excav. Cran. Chase, II, 190), and may have been used for fastening 49.woodW ) (D . C. 10. CHAIN-JUNCTION (?), iron, broken. A small rectangular block of iron 0-9 by 0-9 by 0-6 in., decorated on each face with an incised cross. A slightly curved rod, broke f shortof n , springs asymmetrically froe mon 85c.edgeW (D ) . brokea e b Thiy n junction-piecsma r threefo e chains s suggestea , r dfo an object in the Blackburn Mill find (B. 18) discussed below. C. 11. HANDLE-LOOP, iron. Perforated figure-of-eight-shaped plate riveted to another with three arms having expanded circular ends with rivet-holes, 5-6 by 5-3 ins. overall. The outer face is roughly decorated with incised lines forming chevron pattern 67. arms.e W )th (D n so This is one of a pair of handle-loops from a metal or wooden cauldron, bucketr o probabls i b t I tu . yRomaa n type therd closa an ,s ei e parallel, decorate e samth en i d manner, from Corbridge (Corbridge Mus.)d an , another from Traprain, much corroded. C. 12. HANDLE-LOOP, iron, fragmentary. The upper part, with rivet, of a simila t slightlbu r y more massive object tha e aboventh , 3-y 1-b 45 ins. overall. (DW 66.) (Cf. another fragment (DW 85t) perhaps also of a similar handle.) THREE METAL-WORK HOARDE ROMATH F NSO PERIOD3 3 .
Fig . Object8 . iro f 3-C. so n(C bronz.d 14)an ) , wooe15 . d(C (C. 16-19) from Carlingwark Loch. (J.) VOL. LXXXVII. 34 PROCEEDING E SOCIETYTH F O S , 1952-53. C. 13. PADLOCK-SPRING, iron, fragmentary, 1-9 by 0-8 ins. overaU. (DW 85a.) sprine Th g padloc commoa s kwa n Roman type (cf. Pitt-Rivers, Prim. Locks and Keys (1883); Verulamium, pi. Ixv, 16). C. 14. LATCH-LIFTER, iron, contorted, now 8-1 by.1-3 ins. overall. A rod of iron, flattene tipe flatteneth s ,i t da d int ostria p 1-3 ins.s ins wid5 it f r .o efo length, which is folded over to hold a D-shaped loop. The whole object has been folded upon itself. (DW 48.) This appearlatch-liftea e b o st commof o r n Early typeIroe nAg . For list of examples known to 1917 see Glastonbury, II, 375; many more could now be added. C. 15. HANDLE (?), wood, in two fragments, 2-0 and 0-8 ins. long, 1-0 in. maximum width smalA . l woode f ovao r l sectionnba e endth , s grooved and widened beyond the grooves. The original illustration in P.S.A.S., vii (1866—8), pi. i, 17, shows that a further portion existed beyond one widened end, but does not suggest its purpose. Perhaps a handle of some object. 78.W )(D . BOWLC16 . , bronze, fragmentary, 4-4 ins. diamete t . mout2-a rc 9 d inshan . t bee no bowe high ns Th hammereha l. f sheeo t t s dou casi meta t t bu l or spun. Type and technique of manufacture indicate a Roman date (cf. Hawkes, Aspects of Arch.,) 3. 185)W (D . C. 17. BRONZE SHEET, fragment, with repairs by patching and riveting, 2-8 by 1-2 ins) .6. overall W (D . . BRONZC18 . E SHEET, fragment, with repairperiodo tw f patchiny so b s d gan riveting, 4-3 by 3-0 ins. overall. (DW 7.) Both these fragments are probably from cauldrons of the type of C. 1 above. C. 19. FRAGMENT OF BRONZE CASKET-MOUNTING (?), folded, but originally (as shown in drawing) 4-8 by 3-8 ins. overall. This is a piece of bronze sheet with three rivet- or nail-holes and half a concentric circular design embossed at the broken end. (DW 16.) This seem a piecs likele f sheet-bronzb o e o yt e ornamental mounting fo woodea r n box Romaf ,o n typet Richborouga s a , Woodcutd han s (both centurd 3r y A.D.; Richborough, 142, IV ; Excav. Cran. Chase, , pisI . xix, xx). . SAWC20 . , single-edged, iron, 6-5 ins. tooe long brokes Th i le point .th t na , 4-d 0bladan e insth f .eo survives, with backward-pointin gedgee teeton .n ho The tang is 0-9 in. wide, with two rivet-holes. (DW 17.) The well-known iron saw with its original wooden handle from Glaston- bury is an accurate parallel [Glastonbury, II, pi. Ix, p. 371, with/other examples quoted. Cf. also Bredon (Arch. J., xcv (1938), 78; Maiden Castle, p. 274; Hunsbury (Arch. J., xcni (1936), 66)]Scotlann I . d ther fragmente ear f o s a large saw from the Lochlee Crannog (Munro, Anc. Scot. Lake-Dwellings (1882), 87), and a double-edged saw from Newstead (with original bone handle) is comparable (Newstead, pi. Ixviii, 6). . SCYTHE-BLADC21 . E FRAGMENT, iron, 4-6 ins. lon y 1-b g5 ins. maximum 29.widthW )(D . THREE METAL-WORK HOARDE ROMATH F NSO PERIOD5 3 . C. 22,23.. PEOBABLE SCYTHE-BLADE FRAGMENTS t illustrated 32., no , )30 W (D . All these appea fragmente b o rt largf o s e knife r scythe-bladeso - 1 2 . C ; has been cut at both ends with a cold chisel, and has a heavy thickened back and media l Newsteae ridgth s ea d scythes (Newstead, pi. scythe Ixii)Th .e appear havo st e bee nRomaa n introductio nativo nt e econom Britain yi n as in Scandinavia (Steensberg, Anc. Harvesting Imps. (1943), 101; cf. notes on field-anvil from Bckford, above)B9 1 . . . KNIFE-BLADEC24 . , iron, single-edged, 4-0 ins. lon y 1-gb 4 ins. wide, broken shor handlef 31.o t W ) (D . . KNIFE-BLADEC25 . , iron, single-edged, 7-1 ins. lon y 1-gb 3 ins. wide, broken near handle 28.W ) (D . Both these knives, wit hstraigha t bac curved kan d cutting edge, differ from the curved, flame-shaped knives as at Grlastonbury and surviving at Newstead (Newstead, theo d yr resemblpi No , 13).7 Ix., 2 e ,Roma eth n form (ibid., knifA , 14)5 e . from Lochlee (Anc. Scot. Lake-Dwellings, fig. 129) approximates more nearly in outline. C. 26. TIP OF SWORD-BLADE, iron, 6-9 by 1-4 ins. maximum width. (DW 25.) SWOED-BLADEF O P TI . C27 . , iron, 5-1-y 8b 3 ins. maximum width 23.W ) (D . SWOED-BLADEF O P TI . C28 . , iron, 6-0 ins. ins2 lon . y b gmaximu m width. (DW 18.) C. 29. TIP OF SWOED-BLADE, iron, 5-2 ins. long by 1-7 ins. maximum width. (DW 24.) C. 30-33. TIPS OF SWORD-BLADES, not illustrated. (DW 19, 22, 21, 20.) All these fragments are from the tips of swords, and no blades or hilts are represented, suggesting deliberate selection and not casually accumulated scrap metal. Nonmid-rib a narrowe e s th eha d ,an r example certainle sar y likely to be from native rather than Roman weapons (cf. Newstead, pp. 183 ff.), e wideth t r t bu resemblexampleno o d ) e s28 Roma . (e.g.C n Legionary types, which have mid-ribs. Comparison wit e widehth r typIroe nAg e (Piggott, P.P.S., XVI e possibl(1950) madee th b d y ,an y , 1-28ma com) - parable fragments from Lochlee (Anc. Scot. Lake-Dwelli'ngs, p. 125) may be noted bese Th t . parallels (without mid-rib witd e saman shth e rangf eo width as the Carlingwark fragments) are the Llyn Cerrig swords (Llyn Cerrig, pi. xxxiv), suggesting tha Carlingware th t k serie probable sar y native. . FERRULEC34 . , iron, 2-8 ins. 85gr.longW (D ) . . FERRULEC35 . , iroins7 . n 1- 84. longW ) (D . These indeterminate form probable ar s Romaf yo n origin. C. 36. LOOP or BUCKLE, iron, 2-7 by 2-4 ins. (DW 85/.) This D-shaped loo probabls pi Romaf yo n manufacture. C. 37. SICKLE, iron, tanged, 5-9 ins. long, broken. (DW 35.) This is a "balanced" sickle of well-known Early Iron Age type (Llyn r comment)fo , Cerrig,86 t . appearinp bu , g als Roman oi n contexts (e.g. Newstead, e Blackburth . Cf pi . .n 5) Ixi Mil, below2 ,4 3 l . sickleB .. no , 36 PROCEEDING E SOCIETYTH F O S , 1952-53.
Pig. 9. Iron objects from Carlingwark Loch * THREE METAL-WORK HOARDS OF THE ROMAN PERIOD. 37 . CZKCHTLAC38 . R HOUNTMG, iron, broken t originallbu , 4 ins . c y. internal diameter os i f thit I .n metal plate with remain t leasa f naiso 3 t rivetr o l - holes, and has decorative nicking on the outer edge. (DW 85h.) This may be the rim of a circular shield-boss of Roman type, as that in bronze t similarlbu , y nicked, from Copthall Court, London. (Wheeler, Lond. in Rom. Times (1930), 31, fig. 3.) C. 39. TANGED BLADE, iron, double-edged, 3-2 ins. long. (DW. 26.) Such small blades might be for veterinary or surgical purposes and are likelRomane b o yt . . TANGEC40 . D BLADE, iron, double-edged t illustratedno , 27.)W (D . . HAMMER-HEADC42 . , iron, double-ended, 7-3 ins. lon y 1-gb 3 ins. maximum widt t burreha de hamme bode endsth f Th slia yo . s t narrows i r ha d an , as the opening for the haft. (DW 8.) This is a version of the Early Iron Age hammer-heads discussed in con- nexion with those from Bredon Hill (Arch. (1938)v J.,xc , 73), with character- istic oblong hafting-slit. . HAMMER-HEADC43 . , iron, double-ended, 6-6 ins. long, 1-8 ins. wid t centreea , with circular shaft-hole. (DW 9.) . HAMMER-HEADC44 . , iron, broken, 2-4 ins. lon y 2-gb 3 ins. maximum width t shaft-holea 12.W )(D . C. 45, 46. HAfflfflER-HEADS, iron. Not illustrated. (DW 10, 11.) C. 47-49. HAMMER-HEADS, iron, fragments. Not illustrated. (DW 13, 14, 15.) All these hammers are of Roman type, with markedly expanded centre and circular shaft-hole (Newstead, pi. Ixiii, 3, 5, 11). C. 50. ADZE-HAMMER, iron, 6^8 ins. overall, blade 1-8 ins. wide, tubular shaft- hole of oval plan and heavily burred hammer-butt. Side-clips or projections on upper edge of shaft-hole. (DW 4.) This is an adze-hammer of a type regarded by Collingwood as typically Roman (Arch. Rom. Brit. (1930), 269; fig. 65, p.). It does not occur at Newstead, but is known from Richborough (Richborough, IV, 154), and a similar tool was in the Bckford find (E. 15 above). C. 51. AXE-HEAD, iron, 4-8 ins. long, with squared butt and side-clips at upper and lower edges of shaft-hole. (DW 5.) A typical Roman form (cf. Newstead, pi. Ixi, 1, 4). C. 52. HINGE-STAPLE, iron, 6-1 ins. overall. (DW. 47.) This L-shaped staple presumably supported an iron hinge similar to Eckfore thath n i t d above)2 hoar1 . d(E . C. 53. STAPLE (or HANDLE), iron, 4-8 by 2-1 ins. overall. The top edge is nicked, perhaps with the intention of ornament. (DW 61a.) . STAPLEC54 . , iron, 4-1-y 2b 7 ins. overall 61.W )(D . Both are probably of Roman origin (cf. no. C. 9). C. 55. HOOK, iron, 4-7 by 2-5 ins. overall. (DW 54.) 38 PROCEEDING E SOCIETYTH F O S , 1951-53.
. HOOKC56 . , iron, 3-1-y 9b 2 ins. overall, with adherent corroded chain-mail (cf. C. 74). (DW 59.), C. 57. HOOK, iron, 5-5 by 1-5 ins. overall, with flattened perforated head. (DW 60.) . HOOKC58 . , iron, 3-2-y 3b 3 ins. overall 55.W )(D . C. 59. HOOK, bronze, 3-2 by 1-3 ins. overall. (DW 81.) l presumablal Thes e ar e y Romae meat-hookb d somy nan ma e s (cf. Newstead, p. 288). The possibility of large fish-hooks for some must be considered, however (cf. Richborough, iv, 155, and pi. Ixii, 348). C. 60. DRAW-KNIFE, iron, broken, 8-8 ins. long, blade 1-4 ins. wide. (DW 36.) Paring-knives or draw-knives of this type are acknowledged Roman types: cf. Silchester (Archceologia, Liv (1894), 139 ff. and fig. 17). . LOOPC61 . , iron, 3-y 1-0b 7 ins. overall 68.W )(D .
C. 62. LOOP, iron, 3-4 by 1- 2% ins. overall. (DW 71.) . LOOPC63 . , iron, 3-2-y 5b 5 ins. overall 69.W )(D . Such loops, with the pointed ends beaten out at right angles, are presum- abl attachmenr yfo wooo t hav d dan e good Roman parallels from Newstead (pi. Ixvii, 10, 11, 13). There is one in the Blackburn Mill find (B. 24 below). . PUNCHC64 . , iron, 4-5 ins .41. W long)(D . . PUNCHC65 . , iron, broken t illustratedNo .82. W )(D . . PUNCHC66 . , iron, 4-4 ins 40.. longW )(D . C. 67. PUNCH, iron, 5-3 ins. long. (DW 42.) Thes simple ear e blacksmiths' tools (cf. Newstead, pi. Ixvi , 17)16 , . . FILEC68 . , iron, broken, 3-8 ins. lon 46a. 0-y . widegW b 3 in (D ) . C. 69. FILE, iron, 9-5 ins. long by 0-9 in. wide. (DW 46.) Files are known in British and Continental Early Iron Age contexts (Glastonbury, n, 374; Dechelette, iv, 881, fig. 608), but indistinguishable tools occu Roman o r n sites (Newstead, pi. lix. 5) , . BABC70 . , iron, rectangular section 4 ins1 , . long, 1-5 ins. maximum width, tapering at one end. (DW 83.) This may be a rough-out for a cold chisel or blacksmith's scrap. Cf. e Tiefenae barth th n si u find (Jahrb. Schwiez. Ges. Urgesch., (1929)I XX , fig. 3). C. 71. GRIDIRON, iron, broken, originally c. 6-7 by 6-0 ins. overall, 1-5 ins. high. A common Roman type (cf. Newstead, 274, pi. liii, 2). (DW 86.) . TRIPODC73 . , iron, 4-0 ins. diamete 3-d 5ran ins 53.. highW )(D . This agai Romaa ns i n typ Curis ea e pointe t (P.S.A.S.,dou LXVI (1931-2), 311-3). C. 74. CHAIN-MAIL, iron, fragments (DW 2.) Pre-Roman finds of chain-mail are known on the Continent and in Britain, e.g. Tiefenau, Berne (Jahrb. Schwiez. Ges. Urg., xxi (1929), fig , .10)4 ; Hjortspring, Denmark (Rosenberg, Hjortspringfundet (1937), 47); Lexden, THREE METAL-WORK HOARDS OF THE ROMAN PERIOD. 39
. PigObject10 . irof bronzd so nan 59). e(C , Carlingwark Loch. (J.) 40 PROCEEDING E SOCIETYTH F O S , 1952-53. Essex (Archceologia, LXXVI (1927), 248); Stanwick, Yorks (Brit. Mus. E.I.A. Guide (1925), 142). Diodorus Siculus refers to Gaulish chain-mail (v, 30), and Varro attributes its invention to the Gauls (De ling, lat., V, 24, 116), while a Gallo-Roman sculpture from Vacheres, Basses-Alpes, shows a mail-clad chieftain (Dechelette, iv, 661; Lantier and Hubert, Les Origines de VArt Franqais (1947) s alsi , figt oi characteristicall.t 85)Bu . y Roman, and foun instancr dfo t Newsteaea d (Newstead, . 161) p "e Belgo-Romanth : " fragment from Maiden Castle could conceivably be regarded as native (Maiden Castle, p. 284). For the technique of manufacture, cf. Burgess in Ant. J., xxxill (1953), 48. Further corroded fragments at Carlingwark adhere to a hook (C. 56). A technical report on the Carlingwark mail by Mr Burges gives Appendin si a n i . 50)x(p . C. t 75-10illustratedmann i no e e yar 3 ar instance d an , s nondescrip r brokeo t n pieces of iron. Some small BARS (C. 76, 77, 79, 80, 98) may be pieces of punches or chisels; possible fragments of HINGES (C. 84, 85); two HOOKS of the type of C. 55-59 (C. 86, 91); two small HANDLES (C. 88, 89) to drawers, etc.; thre, o 94)93 tw e;, . 61-6LOOP92 e typC . th f (C 3f eo o S smala probabl d lan RINidentified) e b 100. e97 Gn , AWL(C )ca 96 . S.(C
III. THE BLACKBURN MILL (COCKBTJRNSPATH) FIND.
B. CAULDRON1 . , bronze, globular, 152 ins. diamete t mouthra ins8 t 1 bulge,a . , 10 ins. vessehigh e thif o Th ns . i l iro n beatea m nri d ha n s bronzeha d ; an , now vanished handleo tw d s an ,hel three-rivetey db d attachmentss ha t I . been extensively repaired in antiquity, the whole central area of the base being replaced by patches riveted on. (DW 87.) B. 2. CAULDRON, bronze, globular, 12-4 ins. diameter at mouth, 14 ins. at bulge, 8-5 ins. vesse e highf spuo Th s n.i l bronze wit ha smal l plugged holn ei handleo tw d s an iroheln a m three y ndri d b ha -s centr basee ha e th d th f ,an e o rivet attachments. repairs o Ther88.n W )e e(D ar . Globular cauldron thesf so e types have been discussex listed Fo dan y db (Llyn d Hawke an Cerrig,) 88 s . (Aspectsp 1 . B r o fff.)2 Ou Arch.,17 . . pp represents the pre-Roman type, heavily repaired and evidently old when the hoard was deposited (cf. the Carlingwark cauldron, C. 1 above), and B. 2 the same form made by the Roman technique of spinning on a metal- workers' lathe, incorporateand finthe d in dbefor it needee d patching below)4 1 - (cf. . B . CAULDRONS OF BATTERSEA TYPE.
Site. Reference. Comment.
Abercairney, Perthshire. Perth Mus. Rim-band missing; 27£ ins. max. diameter; patternin f ovao g l hammer punch-marksr -o . Battersea, London. Llyn Cerrig,. 88 Iron band on rim; 14-75 ins. diameter.
Blackburn Mill. Present paper. THREE METAL-WORK HOARDS OF THE ROMAN PERIOD. 41
Cauldrons of Batter sea Type—continued.
Site. Reference. Comment.
Bwartly Shank, Amham Blackgate Mus., New- Rim-band missing, c. 24 ins. max. Moor, Northumberland. castle-on-Tyne. diameter. Ipswich, Suffolk. Llyn Cerrig,. 88 Fragments. Kyleakin, Skye. P.S.A.S., xix, (1884-5), Rim-band missing; 18 ins. dia- 309. meter; found . 7deeft £ n i p peat near keg bog-butterf so . Llyn Oerrig Bach, Anglesey. Llyn Cerrig,. 88 Fragments. Walthamstow, Essex. Ibid. Similar typd dimensionan e s a s Battersea cauldron. Whitehills Moss, Lochmaben, P.S.A.S., xxrv (1889-90), Run-band missing, 15 ins. max. Dumfriesshire. 16. diameter; pattern on base as on Abercairney cauldron.
B. 3. MRE, iron, four fragments, of which the largest is 12-2 ins. long, 1-1 in. wide 0-1d . thickan 5in . (D a-d.)W6 9 B. 4. LSNCH-PUiT, iron, 5-4 ins. long, with spatulate head 2-0 ins. across, and hooke 123.dW top(D ) . This type of linch-pin was recognised by Ward-Perkins as related to his crescent-headed Romano-Belgic type (Ant. J., xxi (1941), 67), but its distribution and associations have never been studied in the detail they deserve. Example knowsare n fro Continenmthe t [e.g. Saalburg (Jacobi, pi. xxxxii, 1, 7) and Zugmantel (O.R.L., xxxn (1909), pi. xv and fig. 44)], and from southern England e Northth n I , . examples -ma notee yb d from Newstead (pi. Ixx, 1, 3, 6), Mumrills (P.S.A.S.,.l3aii (1928-9), 561), and The Law f Monifieto s n Angui h s (Nat. Mus. Ant. n Scotlandi ) ; Corbridge (Corbridge Mus. d Chester)an s (Chesters Mus.) Broughd an ; , E.R., Yorks (Trans. E. Riding Ant. Soc., xxvm (1939), pi. iii, fig. 10, 8). The spatulate- headed probablt s forEckfori a ) s m4 a . d(E y related. B. 5. TBRRET, iron, with ring 2-3 ins. external diameter, broken at base. (DW Ilia). The ring-terret set on a base with spreading wings is a Roman type not uncommo Britainnin . Ther mora eis e elaborate exampl bronzein e from Newstead (pi. Ixxv, 12) (cf. Woodcuts, Excav. Gran. Chase, I (1887), pi. xvii, 6, and xviii, 14 (? part of one object), with other examples cited on p. 60). B. 6. CAP of LINCH-PIN, bronze, one of a pair, 1-9 ins. high and 1-3 ins. diameter t topa , with sub-classical mouldings , 93.92 ) W (D . The significance of these caps, appreciated by James Curie, can now be taken further. They represen a nativt e typIroe neAg derivative th f eo Arras (Yorks) type (Ward Perkin Ant.n si J., XX (1940), 358; xxi (1941), 64 , supplemented by Llyn Cerrig, p. 78; Camulodunum, pp. 329-31). The Belgic types as at Westhall and Colchester come nearest to the Blackburn" Mill pair. They are mid-lst century A.D. (cf. Llyn Cerrig, 20). 42 PROCEEDING E SOCIETYTH F O S , 1952-53.
. LINK7 . B BITf So , iron, broken, 4-2 ins. overall. (DW 111.) . LIN8 . BITf B Ko , iron, broken, 1-8 ins 149.. longW ) (D . r commenFo iron to n bit thif so s type undee ,se Carlingwar e rth k . find5 . C , B. 9. KEY, iron, broken at top end, 4-2 by 1-8 ins. (DW 117.) This is an L-shaped slide-key of Wheeler's classification (London in Roman Times (1930), 69; pi exacr . fo xxx 3 t, parallel ,"well-knowA a s i d )an n Roman type (cf. Newstead, pi. Ixxviii. 2) , 1 , B. 10. HANDLE, iron, semicircular with hooked ends, 9-0 ins. across the chord. (DW 122.) . HANDLE11 . B , iron, with tanged endd centraan s l loop, 6-2 ins. overall. (DW 153.) B. 12. HANDLE, iron, twisted with hooked ends, 5-2. ins. overall. (DW 145.) . HANDLE13 . B , iron, twisted, fragment t illustratedno , 145.W ) (D . Thes hav y likel Romae e eb ar ema yo belonged t nan tubsbronzo do t t r ,o e camp-kettle Romaf so n military type (cf. Newstead, pi. liii). They survive into Dark Ages contexts (e.g. Lagore, . 114)p . B. 14. CAULDRON-RIM, bronze, with iron edge, broken, 4-5 ins. long, with part of handle-attachment. (DW. 148a.) This is probably a fragment of an iron-rimmed cauldron of Santon or Carlingwark type, sinc side eth e mannee doe t bulgth sno n i er m belori e wth of the globular cauldrons B. 1 and B. 2. B. 15, 16. PAIR OF TUB-HANDLES, iron, consisting each of a T-shaped looped attachment, 6-5 ins. in each dimension, "with nails for fastening to a wooden vessel some 0-7 . thickrina 5 in d g an 4-, 6 ins. externald diametean e on n ri 4-3 ins. in the other. (DW 94, 95.) These roughly-made handles difficule ar parallelo t t t arebu , likele b o yt of native workmanship. . PAR17 . f CAULDRON-CHAINTo B , ins0 iron4 . c lon, g overall. e Thith s i s central member, designed to make an inverted Y-junction for the three component chains of the apparatus. It consists of a rectangular block 1-5 by 0-9 by 0-7 ins., from which spring two pairs of hooks and two pairs of decorative spirals, both of bent rods 0-3 and 0-2 in. dia'meter respectively, and at right angles to one another. The hooks are curved round in almost complete circles and their ends are capped with pine-cone-shaped knobs. The upper pair carr ydoubla e loop formin lowese gth sucseriea 5 1 f o thf so links making up a chain: the links are made of a 0-2 in. diameter rod. The lower pair would originally have held the two divergent chains ending in hooks for attachment to the cauldron-handles. The decorative spiral twists form a symmetrical rectangular pattern 4-0 by 3-5 ins. (DW 97.) While unique in itself, this object is recognisably a member of the Type II (Great Chesterford) series of cauldron-chains (above, under E. 6-9). These chains do not employ rods except for the hooked terminals of the two arms, and are composed mainly of links which may be doubled (as in the Blackburn Mill example), or pinched into figure-of-eight form (cf. Llyn Cerrig, pp. 38, 85 on such links). The Blackburn Mill type of link is found Greae onth t Chesterfor Type th d en I chai o chai d nn an fro mI I Oven Fe r (above, under E. 6-9). It appears again on gang-chains of Early Iron Age Lagore typth n ei e crannog (Lagore, Romae th . 115-7) n i pp e nd forth ,an f o t THREE METAL-WORK HOARDS OF THE ROMAN PERIOD. 43
Pig. Object11 . s from Blackburn Mill (all iron excep , bronze6 4 . 1 tB . B ; bronze and iron). (J.) 44 PROCEEDINGS OF THE SOCIETY, 1952-53. Saalburg (Jacobi, pi. Ixvii uppee , Th 10)r .par t comprise largsa e suspension ring, below whic a decorativ s hi e open-work featur f twisteo e d rodd an s hooks. The Y-junction may be elaborated, as at Great Chesterford, where elaboratn a s i t i e four-looped devic rodf eo s intertwine reef-knoa n di d an t havin gpaia spiraf o r l terminals double Th . e link identicale e sar th d an , uppermos e Typt th par I f chainI eo t s formei elaboratsn a y b d e feature incorporatin blocko gtw s from eac f whicho h spring four knobbed hooks of Blackburn Mil e Greal th type n tI Chesterfor. d chain itself, spirale ar s incorporated in the elaborated Y-junction-piece (fig. 2). For a similar use of spirals in iron, cf. the Lochlee flesh-hook (Anc. Scot. Lake-Dwellings, fig. 139), earl centurd y2n y versioA.D A e Typ.th I f cauldron-chaieI no n was foun t Newsteada Agricolan a n di n context (Newstead, pi. Ixiv, 3; from Pit XVI). A fragment, possibly of a hooked block from a Type II chain, came from Carlingwark (C. 10; cf. B. 18 below).
CAULDRON-CHAINS OF TYPE II (GREAT CHESTERFORD TYPE).
Site. Reference. Comment.
Bartlow, Cambs. Arch. , xiJ. n (1856). 5 , Found with "Roman remains," 1854. Blackburn Mill, Berwick- Present paper (B. 17). shire. Oarlingwark Loch. Present pape 10). (C .r Probably broken fragmenf o t hooked block. Cirencester, Glos. Corinium Mus., Ciren- Complete chain very similao t r cester. Great Chesterford. Presum- ably Roman in date. Fenhouse Farm, Brandon, Cambridge Mus f Archo . , Fragments of chain and hooks of Norfolk. and Ethn., unpublished. Great Chesterford type. Great Chesterford, Essex. Arch. J., xin (1856)- 1 , a largPar f o te ironwork hoard 13; pi. 3, no. 32. of Roman date. Ickleton, Cambs. Arch. J., xm (1856). 5 , Found in a Roman building, 1848. Newstead, Roxburghshire. Newstead, pi. Ixiv, 3. Par f chaino t d Y -junctionan s . Fro t XVImPi : Agricolan. Rotherley, Wilts. Excav. Cran. Chase, n Probable example; fragmenf o t (1888). 7 , , picv . Y-junctio d figure-of-eighan n t chain. Foun n bottoo d f mo main circle ditch and so early 1st cent. A.D. and pre-Roman (Hawkes in Arch. J., civ (1948), 36-42). Silchester, Hants. Archceologia, LVII (1901), Fragments of at least three 246. chains fro Insule mth a XXIII hoard, with one of Type I. Roman t beforno , . A.Dec . 100. THREE METAL-WORK HOARDE ROMATH F NSO PERIOD5 4 . . CHAIN-JUNCTIOB18 . N (?), iron, broken. This consist a bloc f o sk originally 0-9 in. square with rough crosses incised on two faces, with remains of two divergent rods springing from one edge. (DW 98.) B. 19. KNIFE KOTOH-OUT, iron, tanged, 5-2 ins. long. (DW 144d.) This appears to be a blank for a knife, with the edges not yet hammered or ground. HIPPO-SANDA20. B. L FRAGMENT, iron, with hook, 4-4-by 00 ins. overall. (DW 115.) B. 21. HIPPO-SANDAL, broken, 6-0 by 4-5 ins. overall. (DW 114.) These both represent objects characteristi f Romao c n Gaud an l Britain, discussed by Wheeler, Verulamium, 220-1. His pi. Ixiii, 3, 4, show very simila wit1 2 r . hexample B centra r ou o lt s openin d side-hooksgan ; latese thesth e t ar etype , date centurh t Verulamiuda 5t r o yh A.D4t mo t . The Blackburn Mill hippo-sandal is furnished with spikes on the bottom in the manner of certain Continental examples—e.g. from Dieppe and Valen- tigney (Rev. Muse.ess de (1928 ,e crampon 141)th . e Continentacf th ; f o s l Iron. Age (Dechelette, IV, 899)—presumably to give the horse a surer footing under ice-bound conditions. B. 22. TWISTED ROD, iron, broken, with expanded and flattened end curved clio t pf i soms a n ei object, 5-0 ins. overall155.W (D ) . B. 23. ROD, iron, broken, with flattened end bent over to form a socket, 5-0 ins. long. (DW 154.) affinitied an e f thes us o srelate o e etw Th d object s unknowni s . They some b y e forma elongatef mo d ox-goads. . LOOP24 . B , iron, 2-6 ins. overall 126.W (D ) . A Roman t typCarlingwara s ea 61-63. k(C , 92-94). B. 25. EYED HOOK, iron, 2-8 ins. overall. (DW 151.) Probably a Roman type. . LAMPSTAND26 . B , iron, 7-5 ins 99.. highW RomaA )(D . n type (cf. Newstead, pi. Ixxix). . STAPL27 . B r CLEAo E T (?), iron, broken, 2-8 ins. overall 136.W (D ) . A Roman type (cf. Carlingwark, C. 9 above.) . HINGE-STAPL28 . B E (?), iron, 8-9 ins. overall 157.W (D ) . A Roman type (cf. Carlingwark, C. 32 above). B. 29. SHEARS, iron, broken, 7-9 ins. long, with blade 0-6 in. wide. (DW 128.) This half-pai f shearo r s represent typa s e commo e Continentath n ni l Early Iron Age (Dechelette, IV, 789), but in Britain there seem few certainly pre-Roman examples except those from Fifield Bavant, Wilts. (Wilts. A.M., XLII (1922-4) ,paie 482)th r d fro,an m Colcheste Periof ro . A.D (c dI . 10-43) (Camulodunum, 343 . Specimen7) , , picv . Romano-Britisn si h contexte sar frequent: e.g. Woodcuts (Pitt-Rivers, Excav. Cran. Chase, , piI . xxii8), 5 ;, 1 , Richborough (Richborough, II, 51, pi. xxiv, 67; IV, 154, pi. Ix, 337); New- stead (pi. Ixviii, 5); Traprain (P.S.A.S., LIV, fig, 3);83 ,.15 . Lochlee (Anc. Scot. Lake-Dwellings, fig. 138) sheare Th . s fro Ashgrove mth e Loch crannog, 46 PROCEEDING E SOCIETYTH F O S , 1952-53.
Fig. 12. Iron objects from Blackburn Mill (B. 34 with wooden handle). (J.) THREE METAL-WORK HOARDE ROMATH F NSO PERIOD7 4 .
Stevenston undatede ar , could an ,Dare db k Age r lateso r (Arch. Coll. Ayr, and Gall., vn, 56-61; cf. Lagore, 112; Ward Perkins, Medieval Cat. London Mus. (1940), 153). . SHIELD-BOS30 . B S (?), iron, fragment, 6-3-y 5b 0 ins. overall, with remainf o s circular flange wit survivine hon g rive r nail-holo t beginnind ean domef go . (DW 148.) This appears to be part of a circular iron shield-boss of a type known in late La Tene Europe (Dechelette, iv, 678; Germania xxx (1952), 334; Bayerische Vorgeschichtsblatter, x (1954)x , 45)probabld an , y representey db e curiouth s iron bosses, some with spikes, from Hunsbury (Arch. J., XCIII (1936), 67, no. 26). Bronze circular bosses are of course well known (Fox, Arch. Camb. 1945, 200) t sucBu .h iron bosse alse sar o Romaa n type (cf. Newstead, p. 181, pi. xxxiv, 3; Verulamium, p. 219, pi. Ixiv, 3, 5), and the Blackburn Mill specimen is as likely of Roman as of native origin. B. 31. PLOUGHSHARE, iron, broken, 6-2 by 4 ins. overall. (DW 113.) Ploughshare thif so s type have been discusse Payny db e (Arch. , ciJ. v (1948), 82) and associated with Belgic and Romano-Belgic agriculture. Cf. comment on the Eckford share E. 10. Other comparable examples come from Traprain and Oxnam in Scotland; Bigbury, Silchester, London and Box (Wilts. southern )i n England. The "other plough fittings" in'the Blackburn Mill hoard, referred to by Payne (loc. cit., 99) are the hippo-sandals, our nos. B. 20, 21, as their Museum catalogue numbers quote footnots hi n di show 6 e5 . B. 32. BUTCHER'S KNIFE, iron, socketed and with triangular blade, 7-1 ins. long. (DW 106.) A common Roman type with straight back, triangular blade, and right- angled junctio o handlent . Newstead,Cf . pi. Ix; Verulamium, pi. Ixiv (sockete tanged dan d forms). B. 33. PEAT-SPADE BLADE, iron, with folded edges to form socket and foot- benresw , 7-no tin t3 ins. long, breadt f bladho e 4-5 ins., foot-rest (broken) 3-3 ins. long, socket 3-2 by 1-2 ins. (DW 108.) There seem identitdoube o sn th f o tthi f yo sblade objecta th f s ei o t i ; turf-spade or slane of a recent type well known in Scotland and Northern Ireland (cf. Colin Sinclair, Thatched Houses of the Highlands (1953), 73; Evans, Irish Heritage, . 136-8pp , figs. 86-89). Genericall t ybeari clossa e resemblanc e Eckforth o et d ploughshar . 10)(E e , wit e additiohth e th f no foot-rest. Its presence in a Romano-British context is surprising and one suspects intrusion. B. 33a (DW 109) is a less complete example of similar type, lacking the foot-rest and in its corroded state resembling the plough-share E. 10 from Eckford. If contemporary with the rest of the hoard, it constitute earliese sth t evidenc peat-cuttinf eo Britainn gi . Peauses wa td s fuea l from Neolithic time Scotlandn i s r evidencfo ; f Iropeate eo nAg - cutting in Denmark, see Becker in Fra Nationalmuseets Arbejdsmark (Copenhagen), 1948, 92-100. B. 34. SICKLE, iron, crescentic blade 1-1 in. wide, with flat tang 2-0 ins. wide, held by two rivets to wooden handle, 4-75 ins. long, slotted to receive it. The handl carve s ei fore phallusa 107.th f m W o n d i (D ) . 48 PROCEEDINGS OF THE SOCIETY, 1952-53.
A "balanced" sickle comparable to the tanged example from Carling- phallie war 37). Th k .(C c handle forth f mo e implie sRomaa n origid nan the association of Priapus with the sickle and harvesting. Naturalistic phallic representation t characteristino e ar s e Earlth f yo c Iroe (cf.nAg Hawkes, Ant. J., xxvm (1948), 166-9). B. 35. FLAT BAE, iron, 4-6 ins. long and 6-4 ins. wide. (DW 144/.) Probably a blank or rough-out for a small knife (?) (cf. B. 19 and B. 36). B. 36. ROUGH-OUT FOR KNIFE or RAKE-PRONG (?), iron, 5-7 ins. long and 0-. maximu5in m width144e.W (D ). This was ingeniously interpreted by Curie as a rake-prong, as Newstead, blan a single-edge a e r b kfo y ma t i dt piknifeBu . Ixiibid.,s . a , 7 , pi. Ix2 , (cf. B. 19). B. 37. SMALL PICK, iron, 6-4 ins. long, 1-3 ins. maximum width, with large oval shaft-hole 102.W )(D . A Roman type as at Bckford (E. 13, 14). B. 38. FIELD-ANVIL, iron, 7-3 ins. long, 1-2 ins. diameter at head. (DW 101.) Another Roman type, also represente t Eckforda . 19)d(E . . ADZE-BLADE39 . B , iron, 7-2 ins. overall, cutting edge broken, socket 1-y 5b 100.W )(D 1- . 1in Adzes are a common Early Iron Age type (Glastonbury, 11, 373), and as adze-hammers are known in Roman contexts (cf. Eckford, E. 15; Carling- warkpresene 50). Th C , . t example, -without hammer-butt, comes neareo rt Iron Age than to known Roman types. B. 40. SOCKETED GOUGE, iron, 9-3 ins. long, broken at edge. (DW 103.) Socketed gouge wood-workinr sfo typee gar s commo Earlo nt y Iroe nAg and Roman carpenters' equipment: e.g. Glastonbury, II, 386, with references to examples from that site, Bigbury, Hod Hill and Woodcuts; Newstead, pi. lix ; als3 , o Silchester (Reading Mus. . 50)B ) . (cf.no B. 41. HINGE FRAGMENT (?), iron, a flat bar 7-2 ins. long, with one end expanded perforatedd an 121.W )(D . . SPIKE42 . B , iron, 9-8 ins. long, rectangular section with head turne t righda t angle. (DW 120.) B. 43. PERFORATED MOUNTING, iron, broken, 8-2 ins. overall. An iron plate of uncertain original form with terminal perforated expansion and curved outling. (DW 156.) This appears to be part of an iron mounting of the type of Newstead, pi. Ixv, 3, though half its size. . TANGE44 . B D CHISEL (?), 5-2 ins. long , 137.bentW )(D . B. 45. PUNCH, iron, 3-8 ins. long. (DW 1446.) B. 46. PUNCH, iron, 3-8 ins. long. Not illustrated. (DW 144a.) B. 47. PUNCH, iron, 4-9 ins. long. (DW 144c.) Cf. Carlingwark, C. 64-67; Newstead, pi. Ixvi, 16, 17. THREE METAL-WORK HOARDS OF THE ROMAN PERIOD. 49 . BAR48 . ,B iron, 5-0 ins. long, D-section, perforated 129.W (D ) . . BAR49 . B , iron, broken, 6-0 ins. long, oval sectio d slightlnan y burred end. (DW 118.) Probably part of a punch or cold-chisel of Roman type.
B50
Fig. Object 13 fro) .bronzd 51 m an iro f . . ) 37-49 so Blackbur en(B (B 50 . B ; n Mill. (J.)
B. 50. SOCKETED GOUGE, iron, 6-4 ins. long. (DW 104.) above0 4 . Cf.B .
. DISC51 . ,B bronze, 2-8 ins. diameter, moulded, with oblong slit slightlf of y centre. (DW. 91.) This is presumably a Roman harness-mount of the type of pi. Ixxiv, especiall. 9 . yno 4 VOL. LXXXVII. 50 PROCEEDING E SOCIETYTH F O S , 1952-53. B. 52. PATEEA, bronze, with handle missing. Not illustrated. (DW 89.) This characteristically Roman objec s discussewa t illustrated dan y db Bosanquet (P.S.A.S., LXII (1927-8), 246-54, fig. 1, 1) and Curie (ibid., LXVI (1931-2), 300, fig. 12, 1), the former assigning the inception of the particular type represented at Blackburn Mill to A.D. 120-150. t illustrated . 52-6no B e 5ar d comprisan , e miscellaneous broken frag- ments including iron HOOKS (B. 52-54); iron STRIP (B. 57, 58, 63); small iron L-shaped STAPLE . 140-142)(B S , ann INGOa d r cako Tf o e bronze, circular, 3-4 ins 90)W . D ,diamete illustrate; 65 . (B ry Curieb d , P.S.A.S., LXVI (1931-2), 315, fig. 22, 33. This is a smaller version of the circular copper ingottype th ef s o illustrate Curiey db , loo. cit., fig , fro.37 m Carleton, Wigtownshire e half-cakth r o , f bronzo e e from Dundonald (Dick Institute, Kilmarnock). APPENDIX. Technical Fragmente th Note n o f Irono Mail from Carlingivark Loch . 74)(C . By E. MARTIN BURGESS.
• . The fragment examined is composed of alternate rows of riveted links and whole links. The whole links are punched from an iron sheet. Holes were punched wit hsmala l punc thed e han nlargea th t ou r t punc uses cu o hwa dt rings; this give whole sth e rings their somewhat square wire section. Wire thicknesses: 0-056, 0-045, 0-039, 0-046, 0-047 inches. External diameter: 0-267, 0-271, 0-273, 0-274, 0-272 inches. These links are quite circular, as would be expected. The riveted links have dimension follows:s sa — Wire thickness: 0-033, 0-042, 0-035, 0-032, 0-039 inches. External diameter (paralle riveo t l t joint): 0-292, 0-286, 0-304, 0-307, 0-323 inches. The wire of these links is rather natter in section than the wire of the whole links. The former have a slight ovality, the major axis being parallel to the rivet joint which to some extent accounts for their greater diameter. The real interes rivetee th f o td link face s th lie t n thas i rivete th tdomee sar d on both sides and are made from cylindrical wire, as opposed to the usual medieval wedge type rivet whic mucs hi h productionmorr fo ed practicaan e us . r lfo Som e e cylindricaoth f l rivets hav e ben ar gone td on ovee an n througr o r fa o hto side, while on the other side hardly any rivet shows. In my experience this is very likel happeo yt n wit cylindricae hth l rivet however much car takes ei n in closing it. There is no visible "water-shed" formation round the rivet area whic bees hha n simply sprea t befordou e punching. The alternate riveteunusualy whold wa dan o n e . n ringi Examplee sar e sar the Newstead fragments (Curie, Newstead, 161; pi. xxxviii, 10) and those from the Thorsbjerg find in Denmark (Engelhardt, Denmark in the Early Iron Age (1866) . , nos6 3) , L 2 . P ,